Color Desktop Printer Technology 082475364X

Color Desktop Printer Technology 082475364X
Founding Editor
Brian J. Thompson
University of Rochester
Rochester, New York
Electron and Ion Microscopy and Microanalysis: Principles
and Applications, Lawrence E. Murr
Acousto-Optic Signal Processing: Theory and Implementation,
edited by Norman J. Berg and John N. Lee
Electro-Optic and Acousto-Optic Scanning and Deflection,
Milton Gottlieb, Clive L. M. Ireland, and John Martin Ley
Single-Mode Fiber Optics: Principles and Applications,
Luc B. Jeunhomme
Pulse Code Formats for Fiber Optical Data Communication:
Basic Principles and Applications, David J. Morris
Optical Materials: An Introduction to Selection and
Application, Solomon Musikant
Infrared Methods for Gaseous Measurements: Theory
and Practice, edited by Joda Wormhoudt
Laser Beam Scanning: Opto-Mechanical Devices, Systems,
and Data Storage Optics, edited by Gerald F. Marshall
Opto-Mechanical Systems Design, Paul R. Yoder, Jr.
Optical Fiber Splices and Connectors: Theory and Methods,
Calvin M. Miller with Stephen C. Mettler and Ian A. White
Laser Spectroscopy and Its Applications, edited by
Leon J. Radziemski, Richard W. Solarz, and Jeffrey A. Paisner
Infrared Optoelectronics: Devices and Applications,
William Nunley and J. Scott Bechtel
Integrated Optical Circuits and Components: Design
and Applications, edited by Lynn D. Hutcheson
Handbook of Molecular Lasers, edited by Peter K. Cheo
Handbook of Optical Fibers and Cables, Hiroshi Murata
Acousto-Optics, Adrian Korpel
Procedures in Applied Optics, John Strong
Handbook of Solid-State Lasers, edited by Peter K. Cheo
Optical Computing: Digital and Symbolic, edited by
Raymond Arrathoon
Laser Applications in Physical Chemistry, edited by
D. K. Evans
Laser-Induced Plasmas and Applications, edited by
Leon J. Radziemski and David A. Cremers
22. Infrared Technology Fundamentals, Irving J. Spiro
and Monroe Schlessinger
23. Single-Mode Fiber Optics: Principles and Applications,
Second Edition, Revised and Expanded, Luc B. Jeunhomme
24. Image Analysis Applications, edited by Rangachar Kasturi
and Mohan M. Trivedi
25. Photoconductivity: Art, Science, and Technology, N. V. Joshi
26. Principles of Optical Circuit Engineering, Mark A. Mentzer
27. Lens Design, Milton Laikin
28. Optical Components, Systems, and Measurement Techniques,
Rajpal S. Sirohi and M. P. Kothiyal
29. Electron and Ion Microscopy and Microanalysis: Principles
and Applications, Second Edition, Revised and Expanded,
Lawrence E. Murr
30. Handbook of Infrared Optical Materials, edited by Paul Klocek
31. Optical Scanning, edited by Gerald F. Marshall
32. Polymers for Lightwave and Integrated Optics: Technology
and Applications, edited by Lawrence A. Hornak
33. Electro-Optical Displays, edited by Mohammad A. Karim
34. Mathematical Morphology in Image Processing, edited by
Edward R. Dougherty
35. Opto-Mechanical Systems Design: Second Edition,
Revised and Expanded, Paul R. Yoder, Jr.
36. Polarized Light: Fundamentals and Applications,
Edward Collett
37. Rare Earth Doped Fiber Lasers and Amplifiers, edited by
Michel J. F. Digonnet
38. Speckle Metrology, edited by Rajpal S. Sirohi
39. Organic Photoreceptors for Imaging Systems,
Paul M. Borsenberger and David S. Weiss
40. Photonic Switching and Interconnects, edited by
Abdellatif Marrakchi
41. Design and Fabrication of Acousto-Optic Devices, edited by
Akis P. Goutzoulis and Dennis R. Pape
42. Digital Image Processing Methods, edited by
Edward R. Dougherty
43. Visual Science and Engineering: Models and Applications,
edited by D. H. Kelly
44. Handbook of Lens Design, Daniel Malacara
and Zacarias Malacara
45. Photonic Devices and Systems, edited by
Robert G. Hunsberger
46. Infrared Technology Fundamentals: Second Edition,
Revised and Expanded, edited by Monroe Schlessinger
47. Spatial Light Modulator Technology: Materials, Devices,
and Applications, edited by Uzi Efron
48. Lens Design: Second Edition, Revised and Expanded,
Milton Laikin
49. Thin Films for Optical Systems, edited by Francoise R. Flory
50. Tunable Laser Applications, edited by F. J. Duarte
51. Acousto-Optic Signal Processing: Theory and Implementation,
Second Edition, edited by Norman J. Berg
and John M. Pellegrino
52. Handbook of Nonlinear Optics, Richard L. Sutherland
53. Handbook of Optical Fibers and Cables: Second Edition,
Hiroshi Murata
54. Optical Storage and Retrieval: Memory, Neural Networks,
and Fractals, edited by Francis T. S. Yu
and Suganda Jutamulia
55. Devices for Optoelectronics, Wallace B. Leigh
56. Practical Design and Production of Optical Thin Films,
Ronald R. Willey
57. Acousto-Optics: Second Edition, Adrian Korpel
58. Diffraction Gratings and Applications, Erwin G. Loewen
and Evgeny Popov
59. Organic Photoreceptors for Xerography, Paul M. Borsenberger
and David S. Weiss
60. Characterization Techniques and Tabulations for Organic
Nonlinear Optical Materials, edited by Mark G. Kuzyk
and Carl W. Dirk
61. Interferogram Analysis for Optical Testing, Daniel Malacara,
Manuel Servin, and Zacarias Malacara
62. Computational Modeling of Vision: The Role of Combination,
William R. Uttal, Ramakrishna Kakarala, Spiram Dayanand,
Thomas Shepherd, Jagadeesh Kalki, Charles F. Lunskis, Jr.,
and Ning Liu
63. Microoptics Technology: Fabrication and Applications of Lens
Arrays and Devices, Nicholas Borrelli
64. Visual Information Representation, Communication,
and Image Processing, edited by Chang Wen Chen
and Ya-Qin Zhang
65. Optical Methods of Measurement, Rajpal S. Sirohi
and F. S. Chau
66. Integrated Optical Circuits and Components: Design
and Applications, edited by Edmond J. Murphy
67. Adaptive Optics Engineering Handbook, edited by
Robert K. Tyson
68. Entropy and Information Optics, Francis T. S. Yu
69. Computational Methods for Electromagnetic and Optical
Systems, John M. Jarem and Partha P. Banerjee
70. Laser Beam Shaping, Fred M. Dickey and Scott C. Holswade
71. Rare-Earth-Doped Fiber Lasers and Amplifiers: Second
Edition, Revised and Expanded, edited by
Michel J. F. Digonnet
72. Lens Design: Third Edition, Revised and Expanded,
Milton Laikin
73. Handbook of Optical Engineering, edited by Daniel Malacara
and Brian J. Thompson
74. Handbook of Imaging Materials: Second Edition, Revised
and Expanded, edited by Arthur S. Diamond
and David S. Weiss
75. Handbook of Image Quality: Characterization and Prediction,
Brian W. Keelan
76. Fiber Optic Sensors, edited by Francis T. S. Yu
and Shizhuo Yin
77. Optical Switching/Networking and Computing for Multimedia
Systems, edited by Mohsen Guizani and Abdella Battou
78. Image Recognition and Classification: Algorithms, Systems,
and Applications, edited by Bahram Javidi
79. Practical Design and Production of Optical Thin Films:
Second Edition, Revised and Expanded, Ronald R. Willey
80. Ultrafast Lasers: Technology and Applications, edited by
Martin E. Fermann, Almantas Galvanauskas, and Gregg Sucha
81. Light Propagation in Periodic Media: Differential Theory
and Design, Michel Nevière and Evgeny Popov
82. Handbook of Nonlinear Optics, Second Edition, Revised
and Expanded, Richard L. Sutherland
83. Polarized Light: Second Edition, Revised and Expanded,
Dennis Goldstein
84. Optical Remote Sensing: Science and Technology,
Walter Egan
85. Handbook of Optical Design: Second Edition, Daniel Malacara
and Zacarias Malacara
86. Nonlinear Optics: Theory, Numerical Modeling,
and Applications, Partha P. Banerjee
87. Semiconductor and Metal Nanocrystals: Synthesis
and Electronic and Optical Properties, edited by
Victor I. Klimov
88. High-Performance Backbone Network Technology, edited by
Naoaki Yamanaka
89. Semiconductor Laser Fundamentals, Toshiaki Suhara
90. Handbook of Optical and Laser Scanning, edited by
Gerald F. Marshall
91. Organic Light-Emitting Diodes: Principles, Characteristics,
and Processes, Jan Kalinowski
92. Micro-Optomechatronics, Hiroshi Hosaka, Yoshitada Katagiri,
Terunao Hirota, and Kiyoshi Itao
93. Microoptics Technology: Second Edition, Nicholas F. Borrelli
94. Organic Electroluminescence, edited by Zakya Kafafi
95. Engineering Thin Films and Nanostructures with Ion Beams,
Emile Knystautas
96. Interferogram Analysis for Optical Testing, Second Edition,
Daniel Malacara, Manuel Sercin, and Zacarias Malacara
97. Laser Remote Sensing, edited by Takashi Fujii
and Tetsuo Fukuchi
98. Passive Micro-Optical Alignment Methods, edited by
Robert A. Boudreau and Sharon M. Boudreau
99. Organic Photovoltaics: Mechanism, Materials, and Devices,
edited by Sam-Shajing Sun and Niyazi Serdar Saracftci
100. Handbook of Optical Interconnects, edited by Shigeru Kawai
101. GMPLS Technologies: Broadband Backbone Networks
and Systems, Naoaki Yamanaka, Kohei Shiomoto, and Eiji Oki
102. Laser Beam Shaping Applications, edited by Fred M. Dickey,
Scott C. Holswade and David L. Shealy
103. Electromagnetic Theory and Applications for Photonic
Crystals, Kiyotoshi Yasumoto
104. Physics of Optoelectronics, Michael A. Parker
105. Opto-Mechanical Systems Design: Third Edition,
Paul R. Yoder, Jr.
106. Color Desktop Printer Technology, edited by Noboru Ohta
and Mitchell Rosen
edited by
Noboru Ohta
Rochester Institute of Technology
Rochester, New York
Rochester Institute of Technology
Rochester, New York
Boca Raton London New York
A CRC title, part of the Taylor & Francis imprint, a member of the
Taylor & Francis Group, the academic division of T&F Informa plc.
Published in 2006 by
CRC Press
Taylor & Francis Group
6000 Broken Sound Parkway NW, Suite 300
Boca Raton, FL 33487-2742
© 2006 by Taylor & Francis Group, LLC
CRC Press is an imprint of Taylor & Francis Group
No claim to original U.S. Government works
Printed in the United States of America on acid-free paper
10 9 8 7 6 5 4 3 2 1
International Standard Book Number-10: 0-8247-5364-X (Hardcover)
International Standard Book Number-13: 978-0-8247-5364-1 (Hardcover)
Library of Congress Card Number 2005050110
This book contains information obtained from authentic and highly regarded sources. Reprinted material is
quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts
have been made to publish reliable data and information, but the author and the publisher cannot assume
responsibility for the validity of all materials or for the consequences of their use.
No part of this book may be reprinted, reproduced, transmitted, or utilized in any form by any electronic,
mechanical, or other means, now known or hereafter invented, including photocopying, microfilming, and
recording, or in any information storage or retrieval system, without written permission from the publishers.
For permission to photocopy or use material electronically from this work, please access
( or contact the Copyright Clearance Center, Inc. (CCC) 222 Rosewood Drive,
Danvers, MA 01923, 978-750-8400. CCC is a not-for-profit organization that provides licenses and registration
for a variety of users. For organizations that have been granted a photocopy license by the CCC, a separate
system of payment has been arranged.
Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only
for identification and explanation without intent to infringe.
Library of Congress Cataloging-in-Publication Data
Color desktop printer technology / edited by Noburu Ohta, Mitchell Rosen.
p. cm. -- (Optical engineering ; 106)
Includes bibliographical references and index.
ISBN 0-8247-5364-X (alk. paper)
1. Color computer printers. 2. Color printing. I. Ohta, Noboru. II. Rosen, Mitchell. III. Optical
engineering (CRC Press) ; 106.
TK7887.7.E635 2005
Visit the Taylor & Francis Web site at
Taylor & Francis Group
is the Academic Division of Informa plc.
and the CRC Press Web site at
For Alma on her early morning bike rides.
M. R.
To Chiaki for her peerless patience
and continuing assistance.
N. O.
Printing has been practiced since the eighth century in China, when wood blocks
were used to reproduce Buddhist scriptures. As early as the 11th century, the
Chinese employed movable type, although modern printing technologies trace
back to Gutenberg’s rediscovery of movable type in the mid-1400s. Centuries of
innovation and improvement have gifted us with high-quality printing affordable
to the masses. In 1984, Apple introduced the Macintosh computer with its graphical interface and paired it with the ImageWriter, a dot-matrix printer used to
map screen pixels to dots on the page. The desktop printing genie was out of its
bottle. The decade of the 1990s opened with the introduction of color to the
desktop, first through the monitor and subsequently through printers, and it closed
with full-color desktop environments ubiquitous. Today’s printers are astonishingly inexpensive, robust, of incredible specifications, and with superb quality.
The introduction of better, faster, cheaper models is constant. Color desktop
publishing, the obvious corollary, has permeated lives and lifestyles worldwide
to the point where the norms for business and personal communications surpass
that which could only have been dreamed of 20 years ago.
While still an analog art prior to its partnership with computers, traditional
printing included the tasks of text production, drawing figures, converting photographs into dot images, and synthesizing the pieces through printing plates.
These steps were labor-intensive and time- and money-consuming, representing
a considerable portion of the printing cost before inks or paper. To overcome the
up-front costs, press runs needed to be high in volume. The publication of
documents was, thus, expensive and required special equipment, limiting its use
to an elite few. The evolution of desktop printing technology has literally eradicated these barriers. Through cost reduction, universal access, and highly intuitive
desktop publishing programs, the desktop printer has truly democratized the
production of high-quality paper documents.
It was foretold not long ago that computer technology would quickly yield
the paperless office. To many, this seemed a fair assumption. With a preponderance of documents being generated and stored online, and with the growth of
networking, e-mail, Portable Document Format (PDF), e-books, and most significantly, the World Wide Web (WWW), what could hardcopy continue to offer?
It appears the answer is: quite a lot. Predicting the future is an inexact science,
even when all signs point in a single direction. Alas, the opposite of the prediction
has come to pass: printed page per capita has continued growth at a ferocious pace.
An all-around bird’s eye view of the present art of the color desktop printer
fills an important gap in the available literature. Given color printers are so
thoroughly dispersed throughout the workplace and into the home, the audience
for this information is enormous. Some who are interested in the subject will be
considering the topic for the first time and will benefit from Part I of this book.
There, Chapter 1 and Chapter 2 give a useful introduction to the basic principles
of color printing and to the concepts of document and image quality. Chapter 3
goes into detail on the business and market of desktop printers, starting with a
historical overview, proceeding through present day, and moving toward the
Chapter 4 through Chapter 7, composing Part II, are devoted to the four major
color desktop printer platforms: inkjet, laser printer, thermal transfer, and film
recording. The technical details will enable a deep understanding of how printers
commonly work. The hope is that these chapters not only teach the current state
of these technologies but also prepare the reader to place new innovations encountered in the marketplace in context. This should prolong the useful shelf-life of
this book because printers, as with all computer peripherals, are faced with everincreasing obsolescence cycle frequency.
Part III looks at contemporary and future means for digital control of color.
Color management systems, particularly the International Color Consortium
industry standard approach, are discussed in Chapter 8. Chapter 9 is devoted to
the potential future direction of spectral printing.
A healthy dose of historical perspective permeates the chapters. This would
have been appreciated by Louis Walton Sipley, author of the book A Half Century
of Color, itself now just more than half a century old. Dr. Sipley opened his 1951
treatise with these words, as true today as they were when written more than 50
years ago:
In the recording of contemporary achievement in any phase of the arts and sciences
there is always a background of pioneering and creative research which must be
taken into consideration. If proper recognition is given to the many whose efforts
have enabled the attainment of modern perfection (so-called), then those occupying
today’s spotlight may be expected to wear their honors with reasonable humility.
This is particularly applicable to the art-science of color photography and color
reproduction on the printed page which has attained its present dominant position
through an evolutionary development.
This book presents to the interested novice, as well as the imaging and printer
technologist, an overview of the basics of today’s color desktop printers and their
history. Sufficient detail is provided while explanations of the fundamentals
should ensure that all levels of reader find satisfaction.
Mitchell Rosen and Noboru Ohta
Munsell Color Science Laboratory
Center for Imaging Science
Rochester Institute of Technology
Rochester, New York
Ross R. Allen
Hewlett Packard
Palo Alto, California, U.S.A.
Eric G. Hanson
Hewlett Packard
Palo Alto, California, U.S.A.
Takesha Amari
Chiba University
Chiba, Japan
Akira Igarashi
Fuji Photo Film Co., Ltd.
Fujinomiya, Japan
Jon S. Arney
Center for Imaging Science
Rochester Institute of Technology
Rochester, New York, U.S.A.
Francisco H. Imai
Pixim Corp.
Mountain View, California, U.S.A.
Roy S. Berns
Munsell Color Science Laboratory
Rochester Institute of Technology
Rochester, New York, U.S.A.
Yongda Chen
Munsell Color Science Laboratory
Rochester Institute of Technology
Rochester, New York, U.S.A.
Tsutomu Kimura
Fuji Photo Film Co., Ltd.
Kaisei-machi, Japan
Toshiya Kojima
Fuji Photo Film Co., Ltd.
Kaisei-machi, Japan
Kenichi Koseki
Chiba University
Chiba, Japan
Gary Dispoto
Hewlett Packard
Palo Alto, California, U.S.A.
Masahiro Kubo
Fuji Photo Film Co., Ltd.
Kaisei-machi, Japan
Atsuhiro Doi
Fuji Photo Film Co., Ltd.
Kaisei-machi, Japan
Nobuhito Matsushiro
Oki Data Corporation
Tokyo, Japan
Yasuji Fukase
Fuji Xerox Co., Ltd.
Sakai Nakai-Cho, Japan
John D. Meyer
Hewlett Packard
Palo Alto, California, U.S.A.
Nathan Moroney
Hewlett Packard
Palo Alto, California, U.S.A.
Fumio Nakaya
Document Product Company
Fuji Xerox Co., Ltd.
Sakai Nakai-cho, Japan
Frank Romano
School of Print Media
Rochester Institute of Technology
Rochester, New York, U.S.A.
Mitchell R. Rosen
Munsell Color Science Laboratory
Rochester Institute of Technology
Rochester, New York, U.S.A.
Lawrence A. Taplin
Munsell Color Science Laboratory
Rochester Institute of Technology
Rochester, New York, U.S.A.
Table of Contents
Chapter 1
Introduction to Printing.........................................................................................3
Takesha Amari and Kenichi Koseki
Chapter 2
Image Quality of Printed Text and Images ........................................................31
Jon S. Arney
Chapter 3
The Business and Market for Desktop Printers .................................................85
Frank Romano
Chapter 4
Inkjet ................................................................................................................111
Ross R. Allen, Gary Dispoto, Eric Hanson, John D. Meyer, and
Nathan Moroney
Chapter 5
Laser Printer......................................................................................................157
Fumio Nakaya and Yasuji Fukase
Chapter 6
Dye Thermal-Transfer Printer...........................................................................195
Nobuhito Matsushiro
Chapter 7
Film-Based Printers...........................................................................................211
Tsutomu Kimura, Atsuhiro Doi, Toshiya Kojima, Masahiro Kubo, and
Akira Igarashi
The Management of Color
Chapter 8
Color Management............................................................................................237
Mitchell R. Rosen
Chapter 9
Desktop Spectral-Based Printing ......................................................................249
Mitchell R. Rosen, Francisco H. Imai, Yongda Chen,
Lawrence A. Taplin, and Roy S. Berns
Part I
Introduction to Printing
Takesha Amari and Kenichi Koseki
Relief Printing .............................................................................................4
1.2.1 Historical Sketch .............................................................................4
1.2.2 Feature .............................................................................................7
1.2.3 Plate Making....................................................................................7
1.2.4 Printing Press...................................................................................8
1.2.5 Flexography .....................................................................................9
1.2.6 Application ....................................................................................10
Planographic Printing ................................................................................11
1.3.1 Historical Sketch ...........................................................................11
1.3.2 Feature ...........................................................................................11
1.3.3 Classification..................................................................................12
1.3.4 Plate Making..................................................................................12 Presensitized Plates .........................................................12 Waterless Plates...............................................................13 Laser Plate Making .........................................................13
1.3.5 Elements of Offset Lithography....................................................14
1.3.6 Printing Presses .............................................................................14
1.3.7 Dampening Systems ......................................................................17
1.3.8 Inking Systems ..............................................................................18
1.3.9 Troubles in Lithography................................................................19
Recess Printing ..........................................................................................19
1.4.1 Classification..................................................................................19
1.4.2 Gravure ..........................................................................................19 Historical Sketch .............................................................20 Plate Making ...................................................................21 Printing Process...............................................................24 Application ......................................................................25
1.4.3 Intaglio...........................................................................................25 Line Engraving................................................................26 Application ......................................................................26
Through-Printing .......................................................................................26
1.5.1 Feature ...........................................................................................27
Color Desktop Printer Technology
Plate ...............................................................................................28
Printing Process .............................................................................28
Application ....................................................................................30
Printing is defined as a permanent, graphic, visual communication medium,
including all the ideas, methods, and devices used to manipulate or reproduce
graphic visual messages. Generally, printing processes are accomplished by
applying an inked image carrier to the substrate as it operates through a highspeed press. A letterpress plate, a lithographic plate, a gravure cylinder, and a
screen used in screen printing are examples of image carriers.
For electrophotographic or inkjet printing, the data representing the images
are in digital form in computer storage and the image must be created each time
it is reproduced. For such digital printing technologies, image carriers are not
required to transfer the printed image to the substrate. Although most people
think of printing as ink on paper, printing is not limited to any particular medium.
Some special printing technologies reproduce the printed image without printing
ink. ISO TC130/WG1 classified the printing technology shown in Table 1.1.
This chapter explains printing processes having a concrete image carrier.
These printing processes can be divided into four categories depending on the
feature of the plate, namely, relief printing, planographic printing, recess printing,
and through-printing, as shown in Figure 1.1.
Relief printing is the natural and oldest printing method, using an image carrier
on which the image areas are raised above the non-image areas. Letterpress
printing is a kind of relief printing from movable type (Figure 1.2). Printing using
movable type appeared in China and Korea in the 11th century. In 1041, a Chinese
printer, Pi-Sheng, developed type characters from hardened clay. Type cast from
metal in Korea was widely used in China and Japan, and by the middle 1200s,
type characters were being cast in bronze. The oldest text known was printed
from such type in Korea in 1397 A.D.
Half a century later in 1440, Johann Gutenberg invented a systematic letterpress technology with movable type using a wine press as a high-speed printing
press. Until Gutenberg’s system of separate characters for printing on a press
with ink on paper, all books were laboriously handwritten by scribes. Gutenberg’s
most notable work was to complete a Bible with 42 lines to the page using
letterpress printing (Figure 1.3). People could read the Bible without a Catholic
priest. This innovation by Gutenberg was in many ways the catalyst for the
Reformation by Martin Luther 50 years later. It is worth noting that this one
Introduction to Printing
ISO Printing Technology
Workflow Stage
Inkless Printing
Photochemical Printing
Silver Halide Printing
Diazo Printing
Thermochemical Printing
Thermal Printing
Electrochemical Printing
Spark Discharge Printing
Ink Jet Printing
Graphic Technology
Plateless Printing
Plate-Based Printing
Post Press
Relief Printing
Planographic Printing
FIGURE 1.1 Basic printing processes.
Continuous Ink Jet Printing
Drop-on-demand Printing
Thermal Transfer Printing
Thermal Wax Transfer Printing
Thermal Dye Transfer Printing
Electrographic Printing
Electrographic Printing
Electrophotographic Printing
Ion Deposition Printing
Magnetographic Printing
Relief Printing
Flexographic Printing
Letterpress Printing
Letterset Printing
Planographic Printing
Lithographic Printing
Recess Printing
Gravure Printing
Intaglio Printing
Pad Transfer Printing
Permeographic Printing
Screen Printing
Stencil Printing
Recess Printing
Color Desktop Printer Technology
FIGURE 1.2 Movable type.
FIGURE 1.3 Gutenberg’s 42-line Bible.
Introduction to Printing
important innovation helped to terminate the Dark Age and bring about the new
era of the Renaissance. High technology is not irrelevant to human life and the
progress of technology does change social structure. Gutenberg’s innovation and
its impact on the Renaissance are good examples regarding this matter.
Any method in which the impression is taken from the raised parts of the printing
surface is described as relief printing. Printing is performed by cast metal type
or plates on which the image or printing areas are raised above the non-printing
areas. Ink rollers touch only the top surface of the raised areas; the surrounding
(non-printing) areas are lower and do not receive ink. The inked image is transferred directly to the paper.
The distinctive feature for recognizing letterpress is a heavier edge of ink
around each letter (ring of ink or marginal zone), as shown in Figure 1.4. The
ink tends to spread slightly from the pressure of the plate upon the printed surface.
Sometimes a slight embossing (caused by denting) appears on the reverse side
of the paper. The letterpress image is usually sharp and crisp.
Plates used for letterpress printing can be original, duplicate, or wraparound.
Original plates are usually photochemically engraved onto zinc, magnesium, or
copper. A wraparound plate is a thin one-piece relief plate wrapped around the
press cylinder like an offset plate. Photopolymer plates are also used as printing
The oldest of the photomechanical processes, photoengraving, pertains to the
production of relief printing plates for letterpress. Photoengraved plates fall into
two categories: line and halftone. Generally, line and coarse screen engravings
are made on zinc and magnesium, and fine screen halftone plates are made on
Conventional etching is done as follows. The plate is coated with a lightsensitive coating, exposed to a negative, and then processed. The exposed coating
(a) Letterpress
(b) Gravure
(c) Offset Lithography
FIGURE 1.4 Characteristics of a letter printed by each printing method.
Color Desktop Printer Technology
serves as a resist for protecting the image areas as the non-image areas are etched
in acid baths. The main problem in conventional etching is to maintain the correct
dot and line width at the proper etch depth, which is accomplished by scale
compression in the negative and powderless etching.
Powderless etching can be used for zinc, magnesium, and copper plates. Zinc
and magnesium use the same process. Copper uses essentially the same principles,
but the chemicals and mechanism are different. The plate is prepared as in the
conventional process, but a special etching machine is used. Zinc and magnesium
are etched in an emulsion of dilute nitric acid, a wetting agent, and oil. During
etching, the wetting agent and oil attach to the surface of the metal, forming an
etch-resistant coating on the sidewalls of the etched elements, thus preventing
In copper etching, the etchant used is ferric chloride, in which certain organic
chemicals are dissolved. During etching, the additive chemicals react with the
dissolved metal to form a gelatinous precipitate, which adheres to the sides of
the image elements and protects them from undercutting.
Photopolymer films and plates are used in the relief printing process in which
a plastic designed so it changes upon exposure to light. The photopolymer plates
are used for wraparound plates and are most often used in flexography.
All printing presses have a feeding system, registration system, printing unit, and
delivery system. The printing unit is the section on printing presses that furnishes
the components for reproducing an inked image on the substrate under pressure.
For halftone relief printing, slight local modifications in pressure are necessary
to obtain proper tone reproduction. When highlights and shadows are at the same
height, the highlights exert more pressure than the shadows so that pressure must
be relieved in the highlights and more pressure is added in the shadows or heavy
printing areas. This procedure is known as make ready operation.
There are three types of presses: platen, flat-bed cylinder, and rotary, as shown
in Figure 1.5. On platen and flat-bed cylinder presses, the type or plates are
mounted on a flat surface or bed. Type and flat plates cannot be used on rotary
presses, where the printing member is a cylinder, and plates must be curved.
Printing is done on sheets of paper on sheet-fed presses or on rolls of paper
on web-fed presses (Figure 1.6).
A platen press carries both the paper and the type form on flat surfaces known
as the platen and the bed. A flat-bed cylinder press has a moving flat bed that
holds the form, and a fixed rotating impression cylinder provides the pressure.
The paper, held securely to the cylinder by a set of grippers, is rolled over the
form as the bed passes under the cylinder. As the bed returns to its original
position, the cylinder is raised, the form is re-inked, and the printed sheet is
delivered. A rotary press is the fastest and most efficient of the three types of
letterpress machines and has been used mainly for long runs.
Introduction to Printing
Plate Cylinder
Flat-Bed Cylinder
FIGURE 1.5 Printing methods used to transfer ink to the substrate in conventional printing
Sheet-Fed Press
Web-Fed Press
FIGURE 1.6 Schematic diagram of printing presses having four units.
Flexography is a high-speed web-rotary press with the relief plates made from
rubber or plastic. Typical web presses for flexography are shown in Figure 1.7.
The inking system consists of an ink fountain, an anilox roller system, and wateror solvent-based inks. The anilox roller is a steel or ceramic ink metering roller.
Its surface is engraved with small uniform cells that carry and deposit a thin,
controlled layer of ink film onto the plate (Figure 1.8). An important feature of
flexographic printing is that a uniform film of ink can be printed even on rough
papers because the surface of the rubber plate is sufficiently resilient so it can
Color Desktop Printer Technology
(b) Stack Type
(a) CIC Drum Type
(c) Inline Type
FIGURE 1.7 Typical web presses for flexography.
Printing Ink
Printing Ink
Surface of Anilox Roller
FIGURE 1.8 Ink metering system in flexo printing.
force ink into the hollows of the substrate. Using solvent type inks has made
flexography possible to print on various plastic films.
Sheet-fed letterpresses on small platen and flat-bed cylinder presses are used for
short run printing, such as for letterheads, billheads, envelopes, announcements,
invitations, and small advertising brochures. Larger sheet-fed letterpresses are
used for general printing, such as for books, catalogs, advertising, and packaging.
Web letterpresses are used for newspapers and magazines.
For nearly 400 years, Gutenberg’s foundry type dominated all the printing
trades. However, in the last century, especially in the last two decades, offset
Introduction to Printing
lithography has assumed the position of importance in the printing industry,
replacing type and letterpress. Only flexography is still used in the packaging
industry as relief printing.
Printing with the flexographic process includes decorated toilet tissue, bags,
corrugated board, foil, hard-calendered papers, cellophane, polyethylene, and
other plastic films. It is well suited for printing large areas of solid color. Inks
can be overlaid to obtain high gloss and special effects. The growth of flexography
parallels the expansion of the packaging industry and the development of the
central impression cylinder press, new ink metering systems, and new photopolymer plates. Halftones as fine as 150 lines per inch can be printed on flexible films.
Lithography was invented by Aloys Senefelder in 1796. The printing surface was
a level slab of a special form of limestone from Bavaria. The litho stone is ground
flat and polished at the top surface. The image is formed by drawing or writing
with a water-soluble fatty acid ink. The whole surface is treated with a weak
aqueous solution of nitric acid and then etched and coated with an aqueous
solution of gum arabic. After moistening the surface with water and inking with
the fatty ink, the image areas accept the ink, whereas the non-image areas remain
wet and reject it. Successive copies are taken by alternately dampening the stone,
applying ink, and printing on paper.
Lithography is based on the principle that grease and water do not mix. On a
lithographic plate the separation between the image and the non-image areas is
maintained chemically because they are essentially on the same plane; the image
areas must be ink receptive and refuse water, and the non-image areas must be
water receptive and refuse ink (Figure 1.9). However, in reality ink and water do
mix slightly. If they didn’t, lithography would not be possible. If they mix too
much, problems such as tinting, scumming, and bleeding occur.
Dampening Solution
Desensitized Layer
Non-Image Area
FIGURE 1.9 Schematic diagram of lithographic plate.
Image Area
Color Desktop Printer Technology
The planographic process is a method of printing from a flat surface. The image
areas accept printing ink, and the non-image areas reject it. The best known and
most widely used planographic process is lithography. Litho printing is divided
into two categories: direct lithography and offset lithography. A. F. Harris invented
offset lithography, observing that a sharp print was transferred to paper after
accidentally first printing onto the rubber blanket of the press cyclinder. A collotype, which resembles lithography in some respects, is another form of planographic printing.
On the lithographic plate, ink receptivity is achieved with inherently oleophilic
(oil-loving) resins or metals such as copper or brass on the image areas. On the
other hand, water receptivity of the non-image areas is usually achieved by using
metals such as aluminum, chromium, or stainless steel, whose oxides are hydrophilic (water-loving). Water receptivity is maintained in plate making and storage
by using natural and synthetic gums. The most widely used gum is gum arabic.
Most lithographic plates use either grained or anodized aluminum as a base.
An advantage of lithographic plates, besides simplicity and low cost, is the ease
of making minor corrections on the press. If corrections are extensive, however,
it is more economical to make a new plate. Much of the growth in the lithographic
industry in recent years may be attributed to this advantage. Automatic processors
for plate making are used almost as extensively as for photography. These processors are an important factor in the use of web offset by newspapers. Some
processors combine exposure with the processing and gumming, and several
include coating and exposing as well. Figure 1.10 shows the schematic representation of lithographic plate making. Presensitized Plates
Presensitized plates (PS-plates) are those in which the light-sensitive coating
becomes the ink-receptive image area on the plate. Most are made from negatives. There are two types of PS-plates: additive and subtractive. On the additive
plate, the ink-receptive lacquer is added to the plate during processing. On the
subtractive plate, it is part of the precoating, and processing removes it from
the non-printing areas. Such plates are also called surface plates. Until recently,
all PS-plates were used for short or medium runs. For many years, albumin
plates dominated this field, but they are now obsolete. PS-plates are currently
diazo presensitized (precoated) or wipe-on (in-plant coated) for short and
medium runs and prelacquered diazo presensitized and photopolymer plates for
longer runs.
Introduction to Printing
P-S Plate (presensitized plate)
Light-Sensitive Diazo Coatings
Aluminum Plate
Negative Film
Aluminum Plate
Aluminum Plate
Gum Arabic
Aluminum Plate
FIGURE 1.10 Planography plate-making process using P-S plate. Waterless Plates
Waterless lithography, also called driography, is a planographic process like
lithography, but it prints without dampening water. The process eliminates all of
the disadvantages caused by the need for an ink–water balance in lithography but
retains all the advantages of low plate costs, ease of make ready, high speed, and
good print quality plus the advantages of letterpress in ease of printing and low
waste. Waterless plates generally consist of ink on aluminum for the printing
areas and silicone rubber for the non-image areas. Silicone rubber has a very low
surface energy and, thus, resists being wet by anything, especially ink. However,
under the pressure and heat of printing, ordinary litho ink has a tendency to smear
over the silicone and cause scumming or toning. Laser Plate Making
Lasers are high-energy concentrated light sources that are used for scanning and
recording images at high speed. Several plate-making systems have been developed in which helium–neon (He/Ne) lasers are used to scan a pasteup, and this
information is processed to expose a plate with an argon–ion laser. Because the
Color Desktop Printer Technology
laser can be controlled by electronic impulses, it can be operated by digital signals
from a computer, thus enabling satellite and facsimile transmission of plate
images to remote printing locations.
Offset printing is an indirect printing method in which the inked image on a press
plate is first transferred to a rubber blanket, which in turn offsets the inked
impression to a press sheet. Letterpress and gravure can also be printed using the
offset principle, but most lithography is printed in this way, then offset printing
is a synonym of the offset lithography. Offset lithography has the following
feature: the rubber printing surface conforms to irregular printing surfaces, resulting in the need for less pressure, improved print quality, and halftones of good
quality on rough surfaced papers. Because the paper does not contact the lithographic plate directly, abrasive wear of the plate is reduced and the plate life
increases considerably. The image on the plate is straight-reading rather than
reverse-reading. Less ink is required for equal coverage, drying speeds up, and
smudging and set-off are reduced.
All offset presses make one impression with each revolution of the cylinders.
As shown in Figure 1.11, conventional offset presses have three printing cylinders
(plate, blanket, and impression) as well as inking and dampening systems. As the
plate that is clamped to the plate cylinder rotates, it comes into contact with the
dampening rollers first and then with the inking rollers. The dampeners wet the
plate so the non-printing area will repel ink. The inked image is then transferred
to the rubber blanket, and paper is printed as it passes between the blanket and
impression cylinders.
There are two types of offset lithography presses: the sheet-fed press and the web
press (Figure 1.12). A sheet-fed press feeds and prints on individual sheets of
paper. A printing press that prints on both sides of the paper using the blanketto-blanket principle in one pass is called a perfecting press. Some sheet-fed
presses are designed as perfecting presses. A web press is a rotary press that
prints on a continuous web of paper fed from a roll and threaded through the
press. Web presses are gradually replacing the sheet-fed press.
Sheet-fed printing has the following advantages: a large number of sheet or
format sizes can be printed on the same press and waste sheets can be used during
make ready, so good paper is not spoiled while getting position or color up for
running. Sheet-fed lithography is used for printing advertising, books, catalogs,
greeting cards, posters, packaging, decalcomanias, and art reproduction.
Much of the growth in the lithographic industry in recent years can be
attributed to web offset, which is used to produce newspapers, magazines, business forms, computer letters, mail order catalogs, gift wrappings, books, and a
variety of commercial printing. The latest innovations in web offset are in
Introduction to Printing
Dampening System
FIGURE 1.11 Elements of offset lithography press.
common impression, keyless tower, and in-line design, which make web offset
competitive with gravure in long run printing. Speed is the main advantage of
web offset. Speeds of 300 meters per minute are common, and new presses are
designed for speeds of 500 meters per minute and faster. Most web offset presses
have in-line folders where various combinations of folds convert the web into
folded signatures. Other in-line operations that can be performed on-press include
paste binding, perforating, numbering, rotary sheeting, and slitting. All of these
make web offset very flexible, and all are done while the presses are running at
high speeds, up to four times faster than sheet-fed presses.
The main disadvantage of web offset and web letterpress is that they have a
fixed cut-off (i.e., all sheets cut off at the same length).
There are four types of web offset presses:
1. The blanket-to-blanket press has no impression cylinders (Figure 1.13).
The blanket cylinder of one unit acts as the impression cylinder for
the other and vice versa. Each printing unit has two plates and two
blanket cylinders. The paper is printed on both sides at the same time
as it passes between the two blanket cylinders.
2. The in-line open press is similar to a sheet-fed offset press, except the
cylinder gap is very narrow. Grippers and transfer cylinders are eliminated. Each unit prints one color on one side; additional units are
required for additional colors. To print the reverse side, the web is
turned over between the printing units by means of turning bars, which
Color Desktop Printer Technology
Sheet-Fed Press
Press Unit
Plate Cylinder
Blanket Cylinder
Impression Cylinder
Web Press
Blanket Cylinder
Impression Cylinder
FIGURE 1.12 Sheet-fed and web-fed offset printing presses.
expose the unprinted side of the web to the remaining printing units.
This type of press is used extensively for printing business forms.
3. The drum or common impression cylinder (CIC) press has all the
blanket cylinders grouped around a large common impression cylinder
(Figure 1.14). Some multi-color web presses designed as two to five
colors are printed in rapid succession on one side, after which the web
is dried and turned, and printed on the reverse side.
4. Keyless tower printing system: In ink supplying systems, anilox rollers
are used to meter the ink quantity instead of the ink adjusting key and
the printing ink emulsifies the dampening solution. Each printing unit
is designed as perfecting and stacks vertically. Usually, this printing
system is used in multicolor news printing (Figure 1.15).
Introduction to Printing
Prints One Side
Prints Back Side
FIGURE 1.13 Blanket-to-blanket–type perfecting press.
Plate Cylinder
Plate Cylinder
FIGURE 1.14 Common impression cylinder web press.
The conventional dampening system on offset presses transfers the dampening
solution directly to the plate. In the Dahlgren type of direct-feed dampening
system, the fountain solution, containing up to 25% alcohol, is metered to the
plate through the inking system, or it can be applied directly to the plate, as in
other systems. In general, this type of dampening system uses less water and
Color Desktop Printer Technology
Printing Paper
Ink Roller
Plate Cylinder
Dampening Solution Spray
Ink Cylinder
Digital Ink Pump
FIGURE 1.15 Keyless tower printing system.
reduces paper waste at the start-up of the press. Because of the cost of isopropyl
alcohol and regulation of volatile organic compounds (VOCs), a number of new
fountain solutions have been developed to reduce or replace the alcohol in this
type of dampening system.
Inking systems are designed to transport ink from the ink fountain to the printing
plate. All systems use composition rollers. Some have plastic coated rollers and
others have copper-plated steel rollers to prevent stripping of the ink on the
distributors. Some inking systems, especially on web presses, are water cooled.
The large number of rollers is needed in inking systems because litho inks are
thixotropic materials, and sufficient kneading by the rollers is necessary for good
fluidity. With old albumin and zinc plates, which require much water for dampening, a large roller surface area is needed to evaporate the water from the ink
to keep it from waterlogging and emulsifying.
Introduction to Printing
Damages to the planographic plate and the lack of water–ink balance are characteristic of the troubles in lithography. Scumming and greasing damage the
planographic plate and tinting, emulsifying, bleeding, washing, and spreading are
due to the lack of water–ink balance. The non-imaging area on the litho plate is
protected by a desensitizing layer to refuse the ink; however, once the layer is
damaged by foreign material or grease, these areas begin to take ink. This
phenomenon is called scumming. Tinting occurs when a lithographic ink picks
up too much dampening solution and prints a weak snowflake pattern. In extreme
cases, the ink actually emulsifies in the dampening solution, causing an overall
tint to quickly appear on the unprinted areas of the sheet.
Recess printing is a method of taking impressions from recesses engraved or
etched in the printing plate or cylinder. The most important application is rotary
photogravure printing, but many other methods in intaglio are still used in important fields such as bank notes and stock certificates. Schematic classification of
recess printing is shown in Figure 1.16.
In rotary photogravure printing, image areas consist of cells or wells etched into
a copper cylinder or wraparound plate, and the cylinder or plate surface represents
Conventional Gravure
Gravure Printing
Lateral Hard Dot (Half Tone)
Direct Transfer
Pad Transfer Printing
Recess Printing
Direct contact wells
vary in opening
size, but all have the
same depth
Copper Plate Engraving
Line Engraving
Intaglio Printing
Steel Plate Engraving
Artistic Engraving
Drypoint Engraving
FIGURE 1.16 Schematic classification of recess printing.
Color Desktop Printer Technology
FIGURE 1.17 Conventional gravure plate.
the non-printing areas, as shown in Figure 1.17. The basic elements of gravure
printing are shown in Figure 1.18. The plate cylinder rotates in a bath of ink. The
excess is wiped off the surface by a flexible steel doctor blade. The ink remaining
in the thousands of recessed cells forms the image by direct transfer to the paper
as it passes between the plate cylinder and the impression cylinder. Gravure
printing is considered to be excellent for reproducing pictures, but high platemaking expense usually limits its use to long runs. A distinctive feature for
recognizing gravure is that the entire image is screened — not only halftone
images but also type and line drawings. Historical Sketch
The history of gravure printing began with the work of creative artists during the
Italian Renaissance in the 1300s. The recognized inventor of modern gravure
Printing Paper
Doctor Blade
Plate Cylinder
Printing Ink
FIGURE 1.18 Basic elements of gravure printing.
Introduction to Printing
printing is Karl Kliche. He began experimenting with photographic copper etching in 1875. He made the revolutionary move from flat printing plates to cylinder
forms. He developed the first doctor blade and even designed a method of printing
color on a web press. He originated the term rotogravure for printing from a
cylinder. Plate Making Chemical Method
In conventional gravure, the image is transferred to the copper cylinder by the
use of a sensitized gelatin transfer medium known as carbon tissue. The process
of plate making of gravure is shown in Figure 1.19. The carbon tissue is first
exposed in contact with an overall gravure screen, which is usually 150 lines per
inch, virtually invisible to the naked eye (Figure 1.20). The gravure screen serves
a purely mechanical purpose and, unlike other processes, has nothing to do with
producing the tones of the picture. It merely provides the partitions or walls of
the cells etched into the cylinder to form a surface of uniform height for the
doctor blade to ride on. Then the continuous-tone positives are exposed in contact
with the carbon tissue. In the highlight tones in the positives, where light passes
through freely, the gelatin on the carbon tissue becomes proportionately harder
with light intensity. The exposed carbon tissue is positioned on the copper plate
or cylinder with precision machines. After removal of paper backing, the tissue
is developed by hot water, leaving gelatin of varying thickness in the square dot
areas between the hardened screen lines. The etching is done in stages using
solutions of ferric chloride at varying concentration levels. Photographic resists
are being developed to replace the carbon tissue. They are more stable, easier to
use, and can be stored for a longer period. Conventional gravure is used for highquality illustrations but mainly for short runs because of doctor blade wear of
the shallow highlight dots.
Gravure cylinders are chromium plated for long runs. On very long runs, the
chromium is worn off by friction of the doctor blade over the cylinder. In such
cases the chromium-plated layer on the cylinder is removed, rechromed, and
replaced in the press for continuing the run. Electromechanical Engraving
Another method of electromechanical engraving is the Helioklischograph. In this
equipment, positives or negatives of the copy made on a special opaque white
plastic are scanned as the cylinder is being engraved electromechanically by
special diamond styli, as shown in Figure 1.21. Research is being done on the
use of electron-beam and laser etching of the copper cylinder, which are considerably faster than electromechanical engraving. The Lasergravure process by
Crosfield uses a plastic-coated, copper-plated cylinder in which helical grooves
with variable depths and cross bridges are etched with a 100-W CO2 laser driven
by a scanner or other front end system.
Color Desktop Printer Technology
Exposed Hardened
Paper Backing
(a) Side view of carbon tissue after exposure to a cross-line screen
MiddleTone Area
Paper Unexposed
Backing Emulsion
(b) Side view of carbon tissue after exposure to a continuous-tone film positive
MiddleTone Area
Gravure Cylinder
Paper Emulsion
(c) Side view of carbon tissue after transfer on the gravure cylinder
MiddleTone Area
Gravure Cylinder
(d) Side view of carbon tissue after development by hot water
FIGURE 1.19 Conventional gravure plate-making process. The Variable Area–Variable Depth Plate
(Figure 1.22)
The process for long run periodical printing differs from the conventional gravure
process just described in that the size of the cells as well as the depth varies to
produce more durable tones in publication printing. The cells of lighter areas are
smaller but deeper. In this process, continuous-tone positives and resists are used
Introduction to Printing
FIGURE 1.20 Cross-line gravure screen.
FIGURE 1.21 Electromechanical engraving with special diamond styli.
as in conventional gravure but, instead of the overall screen, a special halftone
positive is used. In the direct transfer or variable area method, an acid-resistant,
light-sensitive coating is first applied to the copper. The screened positive is
wrapped around the cylinder and exposed directly to it by a strong light source
through a narrow slit as the cylinder turns. The cylinder is then developed, and
the coating, which has not been exposed by light, is removed. The cylinder is
etched, producing image elements that vary in area but not in depth, so the number
of tones is limited. This method is used widely in packaging and textile printing. Pad Transfer Printing
This process is classified to the indirect recess printing process whose inked
image areas are transferred from engraved steel or plastic forms onto substrates
by means of a flexible silicon pad that adapts itself to the surface of irregularly
shaped objects.
Color Desktop Printer Technology
Conventional Gravure Plate
Variable Depth, Constant Area
Halftone Gravure Plate (Dalgian Method)
Variable Depth, Variable Area
Halftone Gravure Plate (Direct Method)
Constant Depth, Variable Area
FIGURE 1.22 Structure of gravure plate. Printing Process
Almost all rotogravure presses are designed for multicolor and web-fed gravure
presses (Figure 1.23). Rotogravure printing units consist of a printing cylinder,
an impression cylinder, and an inking system. Ink is applied to the printing
cylinder by an ink roll or spray, and the excess is removed by a doctor blade and
returned to the ink fountain. The impression cylinder is covered with a rubber
composition that presses the paper into contact with the ink in the tiny cells of
FIGURE 1.23 Multicolor web-fed gravure presses.
Introduction to Printing
the printing surface. Gravure inks are volatile and dry almost instantly. Hot air
dryers are used between printing units to speed up drying. Therefore, in color
printing each succeeding color is printed on a dry color, rather than on one that
is still wet as in letterpress and offset. For color printing, presses use photoelectric
cells for automatic register control. Cylinders are chromium plated for press runs
of a million or more. When the chromium starts to wear, it is stripped off and
the cylinder is rechromed. One disadvantage of gravure for publication printing
has been the inability to change pages on the cylinder. Wraparound printing
cylinder segments have been introduced that give gravure this added capability.
Sheet-fed gravure presses operate on the same rotary principle as rotogravure.
The preparatory work is identical. The image is etched flat on a flexible sheet of
copper, which is then clamped around the plate cylinder of the press. Sheet-fed
gravure is primarily used for short runs and press proofing. Because of the high
quality and plate-making expense, it is used for art, photographic reproductions,
and prestige printing such as annual reports. In packaging, sheet-fed gravure
presses are used for printing new packages for market testing.
Offset gravure has been used for printing wood grains and in packaging. In
this application, a converted flexographic press is used. The anilox roller is
replaced by a gravure cylinder and doctor blade for printing the image and the
plate cylinder of the flexographic press is covered with a solid rubber plate. Application
Gravure printing is used in packaging for quality color printing on transparent
and flexible films in which any cut-off length is possible by changing the size of
the printing cylinder. Gravure printing is also used for printing cartons, including
die-cutting and embossing, which can be done in-line on the press. Most long
run magazines and mail order catalogs are printed by gravure. Among the specialties printed by gravure are vinyl floor coverings, upholstery and other textile
materials, pressure-sensitive wall coverings, plastic laminates, imitation wood
grains, tax and postage stamps, and long run heat transfer patterns.
As shown in Figure 1.16, intaglio printing is divided into two categories: line
engraving and artistic engraving. Copperplate engraving and steel engraving
belong to line engraving, and drypoint engraving and mezzotint engraving belong
to artistic engraving. Drypoint engraving is produced by direct drawing using an
engraving needle without etching solution. Mezzotint is engraving produced on
a roughened copper plate. The highlight and midtone areas are hand-scraped to
reduce ink retention, and shadow areas are burnished to strengthen ink retention.
The printing ink used in intaglio printing is solidlike, whereas in gravure printing,
solvent-type ink is used. This is the distinct point in the difference between
gravure printing and intaglio printing.
Color Desktop Printer Technology Line Engraving Copperplate Engraving
Engraving is a highly skilled art in which lines of varying depth and width are
cut into metal plate with engraving tools. The plates are printed on a copper plate
press, which has a flat iron bed between two rollers. The plate, after warming to
soften the ink, is inked with stiff copper plate ink, and the excess is wiped off
the surface with wiping canvas and finally with the palm of the hand. The inked
plate is laid on the iron bed of the press and the paper (previously damped to
soften the fibers) is placed in position over it and backed with layers of printer’s
blanket. When rolling pressure is applied, the paper is forced into intimate contact
with the plate to receive the ink. Steel-Die Engraving
Steel-die engraving is line engraving in which the die is hand or machine cut, or
chemically etched to hold ink. The plate is inked so that all sub-surfaces are filled
with ink. Then the surface is wiped clean, leaving ink only in the depressed (or
sunken) areas of the plate. The paper is slightly moistened and forced against the
plate with tremendous pressure, drawing the ink from the depressed areas. This
produces the characteristic embossed surface, with a slightly indented impression
on the back of the paper. Mezzotint
In mezzotint engraving, small recesses are formed over the whole plate surface
with a serrated-edge rocking tool that throws up a burr of displaced metal.
Gradation of tone is obtained by using scrapers and burnishers to remove the burr
and reduce the depth of the recesses. This art process is sometimes used for color
printing, the different colored inks being applied locally to the plate for printing
at one impression. Application
Copper plates are used for short runs of one-time use (invitations and announcements). For longer or repeat runs such as letterheads, envelopes, greeting cards,
stamps, bank notes, and stock certificates, chromium-plated copper or steel plates
are used in a die-stamping press. Line engraving is used to produce plates for
some forms of maps and charts, for producing litho transfers, and for printing
invitation cards, visiting cards, and similar work.
Compared to other printing technology, through-printing has a characteristic of
an industrial art object. Practically, screen printing has been used for art prints,
posters, decalcomania transfers, greeting cards, menus, program covers, and wallpaper. Screen printing is important in the printing of textiles such as tablecloths,
shower curtains, and draperies. On the other hand, the remarkable development
Introduction to Printing
of the electronic industry in the 1970s depended on screen printing technology.
Screen-stencil technology is indispensable to large-scale integration (LSI) manufacture. Now, screen printing is a basic technology for integrated circuit (IC)
In 1907, Samual Simon of Manchester, England, got the idea for screen printing
from the traditional dying process of Yuzen-zome in Kyoto, Japan. The screen
printing process originated as a method of printing from stencils supported on
fabric stretched on a frame. Screens made with perforated metal and other materials are sometimes used.
Formerly known as silk screen, this method employs a porous screen of fine
silk, nylon, dacron, or stainless steel mounted on a frame. Figure 1.24 shows
microphotography of the screen fabric. A stencil is produced on the screen, either
manually or photomechanically, and the non-printing areas are protected by the
stencil. Printing is done on paper or other substrate under the screen by applying
ink with a paintlike consistency to the screen, spreading and forcing it through
the fine mesh openings with a rubber squeegee. At the same time, the squeegee
scratches out the excess ink from the surface of screen fabric. The ink pressed
through the open image areas of the screen form the inked image on the substrate.
This printing process is shown in Figure 1.25.
As the name suggests, through-printing is the method of obtaining an inked image
through the openings in a screen fabric. The basic concept of through printing is
simple and is based on the idea of a stencil. By taking a piece of paper, drawing
some outline or sketch of an object on it, and then cutting out the sketch, we can
get a stencil. Two important printing methods are stencil duplicating and screen
FIGURE 1.24 Micrograph of screen fabric.
Color Desktop Printer Technology
Screen Frame
Screen Fabric
Printing Ink
Printing Paper
FIGURE 1.25 Screen printing by hand.
Generally, stencil duplicating is performed by off-contact printing, and screen
printing is done by on-contact printing. In stencil duplicating, after painting,
spraying, or depositing ink on the stencil, ink passes through perforations in a
stencil master to form a printed image on a substrate. After the image is transferred, the frame must be hinged up and the stock carefully peeled away.
As shown in Figure 1.25, in screen printing, the screen is slightly raised away
from the printing material by small shims under the hinge and frame. With this
technique, the screen touches the paper stock only while the squeegee passes
over the screen. Once the image is transferred across the squeegee line, the screen
snaps back away from the substrate. Off-contact printing helps keep the press
sheet from sticking to the screen and usually prevents image smearing. Figure
1.26 shows the schematic diagram of ink transfer in the screen printing process.
1.5.3 PLATE
There are many methods of making plates for screen printing. As previously
mentioned, the screen consists of a porous material, and the printed image is
produced by blocking unwanted holes or pores of the screen. Early screen plates
were made manually by painting the image on silk fabric mounted on a wooden
frame. Masking materials were used to block out unwanted areas. Today, both
hand-cut stencils and photomechanical means are used. In the photomechanical
method the screen is coated with a light-sensitive emulsion; exposure is made
through a screened film.
Some screen printing is done by hand with very simple equipment consisting of
a table, screen frame, and squeegee, as shown in Figure 1.25. Most commercial
screen printing, however, is done on power-operated presses. There are both rollfed and sheet-fed presses, with hot air dryers, which run at speeds up to 400 feet
per minute or more than 5000 impressions per hour.
Introduction to Printing
FIGURE 1.26 Schematic diagram of ink transfer in screen printing. Ink profile in screen
opening at corresponding position as indicated.
There are two types of power-operated presses. One type uses flat screens,
which require an intermittent motion as each screen is printed. Butts and overlaps
require close register, which limits running speed. The latest type uses rotary screens
with the squeegee mounted inside the cylinder and the ink pumped in automatically
(Figure 1.27). These presses are continuous running, are fast, and print continuous
patterns with little difficulty. The amount of ink applied by screen printing is far
greater than in letterpress, lithography, or gravure, which accounts for some of the
Rotary Screen
FIGURE 1.27 Diagram of a rotary screen printing unit.
Printing Ink
Color Desktop Printer Technology
unusual effects in screen printing. Because of the heavy ink film, the sheets must
be racked separately until dry or passed through a heated tunnel or drier before
they can be stacked safely without smudging or set-off. Ultraviolet curing ink has
effective drying and is helping to promote greater use of screen printing.
Versatility is the principal advantage of screen printing. Screen printing prints on
almost any surface, and both line and halftone work can be printed. Any surface
can be printed such as wood, glass, metal, plastic, fabric, and cork, in any shape
or design, any thickness, and any size. In advertising, screen printing is used for
banners, decals, posters, 24-sheet billboards, car cards, counter displays, menu
covers, and other items. Heavy paperboards can be printed, eliminating costly
mounting. Wallpapers and draperies are printed because of the depth of colors
afforded, especially in the short run custom designs of interior decorators. There
are many other specialty uses for screen printing, such as decorating melamine
plastic sheets before lamination and the printing of electronic circuit boards.
Image Quality of
Printed Text and Images
Jon S. Arney
The Meanings of Image Quality ...............................................................32
2.1.1 The Image Quality Circle..............................................................32
2.1.2 Image Quality Metrics...................................................................32
2.1.3 Image Quality Modeling ...............................................................34
The Printing of Text ..................................................................................35
2.2.1 Printer Resolving Power and Addressability ................................35
2.2.2 Ink Density ....................................................................................38
Color Inks and Color Mixing....................................................................41
2.3.1 Primary Colors...............................................................................42
2.3.2 Color Densitometry .......................................................................44
2.3.3 Colorimetry Coordinates ...............................................................44
The Printing of Pictorial Images...............................................................45
2.4.1 Halftone Printing ...........................................................................47
2.4.2 Color Halftone Printing.................................................................49
2.4.3 Halftone Color Calculations..........................................................51
Tone and Color Control.............................................................................54
2.5.1 The Dot-Gain Phenomenon...........................................................54
2.5.2 Dot-Gain Mechanisms...................................................................55
2.5.3 The Yule–Nielsen Correction ........................................................56
2.5.4 Yule–Nielsen Color Calibration ....................................................60
2.5.5 Mechanistic Models and Dot Gain ...............................................61
Making Halftone Images...........................................................................64
2.6.1 Pre-Press Process Photography .....................................................64
2.6.2 The Digital Halftoning Mask ........................................................67
2.6.3 The Noise Distribution Technique: An Alternative to
the Mask Technique ......................................................................70
2.6.4 Classification of Digital Halftoning ..............................................73
Image Quality and the Noise Power Spectrum ........................................75
2.7.1 Fourier Analysis of Noise Power ..................................................76
2.7.2 Noise Power and the Human Visual System ................................76
2.7.3 Objective Measures of Granularity Constants ..............................79
Color Desktop Printer Technology
2.7.4 Granularity and the Printing Device .............................................81
References ...........................................................................................................82
This chapter focuses primarily on the characteristics of printer systems that have
a significant impact on the visual quality of printed images. Characteristics such
as the absorption spectrum of printing inks, the gamma of photoconductors, and
satellite drops in inkjet are examples of system characteristics that impact image
quality. The relationship between system characteristics of this kind and the
ultimate quality of a printed image is of considerable practical importance, but
it is not always easy to quantify the connection. The term image quality is used
in many different ways, so before considering printers and printer characteristics,
it is helpful first to consider the various ways the term image quality is used.
Many different technologies have been developed into commercially successful
printing systems. The traditional printing technologies described in Chapter 1
and the newer digital techniques described in Part II of this text differ significantly
in the way in which colorant is applied to paper. However, all these different
technologies have two attributes in common. They all produce printed images of
high quality, and they do so at a low price. Many other technologies for putting
colorant on paper have been developed and described in the technical literature,
but only those capable of quality and economy are commercially successful. Thus,
both quality and economy are topics essential for anyone in the printing industry
to understand.
The meaning of the term economy is easy to understand quantitatively, but
image quality is a more difficult term to express in a quantitative way. The difficulty
is the subjective way in which customers use and describe the quality of printed
images. Nevertheless, if a printing process does not provide the customer with
enough quality, regardless of the ability to understand and quantify it, the printing
process will not survive commercially. The overall quality of an image is therefore
measured indirectly by the commercial success of the printing process.
To predict image quality, and thus commercial success, other parameters are
measured that are believed to serve as useful quality metrics. The relationship
between overall image quality and these other kinds of metrics can be represented
as shown in the Engeldrum Image Quality Circle shown in Figure 2.1. The
different types of image quality metrics are represented by the boxes, and the
relationships among the types of metrics are represented by the ovals.
The overall quality of an image, represented by box (A) at the top of Figure 2.1,
is generally believed to be made up of a group of perceptual attributes represented
Image Quality of Printed Text and Images
Image Quality
(the “nesses”)
image quality
FIGURE 2.1 The image quality circle. (Modified from P.G. Engeldrum, Psychometric
Scaling: A Toolkit for Imaging Systems Development, Imcotek Press, Winchester, MA, 2000.)
by box (B) on the left side. Examples of these component attributes are colorfulness,
sharpness, graininess, saturation, and crispness. These individual attributes, sometimes called the “nesses,” can be quantified experimentally with well-known techniques of psychophysical testing.1 Thus, these attributes can be expressed as unambiguous, quantitatively measurable metrics. For example, a group of observers
might be asked to judge the relative degree of graininess, colorfulness, or sharpness
of a set of test images. Appropriately designed testing can lead to numerical scales
of graininess, colorfulness, or any of the other “nesses.”1
The set of test images can also be characterized with a variety of instrumental
measurements. These measurements provide the physical image parameters
shown in box (C) at the bottom of Figure 2.1. Examples might include the
reflectance spectrum, the noise power spectrum, and the image histogram. From
these functions, individual metrics such as L*, a*, contrast, resolution, and granularity can be extracted. These physical image parameters are often used as
predictors of the “nesses” in Figure 2.1.
Box (D) on the right side of Figure 2.1 represents the technology variables,
also called the printer device parameters. These are the parameters that describe
the operation of the printing device, and they ultimately control the characteristics of the image in the box (C) at the bottom of Figure 2.1. Examples of
device variables and functions containing device variables include the device
modulation transfer function (MTF), the paper roughness, the colorant concentration, and the device gamma. These are variables over which the printer
manufacturer and the printing service company have some control, and it is
common to search for the optimum values of the variables that will ensure the
highest quality of printed image. The relationship between the variables of the
printing device (box (D) of Figure 2.1) and the overall quality of the image
Color Desktop Printer Technology
(box (D) of Figure 2.1) and the relationship between any two boxes in Figure
2.1 are referred to as image quality models. The different types of image quality
models are represented as ovals in Figure 2.1.
Each oval in Figure 2.1 represents a different type of image quality model for
estimating parameters in one box based on the parameters of another box. Each
type of model in Figure 2.1 has its usefulness.
Device-Dependent Models (D→B): Modeling the relationship between the
nesses and the technology variables has been called device-independent modeling.2,3
This type of modeling is particularly important when a printer must be adjusted to
provide the highest print quality in the shortest possible time. A powerful technique
for constructing such a model is to apply an empirical technique called statistical
experimental design4 in which all of the relevant variables of the printing system
(ink density, photoconductor gamma, paper type, print speed, toner fusing temperature, etc.) are varied in all relevant combinations. For each combination of variables, a print is generated and evaluated by observers in a psychophysical experiment to determine a set of perceptual metrics (sharpness, colorfulness, graininess,
etc.). Then a statistical optimization is carried out to determine the values of the
printer variables that produce the best set of perceptual values, or nesses. This sort
of device-dependent modeling is a powerful way to calibrate a printer for the best
quality in the shortest amount of time. However, it does not provide much guidance
in understanding the underlying nature of image quality or how it relates mechanistically to the printing process. Thus, other kinds of image quality modeling are
also important.
System Models (D→C): The properties of the printing system directly control
the properties of the image it produces. Thus, for example, a change in the gamma
of a photoconductor in a laser printer can change the resolution of an image printed
with the printer. It is important for product development specialists to understand
the relationship between the technology variables and the properties of the printed
images. Models relating these metrics are also useful for trouble-shooting difficulties with a printing system.
Visual Models (C→B): Images can be characterized by a variety of instrumental techniques. Common examples are reflectance spectroscopy, colorimetry,
histogram analysis, microdensitometry, gloss, and goniophotometry. Experimental
metrics derived from such measurements are often called objective quality metrics
and are often correlated with the results of psychophysical analysis. This kind of
modeling does not depend on the properties of an imaging device and involves only
the measurable properties of the image. For this reason, models of this kind are
sometimes called “device-independent models.”4
Combining a visual model with a system model provides a quantitative,
mechanistic link between the perceptual nesses of an image and the underlying
technology variables that control the image-producing device. The advantages of
Image Quality of Printed Text and Images
such a two-step model are clear, but achieving good, predictive models of this
sort is challenging and time consuming.
Image Quality Models (B→A): Modeling the overall quality of an image
based on the collection of perceptual nesses in box (B) of Figure 2.1 is a challenge
that is still a major topic of research among image scientists, experimental
psychologists, and market analysts. As yet, no generally satisfactory solution has
been developed, and empirical trial-and-error is still the principal technique
The remainder of this chapter focuses on the technology variables (Figure
2.1, box (D)) of printing systems and their impact on the so-called objective
image quality metrics of printed images (Figure 2.1, box (C)).
Image quality requirements for document images containing only alphanumeric
characters are fewer than the requirements for pictorial image printing. However,
achieving a high level of text quality is not a trivial matter, and printer characteristics that influence text quality also play key roles in pictorial image quality.
Moreover, changes in the printing system that favor text quality often disfavor
pictorial quality, leading to a need for text/pictorial tradeoff. Thus, before exploring color image quality, it is essential first to examine text image quality.
In general, text characters must have high density relative to the paper and must
have sharp edges, serifs and other fine details must be reproduced repeatably, gloss
differences between printed and not-printed regions should be minimized, and there
should be no stray colorant between the characters. These requirements seem selfevident, but the achievement of these requirements is not a trivial matter. Moreover,
the competitive drive toward higher image quality has led to an overall increase in
consumer expectations for text image quality. During the 1970s, the benchmark for
office text quality was the impact typewriter. Improvements in ink ribbons and the
introduction of the electric typewriter raised the bar for text image quality, and the
best quality achievable with an electric typewriter during the 1970s became known
as a letter-quality text image. The development of desktop computing and printing
during the 1980s raised the bar even higher. The early printers were dot-matrix,
impact printers capable of letter-quality printing, but the introduction of inkjet and
laser printers quickly demonstrated the ability to produce letter-quality prints. By
the end of the 1980s, desktop printers were available for routine office use and
produced prints that were nearly indistinguishable from many offset lithographic
presses. The term letter-quality is no longer used, and offset-litho quality is now
the benchmark for desktop document printing.
Marketing literature for printers and printing systems often describes resolution
in terms of dots per inch (dpi) or dots per millimeter (dpm). However, this is
often misleading. The dpi and dpm metrics are really the addressability of the
Color Desktop Printer Technology
FIGURE 2.2 USAF (left) and NBS (right) test charts.
system. This is the maximum number of locations per unit length on the
substrate paper where the printer can deliver a unit of colorant (dot). Although
addressability is an important metric for a printing system, it is not equivalent
to resolution.
Resolution is generally measured quantitatively by an index of resolving
power. This is done by printing test characters or bar patterns of smaller and
smaller sizes. The smallest pattern size with bars that can be unambiguously
distinguished by someone using an appropriate optical magnification device is
used as the index of resolving power.5 Numerous test patterns and test procedures
have been published for quantifying resolution, and Figure 2.2 illustrates two
very common ones. The United States Air Force (USAF) resolution test chart,
shown on the left, was originally developed during the 1940s to characterize
aerial reconnaissance film,6 and the National Bureau of Standards (NBS) Resolution Test Chart was developed to measure the resolution of optical microscopes.7
These test charts and others have been adapted to the characterization of printing
systems, and a survey of current technical literature is recommended for anyone
involved in this kind of testing.1,5,8 Test charts such as these are available commercially, both for image capture (cameras, microscopes, scanners, etc.) and for
image output. Digital files with resolution charts, for example, are useful for
characterizing the resolution of digital printers.
Printer addressability certainly has an impact on printer resolution. However,
as illustrated in Figure 2.3, addressability is not the only factor controlling
resolution. The grid of addressable points in Figure 2.3 is illustrated with simulated ink dots of one half, one, two, and four times the distance between addressable locations, and it is clear that large dots, or dots that spread significantly on
the printed paper, will not produce images with the same resolution as can be
achieved with small dots. Thus, dot quality is as important to printed image
resolution as is the dpi addressability of a printing system.
Image Quality of Printed Text and Images
FIGURE 2.3 Grid of addressable points with dots at consecutive locations. Dots are
illustrated with diameters that are half, one, two, and four times the distance between
addressable points.
FIGURE 2.4 Examples of different dot shapes produced by different inks, papers, and
printing technologies. (a) 300-dpi electrophotographic laser printer on plain paper. (b)
150-dpi offset lithography on magazine quality paper. (c) 300-dpi inkjet drops forming 4
× 4 dot clusters at 37.5 lines per inch on plain paper.
Printing devices also can have dpi addressability that is different in the
horizontal and vertical directions, and ink dots can be of many shapes other than
round. Moreover, the edges of dots are not always clearly defined, as illustrated
in Figure 2.4. These kinds of effects are quite important both to image resolution
and to pictorial color reproduction. This is the reason for both horizontal and
vertical line patterns in the test charts illustrated in Figure 2.2.
In some technologies, the resolution of the image can be higher than the dots
per inch of the printing system. This is true for the traditional photographic
process for making halftone printing plates for offset lithographic printing, as
illustrated in Figure 2.5 (see also Section 2.4.1). The photographic process controls not only the sizes of halftone dots but also the shapes. This results in partial
dots that align along sharp lines, thus producing images with a higher resolution
than the nominal dpi dot spacing would suggest.9
Color Desktop Printer Technology
FIGURE 2.5 Photographic process halftoning produces printing dots with higher resolution than the dot pitch. (Courtesy of Frank Cost.)
FIGURE 2.6 Illustration of line jags produced by discrete dots on a printing device. The
right-hand image is identical but with significant noise and blur added in a simulated
printing process.
As illustrated in Figure 2.4, there may be a variety of technological factors
that spread ink dots and lead to microscopic blurring of the image structure.
Although this can only decrease resolution of an image, it is not always true that
it decreases image quality. There are cases, as will be described subsequently, in
which micro-blurring can lead to better color reproduction. As shown in Figure
2.6, some amount of blurring can reduce the visual impact of a digital printing
artifact, sometimes called line jags.
For a printing process to produce high-quality text, the text characters must be
very dark. Mechanical typewriters of the mid 20th century print text that is much
lighter than the text produced by an offset web press. The distinction between
letter-quality text and offset-quality text is primarily the darkness of the printed
characters. The lightness and darkness of a printed image can be measured
accurately by comparing the printed image to a photographic step tablet, as
illustrated in Figure 2.7. A photographic step tablet is a series of accurately
produced steps of image density and is readily available at camera and photographic supply stores. The human eye is a very good null detector. This means
Image Quality of Printed Text and Images
Printed Text Quality
The quality of printed text
depends in part on the
density of the ink on the
paper. Density is measured
with a densitometer.
with ad
t on
Photographic Step Tablet
FIGURE 2.7 The visual measurement of print density.
Density: 1.37
FIGURE 2.8 Densitometer showing light sources (a) at 45 degrees, a light detector (b)
at zero degrees, a filter (c), and an output display (d).
that even though we are not able to judge the density of an image accurately by
simply looking at it, we are quite capable of judging whether or not two images
placed side by side are the same or different. Thus, by masking the step tablet
and the printed image, as illustrated in Figure 2.7, it is easy to match a gray level
of the photographic step tablet with the printed image.
Several companies manufacture instruments for measuring image density.
These instruments, called densitometers, range in price from a few hundred to
several thousand dollars. Figure 2.8 illustrates a typical densitometer.
Light of brightness Io is delivered to the paper from light sources typically
placed at 45 degrees from the vertical. The electronic light detector detects the
reflected light intensity, I, and the electronics of the system combines the values
of I and Io to calculate image density, D, according to Equation 2.1 and Equation
2.2. These equations are the definitions of reflectance, R, and reflection density,
D. The density value is then displayed on the output screen.
Color Desktop Printer Technology
R = I/Io
D = –log(R)
The major advantages of a densitometer over visual matching with a step tablet
are convenience and rapidity of operation. Both techniques provide comparable
accuracy and precision when measuring the density of black and white images.
Color image density is much more difficult to measure by the visual matching
technique and requires many more reference samples and more practice to provide
accurate results. Nevertheless, books and charts are readily available and in extensive use for performing visual color matching, particularly in situations in which
an electronic instrument is impractical.10 Forensic pathologists, biologists, and
geologists often measure color by matching to standard color charts.
A color densitometer is not much more complex than a monochrome densitometer. In general, it requires only the addition of red, green, and blue filters.
A monochrome densitometer typically incorporates a filter at location (c) in
Figure 2.8 to make the instrument respond to the spectrum of light in same way
as the average human observer. By adding a red, green, and blue filter, the
densitometer can make accurate and rapid measurements of color density as well
as black and white density. Inexpensive color densitometers make all of the
necessary measurements rapidly and automatically and display four density values as illustrated in Figure 2.9.
The density value labeled V in Figure 2.9 is the visual density used in
monochrome measurements and represents overall impression of lightness/darkness. The R, G, and B density values are the values measured through red, green,
and blue filters. These values correlate approximately with our visual impression
of color. How well they correlate with our impression of color depends on the
specific filters selected for the R, G, and B measurements, as described below.
Often, however, a densitometer is used to measure the amount of cyan, magenta,
and yellow inks printed on paper rather than the color that appears to the eye.
The red, green, and blue density values are therefore often displayed as shown
on the right in Figure 2.9 because the red density correlates strongly with the
RGB Display
CMY Display
R: 0.15
C: 0.15
G: 1.28
M: 1.28
B: 0.08
Y: 0.08
V: 1.37
V: 1.37
FIGURE 2.9 Output displays often found on color densitometers.
Image Quality of Printed Text and Images
G ra y
G ra y
FIGURE 2.10 The Munsell color wheel.
cyan ink, the green density correlates strongly with the magenta ink, and the blue
density correlates strongly with the yellow ink.
Much of the apparent complexity of color printing occurs because of confusion
and inconsistencies in the way colors are named. The basic concepts are actually
not difficult to understand. Many different schemes are used to name colors, and
each is useful in a particular area of application, as described in several introductory texts on color science and technology.10,11 One of best known schemes
for naming colors is the Munsell system, illustrated in Figure 2.10. This system
is based on the theory of trichromatic color vision, which says that humans have
three kinds of color sensors (cone receptors) in the retina of the eye sensitive to
three regions of the spectrum of light. As a result of having three, and only three,
kinds of color sensors, we experience all colors in terms of three characteristics.
In the Munsell system, these characteristics are called hue, value, and chroma.10
The hue is what we commonly call color and involves familiar color names such
as red, orange, and yellow. The Munsell system uses five color names to describe
hue: red, yellow, green, blue, and purple. These five hues can be arranged in a
hue circle, and any hue (or “color”) that human beings can see falls somewhere
in the circle. For example, orange is a hue that falls between red and yellow.
Turquoise falls between blue and green.
Color value is analogous to our gray scale in Figure 2.7. Any given hue may
be a light hue or a dark hue. A light hue of red, for example, is often called pink.
In Figure 2.10, point a is a red of medium lightness, c is the same red hue but
of higher lightness (pink), and d is a green hue with the same lightness as a.
The third attribute of color is called chroma. This is the degree of purity of
the color. The red at point a in Figure 2.10 is a low-chroma red such as brick
red, but point b is a higher-chroma red such as fire engine red. The decrease in
Color Desktop Printer Technology
purity leads all colors toward the neutral gray scale in the center of Figure 2.10,
often called the a-chromatic colors, or colors without hue.
Figure 2.10 and the trichromatic theory of color vision tell us that all colors
can be described as if they occupied a location in the three-dimensional space
defined by the hue circle, the value axis, and the distance from the axis called
the chroma. All colors that human beings are able to see occupy unique locations
in this three-dimensional space, and the total volume of space occupied by all
visible colors is called the total color gamut of human vision.
A consequence of the tri-chromatic theory of human vision is that it is possible
to produce all color hues by making mixtures with only three well-selected
colors. These well-selected colors are generally called primary colors. This is
a somewhat ambiguous term, but it is commonly used by both experts and
laypersons. If you ask the average person to name the three primary colors,
they often will reply red, blue, and yellow. This is entirely correct within the
crayon color scheme of naming colors. However, a much more useful system
of color nomenclature is in use by scientists and engineers in which the primary
colors are called red, green, and blue. Light sources that produce red, green,
and blue light have been found, by trial and error as well as by theory, to have
the ability of mixing to make the greatest number of visible colors.10,11 The
higher the chroma of the three light sources, the more colors one can make.
Blue, green, and red monochromatic light sources (light of only one wavelength)
at 400 nm, 525 nm, and 700 nm, for example, can mix to produce nearly all
colors the average observer is able to see. More practical light sources for red,
green, and blue are used in television sets and computer monitors. It should be
clear that the term primary color is only a general concept and not a rigorously
defined set of colors.
Printing processes also make use of red, green, and blue primary light to
control colors. This is done by starting with the unprinted paper used as the
printing substrate. The unprinted paper reflects red, green, and blue light toward
our eyes so we see the mixture of the light and call it “white.” To control the
amount of red light that reflects from the paper, we use an ink that absorbs
only the red light. The red light is still our primary color of light, but we control
this color of light with an ink that absorbs the red light. When we look at such
an ink, the color we do not see is red. This ink looks exactly unlike red, and
the traditional name that printers give this hue is cyan, as illustrated in Table
2.1. Similarly, ink that absorbs green light is used to control green light in the
printed images. This ink is called magenta. A yellow ink is used to absorb the
blue primary of light.
Red, green, and blue are our primary hues of light, sometimes called additive
primaries. The corresponding inks (cyan, magenta, and yellow) are often called
printer primaries, subtractive primaries, or primary inks. Collectively, these six
hues form a hue circle similar to the Munsell hue circle, as illustrated in Figure
Image Quality of Printed Text and Images
The Primary Colors and Primary Inks
The Light Absorbed by the Ink
Name Given to the Hue of Ink Color
Hue Circle
FIGURE 2.11 The descriptive hue circle used by printers, engineers, and scientists. Red,
green, and blue are additive primary colors. Cyan, magenta, and yellow are printer primary
2.11. By convention, red is always on the lower right of this hue circle, and green
is at the top. Unlike the Munsell system, the printer hue circle is not used as the
basis of an exact color naming system. Rather, the printer hue circle is a tool
used extensively in qualitative descriptions of color printing. It is a convenient
map for orienting oneself in color space, so it is well worth the effort to memorize
the printer hue circle. As an example, one can refer to the printer hue circle and
say instantly that the color blue is made by printing a combination of cyan and
magenta inks, that combining red and green light will make yellow, and that a
mix of cyan and yellow ink make green.
It is easy to understand why many people say the three primary colors are
red, blue, and yellow. The color called cyan in the printer hue circle looks to
most people like a sky blue, the printer blue in common crayon terms is called
violet, and magenta is often called a bubble gum shade of red. Therefore, the
terms red, blue, and yellow are simply the common crayon names for the printer
primary inks called magenta, cyan, and yellow. It is also easy to see why words
are inadequate as a means of naming colors. Color measuring instruments that
report numbers are much less ambiguous than words.
Color Desktop Printer Technology
The tri-chromatic nature of human vision provides a rationale for why color
densitometers measure density with red, green, and blue filters, as described
previously in Figure 2.9. The printer hue circle helps us remember that a red filter
is used in a densitometer to measure the cyan ink, as indicated in Figure 2.9. A
high red density means that much red light is absorbed, which means the image
appears cyan. Red density is useful for measuring the amount of cyan ink on a
printed page. The term cyan density is often used instead of red density because
a high red density appears cyan to the eye. Such terminology is logically inconsistent, but it is in common use. Similarly, green and blue density values are used
to measure magenta and yellow inks. Because printers often are concerned with
the amount of ink used in printing, they often rely on densitometry as a means
of monitoring the ink printed on the page.
Referring to the diagram of the densitometer in Figure 2.8, one uses red, green,
and blue (RGB) filters to measure density that correlates well with the amount of
cyan, magenta, and yellow (CMY) inks printed on a page. However, if one is more
concerned with the visual appearance of the printed color than with the amount
of ink used, then a special set of filters must be used instead of the RGB filters.
These special filters adjust the spectral response of the instrument so it correlates
exactly with the spectral responses of the three cone sensors in the average human
eye. The instrument then translates the density measurements into a special color
space that correlates well with the visual appearance of the printed image. Such
an instrument is called a colorimeter instead of a densitometer.
A colorimeter looks just like an RGB densitometer. However, the colorimeter
has special filters so that instead of measuring three density values (red, green,
and blue), it measures three numbers called color coordinates [X,Y,Z]. These
color coordinates correlate with our visual perception of colors. No two colors
will have the same [X,Y,Z] values, and two colors that look identical will have
identical [X,Y,Z] values. This is not the case with RGB densitometers. If one
is concerned with the behavior of inks on paper, a densitometer is the preferred
instrument, but if one is concerned with the measurement of the visual quality
of the printed colors, a colorimeter is the preferred instrument. Because the
two instruments use two different kinds of filters, it is not possible to convert
the results of one instrument into the other. RGB density values cannot be
converted to exact [X,Y,Z] values, and [X,Y,Z] values cannot be converted to
exact RGB density values. An approximate correlation can be made between
the two, but intrinsic mathematical ambiguities limit the accuracy of such a
The three numbers produced as output from a colorimeter map uniquely and
unambiguously onto our perception of color. However, there are several color
coordinate systems that do this, and several are in common use. The [X,Y,Z]
Image Quality of Printed Text and Images
numbers are commonly used by most commercial colorimeters. These three
numbers are defined according to the integral functions shown below. The term
P(λ) is the spectral energy distribution of the light under which the printed image
is viewed, and R(λ) is the spectral reflectance of the printed image.
100 =
∫ P (γ ) ⋅ y (γ ) d γ
X= K⋅
∫ R (γ ) ⋅ P (γ ) ⋅ x (γ ) d γ
Y = K⋅
∫ R (γ ) ⋅ P (γ ) ⋅ y (γ ) d γ
Z = K⋅
∫ R (γ ) ⋅ P (γ ) ⋅ z (γ ) d γ
The terms x ( γ ) , y ( γ ) , z ( γ ) are called color-matching functions and are functions that are linearly related to the spectral response of the three color sensors
in the human eye.
With the correct filters, the colorimetric instrument automatically performs
the calculations indicated in Equation 2.3 through Equation 2.6, and one only
has to note the X, Y, and Z color values in just the same way one would note the
RGB density values when using an ordinary densitometer. The [X,Y,Z] values
are more like reflectance values, R in Equation 2.1, than density values. A high
value of Y, for example, means a low green density and a high reflectance of
green light. Indeed, a densitometer that measures the visual density, V in Figure
2.9, actually measures Y and calculates V = –log(Y).
Most colorimeters can present the color coordinates [X,Y,Z] and also other
useful color coordinates. Many commercial instruments convert the [X,Y,Z] coordinates into [L*,a*,b*] coordinates, which are more useful for describing the color
difference we see between two printed colors. The reader is referred to texts on
colorimetry for more information about color theory and colorimetry.10,11
The quality of a pictorial image depends on the quality of the reproduction of
gray and color tones. The reproduction of gray and color tones is influenced by
the same printing factors that influence text printing. However, the optimum
Color Desktop Printer Technology
values of these factors for text printing and pictorial printing are generally not
the same. Factors that produce high-resolution, high-density letters on clean,
white paper often lead to pictorial images of poor tone and color quality.12,13
Most printing technologies are inherently incapable of printing gray tones.
They do well when printing a text letter at full density or when printing text at
all. The 18th-century printing process used by Benjamin Franklin, for example,
employed a relief printing plate with raised letters that pressed ink into paper.
Text quality was good, but no gray tones could be produced. This was the case
for all printing processes until the development, during the 19th century, of the
photographic process for printing film negatives onto photographic print paper.
The photographic process is intrinsically able to print a continuous range of gray
and is thus often referred to as a continuous tone process. Mechanical printing
processes, including those used for desktop digital printing, are intrinsically bilevel processes. They either deliver ink at a selected location or they do not. With
few exceptions, they are unable to print continuous tones.
To simulate the appearance of gray tones, printers have long used a variety
of techniques for fine-scale distribution of ink, as illustrated in Figure 2.12. Many
different techniques for simulating tone have been developed over the past several
centuries.14 The most common technique currently in use for both commercial
and desktop printing is the halftone technique, illustrated in Figure 2.13.
By converting a continuous tone image into a halftone image, as illustrated
in Figure 2.13, the printing process does what it intrinsically does best. It prints
solid ink on white paper, simulating gray tone by the size or closeness of the
printed dots. It would seem that the ability to print crisp, well-formed dots,
would depend on the same factors as those for printing text. Indeed, if the
halftone dots are as large as in the illustration of Figure 2.13, this is so. However,
high-quality halftone images require the printing of very small halftone dots.
The continuous tone illustration in Figure 2.13(A) is actually not a continuous
FIGURE 2.12 A print by William Downing, printer of Juvenal (tr. to B. Holyday)
Decimus Junius Juvenalis, (Oxford, 1673). (Used with permission from the Melbert B.
Cary, Jr. Graphic Arts Collection, Rochester Institute of Technology.)
Image Quality of Printed Text and Images
Continuous Tone Photograph
Halftone Copy
FIGURE 2.13 Illustration of a continuous tone image and a corresponding halftone image.
(Courtesy of Frank Cost.)
tone photograph at all. It is a halftone printed at 150 halftone dots per linear
inch. This can be seen by examining Figure 2.13(A) with a strong magnifying
glass (at least 6× recommended). It is clear that high quality halftone images
require that dots be printed well and also that they be printed much smaller
than the text characters used in document printing. The need to print very small
dots makes the requirements for halftone pictorial quality different from the
requirements for text quality.
A halftone image can be indistinguishable from a continuous tone photograph if
the halftone dots are smaller than the eye can resolve. In this case, the eye
perceives only the average reflectance of light from the image, as illustrated in
Figure 2.14. Each ray of light from the continuous tone image is equally bright,
but the halftone reflects rays of two types. Some of the rays are reflected from
the paper and are very bright. Some of the rays are reflected from the ink dots
and are very dim. If the dots are small enough, the observer sees only the average
brightness from the combined ink and paper rays of reflected light.
Quantitatively, the lightness of the image can be described in terms of its
reflectance, which was defined in Equation 2.1 as the ratio of the reflected light
to the incident light, R = I/Io. The continuous tone image on the left side of Figure
2.14 has a single reflectance value, R. The halftone image on the right has an
average reflectance, R, given by the weighted average of the ink and paper
reflectance, as illustrated in Equation 2.7, where F is the fraction of the area of
Color Desktop Printer Technology
FIGURE 2.14 Viewing a continuous tone photograph and a halftone print of the same
average reflectance.
the image that is covered by dots and (1 – F) is the area of the bare paper between
the dots.
R = F·Ri + (1 – F)·Rp
The human visual system responds approximately logarithmically to changes
in reflectance. Thus, density, which is the negative log of reflectance (see Equation
2.2), is often preferred when evaluating the appearance of a printed gray scale.
In terms of density, F can be solved from Equation 2.7 as shown in Equation 2.8.
10 − Dp − 10 − D
10 − Dp − 10 − Di
where Di is the density of the ink, Dp is the density of the paper, and D is the
average density of the image.
Densitometer readings are often calibrated in such a way that the paper is
assigned the relative density value of 0.00. In that case, Equation 2.9 is the
relationship between the density of the image and the area fraction of the image
that is covered by dots.
1 − 10 − D
1 − 10 − Di
Equation 2.7 through Equation 2.9 are each variations of what is often called
the Murray–Davies equation.11,13 The Murray–Davies equation shows how one
can control a printing device to produce any desired gray level by controlling the
dot area fraction, F. One can calibrate the printing system by measuring the ink
and paper reflectance values and then control the output gray level by controlling
the dot area fraction, F, sent to the printer. In practice, when a user selects the
Image Quality of Printed Text and Images
desired gray level, R, the software that controls the printer translates the value
of R into the dot area command, F, that is sent to the printer. This software process
is called raster image processing, or RIP, and we often refer to the process as
ripping the image.15
The Murray–Davies equation is a very simple description of halftone imaging,
and, not surprisingly, it is a good description only for ideal halftones. Real
halftones suffer from a variety of physical and optical effects that render Equation
2.7 through Equation 2.9 only poor approximations of printed images. Failure of
the Murray–Davies equation, regardless of cause, is often called dot gain.12,13,15
This is described in more detail in Section 2.5.
The three primary printing inks cyan, magenta, and yellow are used to print color
images. Mixtures of these three inks can produce any desired hue over a wide
range of shades. There are two ways one can mix inks in a printed image, and
they produce significantly different results. Color photography and color photographic printing involve the physical mixing of cyan, magenta, and yellow dyes
within a single gelatin layer. The mechanical processes of printing ink on paper
involves printing three overlapping halftone patterns, one for each of the primary
inks. This is true for the processes of printing as well as for digital desktop
printing. The way in which the separate dot patterns for cyan, magenta, and yellow
are combined on the printed paper has a significant effect on the quality of the
printed image. One major phenomenon that must be considered is the moiré effect.
The Moiré Effect: When more than one halftone pattern is printed on a paper,
the way the two patterns are arranged has a major impact on color and tone
reproduction. For example, suppose the halftone pattern shown on the left side of
Figure 2.15 is to be printed along with a second, identical pattern. If we print the
second pattern so that the second set of dots falls exactly between the first set, we
have the much darker image on the right side of Figure 2.15. However, if the second
set of dots is printed in perfect registration with the first set, the result is indistinguishable from the original pattern on the left. It is clear from this illustration that
the location of the dots must be controlled appropriately to control tone and color
FIGURE 2.15 Dots on dots and dots beside dots with perfectly black ink.
Color Desktop Printer Technology
FIGURE 2.16 Illustration of moiré caused by superimposing two halftone patterns that
differ by 2% in their LPI.
The example in Figure 2.15 illustrates the extreme difference in appearance
between printing dots-on-dots and dots-beside-dots. To do either, the registration
of the dot patterns must be in perfect control. Precision dot placement of this
kind is not achievable in most printing systems for a variety of reasons, including
the intrinsic elasticity of paper itself. For example, if the paper stretches a small
amount between printing the first and second dot pattern, a very severe nonuniformity can result, as illustrated in Figure 2.16. This effect is called moiré.
With 150-lines-per-inch (LPI) halftones, this effect can be very severe and objectionable, with less than a percent difference in the halftone LPI when attempting
to register two halftone patterns exactly. For this reason, it is generally impractical
to attempt to print either the dot-on-dot or the dot-between-dot patterns.
Another example of moiré is observed when two identical dot patterns are
superimposed with a rotation angle between them, as illustrated in Figure 2.17.
FIGURE 2.17 The moiré effect and screen rotation.
0, 30°, 60°
Image Quality of Printed Text and Images
Black & White
CMY Color
CMYK Color
M 75°
M 75°
K 45°
K 45°
C 15°
C 15°
Y 0°
Y 0°
FIGURE 2.18 Commonly used angles for halftone printing.
The rotation angle is called the screen angle, a term originating from the old
photographic process of using a halftone screen to convert photographic images
into halftone images.13–15 The greater the screen angle, the smaller the moiré
pattern is; thus, to minimize the moiré effect, halftone patterns are superimposed
with a large screen angle.
Three patterns are required to print color images, so a typical technique is to
print the yellow at zero degrees, the cyan at 30 degrees, and the magenta at 60
degrees, as illustrated in Figure 2.18. The resulting pattern of rosettes, illustrated
in the lower right configuration of Figure 2.17, is too small for the average
observer to see for 150 LPI halftone patterns at a typical reading distance of 35
cm. If black ink is printed in addition to the three primary inks, the typical scheme
places yellow at zero degrees, cyan at 15 degrees, black to 45 degrees (30 degrees
from cyan), and magenta at 75 degrees (30 degrees from black and 15 degrees
from yellow). Using 15 degrees between yellow and cyan and between yellow
and magenta would be expected to lead to undesirable moiré, but the human eye
is less sensitive to yellow patterns or moiré involving yellow.12
As illustrated in Figure 2.15 and Figure 2.16, the tone and color reproduced by
superimposing halftone patterns depend strongly on the way they are superimposed
and on the quality of the registration between the patterns. Because exact registration
is not a practical goal in color printing, the alternative for color reproduction is to
print the halftone dots in random registration. This random dot placement is approximated by the same rotated screen technique used to avoid moiré. As shown in Figure
2.19, using 0 degrees, 30 degrees, and 60 degrees for yellow, cyan, and magenta dots,
all degrees of dot overlap occur. This quasi-random pattern of overlap is the key to
accurate color calculations with halftones. To calculate the total integrated color one
sees when viewing a halftone image at a normal reading distance, we need to add
up all the microcolors in the microregion illustrated in Figure 2.19. This is a threestep process first described by Neugebauer in 1937.16
Step I: Calculate the Color Area Fractions. Three signals (Fc, Fm, and Fy) are
formed in the color RIP process and are sent to the printer. These three signals are
the fractional dot area (F) values in Equation 2.7 applied to each of the three primary
Color Desktop Printer Technology
Green Overlap Color
Red Overlap Color
Black Overlap Color
Blue Overlap Color
FIGURE 2.19 The quasi-random registration of CMY dots at 30o, 60o, and 0o screen
rotations. Note the regions of RGB and black overlap.
inks: cyan, magenta, and yellow. This causes the printing device to print three halftone
patterns: a pattern for cyan at dot area fraction Fc, a magenta pattern at Fm, and a
yellow pattern at Fy. The random overlap of these three patterns results in eight color
area fractions, shown in Table 2.2. A region with no cyan (1 – Fc), no magenta (1 –
Fm), and no yellow (1 – Fy) is white. The fraction of the image that is white is the
product fw = (1 – Fc).(1 – Fm).(1 – Fy). The area fractions of the other seven colors
are similarly calculated, as shown in Table 2.2. These eight colors are called the
Neugebauer primary colors.
Step II: Measure the eight Neugebauer primary colors. We print test patches
of each of the eight Neugebauer primary colors. We can measure the reflection
spectrum of each primary sample, Rw(λ), Rc(λ), Rm(λ), Ry(λ), Rr(λ), Rg(λ), Rb(λ),
and Rk(λ). We also can measure the color coordinates for each sample,
[Xw, Yw, Zw], [Xc, Yc, Zc].... [Xk, Yk, Zk].
Step III: Calculate the Printed Color. We apply an expanded version of the
Murray–Davies Equation 2.7 to calculate the spectrum of the printed image, R(λ)
as shown in Equation 2.10. Equation 2.11 is a shorthand way to represent this
calculation, in which the subscript 1 through subscript 8 stand for colors w through k.
Image Quality of Printed Text and Images
The Eight Neugebauer Colors Produced by
Printing Random or Pseudo-Random Dots
of Cyan, Magenta, and Yellow Ink. The
Three Area Fractions of Ink (Fc, Fm, and
Fy) Produce the Color Area Fractions
(fw, fc, fm, fy, fr, fg, fb, fk)
Area Fraction
fw = (1 – Fc).(1 – Fm).(1 – Fy)
fc = Fc.(1 – Fm).(1 – Fy)
fm = (1 – Fc). Fm.(1 – Fy)
fy = (1 – Fc).(1 – Fm). Fy
fr = (1 – Fc). Fm. Fy
fg = Fc.(1 – Fm). Fy
fb = Fc.(1 – Fm). Fy
fk = Fc. Fm. Fy
R ( γ ) = fw ⋅ R w ( γ ) + fc ⋅ R c ( γ ) + fm ⋅ R m ( γ ) + fy ⋅ R y ( γ ) +
fr ⋅ R r ( γ ) + fg ⋅ R g ( γ ) + fb ⋅ R b ( γ ) + fk ⋅ R k ( γ )
R (γ ) =
∑ f R (γ )
Color Equation 2.4 through Equation 2.6 can be applied to the reflection
spectrum, R(y), to calculate the color coordinates of the image, [X,Y,Z]. Alternatively, the measured color coordinates of the eight Neugebauer primary samples
can be combined to find the coordinates of the printed image, as shown in
Equation 2.12 through Equation 2.14.
∑f X
i =1
∑f Y
i =1
Color Desktop Printer Technology
∑f Z
i =1
Equation 2.10 through Equation 2.14 are collectively known as the Neugebauer equations. They are the color equivalent of the Murray–Davies Equation
The practical application of this kind of calculation involves selecting the
color one wishes to print, [X,Y,Z], and inverting the calculations shown above to
determine the three ink area fractions, Fc, Fm, and Fy, that must be sent to the
printer. The simple Murray–Davies equation is easily inverted to calculate F given
R, Rp, and Ri. However, the Neugebauer equations are not straightforward to
invert. One must rely on numerical and statistical techniques to do this. One
straightforward method is to define a wide range of values for Fc, Fm, and Fy and
calculate the color coordinate [X,Y,Z] for each combination. The result is then
analyzed statistically to construct a set of empirical equations for predicting Fc,
Fm, and Fy from [X,Y,Z].
The Murray–Davies equation and the Neugebauer equations are idealized descriptions of perfectly well-behaved halftones. However, real printed images almost
never behave exactly as described by Murray–Davies and Neugebauer, and some
amount of correction is needed to achieve acceptable image quality. The most
common problem encountered is the dot-gain phenomenon.12,13,15
When printing a halftone image with a single ink, the actual printed image is
usually darker than predicted by the Murray–Davies equation. This is illustrated
in Figure 2.20, where Fn is the dot area command sent to the digital printer.
The printer responds by printing halftone dots, but the measured reflectance of
the printed image (Figure 2.20b) is darker than expected. The phenomenon is
not unique to digital printing. It was originally observed by commercial printers
more than a century ago, before much was known about the underlying physical
and optical processes involved. Early printers assumed the image was printed
too dark because the halftone dots spread out in the printing process. If this
occurred, the size of the printed dots, Fa, must be larger than the nominal dot
size, Fn, that was intended. Thus, the darkening phenomenon illustrated in
Figure 2.20 was called the dot-gain effect. This term is still in use, even though
the underlying causes are more involved than a simple increase in the physical
size of the printed dot.
If one assumes the Murray–Davies equation would be correct if one used
the correct dot size in Equation 2.7, then one can rearrange Equation 2.7 and
solve for the apparent dot size, Fa, as shown in Equation 2.15. Thus, by
Image Quality of Printed Text and Images
FIGURE 2.20 The dot-gain phenomenon. For a series of nominal dot area fractions, Fn,
sent to a printer, line (a) is the result predicted by the Murray–Davies equation and data
points (b) are the actual printed image.
measuring the actual printed reflectance, R, the reflectance of the paper, Rp,
and the reflectance of the solid ink, Ri, one may calculate the apparent size of
the halftone dot, Fa.
Fa =
Rp − R
Rp − Ri
This leads to the quantitative definition of dot gain given in Equation 2.16.
Dot gain, defined in this way, is a useful metric for characterizing and calibrating
the printing process, and knowledge of dot gain is essential to control the printing
process for maximum pictorial image quality.
DG = Fa - Fn
There are two primary causes of the dot-gain phenomenon. The physical spread
of ink on paper is one cause. In general, the decreased reflection density that
results when the ink spreads out on the paper is more than offset by the increased
area coverage, F, and the overall image is darker.
The second major cause of dot gain is an optical effect. As illustrated in
Figure 2.21, light used to view the halftone image can enter the paper by passing
through a halftone dot or by entering between the dots. The light that enters the
paper at (B) in Figure 2.21 scatters around in the paper and eventually emerges
Color Desktop Printer Technology
FIGURE 2.21 Illustration of optical dot gain.
as reflected light between the dots. Some fraction of the light, however, enters as
shown in (C) and scatters under a dot and is absorbed. This increases the fraction
of light that is absorbed by the halftone dots and results in a darker image. Thus,
the effect is similar in appearance to a physical dot gain and is often called an
optical dot gain. It is also called the Yule–Nielsen effect.17
Much research has been devoted to both physical and optical dot gain and to
the characteristics of the paper, the ink, and the dot patterns that influence it. In
general, most common non-coated papers show about the same amount of dotgain, both physical and optical. Coated papers show less of an effect, and special
substrates designed to minimize these and other artifacts of the printing process
show much less dot gain. Inkjet, in particular, benefits significantly from the use
of special substrates. Nevertheless, some amount of dot gain is always observed,
and systems in common use for digital desktop printing have been well optimized
to compensate for these effects.
There are several ways to calibrate a printer to correct for the dot-gain phenomenon. The most general approach is empirical. One prints a set of known samples
with a set of known print commands, Fn, and measures the resulting output
reflectance, R. Then a statistical polynomial regression can be used to fit the data.
The resulting polynomial is called a printer model, or sometimes a look-up-table
(LUT), and provides the necessary tone correction in the printing process. The
process that does this is part of what is called “color management.”18
When multiple colors are printed, the polynomial regression process can
become quite complicated. Before the advent of powerful desktop computers,
such complexity required a much simpler solution. In the early 1950s a simple,
empirical equation was suggested as a correction for dot gain. This is called the
Yule–Nielsen equation, shown below as Equation 2.17.13
R = F ⋅ R1i / n + (1 − F ) ⋅ R1p/ n 
The Yule–Nielsen equation is a simple polynomial that looks much like the
Murray–Davies equation but with a single power factor called the Yule–Nielsen
Image Quality of Printed Text and Images
Halftone Data for a 60-LPI Halftone Printed with a 600-dpi Printer.
Fn = Nominal Dot Fraction Command. R = Reflectance of Printed
n-factor. By fitting this equation to measured data, the value of n is determined.
The n value is then a single calibration constant for the printing process. The dot
gain, DG, defined in Equation 2.16 varies with the gray level of the printed image.
However, the Yule–Nielsen n factor is useful as a single index of dot gain for the
printing process at any gray value. A value of n = 1 means zero dot gain, and
any n > 1 means non-zero dot gain. The Yule–Nielsen equation, and variations
of it, are often still used to characterize both commercial and desktop printers.
Using the Yule–Nielsen n factor is illustrated by the data in Table 2.3. These
data are the experimental data shown in Figure 2.20. The halftone dots were
printed using an inkjet 600-dpi printer to form the halftone dots at 60 halftone
dots per inch. The printer was sent 11 nominal dot fractions (Fn in Table 2.3),
and the printer responded by printing 11 gray patches. The reflectance, R, of each
patch is also shown in Table 2.3.
The value of the Yule–Nielsen n factor used in Equation 2.17 can be found
experimentally by a spreadsheet calculation using the data in Table 2.3. The
first step is to note the values of Ri and Rp. The value of Rp is shown in Table
2.3 as the value of R at Fn = 0. Thus, Rp = 0.85. Similarly, the solid ink
reflectance is Ri = 0.05. The values of n typically observed for most halftone
images is in the range 1 < n < 5. In the spreadsheet, we can define a trial range
of n values we believe will include the correct value. Later, if the calculation
fails to identify the correct value of n, we can adjust the range. In this example,
the trial range of n values was chosen from 1.70 to 1.90, as shown in Table
2.3. The nominal values of the dot area fraction, Fn, range from 0 to 1, as shown
in the table. For each combination of n and Fn, Equation 2.17 was used to
calculate the reflectance, R.
The values of R shown in Table 2.3 were compared with the values of R
measured experimentally. For each value of n and Fn, the deviation, δ, between
the calculated and experimental value of R was determined. The squared deviation
values, δ2, are shown in Table 2.4.
For each value of n, the values of δ2 were averaged over the range of Fn
values. The standard deviation, σ, is the square root of the average value of δ2.
These values are shown on the right of Table 2.5. Figure 2.22 shows the value
of σ vs. the trial value of n. The standard deviation reaches a minimum at n =
1.82. This is our experimental estimate of the n value characteristic of this printing
Color Desktop Printer Technology
Values of R Calculated with Equation 2.17 Using Rp = 0.85 and Ri = 0.05.
Experimentally Measured Values of R for Each Value of Fn are Shown as
the Last Row of the Table
Minimum at n = 1.82
FIGURE 2.22 Standard deviation in reflectance vs. trial value on n from the data in Table
2.1 through Table 2.4.
The Values of Squared Deviation, δ2, for Each Combination of Fn and n
Image Quality of Printed Text and Images
Color Desktop Printer Technology
This technique for determining the value of n is simple and can be performed
with a spreadsheet. Once the value of n is found, it is easy to calculate Fn. The
dot area fraction, Fn, one must send to the printer to print any chosen value of R
is easily calculated using the rearranged form of Equation 2.17. For example, if
we want to print a reflectance of R = 0.234, we would use Equation 2.18 to
determine that we should send the command Fn= 0.643 to the printer.
p −R
Fn =
p − Ri
Dot gain also occurs in color printing, and the Yule–Nielsen n-factor can be
applied to the Neugebauer color Equation 2.11 through Equation 2.14, as shown
in Equation 2.19 through Equation 2.22 below. One may choose to apply the nfactor to the reflection spectra for the eight Neugebauer primary colors, as illustrated in Equation 2.19, or one may apply the factor to the individual colorimetric
numbers, as shown in Equation 2.20 through Equation 2.22.
R (γ ) = 
∑f R
∑f X
i =1
∑f Y
i =1
∑f Z
i =1
( γ )
It is important to remember that the n-factor is an empirical correction and
is not based on physical theory. If the color coordinates [X,Y,Z] are calculated
from the reflection spectrum R(λ) (see Section 2.4.3), the results will not be the
same as the [X,Y,Z] values found by applying the n-factor in Equation 2.20
through Equation 2.22. One must determine experimentally which correction
technique works best for any given printing system.
Image Quality of Printed Text and Images
FIGURE 2.23 Geometric model of ink spread. Adapted from Pappas and Pappas, Dong,
and Neuhoff.19,20
In general, the n-factor is a useful tool for characterizing a printer. However,
it is not always used as a means of calibrating a printer and managing practical
color reproduction. Instead, more extensive experiment data are collected for the
printer over a wide range of colors, and an empiral look-up table is developed
through advanced statistical techniques of color management.18
Two types of models are used to describe halftone printing. One type of model,
the empirical model, is based on statistical analysis of a large set of printed
samples. These empirical models are useful for the practical calibration of a
printer and for developing methods for color management. The other type of
model is the mechanistic model. A mechanistic model attempts to describe the
underlying physical or optical mechanisms of the printing process. Mechanistic
models are often used to guide scientists and engineers in the design of improved
printing devices. However, mechanistic models generally tend to be more complex
than empirical models and therefore are not often used for printer calibration or
color management. Many examples of mechanistic models have been published.
Some describe the way inks penetrate and spread in paper. Others describe the
scattering of toner. Numerous models have been published to describe the optical
scattering of light in paper. Three examples of mechanistic printer models are
presented below.
Geometric Model of Physical Dot Gain: An example of a simple geometric
model of physical dot gain is shown in Figure 2.23.20,21 This model compares the
idealized matrix of printer locations to the circular shape of ink dots. For such
an idealized process, the geometry of physical dot gain can be solved exactly.
The printer is able to address the center of each square of dimension T, but the
dot that is printed is of radius r > T. This leads to exactly three geometric shapes,
α, β, and γ, each with an area that can be calculated by knowing only T and r.
These three types of regions can be used to add up all of the dot gain and dot
overlap in the idealized printing process. Such models can provide a very accurate
Color Desktop Printer Technology
Point Spread Function
FIGURE 2.24 Illustration of light scattering and the point spread function.
calibration for physical dot gain for a printing process that behaves exactly as
described by the model. However, this kind of model is seldom a practical
improvement over empirical statistical techniques of printer calibration because
most real printers do not behave exactly in the way described by the model. This
kind of model is much more useful as a guide in product development than it is
for applied color management.
Convolution Model of Optical Dot Gain: Mechanistic models of optical
dot gain, illustrated in Figure 2.21, have been explored extensively in recent
years.17,22 The spreading of light can be characterized by a point spread function,
illustrated in Figure 2.24. The point spread function is a description of the average
distance, , that light scatters in the paper before it returns to the surface as
reflected light. Some light travels a long distance and some a short distance. The
bell-shaped curve illustrated in Figure 2.24 is a summary of the distances traveled
by all the light entering at location x = 0. This bell-shaped curve is called the
point spread function for light scattering in the paper and is represented mathematically as P(x).
The point spread function is used to calculate the magnitude of optical dotgain by combining it with a function that describes the distribution of halftone
dot pattern, T(x), illustrated in Figure 2.25. This function describes the fraction
of the light that is transmitted through an ink dot, T, as a function of the dot
FIGURE 2.25 Transmittance (T) of halftone dot pattern vs. position (x). A dot is present
at position (a) and the transmittance is low. There is no dot at position (b) so the transmittance is that of the paper.
Image Quality of Printed Text and Images
location, x. Location (a) is an ink dot with T<1, and location (b) is the paper
between dots where the transmittance is 1.
The point spread function and the halftone dot function are actually functions
in terms of the two dimensions, P(x,y) and T(x,y). The theory for combining
these two functions is based on a mathematical operation known as convolution,
illustrated in Equation 2.23, where the symbol * is used as a shorthand notation
for the convolution integral on the right side of Equation 2.23.
( ) ( )
T x, y * P x, y =
∞ ∞
∫ ∫ T ( x, y ) ⋅ P ( x− a, y− b ) dadb
−∞ −∞
This convolution operation is combined with the reflectance of the paper, Rp,
to calculate reflected pattern of light, R(x,y), from the halftone image. This is
done with Equation 2.24. Note the * in the expression means the convolution
integral, not multiplication.
( )
( )
( ) ( )
R x, y = R p ⋅ T x, y ⋅  T x, y * P x, y 
Semi-Empirical Model of Optical Dot Gain: Equation 2.24 is a concise way
of expressing the physical optical dot gain, but the practical application of this
equation can be very complex. Thus, simplifications have been suggested that
provide useful approximations to the optical dot-gain mechanism. One example is
a semi-empirical model based on a combination of optical theory and experimental
measurements of micro-reflectance, illustrated in Figure 2.26.23–25 The terms Rg and
Rk represent the reflectance of the unprinted paper and of the ink printed at F = 1,
respectively. The observation is that the reflectance values of the ink and paper do
not remain constant at these values. Rather, they vary with changes in the dot area
fraction. We represent these non-constant reflectance values as Rp vs. F and Ri vs.
F in Figure 2.26.
The variation in the paper and ink reflectance values reflect the observed
failure of the Murray–Davies Equation 2.7. When Equation 2.7 is used with
constant values of Rg and Rk the result is a calculated predicted value of R is
smaller than the experimental value. However, if the values Rp and Ri are used
in Equation 2.7, where Rp and Ri are the values observed at each value of F, then
the correct value of R is calculated. This means that the Murray–Davies equation
is the correct equation to describe a halftone image, provided the correct values
of values Rp and Ri are used instead of the constant values Rg and Rk.23 The
problem then is to model the way Rp and Ri vary with F. Such models have been
developed and reported by several researchers, and the results have led to useful
insights into the optical and physical behavior of halftone systems.
Mechanistic models such as those described above have contributed significantly to a better understanding of the printing processes,23 and much work
Color Desktop Printer Technology
Dot Area Fraction, F
FIGURE 2.26 Example of micro-reflectance behavior of 30-LPI halftone images printed
with a 600-dpi inkjet printer at dot fractions ranging between 0 and 1. Ri is the reflectance
halftone dots, and Rp is the reflectance of the paper between the dots. R is the overall
average reflectance of the image.
continues to be done in this area. However, routine calibration and control of
printers and for practical color management statistical models are generally found
to be more useful.
The method by which continuous tone images are converted into halftone images
was first developed early during the 19th century concurrently with the development
of photographic technology. During the late 20th century, digital techniques were
developed to convert continuous tone images into halftone images. These digital
techniques initially were designed to mimic the original photographic process. Thus,
it is worthwhile to review the original photographic process for generating a halftone
printing plate. This process is called pre-press process photography.22
The basic photographic process for printing a film negative is illustrated in Figure
2.27. The key to the process is the relationship between the transmittance, T, or
the film negative and the reflectance, R, of the photographic paper. This relationship is the tone transfer function of the print paper. Figure 2.28 illustrates a tone
transfer function for a paper that produces a full range of copy gray values and
one that produces a very high contrast copy with very few mid-tone grays.22
Process photography looks much like ordinary photography, but a special
screen is added to convert grayscale to dots. This is illustrated in Figure 2.29. In
Image Quality of Printed Text and Images
Photographic Copy
Light Source
Continuous tone
FIGURE 2.27 Printing a continuous tone film negative onto a continuous tone photographic paper. (Courtesy of Frank Cost.)
FIGURE 2.28 Examples of normal- and high-contrast tone transfer functions and the
images they produce. (a) Normal contrast; (b) high contrast. (Courtesy of Frank Cost.)
addition, the photographic paper is replaced with a photographic film of very
high contrast, such as in Figure 2.28(b).
Color Desktop Printer Technology
Photographic Copy
Light Source
Continuous tone
FIGURE 2.29 Halftone process photography. (Courtesy of Frank Cost.)
Film Negative
Threshold Value
Tf Ts = 0.25
FIGURE 2.30 The effect of the halftone screen.
The high-contrast copy film acts as a threshold system that produces either
black or white, but no gray values. A mid-tone gray in the original negative image
(R = 0.5, for example) is at the threshold, so any value greater than that is copied
as black (T = 0) and any value less than that is copied as white (T = 1). The
halftone screen modulates the image, as illustrated in Figure 2.30, to produce
dots of a size proportional to the gray value. The film and the light source are
chosen, in this example, for a threshold gray value of 0.25. When the original
film transmittance, Tf, times the halftone screen transmittance, Ts, is less than
0.25, there is too little light to have an effect on the copy film. Under that
condition, the copy film remains at a high transmittance, Tc = 0 (white).
Image Quality of Printed Text and Images
At point (A) in Figure 2.30, the negative film image has a transmittance of
Tf = 0.313 and the screen has a transmittance of Ts = 0.80. The product is 0.25,
so the copy film transitions to black, Tc = 0. Then at point (B) the original has
Tf = 0.35 and the screen has Ts = 0.714 for a product of 0.25, so the copy film
transitions back to white (Tc = 1). If we continue in this way, we note that the
size of the region that is black increases as the transmittance of the con-tone film,
Tf, increases. In other words, dots are produced with areas that are proportional
to the con-tone gray value, Tf. Note that the size of the dot can be no greater than
one cycle of the screen. The screen frequency, therefore, translates directly to the
LPI halftone frequency of the image.
Both negative and positive types of process photography are practiced. Figure
2.30 is a negative process, just like regular photographic film. A positive halftone
process involves a high-contrast copy film that turns clear on exposure and black
in unexposed regions. In this case the black dots are where the white is in the
halftone copy of Figure 2.30, and the white becomes black. Regardless of the
technique, the result is a photographic transparency that is made of halftone dots
rather than continuous tones of gray. This halftone film is then used as a mask
to expose photosensitive material used to form the printing plate, as illustrated
in Figure 2.31.
A digital simulation of the photographic process is often used to generate digital
halftone images. This is done by dividing the original image into pixels at the
same dpi as the printer, as illustrated in Figure 2.32. This process is called
Light Source
Printing Plate
FIGURE 2.31 Exposure through the halftone photographic transparency to form an offset
lithographic printing plate. (Courtesy of Frank Cost.)
Color Desktop Printer Technology
Pixels at printer dpi
Halftone Cell
FIGURE 2.32 Digital halftoning begins by sampling the original image at the printer dpi
and constructing digital halftone cells.
sampling, and the digital halftone is generated by an algorithm that determines
whether to print ink or not to print ink at each pixel location. The halftone
algorithm begins by grouping pixels into halftone cells. Each cell measures N ×
N pixels and will contain a digital halftone dot with a size proportional to the
gray level in the original image. The cell size illustrated in Figure 2.32 is N = 6
for a 6 × 6 halftone cell.
Next, a matrix is designed as illustrated in Figure 2.33. This matrix, also
called a digital halftone mask, is an N × N array of threshold numbers (t = 0, 1,
Threshold Matrix
20 25 29
32 34 35
14 18
27 30
12 16 21
26 31
15 22
11 17
13 19
Halftone Cell
FIGURE 2.33 The halftone algorithm compares each halftone cell with the value, t, in
the threshold matrix. The threshold matrix is also called a digital halftone mask.
Image Quality of Printed Text and Images
Threshold Matrix
20 25 29
32 34 35
14 18
27 30
12 16 21
26 31
11 17
15 22 28
13 19
Halftone Cell
FIGURE 2.34 If the gray value is less than the t/(N2) number divided by the matrix size,
the pixel is black.
2 ... N – 1) used to determine whether or not ink is printed at each pixel location.
This is done by comparing the gray levels in the original image with the corresponding threshold numbers.
For example, the halftone cell in the original image just to the right of the
eye in Figure 2.33 is a uniform area of reflectance R = 0.92. This is compared
to the value t/(N2) as illustrated in Figure 2.34. Thus, R = 0.92 is compared with
0/36, 1/36, 2/36 … 34/36, 35/36. In all cases except the last two, R is larger, so
the algorithm tells the computer to print no ink. In the last two cases. R < 34/36
and R < 35/36, so the algorithm tells the printer to print ink in the last two
locations of the halftone cell. The resulting halftone does not look much like an
eye, but when we zoom out to a reasonable distance, as illustrated in Figure 2.35,
the appearance of gray scale becomes more apparent.
The quality of the halftone image is significantly influenced by the arrangement of the threshold numbers in the halftone cell. Figure 2.36(b) was generated
using the threshold matrix shown in Figure 2.23. Figure 2.36(c) and (d) were
done with the same 6 × 6 cell size but with the threshold matrices shown at the
bottom of the figure. The halftone dots in (c) are more oval in shape and are
much more pleasant in appearance. The threshold values for both (b) and (c) are
clustered to form discrete dots. However, the threshold values for (d) are quasirandom. They form a dispersed halftone dot rather than a clustered halftone dot.
In this illustration, the dispersed halftone dot makes the better-looking image.
In the early days of desktop printing, printers were limited in their addressability. A 100-dpi printer was considered to be a high-resolution printer for text
imaging a decade ago. For such a printer to print halftones with 37 different gray
levels, a 6 × 6 halftone cell was needed. This made a halftone cell of 6/100 inches
or 6/4 = 1.5 millimeters in size. Such halftone cells were quite easy to see at an
ordinary reading distance, so the design of the halftone cell was quite important
to the quality of the printed image. For this reason, the dispersed type of halftone,
Color Desktop Printer Technology
Pixels at Printer dpi
Halftone Cell
FIGURE 2.35 Zooming out from the halftone pattern makes the image look more like
the original image.
as illustrated in Figure 2.36(d), was far superior to the traditional clustered dot
halftone used by commercial printers. Other factors that influence image quality
in digital halftone printing will be explored subsequently.
An alternative to the halftone mask technique for generating a digital halftone is
a technique called noise distribution. Photographers and printers have long known
that the addition of random or quasi-random noise to an image can improve the
appearance of tone. Figure 2.37 is an example of stipple point engraving.14 The
artist used a tool to add random texture to the engraved image. The result is a
fine set of random ink dots that are more closely spaced in shadow regions and
more widely spaced in highlight regions.
Robert’s Method: One of the earliest digital techniques for using noise to
simulate grayscale was developed by L. G. Roberts in 1962.26 Robert’s method
is illustrated in Figure 2.38(a).
This technique involves sampling an original image as described in Figure
2.32, but the sampled pixels are not combined into halftone cells. Instead, the
decision about whether or not to print ink at a pixel location is based on a comparison between the original gray value and a random number, Rnd, between 0
and 1. For example, if R = 0.5, there is a 50% change that R > Rnd. Note also that
(1 – R) < Rnd 50% of the time. If we print ink when (1 – R) < Rnd, then we will
print ink 50% of the time. Similarly, if R = 0.2, then R < Rnd only 20% of the
time, and (1 – R) < Rnd 80% of the time, so 80% of the pixels are printed with ink.
Floyd–Steinberg and Error Diffusion: In 1975 Floyd and Steinberg introduced a halftoning technique they called an adaptive algorithm.27 This technique
Image Quality of Printed Text and Images
(a) Original
(b) Clustered Dot
(c) Better Clustered Dot
(d) Dispersed Dot
Matrix for (c)
Matrix for (d)
FIGURE 2.36 Three halftone patterns.
distributes quasi-random noise in the image in a way that significantly improves
the visual quality of the image.
The technique starts with the first pixel at location (i,j) = (0,0) in the upper
left corner of the image and compares the original gray value at that location,
RO0,0, with a threshold value of 0.5. If (1 – RO0,0) < 0.5, then ink is printed;
Color Desktop Printer Technology
FIGURE 2.37 Since the Renaissance, printers have applied a variety of techniques to
induce random textures to mid-tone gray regions of printed images.14 This is an example
of a technique called stipple point engraving. (Courtesy of Melbert E. Cary, Jr. Graphic
Arts Collection, Rochester Institute of Technology.)
FIGURE 2.38 Robert’s method of error addition and Floyd–Steinberg’s method of error
diffusion. (a): Robert’s method; (b): Floyd–Steinberg method.
otherwise, ink is not printed. This first pixel in the copy image should therefore
have an ideal gray value of RC0,0 = 0 or 1. The difference between the original
and the ideal copy reflectance (E00 = RO00 – RC00 ) is a measure of the amount
of error that occurs in the threshold process. For example, if RO00 = 0.2, then
RC00 = 0, so E00 = 0.2 – 0 = 0.2, and the pixel is too dark by 0.2. To correct for
this error, we propagate this error to the next pixel. We do this by adding the
error to the threshold value of the next pixel. Thus, the value of E00 = 0.2 is added
to the threshold value for the next pixel at location (i,j) = (0,1). Using the new,
adjusted threshold, we examine whether or not (1 – RO0,1) < 0.5 + 0.2. If it is,
Image Quality of Printed Text and Images
Direction of Processing
i, j
Diffusion Kernel
i, j
i, j+1
i +1, j − 1 i + 1, j i + 1, j + 1
FIGURE 2.39 The traditional Floyd–Steinberg diffusion kernel.
then we print ink; otherwise, we do not. Note that by propagating the error, we
increase the probability that ink will not be printed because the threshold value
is increased.
The error can be either positive or negative, so the threshold value can be
increased or decreased. In addition, we generally do not propagate the error only
to the next pixel. We diffuse a fraction of the error to several neighboring pixels.
The most often used pattern for error diffusion is illustrated in Figure 2.39.
In this example, the total error, Ei,j, made in the thresholding process at any
location (i,j), is divided up and added to four other threshold values at four other
locations. Then (7/16)Ei,j is added to the threshold value that will be used for
pixel (i,j+1); (3/16)Ei,j is added at location (i+1,j-1); (5/16)Ei,j is added at (i+1,j);
and (1/16)Ei,j is added at (i+1,j+1). Figure 2.38(b) is an example of the
Floyd–Steinberg technique of error diffusion.
The two major techniques for generating digital halftones may be classified as
shown in Table 2.6. These are digital masking, as described in Section 2.6.2, and
noise distribution, as described in Section 2.6.3. This is an arbitrary classification,
and as with any classification scheme, there are cases reported in the research
Techniques for Digital Halftoning
Techniques for
Digital Halftoning
Example #1
Example #2
(A) Masking
(B) Noise Distribution
Clustered Dot
Robert’s Method
Dispersed Dot
Error Diffusion
Color Desktop Printer Technology
FIGURE 2.40 Illustration of spatial frequency. (a): A low-frequency image; (b): a highfrequency image.
literature that do not clearly fit one or the other category. Nevertheless, halftone
masking and noise distribution are the major techniques in common use.
Digital halftoning has developed rapidly over the past 3 decades, and this
has led to confusing names for the different techniques. For example, the term
dither is often used to describe noise distribution techniques such as error
diffusion and Robert’s method. Unfortunately, the term dither mask is often
used to refer to the digital halftone mask (threshold matrix) described in Section
2.6.2. As a result, dithering has evolved to mean any type of digital halftoning
and thus has lost its utility for distinguishing between different techniques for
generating halftones.
Another classification scheme is based on characteristics of the halftone
pattern rather than the technique used to generate it. The AM/FM classification
scheme is an example. All halftones control grayscale by controlling the dot area
fraction, F, but there are two different kinds of patterns for controlling F. One
pattern, called the AM pattern, is illustrated in Figure 2.36(b) and (c). The AM
halftone varies F by varying the size of the halftone dot. The other pattern is
called the FM pattern and is illustrated in Figure 2.36(c) and Figure 2.38(a) and
(b). In these halftones, F is varied by varying the number of printer dots in a unit
area, but the dots do not change in size. Note that the halftone mask technique
can produce either AM- or FM-type halftone patterns. The noise distribution
technique produces only FM halftones.
More recent work has shown that the concept of AM vs. FM halftones
represents two extremes and that many halftones have attributes of both kinds.
The respi halftone, shown in Figure 2.40, was designed to change from 220 LPI
to 110 LPI in the highlights.28 Most printing devices print large dots more reliably
than small dots. When the size of the AM dots at 220 LPI becomes too small for
the printing device to print reliably, a frequency change (FM effect) down to 110
LPI allows highlights to be printed with fewer large dots.
Image Quality of Printed Text and Images
Conversion Table between
Frequency and Size.
Frequency Is in LPI (line pairs
per inch) and cy/mm (cycles
per millimeter). Size Is in Mils
(1/1000 inch) and Millimeters
A more general, but more complex, technique for characterizing halftone
patterns involves measuring the so-called noise power spectrum of the halftone
image. This provides a more general understanding of the halftone pattern than
the simple concept of AM vs. FM. Understanding the noise power spectrum is
very helpful for understanding the image quality differences between different
halftone patterns.
The concept of spatial frequency is illustrated by the two images in Figure 2.40.
Frequency refers to how often or how quickly something happens in an image.
In Figure 2.40(a), only one person is shown, but in image (b) people are
shown. People appear in (b) more frequently than they occur in (a), so in terms
of people, (b) is a higher frequency image than (a). This also means the people
in (b) are smaller than in (a), with both images completely filled with people. As
the frequency of an image element increases, its size decreases. Halftone dots
are another example. A 35-lpi halftone is composed of halftone cells that are
0.0286 inches (28.6 mils) diameter, and a 100-lpi halftone has cells of 10 mils.
Color Desktop Printer Technology
Frequency in Relative Units
FIGURE 2.41 Noise power spectra of images in Figure 2.40(a) and (b). The axis units
are relative units for illustration purposes.
The lpi is an index of halftone frequency. The relationship between frequency
and size is summarized in Table 2.7.
Frequency is a useful concept for describing how rapidly individual elements
of an image change. Consider the element “people” in Figure 2.40a. The number
of people changes slowly as we move across the image, but in (b) the people
change rapidly. Another example is the element “edge.” For example, the edge
of the man’s collar is a high-frequency element because it changes rapidly as we
move from left to right over the edge. Sharper edges are high frequency, but
blurred broad edges are low frequency. Frequency is a useful metric of image
quality for describing the intrinsic resolution characteristics of an image.29
An 18th-century mathematician named Jean Baptist Joseph Fourier demonstrated
that any shape can be described as the sum of all of the frequencies in the image.
He also developed the mathematical process for extracting all the frequencies in
the shape. This process is called the Fourier transform of the image and is done
easily with mathematical applications such as spreadsheets.29 As an example, the
Fourier transforms of the images in Figure 2.40(a) and (b) were calculated and
are plotted in Figure 2.41. The horizontal axis is the frequency and the vertical
axis is the square of the magnitude of the Fourier transform. It represents the
relative importance of image features at each frequency. Thus, the Fourier analysis
shows, in a quantitative way, that image (b) has more power at higher frequency.
In other words, (b) is a busier image than (a).
Halftoning is a very useful way to control not only the gray level in an image,
but also the noise in an image. Halftones are a type of granularity pattern and
can be an objectionable source of noise in the image. To minimize the visual
impact of this halftone noise, we prefer to print dots that are smaller than the eye
can see. The resolving power of the human visual system reaches a maximum
Image Quality of Printed Text and Images
for features of about 0.8 to 1.0 mm in size for a normal reading distance of 35
cm. This means we are most sensitive to features about the size of printed letters.
Features that are approximately 0.23 mm are on the threshold of resolution of
normal human vision. As shown in Table 2.2, this corresponds to a 110-LPI
halftone. A 150-LPI clustered dot halftone uses halftone cells that are not resolvable at all by most people at a normal reading distance. For this reason, most
magazine images are printed at 150 LPI in order to look indistinguishable from
a continuous tone photograph.
Desktop printing in the 1980s involved devices capable of only limited addressability. A simple dot matrix printer, for example, with an addressability of 72
LPI could construct letters and numbers of reasonable quality. However, a 6 × 6
halftone cell able to print 37 levels of gray resulted in halftones cells 2.1 mm on
a side. Figure 2.42(c) is an illustration of a 25% gray level printed in this way,
and the halftone pattern is quite visible and objectionable. The other patterns in
Figure 2.42 also illustrate a 72-LPI printer but with different halftone patterns.
Some are clearly preferable to others.
FIGURE 2.42 A gray level of R = 0.40 printed with a 72-dpi printer using four different
halftone patterns. (a): Roberts method, (b) Floyd–Steinberg method. (c) and (d) are 6 × 6
mask halftones made with the matrices shown in Figure 2.36.
Color Desktop Printer Technology
35 cm
35 cm
1 meter
2 meter
5 meter
5 meter
Frequency, LPI
2 meter
Frequency, LPI
35 cm
1 meter
5 meter
Frequency, LPI
1 meter
2 meter
2 meter
35 cm
0 5 meter
Frequency, LPI
FIGURE 2.43 Noise power spectra of the halftone patterns shown in Figure 2.42. Dotted
lines are the resolution of the human visual system at viewing distances of 0.35, 1, 2, and
5 meters.
Fourier analysis provides an important technique for measuring the severity
of noise patterns of different types. Figure 2.43 shows the Fourier noise power
spectra for the patterns in Figure 2.42. One can think of the noise power, W, as
the statistical variance (squared standard deviation) of the reflectance in the image
measured for each frequency contained in the image (frequency = 1/size of image
element). Both of the masking techniques of Figure 2.42(c) and (d) show large
peaks at 0.47 cy/mm (12 LPI), which is the frequency of the halftone cells.
However, the dispersed dot image of Figure 2.42(d) looks better to most people.
Part of the reason it looks better is that some of the noise power has been shifted
from the dominant 12 LPI into new peaks at higher frequency, where our eyes
are less sensitive. Robert’s method in Figure 2.42(a) uses white noise to simulate
gray scale. As the name implies, white noise contains equal noise power at all
frequencies. The best-looking pattern is in Figure 2.42(b), the Floyd–Steinberg
pattern. The Floyd–Steinberg process shifts most of the noise to high frequency
(small feature sizes) so the eye does not resolve it as well.
The usefulness of noise power as a method of evaluating granularity in printed
images can be illustrated by examining Figure 2.42 at different reading distances.
If we slowly back away from Figure 2.42, we will notice a distance at which the
Floyd–Steinberg image of Figure 2.42(b) blends to a uniform gray. For most
people, this occurs at about 2 to 3 meters of viewing distance. As shown in Figure
2.43, the sensitivity of the human visual system at a distance of 2 meters drops
to nearly zero at 1 cy/mm. Because most of the noise in the Floyd–Steinberg
Image Quality of Printed Text and Images
image is above 1 cy/mm, the noise in Figure 2.43 nearly disappears at 2 meters
viewing distance.
By slowly moving back farther than 3 meters, one will notice that the noise
power of the other images decreases. The clustered dot and dispersed dot images
show a sharp decrease in visual noise at about 4 to 6 meters of viewing distance.
Most people observe that the dispersed dot image of Figure 2.42(d) blends to a
noiseless gray at a slightly closer distance than the clustered dot image of Figure
2.42(c). Again, this is a result of shifting some of the noise power to higher
frequencies. Because Robert’s method uses white noise with noise power present
equally at all frequencies, it retains noise power at all viewing distances.
It is convenient to define a single number to use as a comparative index for the
granularity of different images. Of course, a single number index is insufficient
to capture all of the facets of image noise and the visual impact of regular vs.
random patterns. However, a single index of granularity can be useful if we
remember that it is only part of the story.
A simple index of visual granularity is the minimum viewing distance at
which an image blends to a constant gray, as described above. This technique
can be carried out quite well by a single observer, and the results are quite useful
for the comparison of different halftone patterns. This technique was one of the
first quantitative measurements of granularity used by the photographic industry
to measure the granularity of silver halide photographs.30
The most common instrumental technique for measuring the granularity of
printed halftone images is to capture a digital copy of the printed image with a
camera or scanner that has a resolution well beyond the dpi addressability of the
printer. Once the image of the printed sample is captured, a granularity constant
can be calculated by digital image analysis. The most common index of granularity is the standard deviation of the captured image, symbolized σ. However,
more is required than just a captured image. The image capture device should be
calibrated, and it is important to know whether the imaging device is calibrated
to reflectance, R, or to reflection density, D = –log(R) or to some other function
of reflectance. The standard deviations σR and σD are quite different.
Equation 2.25 and Equation 2.26 are used to calculate σR and σD; R and D are
the average values of reflectance and density of the image; Rj and Dj are the
individual reflectance and density at each pixel location, j, in the image; and N is
the total number of pixels in the captured image. The values of σR and σD are
sometimes called root mean squared (RMS) granularity constants.
σR =
∑ (R − R )
Color Desktop Printer Technology
σD =
∑ ( D − D)
Ideal printed samples of all of the halftones in Figure 2.42 have the same
RMS granularity, given by Equation 2.27. This is because the calculation of
granularity does not take into account the filtering that occurs by the human visual
σ R = F ⋅ (1 − F ) ( R p − R i )
To account for the filtering effect, we need an equation for the human visual
system. Equation 2.28 is an approximate equation for describing the spatial
sensitivity function of the human visual system, the dotted line in Figure 2.43(A),
for a 35-cm reading distance. We call this function the visual transfer function
(VTF). Notice that frequency in both the horizontal and vertical directions, ω,ν,
is involved in the complete analysis.
for ω 2 + ν2 > 1 cy / mm, VTF ( ω, ν ) = 5.05 ⋅ e −0.84
ω 2 + ν2
− e −1.45
ω 2 + ν2
for ω 2 + ν2 ≤ 1 cy / mm, VTF ( ω, ν ) = 5
where cy means cycles.
The image has two directional dimension, R(x,y), so the complete Fourier
analysis actually leads to a two-dimensional noise power spectrum, W(ω,ν). The
filter calculation is performed with the integral Equation 2.29. The result, σ V2 , is
the squared RMS granularity that the eye is able to see.
σ V2 =
∫∫ W ( u , i) VTF ( u , i) d u d i
u ,i
Integral Equation 2.29 does two things. First, it is the multiplication of the
noise power spectrum, W, and the VTF. The result of this multiplication is shown
in Figure 2.44 for Robert and Floyd–Steinberg images from Figure 2.42. The
integration process simply calculates the area under the curve. Thus, the square
root of the area under each curve is a measure of the RMS image deviation, σV,
which the eye can detect at the viewing distance shown. Recall that σ from
Equation 2.25 is the same for all four images of Figure 2.42 at all reading
distances. However, σV changes with both reading distance and with the halftone
Image Quality of Printed Text and Images
35 cm
2 meter
1 meter
1 meter
35 cm
5 meter
Frequency, LPI
Frequency, LPI
FIGURE 2.44 Illustration of the attenuation of noise power by the human visual system
at various viewing distances for Roberts and Floyd–Steinberg methods, Figure 2.42(a) and
pattern. Thus, σV correlates with our visual experience of observing different
halftone patterns at different viewing distances.
The filtering process illustrated with Equation 2.29 is an example of noise filtering
in general. Many things can filter out noise, and the process of printing an image
is one example. In the above discussion, a perfect printing process was assumed.
That is, it was assumed that the halftone dots are perfectly sharp edges. However,
the dot-gain effects described previously involve spreading and blurring, and these
effects can reduce the noise power, as illustrated in Figure 2.45.
Another printer effect that often plays an even more important role in the
quality of halftone printing is the stability of the printing process itself. In addition
to blurring the noise that is intrinsic to the halftone pattern, the printing process
can contribute noise of its own. Figure 2.46 illustrates a printing process that
introduces both random noise and a regular banding pattern. Both types of noise
are examples of what is often called printer instability.
Original Halftone Image
in the Computer
Printed Copy
FIGURE 2.45 The printing process can blur the halftone dots and reduce noise.
Color Desktop Printer Technology
Original Halftone Image
in the Computer
Printed Copy
Printer with
Internal Noise
FIGURE 2.46 The printing process can blur input noise and add noise of its own. Random
and banding noise is illustrated.
The degree of instability manifested by the printing process often depends
on the pattern that is printed. Some patterns are easier to print than others are.
In particular, many printing devices show more printer instability when attempting
to print small dots than when printing large dots. For this reason, it is often better
to print a traditional clustered dot halftone than an FM pattern such as
Floyd–Steinberg. This is particularly true for printers capable of addressabilities
above 600 dpi. A 6 × 6 clustered dot halftone would produce a pattern of dots
that is barely visible at an ordinary reading distance. However, a Floyd–Steinberg
image in the presence of significant printer instability may actually be more
objectionable than a uniform, stable clustered dot image. For this reason, FM
halftones are rarely preferred over clustered dots for high dpi printers.
1. P.G. Engeldrum, Psychometric Scaling: A Toolkit for Imaging Systems Development, chapt. 2, Imcotek Press, Winchester, MA, 2000.
2. R.R. Buckley, The History of Device Independent Color — Ten Years Later, Tenth
Color Imaging Conference: Color Science and Engineering Systems, Technologies, Applications, Scottsdale, AZ, November 12, 2002, pp. 41–46.
3. P.A. Crean and R. Buckley, Device independent color: Who wants it? in K. Braun
and R. Eschback, Eds., Recent Progress in Color Science, IS&T, Springfield, VA,
1997, pp. 230–232.
4. R. Brook and G. Arnold, Applied Regression Analysis and Experimental Design,
Marcel Dekker, New York, 1985.
5. W. Thomas, Jr., Ed., SPSE Handbook of Photographic Science and Engineering,
chapt. 17, John Wiley & Sons, NY, 1973.
6. USAF Test Target, U.S. Government specification Mil-Std-150.
7. NPS 1963A Microcopy Target. Also known as NBS 1010A Microcopy Test Chart
and ANSI/ISO Test Chart #2.
Image Quality of Printed Text and Images
8. P.G.J. Barten, Contrast Sensitivity of the Human Eye and Its Effects on Image
Quality, SPIE Press Monograph Vol. PM72, Bellingham, WA.
9. F.D. Kagy, Graphic Arts Photography, Delmar Publishing, NY, 1983.
10. R. Berns, Principles of Color Technology, 3rd ed., chapt. 2, p. 31, John Wiley &
Sons, NY, 2000.
11. R.W.G. Hunt, Measuring Colour, 2nd ed., Ellis Horwood, London, 1995.
12. R.W.G. Hunt, The Reproduction of Colour, Fountain Press, England, 1987.
13. J.A.C. Yule, Principles of Colour Reproduction, John Wiley & Sons, NY, 1967.
14. B. Gascoigne, How to Identify Prints, p. 54, Thames and Hudson, NY, 1991.
15. F. Cost, Pocket Guide to Digital Printing, Delmar Publishers, NY, 1997.
16. H.E.J. Neugebauer, Die Theoretischen Grundlagen des Mehrfarbenbuchdrucks, Z.
Wissenschaft. Photograph. Photophys. Photochem., 36, 22, 1937.
17. P.G. Engeldrum, The color between the dots, J. Imag. Sci. Technol., 38, 545, 1994.
18. E.J. Giorgiann and T.E. Madden, Digital Color Management, Addison-Wesley,
Reading, MA, 1998.
19. J.A.S. Viggiano, Modeling the Color of Multi-Colored Halftones, TAGA Proceedings, TAGA, Vancouver, BC, Canada, 1990.
20. T.N. Pappas, Digital Halftoning Techniques for Printing, IS&T 47th Annual Conference, p. 468, 1994, and references contained therein, IS&T, Springfield, VA.
21. T.N. Pappas, C.-K. Dong, and D.L. Neuhoff, Measurement of printer parameters
of model-based halftoning, J. Electronic Imag., 2, 193, 1993.
22. G.L. Rogers, Optical dot gain: lateral scattering probabilities, J. Imag. Sci. Technol., 42, 341, 1998.
23. J.S. Arney, P.G. Engeldrum, and H. Zeng, An expanded Murray–Davies model of
tone reproduction in halftone imaging, J. Imag. Sci. Technol., 39, 502, 1995.
24. J.S. Arney and M. Katsube, A probability description of the Yule–Nielsen effect,
J. Imag. Sci. Technol., 411, 633 and 637, 1997.
25. J.S. Arney, T. Wu, and C. Blehm, Modeling the Yule–Nielsen effect on color
halftones, J. Imag. Sci. Technol., 42, 335, 1998.
26. L.G. Roberts, Picture coding using pseudorandom noise, IRE Trans., IT-8(2),
145–154, 1962.
27. R.W. Floyd and L. Steinberg, Adaptive algorithm for spatial gray scale, SID Int.
Sym. Dig. Technol. Papers, pp. 36–37, 1975.
28. M.H. Bruno, Imaging for Graphic Arts, Image Processes and Materials, p. 472,
Van Nostrand Reinhold, NY, 1989.
29. M.A. Kriss, Image structure, in T.H. James, Ed., The Theory of the Photographic
Process, 4th ed., MacMillan, NY, 1977.
30. C.E.K. Meese, The Theory of the Photographic Process, 1st ed., chapt. 21, MacMillan, NY, 1941.
The Business and
Market for Desktop
Frank Romano
3.1.1 Epson .............................................................................................89
3.1.2 Canon.............................................................................................89
3.1.3 Hewlett Packard (HP)....................................................................90
3.2 Computer Printing Evolves .......................................................................90
3.2.1 Color Laser Printers ......................................................................92
3.2.2 Thermal Wax Transfer...................................................................94
3.2.3 Inkjet Printing................................................................................95 Continuous Inkjet Printing..............................................99 Technological Developments and Considerations........102 Inkjet Papers..................................................................103 Desktop Inkjet Gets Down to Business........................105
3.3 The Printer Market ..................................................................................105
3.3.1 Market Sizing ..............................................................................105
3.3.2 Conditions in Motion ..................................................................107
3.3.3 Forecast........................................................................................107
References .........................................................................................................107
In 1878, James Clephane, a stenographer and developer involved in the typewriter
and the linotype typesetter, said, “I want to bridge the gap between the typewriter
and the printed page.” 1 There is no doubt that the current breed of digital printers
provides a performance level equal to low-level printing presses. The concept of
desktop publishing is based on the ability to compose typographic pages and
output them at a quality level competitive with the graphic arts. The typewriter
went down one road to business communication and the linotype went down
Color Desktop Printer Technology
another road to the printing industry. It would be the digital printer where both
roads would meet.
The desktop printer market began because Robert Howard wanted to personalize poker chips. He sought a method to print around the periphery of a gambling
chip and came up with the concept of impacting the surface of the chip through
a ribbon with small pins that would image the letters, very much as a television
image was composed of raster lines. (Howard had been a pioneer in TV manufacturing.) He and Dr. An Wang became partners in a company to market an early
digital computer co-invented by both. When they decided to separate, Wang
formed Wang Laboratories and Howard founded Centronics Data Computer Corporation in 1969. Howard developed the first dot matrix printer, and Centronics
became one of the world’s leading computer printer manufacturers. Today’s
popular laser printers originated from a subsequent Canon/Centronics joint development program in which Mr. Howard served as a seminal innovator. Howard
served as chairman of Centronics from 1969 to 1982, when Control Data Corporation acquired a controlling interest.
The Centronics parallel plug became the single common denominator in the
computer world. Howard tells the story of how it came to be. “Dr. Wang had
ordered a large number of these connectors for another project. Since there were
so many we decided to use it for the printer cable. The rest is history.” The desktop
dot matrix printer was an important enabler in the evolution of desktop computing.
A computer printer produces hard copy (text or graphics usually on paper) from
data stored in a computer connected to it. The first of the printers were connected
via telephone lines to mainframe computers. They were used by hotels, airlines,
rental agencies, and other organizations.
As mainframe computers, and then minicomputers, came into use, there was
a need to print their output. The first computer printers were typewriter adaptations. They evolved into higher- and higher-speed printers, all using impact and
ribbon technology. These printers may be broadly characterized as:
Character printer
Typewriter printer
Daisywheel printer
Chain or band printer
matrix printer
Character matrix printer
Line matrix printer
Page printer
Thermal printer (used in fax devices)
There are two major mechanisms used for dot matrix printing:
Ballistic wire printer
Stored energy printer
The Business and Market for Desktop Printers
Ballistic wire computer printers have a printhead with holes drilled through
it. Thin wires go through the printhead and pawls are actuated by solenoids. The
pawls strike the wires, causing them to go out and hit an inked ribbon, which
hits the paper. After about a million characters, even with tungsten blocks and
titanium pawls, the printing becomes too unclear to read. The print resolution
ranges from a low of about 50 dots per inch (for receipt printers) to a high of
about 300 dots per inch (better than the human eye can see at 14 inches), used
on premium graphic printers.
A stored-energy printer is a type of computer printer that uses the energy
stored in a spring or magnetic field to push a hammer through a ribbon to print
a dot. These printers print millions to billions of dots per hammer. The most
common printers to use this have been the line-matrix printers made by Printronix
and its licensees. In these, the hammers are arranged as a hammerbank, a sort of
comb that oscillates horizontally to produce a line of dots. In a character matrix
printer, the hammers are machined from an oval of magnetically permeable
stainless steel, and the hammer tips form a couple of vertical rows. The original
technology was patented by Printronix in 1974.
Although character printers provided higher quality, most development was
applied to dot matrix printers. 5 × 7, 9 × 14, 11 × 16, and other matrices evolved
to increase the quality and allow a larger character set. In 1979 the Sanders allpoints addressable printer tried to bridge the gap between the limitations of
mechanical pin technology and the emerging world of non-impact printing.
Throughout the 1970s, the quest for higher speed, larger character sets, and
the ability to handle graphics led to the development of non-impact printing
systems. These can use any of the following technologies:
Laser (toner) printer
Inkjet printer
Dye-sublimation printer
Thermal printer
The introduction of the desktop computer in the late 1970s accelerated the
need for desktop printers. Most were still impact printers at this time. Printers
were available for use with home computers at prices of more than $1000. Most
printers offered 96 characters in the standard American Standard Code for Information Interchange (ASCII) set. There were two main types of printers available:
character printers and dot matrix printers.
Character printers operated like a typewriter by striking a piece of metal type
against a ribbon and onto the paper. This type of printer was often called an
impact or letter quality printer. It used either a type ball, such as IBM’s Selectric
typewriters, or a wheel with spokes that radiated out from the center, with the
type characters at the end of the spokes, also called a daisy wheel.
Dot matrix printers formed characters with a series of pins in a vertical row
that struck the ribbon and produced dots on the paper. As the printhead moved
across the paper, the dots were printed in patterns that formed letters and numbers.
Color Desktop Printer Technology
The matrix used to form a character was usually referred to as the number of
horizontal dots by the number of vertical dots. A 5 × 7 matrix, for example, used
up to five dots across and up to seven dots down.
The Centronics 730 was the first standard printer. It used a parallel cable
whose pin layout went on to also become a standard for use with personal
computers. That pin layout on parallel cable plugs is still in use today. Centronics
also had several other models, including the 737 and 739. A less expensive printer
made by Centronics, the 779, used 5 × 7 dot matrix characters and could print
in sizes from 10 to 16.5 characters per inch (cpi), ranging from 60 characters per
second (cps) at 10 cpi to 100 cps at 16.5 cpi. It also had a one-line buffer (which
held up to 132 characters) but printed a limited 64-character ASCII set, all
uppercase plus some special characters.
A company named Trendcom used mechanical solenoids that drove heated
pins into a printhead — these were thermal printers that needed a special heatsensitive paper. Their operation was very quiet, about as loud as sliding a finger
across a piece of paper. The Model 100 printed 40 characters per line on paper
that was about 4 1/2 inches wide. The Model 200 could print 80 columns per
line on paper 8 1/2 inches wide. The first printer offered by Radio Shack for its
TRS-80 computer also used a thermal printer.
Epson began in the printer business with the Epson MX-80, one of the first
dot matrix printers that sold for less than $1000. A later version of this printer,
the Epson MX-100, became available in early 1982. The MX-100 was a wide
carriage model and could print high-resolution (hi-res) graphics without the
need to add any special hardware. Epson printers were unique because they
had a special feature called a double print mode in which a line was printed
normally and then the paper was advanced 1/216 of an inch and the same line
was printed again. This filled some gaps between dots on individual letters and
made printouts more pleasing to the eye. In print enhancement mode, the pins
hit the ribbon harder and made it possible to make multiple copies using carbon
After the Silentype printer was released in 1979, Apple looked for another
printer that would produce better, more permanent output than could be achieved
with a thermal printer. One of the main problems with thermal paper was that,
with time, the printing could fade. The Apple Dot Matrix Printer was released in
1982 for $699. Made from a modified C. Itoh printer, it was one of the first dot
matrix printers that sold for under $1000. The Apple ImageWriter, released in
1983, was also made by C. Itoh; it had a faster print speed (120 cps) and could
print in eight different pitches (character widths). It was very reliable and sold
originally for $675. Later, a wide carriage version whose abilities were otherwise
identical was made available. It was replaced by the ImageWriter II in September
In 1984, Hewlett Packard introduced the LaserJet laser printer based on the
Canon LBP-CX 8-ppm (pages per minute) 300-dpi (dots per inch) engine. This
was a significant breakthrough in printer quality and was capable of producing
documents that looked professionally typeset. Apple decided to develop its own
The Business and Market for Desktop Printers
laser printer and in 1985 released the LaserWriter. The LaserWriter was supported
only on the Macintosh and did its work through a page description language
called PostScript. Apple entered the inkjet printer market in 1991 when it released
the abortive Apple StyleWriter.
3.1.1 EPSON
Epson was established in 1961 under the name of Shinshu Seiki (Shinshu Precision Manufacturing Company) to provide precision parts for Seiko watches. The
company was awarded a contract to make precision timers for the 1964 Olympics
and also picked up work to build a printer. Thus, the EP-101 printer came about,
and in 1968 it became one of the first printers for electronic calculators to hit the
commercial market. Until the 1980s, approximately 100,000 of these printer
mechanisms were being produced each month — about 90% of the world market.
Electronic watches using relatively high-current light-emitting diodes first
appeared on the market in 1970, and Shinshu Seiki began researching a lowcurrent alternative. It came up with the liquid-crystal display in 1974, which, in
turn, led to LCD watches. Epson America was incorporated in 1975, with offices
set up in Torrance, California, to distribute Shinshu Seiki products.
Epson’s first dot matrix printer, TX-80, was introduced in 1978. However,
this didn’t attract much attention except from Commodore, which used it as the
system printer for its PET computer. (The “80” in “TX-80” refers to the number
of columns it printed per line.) An improved version, the MX-80, began development later that same year. The TX-80 took 3 months to develop. The MX-80
took about 2 years, was introduced in late 1980, quickly became the best-selling
printer in the United States, and eventually became the industrial standard for
microcomputers. This occurred despite its being designed not to produce graphics.
Within a year, the Graftrax version, with graphics, had hit the streets.
3.1.2 CANON
Canon had the advantages of large investments and experience in copier technology. The company was able to leverage this base into fax machines, copiers, and
later printers. In the latter half of the 1970s, engineers at Canon’s Product Technology Research Institute conducted research on printing technologies for the
next generation of copying machines. Initial work was devoted to producing
piezo-elemental data necessary for inkjets, but this pursuit led to the discovery
of a new technology. During testing, a hot soldering iron accidentally touched
the needle of an ink-filled syringe, causing ink to spray from the needle’s tip.
Witnessing this, a member of the research team realized that heat, instead of
pressure, could perhaps be used to induce the spray of ink. In 1997, technical
concepts born of this discovery were combined with thermal-head technologies
under development at the time. This enabled the number of ink nozzles to be
multiplied, a notion that was previously inconceivable, opening the way to a new
high-speed printing technology.
Color Desktop Printer Technology
Numerous tests and refinements were made until Canon succeeded in developing the world’s first Bubble Jet printing method in 1981. Canon continued to
make refinements, finally unveiling the BJ-80 Bubble Jet printer 4 years later.
This huge technological development was the product of 8 years — the period
from the discovery of the initial principle to commercialization.
HP became a major player in the computer industry in the 1980s with a full range
of computers, from desktop machines to portables to powerful minicomputers. HP
made its entry into the printer market with the launch of inkjet printers and laser
printers that connect to personal computers. The HP LaserJet printer line debuted
in 1984 to become the company’s most successful single product line ever.
In the 1980s, HP fueled the desktop publishing revolution with the introduction of high quality and reliable laser and later inkjet printers that connect to
personal computers. The LaserJet was its most successful single product line ever
and the DeskJet inkjet printers spelled the death of dot-matrix printers.
In the 1990s, HP became a follower in the desktop printer market because
the company was slow with a color LaserJet, it used its own page description
language instead of PostScript, and its printers were expensive. The desktop
printer market is extremely competitive, and Epson and Canon gained ground in
the inkjet market at the expense of HP.
HP announced a $1.2 billion investment in the imaging and printing market,
including a planned roll out of 50 new products, in 2003. In 2005, HP announced
a $2 billion investment in new printer technology and over 100 new products.
HP is capitalizing on the ever-expanding digital photography imaging market.
According to analysts, billions of digital images will need to be printed and inkjet
printing will be the technology of choice.
HP has finalized its acquisition of Indigo N.V., an Israel-based manufacturer
of specialty and commercial digital color presses. The digital printing market has
been maturing for a decade, and given its successes with LaserJet technologies,
HP would be a logical contender in the arena. By acquiring Indigo, which has
established digital presses and an established customer base, the company has
rolled out a centralized workflow management tool called HP Production Flow
as well as a Web-based integrated publishing tool called HP Custom Publishing.
In the early days of the personal computer (PC), printing was simple. The PC
owner bought a cheap printer, usually a dot matrix that minimally supported
ASCII, and plugged it into the computer with a parallel cable. Applications would
come with hundreds of printer drivers to output DOS (Disk Operating System)
or ASCII text. The few software applications that supported graphics generally
could only output on specific makes and models of printers. Shared network
The Business and Market for Desktop Printers
printing, if it existed, was usually done by some type of serial port (A-B boxes)
When the Windows operating system was released, application programmers
were finally free of the restrictions of how a printer manufacturer would change
printer control codes. Graphics printing, in the form of fonts and images, was
added to most applications, and demand for it rapidly increased across the corporation. Large, high-capacity laser printers designed for office printing appeared
on the scene. Printing went from 150 dpi to 300 dpi to 600 dpi for the common
desktop laser printer.
Inkjet printers have become widespread only in the past decade, yet the
technology has been under development for more than 50 years. Inkjet recorders
appeared as early as 1950 and inkjet typewriters in the 1960s. In the 1970s
virtually every major printer manufacturer invested in inkjet development in an
attempt to replace impact (matrix) technology. Reliability and print quality have
long been issues because it is difficult to control the ink flow and to prevent the
ink from drying and clogging the printhead. The print quality depends heavily
on the complex relationship between ink, printhead, and receiver material. By
the end of the 1980s, Canon and Hewlett Packard had mastered both the ink
chemistry and the hydrodynamics required. In 1998, the total narrow format
coated inkjet media market was good for approximately 360 million square
meters. Today it is ten times that.
Liquid inkjet printers generally fall into two classes: continuous and dropon-demand. In a continuous inkjet printer, a continuous spray of ink droplets is
produced and the unneeded droplets are deflected before they reach the paper.
Continuous inkjet technology provides high-speed drop generation at 1 million
drops per second or faster. Two classes of continuous inkjet products are available:
High-speed industrial printers for carton and product marking, addressing, on-demand short-run printing, and personalized direct mail.
Proofing printers for verification of materials prior to printing on printing presses. These printers offer the best print quality of any nonphotographic device but are much slower (less than an inch per second).
Resolutions are not high (e.g., 300 dpi) but variable-sized dots make
photographic quality possible.
Drop-on-demand inkjet printers produce ink droplets when needed. The two
technologies to drive the droplets out of the printhead are thermal (used by HP,
Lexmark, Canon, Olivetti, Océ, and others) and piezoelectric (used by Epson).
Thermal printers have been successful because they are inexpensive. The biggest
challenge for piezoelectric inkjet technology is the cost and difficulty of producing
printheads. Today Epson has a very successful line of piezoelectric color printers,
namely, the Stylus color and Stylus photo family of printers offering photographic
Color Desktop Printer Technology
In 1938, Chester Carlson, a patent attorney and a graduate of Caltech, discovered
a dry printing process called electrophotography. For 9 years, Carlson tried to
sell his idea to more than 20 companies including RCA, Remington Rand, General
Electric, Eastman Kodak, and IBM. They all turned him down, wondering why
anyone would need a machine to do something you could do with carbon paper.
Enter Haloid, which became Xerox and the rest is history.
In 1969, Gary Starkweather, at Xerox’s research facility in Webster, New
York, used a laser beam with the xerography process to create a laser printer. In
1979, IBM introduced the IBM 3800 laser printer, capable of printing 20,000
lines per minute.
In 1978, Xerox introduced the 9700 laser printer. This was the first laser
printer commercially available in the United States and in the world. It could
output 120 pages per minute. It is still the fastest commercial laser printer.
However, the 9700 was physically too large and carried a large price tag as well.
In 1984, Canon launched the LBP-CX laser beam printer. Having many
unique advantages over other processes, xerography was adopted for computer
output printing. Xerox was the forerunner in this endeavor.
In 1984, HP marketed the LaserJet printer (8 pages per minute). The major
feature of this printer was its use of an operator-replaceable all-in-one toner
cartridge. The entire development subsystem was built into this toner cartridge.
Canon had used this concept in a line of desktop copiers.
The first desktop color laser printer was the QMS Colorscript 1000 in 1993,
selling for $10,000 and based on a Hitachi engine using a dual-component toner
design. Only a few thousand machines were sold. Annual shipment volumes of
color lasers have steadily increased since then. In 1995, QMS cut the price of
the Magicolor LX to $4995. By late 1997, QMS was offering the PostScriptenabled Magicolor 2 CX for $3500 and, by early 1998, the Windows-only Desklaser 2 for $2500.
Two engines dominated the installed base of color lasers: the Konica engine
in the original HP Color LaserJet and LaserJet 5, and the Canon EP-H engine
used in many models, including the Lexmark Optra C. In 1997, QMS introduced
the Magicolor 2 line based on a Hitachi engine and Minolta brought a new engine
to market. The Minolta Color Pageworks saw models based on this engine priced
under $2000. In 1998, Tektronix introduced the $1800 Phaser 740 based on a
Matsushita (Panasonic) engine. Canon introduced two new engines sold by HP
as the Color LaserJet 4500 and 8500.
Pricing for color printers has fallen as volume drives manufacturing costs and
as shipment volumes increase; cost is lowered and the end-user price is lowered,
which drives shipment volumes even higher. Cost per page has constantly dropped
as original equipment manufacturers (OEMs) adjust the pricing of their supplies
so that color is finally affordable for corporate buyers. Monochrome printing
costs for color lasers are an issue. The cost of color is only a little more than the
The Business and Market for Desktop Printers
cost for monochrome, and OEMs are positioning color lasers as economical as
monochrome-only printers.
Quality continues to improve and speeds increase with each generation of
engines. Color laser generations are shorter than monochrome generations. This
is driven by a combination of declining printer pricing, declining costs per page,
higher print speeds, and improved quality. These rapid advancements make
engines obsolete more quickly than in the monochrome market. Every OEM
desires to establish its brand in the color laser printer market, and no one has
achieved in color laser printers what HP/Canon achieved in monochrome laser
printers. The color laser market is a chance to command a significant market
share of machine shipments that will drive long-term profitability in consumable
supplies. HP probably will not dominate the color laser printer market to the level
it did in the monochrome market. From the late 1980s to early 1990s, HP held
more than 80% of the market share in monochrome lasers. In 2000, HP held
about 50% of the market share in color laser printers.
The value of supplies for color printers is significantly higher than monochrome, and OEMs understand that the path to higher profits lies in increased
market share. Driving this move toward color lasers is the closing gap between
the pricing of monochrome-only printers and color printers. As the price and
speed of color lasers become comparable to monochrome lasers, users perceive
that they are getting color for free if they buy the color laser.
Because the color laser prints both monochrome and color documents, essentially doing the work of two printers, when the purchase prices of color printers
drop to about 150% of the cost of comparable monochrome systems, you get
color for free. The market for letter/legal size–format color lasers has broken
through this 1.5× price point level for several models. Monochrome lasers (letter/legal size) with print speeds of 16 ppm to 20 ppm range in pricing from $1000
to $1500. Pricing levels at 1.5 times this range would be $1500 to $2250. Entrylevel color lasers offering monochrome print speeds from 12 ppm to 16 ppm are
priced from a low of $1300, with many models in the $1500 to $2000 range, and
HP’s Color LaserJet 4500 sells for $2200 to $2300. Color is becoming a viable
option for multiple business scenarios.
In 2002 prognosticators consistently overestimated how quickly the corporate
market would adapt color lasers. Annual U.S. shipment volumes for color lasers
range from low-side projections of 800,000 units to high-side projections of
1,800,000 units for 2002 and 3 million units in 2005. The large range in future
shipment projections is due to uncertainty in the speed of color adaptation by
businesses. If the color laser market does meet these high-side projections, the
overall installed base of color lasers will explode by a factor of ten over the next
3 years compared to the current installed base. There are more than 20 desktop
color laser engines in the market, and additional engines are being released as
this market accelerates. The gap between HP/Canon and the rest of the pack is
much smaller in the color laser printer market than it was in the early years of
the monochrome laser.
Color Desktop Printer Technology
Color lasers use the same fundamental electrophotographic steps as monochromes. The primary difference between monochrome and color is how to
perform this process four times for a single sheet of paper, once for each of the
four colors of cyan, magenta, yellow, and black (CMYK). All monochromes
directly transfer the toner from the organic photoconductor (OPC) device (drum
or belt) to the paper. Color lasers use this direct transfer design and some also
utilize an intermediate transfer mechanism.
Another variation in color laser printer design is in how the four process
colors are positioned to develop toner to the OPC device. Some use fixed positions
for the four toners, whereas other designs rotate the colors on a carousel device.
The majority of color lasers build a complete four-color image either on the OPC
device or the intermediate transfer device. It is then transferred to the paper in a
single operation or pass. An exception is the Optra C (Canon P320/ EPH) design,
which transfers each of the four colors in succession to the paper.
The toner composition of most of the color lasers introduced since 1994 have
been a monocomponent, non-magnetic design. Image development is accomplished via electrostatics. The Konica (HP Color LaserJet), introduced in 1994,
was a dual-component design. A recent development has been the use of monocomponent, magnetic black toner in combination with monocomponent nonmagnetic cyan, yellow, and magenta toners, first introduced in the HP 4500.
In color laser printers, there are at least eight replacement supplies: four toner
units and an OPC unit, plus a transfer unit, fuser oil, fuser cleaning units,
maintenance kits, multiple charging units, four developer units, and other enginespecific replacement units. Monocomponent toner designs reduce user replaceable units.
Since they were introduced by QMS in 1988, thermal wax printers have captured
a significant share of the low-end color output market, in part because of falling
prices. The first such devices were in the $30,000 range, whereas recent thermal
wax printers are less than $2,000. Among the leading vendors of thermal wax
printers are QMS and Tektronix.
Thermal wax transfer is a printing process that transfers a waxlike ink onto
paper. A Mylar ribbon is used that contains several hundred repeating sets of full
pages of black, cyan, magenta, and yellow ink. The ink applied in the thermal
wax transfer methods is solid at normal temperature. A sheet of paper is pressed
against each color and passed by a line of heating elements that transfers the
spots, or pixels, of ink onto the paper. The back of the color ribbon is in direct
contact with the heated surface of the thermal head, which reaches temperatures
of 350°C. The image is built up of the three subtractive primary colors, cyan,
magenta, and yellow, with black as an optional extra. The colors are carried on
the ribbon as sequential panels, and in the printing process, the ribbon moves
forward continuously while the receiver is recycled underneath it, so that the
The Business and Market for Desktop Printers
whole image is written in yellow, then in magenta, and finally in cyan. The
principle of the thermal transfer is:
1. Send print paper together with an ink ribbon into the space between a
head and a platen roller.
2. Run electric current into the thermal head to heat it up.
3. The head and roller put pressure on the print paper and ink ribbon as
they go through.
4. The heated thermal head fuses the solid ink applied on the ribbon into
5. The liquid ink is transferred from the ink ribbon to the paper and
absorbed by the paper.
The amount of ink absorbed by the paper is controlled by the amount of
electric current running into the thermal head. By controlling the amount, gradation reproduction by ink density is possible in this printing method. The major
advantages of thermal wax printers include their relatively low cost and the high
opacity of the wax, which makes them ideal for creating overhead transparencies.
The disadvantages of some thermal wax printers include speed and quality, plus
the need for special paper.
These color printers were used primarily for proofing design and artwork in
the graphic arts market.
The principles of inkjet printing have been known for hundreds of years, and
inkjet devices have been constructed for more than 100 years. However, the
technology has only been applied commercially since 1970. Since then, datecoding requirements and the move toward a databased society have driven development in this area. Inkjet printing is a form of non-impact printing. The first
inkjets were created in the 1970s by Dr. C. Hellmuth Hertz, a physics researcher
at the Lund Institute in Sweden. Inkjet printers have become increasingly essential
in the wake of desktop publishing because of the great demand for the highquality printers for text printing and full color printing. There are two kinds of
inkjet printing: drop-on-demand (DOD) and continuous. Continuous inkjet printers have an advantage over DOD printers because of their ability to run at high
speed; DOD printers produce high-quality images that closely resemble those of
a photograph.
Inkjet printing has a distinction that sets it apart from most other printing
technologies: it is a plateless process. The general principle behind inkjet is that
the ink is sprayed onto the paper. The device that sprays the ink onto the page
is generally referred to as a head. Inkjet systems can be expanded (increasing the
size of the paper or their speed) by increasing the number of heads they print
with. The way the ink is sprayed onto the paper is generally how the different
types of inkjet systems are classified. One approach is to use a constant stream
Color Desktop Printer Technology
of ink from the jet known as continuous systems. These systems use water-based
A second approach is to send ink through the jet only if a dot is meant to be
on the page; these systems are known as DOD. Within the DOD family, the
systems can be divided further into what type of inks they use. Most DOD systems
use water-based inks, but some use inks that are solid at room temperature. These
systems are referred to as phase-change printers because the ink starts as a solid,
is melted, and then solidifies on the page. Continuous inkjet, as the name would
suggest, uses a continuous flow of ink. With this constant stream of droplets, the
system simply has to control whether or not they hit the page. This task is
accomplished with the use of electric forces. As the droplet leaves the orifice, a
charge is placed on it. The drop then travels between two oppositely charged
plates. The charged drop is attracted to one plate and repelled from the other and
its path is controlled. The drop either travels to the page or to a gutter that returns
it to the ink reservoir. An advantage to this system is that the continuous flow
does not allow ink to dry and clog the jet.
DOD printing devices control whether or not the ink hits the page by controlling whether a droplet is formed or not. The ink is contained in a reservoir
and a force is required to push it out onto the page. Some systems use heat and
others use the deformation of piezoelectric crystals. In systems that use heat to
generate the droplet, the ink supply is heated until a portion of the ink vaporizes
and this expansion within the reservoir forces a droplet out of the reservoir onto
the page. In systems that employ a piezoelectric crystal, an electric current is run
through the crystal, which causes it to change shape. This change of shape in the
crystal forces ink from the jet onto the page.
Another form of DOD inkjet printing is known as phase change inkjet. In
phase change inkjet the ink starts out as a waxy solid, is melted by the printhead,
and returns to a solid state once it reaches the page. The system used to spray
the ink onto the page is similar to the DOD process; the only difference is the
characteristics of the ink. In phase change inkjet, the fact that the ink becomes
a solid when it hits the page gives the process some unique advantages. This ink
can be used with a very broad range of substrates, because it bonds with the
surface instead of depending on absorption. The colors of these inks may be
vibrant because of their position on the paper.
There is a tremendous variety of inkjet printers available. This variety includes
systems with different qualities, speeds, and prices. The simplicity of the principles
behind inkjet is what leads to its many different forms. Basically, all inkjet does is
spray ink onto paper. There are almost limitless ways to do this. The ink, paper,
number of heads, size of jets, type of ink, and configuration of the printer are only
a few of the options that have already been explored. An example of some of the
variations can be seen in the continuous process. Some continuous inkjet printers
use large droplets, producing low-resolution results but at a very fast rate, and others
use very tiny jets and produce photo-quality images. The configuration of the printer
may also be specialized. The paper, heads, or both may move as the image is written.
Some printers are only large enough to address an envelope; others are large enough
The Business and Market for Desktop Printers
to print a billboard. The freedom from having a fixed-image carrier and the ability
to be configured in so many ways are inkjet’s greatest assets.
Inkjet technology encompasses a wide spectrum of systems — at one end of
the market is a Canon desktop color printer that challenges Epson’s position as
the quality inkjet supplier at under $500. At the other end of the market, there is
the Scitex Digital Printing’s (now Eastman Kodak) VersaMark press for ondemand book printing at 3,800 book pages per minute at more than $1 million.
Also consider Agfa’s agreement with Xaar to develop Xaar’s page-wide inkjet
printheads for potential use in future Agfa digital color presses, and it appears
that inkjet is on the move. Inkjet technology is an established technology for
much of the digital contract color proofing market and is found in products such
as the Creo Iris (now Kodak Veris) proofers and the DuPont Digital Cromalin.
The concept proofing markets, with products from Canon, Epson, Hewlett
Packard and Lexmark that sell for less than $2000, are likely to take over the
proofing role currently filled by dye-sublimation products and the contract proofing inkjets mentioned above. The quality of these devices is exceptional for their
price and getting better.
Wide-format inkjet units from companies such as Encad, Hewlett Packard,
ColorSpan, Raster Graphics, Roland, and Xerox are changing the point-of-sale
and sales display markets, and impacting technologies such as screen printing.
They are also providing full-color imposition proofs to help speed the acceptance
of computer-to-plate in the marketplace.
For the very highest speed, there is continuous inkjet, which uses an array
of heads, as in the Scitex VersaMark system. This technology is just now reaching
600 dpi. The technology for the future may be DOD inkjet using piezo or thermal
technology. This is already operating at resolutions of up to 1440 dpi, with
individual drop sizes comparable to that of film imagesetters. In the low-cost
desktop printers, these units have heads with a limited number of inkjet nozzles,
and the head is moved across the paper to print a line. This gives high quality
but restricts the speed. In desktop color proofing, this is not a problem. In
applications using technology from Agfa and Xaar, speed and quality are required,
and page-width printheads of either 9 inches or 12.6 inches, each with thousands
of nozzles, are necessary.
For such applications, it is anticipated that there will be four printheads, one
for each process color, printing in a continuous mode on one side of a sheet of
paper. Using a page-width printhead allows a significantly greater number of
droplets of ink to be generated per second. The potential is to develop a far
simpler 100-page-per-minute-plus digital color press than can be achieved using
electrographic toner-based technology.
Piezo inkjet technology is a very simple process compared with the complexity of continuous inkjet technology. The difference is speed. Large-format inkjet
printers will become the standard proofing devices for printers. A number of
large-format engines and desktop devices use similar technology (such as the
new Epson 9000), and large-format devices could also be used for contract-quality
color proofing. Very-large-format printers from Matan, Idanit, and others for
Color Desktop Printer Technology
billboard and poster printing will impact lithographic- and screen-based technologies. For high-speed digital color printing, inkjet will challenge the position of
the electrophotographic systems such as those from Indigo and Xeikon and
forthcoming systems from Xerox, NexPress, and others.
Canon’s bubble jet, desktop color printer challenges Epson’s position as the
dominant printer vendor in the low-cost color proofer market. The A-3 size, 1200dpi printer called the Aspen first appeared more than a year ago when it was
previewed as a concept printer at Comdex. It appears to break through in printing
technology with its six colors plus coating capability. The BJC-8500 uses Canon’s
bubble jet inkjet technology; Epson uses piezoelectric inkjet technology. The
printer also features a Canon-trademarked technology called MicroFine Droplet
Technology. This technology includes a new printhead design and enhanced ink
formulations that make a microscopic drop size possible at the unit’s maximum
output resolution of 1200 × 1200 dpi. This technology ensures that each drop not
only is consistent in size but also is positioned with precision accuracy and high
The BJC-8500 does not lose printing speed to achieve this quality. The new
printhead has 1536 nozzles (which include among others 256 for all colors, 512
for black, and 256 Ink Optimizer nozzles). The volume of nozzles allows the
head in a single pass to print two or three lines of text, depending on the type
size. The printhead is made up of two modules and works in two print modes.
In CMYK mode, one cartridge has the cyan, magenta, and yellow capability and
the other houses the black and print optimizer.
In photo-imaging mode, the black and print optimizer cartridge is removed
and a second three-color cartridge is installed (light cyan and magenta, plus
black). In every case, all cartridges are fed by separate independently installed
ink tanks, one for each color. Individual ink tanks can be replaced as they run
out, instead of replacing an entire cartridge.
A key feature of the BJC-8500 printer is Canon’s P-POP (Plain-Paper Optimized Printing) system. This treats the print surface of the paper with a clear
liquid solvent, Ink Optimizer, just milliseconds before the ink is ejected from the
printhead. The feature, in effect, transforms the plain paper into coated paper for
water resistance, sharper detail, and more vibrant colors. If the unit is printing
onto a coated stock, then the six-color photo-imaging mode will give a wider
gamut for printing without the need for the Ink Optimizer. The sheet feeder of
the unit handles stocks up to a maximum of 13 × 19 inches, allowing an oversize
A3 page to be printed with full bleed.
The result on a high-gloss stock using six-color printing is outstanding. The
1200 × 1200 dpi resolution and MicroFine Droplet Technology provide very fine
control, probably greater than Epson’s 1440 × 720 resolution, which has a larger
droplet size.
Inkjet printers have a wide range of applications in the printing and packaging
industries. Inkjet printers can be used for marking products with dates, such as
“best before,” as well as coding information such as prices and product tracking.
There are hundreds of different products that can be coded using inkjet printers:
The Business and Market for Desktop Printers
food and beverage containers, cosmetics, pharmaceuticals, electronic components, cables, wiring, PVC and P. E. pipes, glass and plastic bottles, and industrial
components. Ticket numbering and high-speed addressing for magazines and
direct mail are just two of the applications in the printing industry for high-speed
inkjet printing.
The Hertz technology that created inkjet printing is known for the high-quality
images it can create. The main reason that the quality is so high is that the inkjet
can produce true halftones, such as different gray levels, or color tones, which
can be generated with every single pixel. This true halftone printing is achieved
by varying the number of drops in each pixel. The number of drops may vary
from 0 to about 30 for each color, which means that a number of different density
levels for each pixel and color may be obtained. It is possible to increase the
number of density levels per pixel from 0 to 200 for each color. Continuous Inkjet Printing
Continuous printing speed is approximately 45 square inches per minute or even
higher. Since the introduction of the continuous inkjet printer, there have been
several modifications and improvements to the system. The two main concerns
of the early inkjet printers, nozzle clogging and uncontrolled ink mist, have been
addressed by the creation of the IRIS continuous inkjet printer and others.
After each cycle is complete and the nozzle has stopped firing, the nozzle
tips are vacuumed to remove any residue ink. An automatic nozzle maintenance
cycle is built into each system. When the printer is not in the print mode, the
system powers up on a timed cycle and fires ink through the nozzles for a few
seconds, shuts down, and vacuums the tips again.
Uncontrolled ink mist is a result of the reaction between the dropping ink
and the printing surface. Ink droplets are forced out of the nozzle at about 650
pounds per square inch of pressure. This means that the droplets travel at about
30 mm from the nozzle tip to the print surface at a speed of 20 meters per second,
or 50 miles per hour. The mist develops from the millions of drops that are hitting
the paper every minute. A mist shield was created to control random ink spots.
The mist shield consists of an absorbent material positioned near the substrate
printing surface that catches the ink as it bounces back toward the ink nozzle.
This allows for a clean print surface with fewer random background spots as well
as clean internal surfaces.
In the IRIS system, there is a print resolution of 240 × 240 dpi with lateral
printing speeds of about 1 or 2 inches per minute. The IRIS 2044 is a largeformat system, and images covering the 34- × 44-inch maximum size can be
created in about 30 minutes. There is also an IRIS 2024, a medium format that
can produce images covering a maximum of 24 × 24 inches in a printing time
of 15 minutes. The substrate is manually sheetfed so that the printing sequence
can always be varied.
The quality of the color image produced on an IRIS system, in terms of the
resolution and the color shades that are possible, depends on the number of
Color Desktop Printer Technology
separate color dots per square inch that compose the final page. This technology
uses more than 230,000 dots per square inch to compose a full-color image at
240 dpi per color.
Continuous inkjet printers are used for a variety of useful applications, including bar coding, pharmaceuticals, and food packaging. They have a wide range
of specifications that make them very flexible for printing on various substrates
of all shapes and sizes. Bar coding seems to be one of the most popular applications of continuous inkjet printers. They can produce medium-density codes
and are able to print up to 17 mm high, which means that the bar codes produced
may only be read in what is called a closed system.
Closed systems are controlled systems in which reading or scanning devices
are necessary, such as the scanners found in grocery stores. Bar-coded products
can be scanned by these systems. If alphanumerics are going to be printed under
a bar code, the option exists of using an inkjet with two printheads: one to print
the bar code and one to print the alphanumerics. In general, inkjet printing is
used to print on the product directly. The unique construction of the jet valve
allows for printing on non-absorbent materials such as PVC tubing, shrink films,
or ceramic tiles.
Scitex Digital Printing has a digital printing system called VersaMark for
books capable of 3800 book pages per minute at 300 × 600 dpi. It uses continuous
inkjet on a web printer at three times the resolution of the 240-dpi earlier models.
The speed is achieved by printing three 6- × 9-inch pages side by side on a 20inch web. For 8.5- × 11-inch pages, the rated speed is 2100 ppm.
There are two 9-inch imaging heads, each with 2600 inkjet nozzles for
monochrome printing, but the system is also applicable for spot-color applications. Binding is a separate step. Scitex is not attempting to raster image process
(RIP) incoming jobs at the speed of the imager. The RIP supplier is Varis and
portable document format (PDF) is the standard format for input to the RIP. The
cost per page is less than half a cent, or about $1.30 to print a 300-page, 8.5- ×
11-inch book, one of the least expensive production approaches. The VersaMark
is priced at $800,000 to $2 million, based on configuration, with the first customer
site in place.
In contrast to continuous inkjet printing, in which the droplet stream is
continuously expelled from the nozzle (with the unwanted drops caught and
recycled into the ink source), DOD forms and expels droplets only at the moment
needed. DOD printers provide a larger letter size, the ability to use more than
one printhead, and a less critical viscosity requirement for the ink. Most use
multiple nozzles. Each system has limitations because of the inter-relationship
between image height, image quality, and print speed, determined by the size of
ink drops and the rate of production. There are three subtypes of DOD: piezoelectric liquid, bubble jet/thermal liquid, and solid ink. In piezoelectric crystal
inkjet printing, stress on the piezoelectric crystal produces an electric charge that
causes the droplet to be expelled. The advantages of this method include reliability
and the potential for high speed. The disadvantage is the relatively high cost of
The Business and Market for Desktop Printers
manufacturing. These systems are sometimes supplemented with a jet of air to
impel the ink drop.
The bubble jet printer is a development of HP and Canon. The first commercially available bubble jet printer was the Hewlett Packard Paint Jet, a letter-size,
monochrome printer with 180-dpi resolution. It was capable of printing on coated
paper and transparency film. In 1991, HP introduced the DeskJet 300C, the first
300-dpi color printer. A three-color printer, the DeskJet 300 could not print a true
black. This was remedied in 1992 with the four-color HP 550. Canon entered the
market in 1992 with its line of BubbleJet printers. Today, most bubble jet printers,
including those made by HP, Apple, Star, and Lexmark, use the Canon engine.
The Epson Stylus line of color printers offers resolutions of up to 1440 dpi.
In bubble jet/thermal liquid inkjet printing, an electric charge is applied to a
tiny resistor, which causes a minute quantity of ink to boil and form a bubble.
As the bubble expands, a drop is forced out of the inkjet nozzle. The current
applied to the resistive heating element causes bubbles to form; small bubbles
consolidate, and the pressure begins to expel the drop. The bubble continues to
expand and push out the drop; as the drop is expelled, the heating element cools
and the pressure of the bubble is countered by the pressure of the ink. The bubble
is expelled and the system returns to the wait state for the next bubble.
Advantages of the bubble jet/thermal liquid ink method include low cost,
excellent print quality, and low noise. However, liquid ink, for both piezoelectric
and bubble jet, generally requires special coated paper, which is expensive and
limits the applications of these printing methods.
The third type of inkjet printer uses a solid ink (also called hot-melt). A waxbased solid ink, similar to a crayon, is quickly melted and then jetted to the paper.
The ink solidifies on contact, preventing smudging. The advantages of hot-melt
include the ability to print on a variety of substrates, excellent print quality, and
the potential for high speed. Experts cite advantages of hot-melt as low cost, high
reliability, flexibility, media independence, user safety, and environmental friendliness. Color and image quality are independent of substrate properties, and the
prints are water resistant and lightfast.
One main disadvantage of hot-melt is clogging of the nozzle by dried ink.
Hot-melt inks perform poorly on transparency film, because they scatter light too
much. This can be controlled by precise temperature control within the printer,
on a platen. Also, because the wax ink sits above the printing substrate, it has
reduced resistance to abrasion, cracking, and peeling. Different manufacturers
use different methods to compensate for this.
The first solid ink printer was the Phaser III PXi tabloid printer, costing
roughly $10,000. Introduced in 1991, it could print with a resolution of 300 dpi
on a page up to 12 × 18. It featured a 24-Mhz reduced instruction set computer
controller and used Adobe PostScript Level 2 page description language. It could
print full-color pages in 40 to 60 seconds and monochrome pages in just 20
seconds. Dataproducts, owner of several key patents in the solid ink area, offered
its Jolt printer in 1991. Brother is also a player in this field, with its HS 1PS.
Color Desktop Printer Technology Technological Developments and Considerations
Continuous inkjet was the first form of inkjet printing and still predominated as of
1994. It boasts higher resolution and higher speed than DOD. However, over the
past few years, developments in DOD technology have given rise to a wide range
of new applications, from photo printing to wide format printing to package printing.
The advantages of DOD include low cost, compact size, quiet operation (only the
ink strikes the paper), excellent print quality, the ability to produce excellent color,
and the potential for high speed.
It is essential to consider the DOD printer as part of a larger system of printer,
ink, and substrate. To that end, considerable research into the physics and chemistry of DOD printing has been conducted. The frequency response and print
quality of a DOD inkjet system are determined principally by the time taken to
replenish lost ink after ejection of a drop. Mathematical analysis was extended
to multijet, systems connected to a common reservoir, and it was found that if
all jets were fired simultaneously, refill was slow and drops were large, due to
the inertia of ink and overfilled tubes. Sequential firing provides quicker refill
due to uniform ink flow rate and minimum inertia.
Much attention has been given to the development of ink. Considerations
include temperature dependence of ink properties, aging, evaporation loss, acidbase resistance, corrosion, and ink mixing. Also important are viscosity, pH,
surface tension, dielectric properties, optical density, and environmental impact.
Aqueous inks were the first to be used, but their limitations (mentioned above)
led scientists to research both improvements and alternatives. Clogging has been
another sticky problem. In DOD thermal inkjet printers, nonvolatile fluids are
mixed with the primarily water-based inks to minimize precipitation in the nozzles, which can result from evaporation of the water.
This problem has been addressed by using highly water-soluble, low-molecularweight additives that are claimed to be 3 to 5 times more efficient than glycols in
reducing water evaporation and useful for prevention of nozzle failure in DOD
printers. It is believed that the additives form a readily ruptured membrane at the
surface of the nozzle during the dormancy period of the jets. Questions of droplet
air drag and the effect of electrical charge on the ink have also been studied. For
most liquid ink printing, the substrate is still more critical than it is for hot-melt. The
transparent inks appear to better advantage if they are not absorbed into the substrate.
Accuracy of reproduction depends on constant diameter, regular contour,
maximum intensity, and instantaneous drying. In 1985, the French firm of Aussedat-Rey produced its Impulsion paper, which was claimed to be of the correct
absorbency for accurately timed drying.
Many printer manufacturers sell their own lines of coated stocks. Paper
manufacturers also have their own lines of inkjet papers. Some printers now claim
high resolution on plain paper. It is the paper that gives the color output the look
and feel of a photographic paper.
The Business and Market for Desktop Printers
103 Inkjet Papers
Papers that are created specifically for inkjet printers have two elements, the
foundation and the coating or treatment. The paper foundations are composed of
fiber, sizing, filler, and colorant. The coating or treatment is added to the foundation of the paper to further make it optimized for inkjet usage.
Short paper fibers are the best inkjet papers. The paper fibers affect how
smooth the paper is and the overall image quality. Paper smoothness is measured
in sheffields; the lower the number, the smoother the paper. With smoother paper
you can help maximize the printers output characteristics. If paper has a rough
surface, the high-dpi equipment’s printing capabilities will be lost. Think of it as
a landscape: one dot will be on a hill, another will be on the side of the hill, and
a third will be in a valley. Paper fiber can be compared to sandpaper grit. The
smaller the grit on the sandpaper, the smoother the surface will be. The same is
true with paper fiber: the smaller the paper fiber, the smoother the paper.
With shorter fibers, there will be less ink migration and the ink will not travel
as much on the surface of the paper. The ink migrates/bleeds because it is being
absorbed by other parts of the paper, i.e., osmosis. The result of having shorter
paper fibers is a decreased amount of bleeding. With less ink bleed, the colors
will not appear as washed out, because the inks will be in a more concentrated
location. Smaller fibers and smoother papers have less air between the fibers. The
ink will fill these pockets of air and increase the transparency and reduce the
reflection density of the image.
Another aspect of preventing ink migration/bleed is using sizing on the
substrate. It is added to prevent water absorption into the paper, prevent ink
smudge, and aid in drying. Sizing can be internal like rosin or alum, or it can be
applied to the surface like liquid starch. The basic principle behind sizing is that
it is hydrophobic so once it has reached the saturation point, it will not permit
any more absorption into other parts of the paper. Without sizing, there would
be an increased amount of bleeding in the paper’s surface. The sizing further
inhibits inks from migrating to other parts of the paper.
Fillers affect the physical and optical appearance such as brightness, color,
opacity, and the feel and stiffness of the paper. Fillers may be clay, talc, titaniumdioxide, calcium carbonate, or other materials. To make the paper appear brighter
and have a whiter color, a filler of titanium dioxide can be used. The opacity of
the paper will also determine now much light is reflected or transmitted through
the paper. The paper surface should be more opaque so that light will not be
transmitted but, rather, reflected off the paper.
The combination of paper brightness, whiteness, and opacity contributes to
the color quality of the image as well. With a brighter, whiter, and more opaque
paper, you have an increased color gamut, more visual contrast, increased saturation, and increased detail and sharpness.
The colorant is used to give the paper a color or tint, such as publicationbased commercial stock. In some cases, such as during proofing, the press
Color Desktop Printer Technology
condition should be mimicked. If the paper has a yellow tint, that effect with the
inkjet paper should be reproduced.
After the foundation of the paper is made, it is then put through a treatment
or receives a special coating. Do not confuse sizing with the special coating;
sizing is one aspect of the coating inkjet paper receives. The treatment/special
coating does help with ink absorption at the location where the ink hits the paper.
It helps prevent ink migration and bleeding and adds florescence to the paper,
making it appear whiter. The paper under a UV light will have a bluish/purplish
glow. An uncoated piece of paper will appear white. Sometimes, however, with
a coated paper, the surface becomes slick. This sometimes causes problems with
the loading and unloading of the paper, because the feed rollers slip on the paper’s
Inkjet papers should have proper porosity, be smooth, and have uniform
weight and thickness. Porosity is the product of the combination of paper fiber
and filler. The proper porosity will allow the right amount of ink to be absorbed
into the paper. If too much ink is absorbed, the paper will have too much water
in it and the fibers will expand and cause waviness and feathering at the image
edge. The resulting image will not appear sharp and may become oversaturated
in the center of solids. If the paper is not porous enough, the ink may not be
absorbed enough into the paper and cause mottling. Mottling occurs when the
ink does not dry because it is not being absorbed into the paper. The resulting
image will not look correct where the solid areas are not as solid as they should be.
The smoothness of the paper is a product of paper fibers and treatment and/or
special coatings. The smoother the paper, the more directly light will be reflected.
The resulting image will appear sharper and brighter. If the paper is rough, the
light being reflected will be more diffused. The resulting image will appear flat
and dull.
Uniform weight and thickness will reflect the quality and smoothness of the
paper. If the paper has much variation internally and on the surface, the print
quality will be poor. This is because the ink placed on the paper will be of
inconsistent densities of like colors. This can be seen in printing solids. It is
similar to the printing of solids on toner-based printers. The solid appears splotchy
and looks as though it has high and low points.The more uniform the weight and
thickness of the paper, the more consistent the output will be.
A large selection of inkjet papers is available, with varying paper thicknesses;
tints; matte, semi-matte, and glossy surfaces; watercolor; textures; and also available are transparency sheets and cloth sheets. The application will determine what
paper to use.
The basis of good inkjet paper comes down to good dot structure. Because
fluids form spheres in a free state, when the ink hits the paper a nice circular dot
should result. Under ideal conditions, the diameter of the sphere should be the
diameter of the dot. Even though the dot is an area and the sphere is a volume,
the excess should be absorbed into the paper.
Average paper thickness is 4 to 7 mils, and the diameter of the ink drop
coming out of the printhead is approximately 10 microns. This means that an ink
The Business and Market for Desktop Printers
drop is less than 1/500 the width of the paper. When the ink hits the paper, a dot
forms. This dot should be circular and have minimal bleed, dot gain, and feathering. If these three occur, the image should have nice dots, good highlights, midtones and shadow detail, good tonal range, good saturation, good detail, sharp
graphics and text, increased contrast, and superior reproduction when compared
to a print down on regular paper (such as copier paper).
A technique used in some high-end inkjet printers is to place a charge on the
ink drop and have the paper be oppositely charged. The paper itself does not have
a charge, but it is given a charge during printing. It is important to keep in mind
that inkjet printers use dyes instead of pigments. Even though pigments have a
stronger hue than dyes, pigments are not very soluble in liquids; in return, they
are too large to fit through a nozzle or piezo of inkjet printhead. The orifice of
the printhead is usually smaller than 10 microns. (Iris Graphics, later Creo and
now part of Kodak, pioneered the under-10 micron orifice.)
Canon has come up with an interesting concept with its inkjet printers using
copier paper. Instead of having the standard four CMYK ink reservoirs, Canon
has five reservoirs. The additional reservoir is a liquid that jets before the ink and
acts as a special coating. It essentially traps the ink and prevents ink migration. Desktop Inkjet Gets Down to Business
The most frequently cited uses for DOD printing have been low-end office printers
and case/carton marking. Other applications include direct mail personalization,
labeling, ticketing, bar coding, dating, and marking.
Digital photography and digital printing were made for each other. Whether
one uses a desktop inkjet printer for outputting the equivalent of photographic
prints or a high-end toner-based printer for document reproduction, digital photography allows for image inclusion more easily than ever before.
The inkjet color printer has become the major printout device for the average
HP has a large global installed base of printer users. The company’s first LaserJet
and inkjet printers were introduced in 1984, and since then it has shipped 230
million printers, with 130 million inkjet printers and 60 million LaserJet printers.
HP is adding 30 million inkjet printers and 7 million LaserJet printers each year.
The company has predicted more convergence at the high end of the imaging
and printing markets, where enterprise customers will be buying more multifunction products.
HP inkjet printers make up about 52% of the market, IDC estimates, whereas
Xerox accounts for less than 4%.
Color Desktop Printer Technology
Monochrome page printer sales have declined 5% to fewer than 3 million
units shipped yearly, but some segments are growing. Volume growth of all U.S.
electronic printers was flat in 2001, but color laser printers and MFPs are strong
growth segments.
The worldwide printer market declined by 5.3% in the first quarter of 2002.
A total of 18.9 million units were shipped, representing about $5.4 billion in enduser spending.
Gartner Dataquest reported that U.S. color page printer hardware shipments
grew 24.6% in 2002, to more than 315,000 units, up from 252,000 units in 2001.
U.S. color copier hardware shipments reached 44,000 units in 2003, an increase
of 3.5% over 2001 shipments of more than 42,000.
Peter Grant, principal analyst for Gartner Dataquest’s Digital Documents and
Imaging Worldwide Group stated at a Seybold Symposium: “New color page
printer products will span the range from entry-level color lasers for the small
office/home office to higher-speed color LED printers that will bring affordable
color into the workgroup and enterprise. Vendors competing in the low end color
printer segment need to balance their low hardware price with an acceptable
supplies price. Users are aware of the high cost to print color and to reach the
magic cost per page for color that removes the barrier to crossover from monochrome to color. The cost for a color page needs to equal that of a monochrome
with the same toner coverage.”1 His prediction was correct.
In 2002, copier vendors moved their products to higher speeds, where they
will not sell large numbers of units but, rather, hope to capture considerable page
volumes. Gartner Dataquest analysts forecast significant price erosion for color
laser copiers at the low end to maintain market share against color laser printer
Gartner Dataquest analysts have said that color copier vendors must respond
more quickly to color desktop printers that will compete aggressively for office
output. Organizations will have multiple options to buy, lease, rent, or outsource
and buy only the pages they print.
Color inkjet printers now make sense as laser replacements for most users.
Inkjet color printers can reproduce photography of stunning quality. The Epson
2000P was the first desktop printer capable of reproducing photographs of archival
quality; Epson claims photograph quality of up to 100 years.
With its acquisition of Compaq, HP saw a break in its printer relationship
with Dell. Dell’s is not a profitable one; therefore, current strategy is merely
reselling HP and Epson printers — with no ink to follow. Dell will make its own
printers. Most printers are sold at a loss, and it takes a long time to penetrate an
installed base as large as HP’s, but Dell will be able to bundle PCs and printers.
In addition to meeting the changing needs of large businesses, paper suppliers
are increasingly turning their attention to the SoHo sector. Although this market
is relatively small (about 2 to 3% of the total market) at the moment, predictions
of potential growth vary widely from 5 to 20% of the overall market in 2005.
Optimistic projections state that SoHo users will represent over 40% of the total
cut size market by 2007.
The Business and Market for Desktop Printers
The U.S. printer industry is going through a transition as inkjet printer shipments
were on the decline in 2004, whereas all-in-one (AIO) multifunction product
shipments were on the rise, according to Dataquest, Inc., a unit of Gartner, Inc.
Gartner Dataquest has said that U.S. inkjet printer shipments decreased 5.8% in
the first quarter of 2002, whereas AIO multifunction product shipments increased
98%. Total inkjet printer shipments reached 4.1 million units in the first quarter
of 2002, whereas AIO (both sheet-fed and flatbed multifunction products) shipments totaled 1.3 million units.
The color laser copier and color laser printer markets suffered from the softer
economy in 2001, but both markets show growth potential if vendors deliver the
proper mix. This has proven true and growth from 2002 to 2005 has been robust.
Users are aware of the high cost to print color. Gartner Dataquest analysts
have said that offering a standard cartridge and a high-capacity cartridge has
worked well for some vendors in helping users over the price hurdle.
At a Seybold Symposium in 2002, Peter Grant said, “More important is to
reach the magic cost per page for color that removes the barrier to cross over
from monochrome to color. The cost for a color page needs to equal that of a
monochrome with the same toner coverage.”
Inkjet and toner-based (laser) printers will compete. For the next 5 years inkjet
printers will own the low end (high unit placement) of the market and lasers will
own the high end (lower unit placement). However, in terms of volume, both
markets may be equal.
The fight will be in the 10- to 20-ppm market, where inkjet and laser printers
overlap. The main drivers will be photographic output and distributed documents
(publications sent as files and then wholly or partially printed out remotely). This
will lead to higher levels of quality — which means more than four inks — and
higher speeds.
Printers are sold through two main approaches: retail and mail order and
bundled with PCs. We do not expect any major changes in channels but there
will be fierce competition in terms of price and performance.
1. Romano, F. Machine Writing and Typesetting. GAMA, Salem, NH, 1978.
Part II
Ross R. Allen, Gary Dispoto, Eric Hanson,
John D. Meyer, and Nathan Moroney
History and Introduction .........................................................................112
4.1.1 Major Technologies .....................................................................113
4.1.2 Years of Color .............................................................................114
4.1.3 Advancements in Thermal and Piezo Inkjet...............................115
4.1.4 Advances in Inks and Print Media..............................................117
Inkjet Printing Technologies ...................................................................118
4.2.1 Thermal Inkjet .............................................................................118
4.2.2 Piezoelectric Inkjet ......................................................................125
Ink Storage and Delivery ........................................................................128
Printhead Service and Maintenance........................................................131
Inkjet Inks................................................................................................133
4.5.1 Liquid Inks ..................................................................................133
4.5.2 Dyes .............................................................................................135
4.5.3 Pigments ......................................................................................135
4.5.4 Solid Inks.....................................................................................137
Inkjet Media.............................................................................................138
4.6.1 Mechanical Properties .................................................................139
4.6.2 Imaging Properties.......................................................................139 Plain Papers...................................................................142 Coated Inkjet Papers .....................................................142 Photo Papers..................................................................142 Overhead Transparency Films ......................................143 Specialty Media.............................................................143
Print Modes .............................................................................................143
High-Fidelity Color .................................................................................145
4.8.1 Overview......................................................................................145
4.8.2 Color Separation..........................................................................147
4.8.3 Printing with Light Inks ..............................................................148 Light Ink Strategies: Halftoning ...................................148 Light Ink Strategies: Light Ink Color...........................150
4.8.4 Printing with Dark Inks...............................................................151
Color Desktop Printer Technology Dark Ink Strategies: Selecting Additional
Primaries........................................................................151 Dark Ink Strategies: Image Processing ........................153
4.9 Closure.....................................................................................................154
References .........................................................................................................154
It is traditional for historical discussions of inkjet technology to begin with
references to the 19th-century studies made by Lord Rayleigh on the stability of
liquid jets. This work formed the practical basis for the development of continuous
inkjet printers that have found broad application in industrial marking processes
and the graphic arts. These devices make use of the Rayleigh instability, which
causes a continuous liquid jet to break up into a stream of droplets, which then
are electrostatically deflected between trajectories allowing them to print or to
collect for recirculation.
Desktop color inkjet printers are all drop-on-demand devices: a droplet of
ink is ejected only when a pixel or portion of a pixel is to be printed. Given this
chapter is concerned with desktop color printers, a reference to Rayleigh might
seem out of place. However, it is worthy to note that Rayleigh studied the
dynamics of bubble growth and collapse, and in the process discovered cavitation,
the damage mechanism associated with bubble collapse. All inkjet printers must
contend with unwanted gas bubbles in the ink supply. In thermal inkjet, a bubble
of superheated ink vapor drives the ink droplet out of the nozzle. The subsequent
collapse of that bubble on the heater can produce cavitation damage, which, left
uncontrolled, can destroy the heater after only a few hundred drop ejection cycles.
The successful development of thermal inkjet drop generators lasting hundreds
of millions of cycles is based on fluidic and materials solutions to control cavitation. Rayleigh’s bubble model was the starting point for computer simulations
developed to understand and control the growth and collapse of vapor bubbles.
It is appropriate, therefore, to begin this chapter with an acknowledgment of Lord
Rayleigh’s contributions to inkjet technology in all its forms.
Desktop color inkjet printers first appeared in the early 1980s. These printers
used piezoelectric (piezo) transducers to generate ink droplets, had only a few
nozzles, and offered resolutions less than 240 dots per inch. Recognizing the
future potential of low-cost color printing, a number of companies engaged in
early experiments with color images using these devices. Affordable desktop color
scanners did not appear until the 1990s, and this meant that access to color digital
image data was mainly through the graphic arts. In these early days of color
desktop printing, the scarcity of digital color images meant that many people
used and published results based on a set of common ones, such as the wellknown photo called “The Mandrill.” The occasional sight of this baboon today
brings back memories to the founders of desktop color printing.
The results were promising, but ahead of their time: the supporting infrastructure of powerful desktop computers, image-capable application software,
desktop color scanners, digital cameras, and the Internet were not yet available.
Market opportunities for color printers were limited, and many industry observers
stated that desktop color printing would not succeed until color copiers became
commonly available in the office. In the 1980s, this led to the much-debated but
never-proclaimed “Year of Color,” when color printing was finally accepted in
the office. The first year of color passed unheralded in the early 1990s.
Today, desktop color is so commonplace that inkjet printers compete in a
commodity market, and this environment demands that all printers offer photographic-quality color as well as high-speed, high-quality black text on plain paper.
Two technologies for producing drops-on-demand have dominated desktop color
printing since the mid-1980s: thermal inkjet and piezo inkjet. An overwhelming
number of inkjet printers use liquid, water-based inks. Solid or hot-melt ink is
used in some printers found in graphic arts and office applications.
Thermal inkjet, also called bubble jet by Canon, rapidly heats the ink to produce
a tiny bubble of superheated ink vapor. The expansion of this bubble ejects an ink
droplet, and its collapse refills the ink chamber. Thus, the only moving part in a
thermal inkjet printhead is the ink itself. Hewlett Packard (HP) introduced the first
printer based on this principle in 1984, called ThinkJet. ThinkJet had 12 nozzles,
each producing up to 1200 drops per second, and used special paper.
Thermal inkjet devices are fabricated from thin metallic and dielectric films
on silicon using photolithographic techniques developed in integrated circuit
manufacture. This enables many drop generators and their control electronics to
be integrated on a printhead. HP’s ThinkJet print cartridge is shown in Figure
4.1. ThinkJet’s printhead and the electrical interconnect were very compact, and
FIGURE 4.1 HP’s ThinkJet print cartridge.
Color Desktop Printer Technology
ThinkJet offered the innovation of a disposable printhead and ink supply. Perhaps
more than any other feature, this concept legitimized inkjet by overcoming the
reliability and (messy) ink-handling issues that gave earlier inkjet products a poor
reputation. Thermal inkjet technology has undergone significant development
primarily by HP, Canon, and Lexmark.
Piezo inkjet technology benefited from substantial investment, development,
and innovation by Epson, who introduced the Stylus 800 office text printer in
1993. The technology behind the Stylus printer is a push-mode, multi-layer, piezo
actuator. This made possible nozzle generators of compact design, although not
small enough to match the high linear nozzle density of thermal inkjet. Piezo
inkjet also benefited from photolithographic fabrication of its silicon and metal
layers, and this enabled the mass production of printheads at modest cost.
Solid inks have been investigated in both thermal and piezo inkjets, but only
piezo solid inkjets have seen systematic product development by Tektronics (now
Xerox), Brother, and others. The Tektronics Phaser III, introduced in 1991, had
96 nozzles and operated at 8 KHz. By convention, operating frequency is the
number of drops generated per second from each nozzle and is expressed in hertz
(Hz). After a decade of technology development, the Phaser 8200 offered a
printhead with 448 nozzles operating at 36 KHz. These printers have focused on
business and graphics markets offering A-size (A4) and B-size (A3) formats.
Although much larger than their liquid ink counterparts, solid inkjet may be
placed on a tabletop in office environments.
4.1.2 YEARS
The introduction of HP’s PaintJet Printer in 1987 was the logical precursor to
the explosive technological development of desktop inkjet color printing. Prior
to PaintJet, inkjet products competed to produce black text on plain paper with
the goal of meeting the rapidly advancing quality standards set by laser printers.
PaintJet offered 180 dots per inch (dpi) color graphics on a special coated paper
and overhead transparencies. It used 30 nozzles for its dye-based black ink and
10 each for its cyan, magenta, and yellow inks. The black and color printheads
were integrated into two user-replaceable print cartridges.
Color displays for desktop computers were in common use by the time
PaintJet was introduced, and the issue of matching the display to the printed
image quickly became a key user need. This coincided with the beginning of
industry-wide adoption of color management, which drove standards for color
space definitions, file formats, and color data conversion methods making possible
device-independent color across display, capture, and hardcopy devices from
different manufacturers.
In 1989, the HP DeskJet 500C offered 300-dpi resolution, plain paper color,
and a printer driver that provided color matching between display and print.
Market response to the DeskJet 500C was dramatic, and in the decade of the
1990s, HP produced and delivered more than 100 million color inkjet printers
worldwide. The success of desktop color inkjet printing involved much more than
higher resolution, smaller drops, and faster throughput. The practical adoption of
the printing technology was made possible by research in color science, image
processing, color data interchange standards, and the availability of user-friendly
color management software.
The 1990s saw the quality of images on special media increase dramatically
with every product introduction cycle, roughly one every 6 months. These were
the years of the “dpi wars,” in which some manufacturers claimed the highest
possible resolution in photo printers often at the expense of throughput and the
ability to print high-quality black text on plain paper. The introduction of the HP
PhotoSmart printer in 1997 with six photo inks demonstrated that a desktop inkjet
printer could achieve the goal of photographic print quality.
By 2000, desktop color inkjet printers had created a digital imaging revolution
by printing photos on special media that were indistinguishable from conventional
photographic prints from 35-millimeter (mm) film. Developments quickly followed in consumer digital photography and desktop color scanners to make highquality digital capture pervasive and affordable. For example, after 2001 consumers purchased more new digital cameras than new 35 mm cameras (excluding
one-time use 35 mm cameras).
Today, the inkjet printing process is no longer the limiting factor that determines the grain, color quality, and color fidelity of the printed image, it is the
process of capturing the digital image with either a digital camera or a color
scanner. For scanners, the quality of the original image is often the limiting factor.
The reproduction of high-quality color images requires sufficient dynamic range
along with precise control of tone reproduction, neutral gray balance, image granularity, color gamut, detail, and sharpness. These needs have been addressed by
technologies allowing higher resolution and addressability, decreases in dot size,
multi-level printing with dark and light primaries including grays, and halftoning
techniques. Fade resistance, driven by the durability of conventional color photographs, has been achieved by cooperative design of dyes and special inkjet papers
and by the introduction of black and color pigment inks. Today, color inkjet prints
can resist fade for more than 100 years and exceed the fade resistance of color
photographs produced by silver halide and color-coupling chemistry.
Since 1989, the printing resolution claimed by liquid inkjet manufacturers
has increased from 300 to 5760 dpi. This rapid development was driven by a
highly competitive marketing situation in which dpi was promoted as a primary,
and sometimes the only, image quality specification.
A significant amount of confusion and misinformation surrounds dpi. This
is because resolution has been used in a number of different ways since it became
a key specification used to guide purchasing decisions. Initially, resolution meant
the (horizontal or vertical) spacing of printed dots with sufficient overlap to
achieve 100% area coverage. This is what could be called true resolution. Resolution has also been used in place of addressability, the horizontal and vertical
Color Desktop Printer Technology
grid at which an ink droplet can be placed. High addressability has value in
achieving smooth, sharp-edged characters in text printing. The use of multiple
drops per pixel (color layering) or drop volume modulation has made it possible
to vary the final dot size while printing at a fraction of true resolution based on
the smallest dot. When inkjet printers print photos using halftone pixels offering
72 million addressable colors, as claimed by HP for its 8-ink PhotoREt Pro
printers in 2003, it is meaningless to characterize a printer’s image quality by a
single number such as dpi. Studies based on typical observers conclude that 4by 6-inch prints produced by a desktop inkjet at 300 pixels/inch using a 6-color
ink system are indistinguishable from conventional photographs in terms of image
grain, color gamut, smooth color gradation, sharpness, and detail.
As resolution has increased, drop volumes have also decreased, from 32
picoliters (pl) in 1997 to 2–6 pl in 2002. Ink drops are commonly specified by
their volume (picoliters, 10-12 liters) or weight (nanograms, 10-9 grams). As most
liquid inkjet inks for the desktop have a high water content, the number of
picoliters in a drop is equal to its weight in nanograms within a few percent. This
trend up to 2002, shown in Figure 4.2, has been driven by competitive forces to
deliver photographic image quality. At 2 pl, drop volumes are approaching a
practical limit where aerodynamic forces limit accurate drop placement. These
small drops produce dots that are virtually invisible, but used in halftoning and
edge-enhancement, algorithms give fine control over the tone reproduction curves,
especially in the highlights, and smooth edges on printed characters.
Because smaller drop volumes make smaller printed dots, more drops per
second from a printhead are required to cover the same area in a given time.
The time to print an image or page is a key competitive specification, and this
translates into higher drop rates and more nozzles per printhead. The potential
number of drops per second from all the printheads taken together is a useful
measure of design capabilities, although such rates cannot be achieved in any
practical print mode. By this measure, before 2000 desktop color inkjet printers
could not produce more than 20 million drops per second. By mid-2005, an HP
Photosmart 8250 printer could deliver over 100 million drops per second from
3900 nozzles integrated onto a single silicon chip. A trend observed over the
past 20 years of desktop inkjet printing shows that the potential number of drops
per second doubles every 18 months. As practical mechanical and fluidic limits
in drop generator design are approached, this trend can continue only by increasing the number of nozzles.
Practically speaking, the throughput of desktop printers is limited now not
by drop rates but by the time needed to dry each sheet, and this has driven the
development of faster-drying inks and print media. Most desktop inkjet printers
do not use heaters to dry the printed sheet because competitive forces in the
marketplace cannot support the additional manufacturing costs.
Dot Diameter
(microns, average of horizontal and vertical)
HP DJ970
Epson 760/860
‘98−‘02 ‘00
Epson 980
Epson 900/
Lexmark Z65
HPDJ cp 1160
Canon 8200
Epson 740
HP DJ722
Xerox M750
Lexmark Z51
HP DJ2000
Canon 7000
HP Photosmart
Lexmark 5700
FIGURE 4.2 (See color insert following page 176.) Evolution of drop volume and drop
Alongside developments in printhead technology, a major effort has been made
in the development of inkjet inks and print media.
The early days of desktop inkjet were dominated by the goal of laser-quality
printing on plain paper, which meant copier and bond (typewriter) papers. A
restriction to special inkjet papers, especially to achieve acceptable text quality,
was understood to be a significant barrier to the adoption of inkjet products in
the office. This drove the development of black inks, especially pigment-based
black inks, with high optical density and minimal spread to compete with laser
printers on text quality.
As desktop color inkjet printing spread into the office and home, markets
developed for special media offering the highest image quality. Today, several
major requirements dominate the development of new print media: rapid dry time,
customer choices for print gloss, predictable and stable color, and durability in
the form of waterfastness and fade resistance.
Inkjet printing is now so highly developed that the consumer can purchase
media for almost any application: specialty materials for iron-on transfers and
creative projects, greeting cards, business cards, labels, brochures, fine art papers,
and papers that enable an inkjet printer to be used for pre-press proofing in
commercial printing applications.
Color Desktop Printer Technology
About 75% of inkjet printers sold today employ the thermal inkjet principle. In
a thermal inkjet printhead, thousands of drop ejection chambers may be placed
on a single silicon chip. Each drop generator consists of a heater, a nozzle, and
an ink refill channel, as shown schematically in Figure 4.3. It is common to have
one, two, or three colors printed by a single monolithic printhead, and recent
developments offer six, eight, and nine colors. With hundreds of millions of
thermal inkjet printers in use worldwide, tens of millions of print cartridges and
ink cartridges are manufactured every month.
There are two configurations of the thermal inkjet drop generator, called topshooter and side-shooter, depending on the orientation of the nozzle with respect
to the planar substrate. HP, Lexmark, and Canon products employ the top-shooter
configuration, whereas the edge-shooter is used primarily by Canon.
In thermal inkjet printheads, nothing moves except the ink itself. This process
is shown schematically in Figure 4.4. To eject an ink drop, an electrical pulse is
applied for about two microseconds to a thin-film resistor raising its temperature
at more than 100,000,000°C per second. Near the heater surface, ink is heated
from about 60°C to above 300°C as a metastable liquid, and in a unique thermodynamic process it undergoes a superheated vapor explosion. This generates an
energetic vapor bubble that grows and collapses to eject a repeatable quantity of
ink from the nozzle.
Superheat is a condition where a liquid contains more thermal energy than
required for boiling at a given pressure. A superheated liquid can explosively
change into a gas when a site is available for a bubble to form (nucleate). This
Inkjet Technology Performance Trends
DesignJet 10ps
C F900
L J110
L Z65
FIGURE 4.3 (See color insert following page 176.) Thermal inkjet configuration.
E PM900C S600
L Z53
E C80
L Z23 E C60 E C40
L Z52
L Z51
10 HP895
N1000 ?
Million drops per second (KCYM)
Top Plate
Ink Inlet
Ink Inlet
FIGURE 4.4 (See color insert following page 176.) Thermal inkjet drop ejection process.
can be a tiny vapor-filled pore in the surface wet by the liquid. For water-based
inks, the superheat limit is about 340°C. Above this temperature, the ink cannot
exist as a liquid and spontaneously becomes a gas. The actual temperature of
vapor bubble nucleation depends on the ink composition, the microstructure of
the heater surface, and the size of gas bubbles trapped in microscopic pores on
that surface. This process is not boiling in the everyday sense. Boiling in water
(and water-based inks) occurs when there are plenty of “large” nucleation sites
available to allow the phase change of water from liquid to gas at about 100°C.
The high heating rate assures consistent nucleation only from the high temperature
sites because the low-temperature sites (activating at 100°C) do not have time to
activate. This gives vapor bubbles that are highly consistent in terms of energy
release and timing.
The vapor bubble expands rapidly and then collapses in a few microseconds.
In this process, it acts like a tiny piston that strokes upward from the surface of
the heater to fill the drop generation chamber and then retracts. The expansion
phase induces a bulk velocity to the ink, causing a jet to leave the nozzle at 10–15
meters/second. The initial pressure inside the vapor bubble is momentarily greater
than 10 atmospheres, but its expansion is so rapid and the quantity of vapor is
so small that the positive pressure phase of expansion lasts only about 1 microsecond. During most of the bubble’s lifetime of 20 microseconds or less, the
pressure inside the bubble is subatmospheric, but the bubble continues to expand
to fill about 50% of the chamber carried by the ink it has set into motion.
With no check valve on the drop generator chamber, some ink is forced back
through the ink refill channels during bubble expansion. This is useful to dislodge
particles and air bubbles in the ink that may accumulate near the inlet structure.
Fluidic design of the inlets and a staggered firing order minimize fluidic crosstalk
between adjacent nozzles. The fluidic impedances of the nozzle and inlets are
balanced to maximize drop ejection efficiency while tuning the drop generator
for optimal refill and frequency response.
Color Desktop Printer Technology
The vapor inside the bubble has much lower thermal conductivity than the
liquid ink, so it insulates the ink from the heater. As a result, little additional
vaporization occurs once the bubble covers the heater surface. The electrical pulse
is timed to end shortly after vaporization occurs, and this limits the peak temperature of the heater surface and allows it to cool by conduction into the substrate
as the bubble expands and collapses. When bubble collapse places ink back into
contact with the heater, the surface temperature must be below that required to
form a new vapor bubble. This ensures that each electrical pulse produces only
a single bubble. There is a thin-film insulating layer, called a thermal barrier,
between the heater resistor and the silicon substrate that controls heat flow into
the substrate. If it is too conductive, too much heat is lost during the heating
phase and this increases the power required to form a vapor bubble and produces
excessive printhead heating. If it is not sufficiently conductive, the heater does
not cool quickly and multiple vapor explosions can occur. Three or more successive nucleations from a single heating pulse have been observed in this situation.
As the vapor bubble reaches its maximum volume and begins to collapse,
the nozzle meniscus retreats into the drop generator, causing the jet of ink to
elongate and break off. It takes about 100 microseconds for the drop to travel
between the orifice plate and the print medium, where the image is recorded by
the ink drops. The carriage holding the printheads typically scans at 0.5–1
meter/second in desktop inkjet printers, so aerodynamic effects can affect drop
placement, especially for small drops.
As the bubble collapses, the nozzle meniscus retracts deeply inside the drop
generator. The final stage of bubble collapse is so rapid that very high pressures
can be created in a cavitation process. Fluidic design of the chamber and selection
of materials in the thin films covering the heater ensure that the energy released
during final collapse does not damage the heater.
The process of ejecting a drop is also the mechanism to refill the drop
generator. Once the bubble collapse is complete, the meniscus is deeply retracted
into the drop generator. Surface tension in this extended meniscus creates a
subatmospheric pressure in the drop generator to draw ink from the supply
reservoir to refill the chamber. As refill progresses, the meniscus returns to a
stable position in the nozzle bore near the outer surface of the orifice plate.
Ink in the drop generator is maintained at a subatmospheric pressure equivalent to supporting a column of 2–5 inches of water. This prevents ink from
drooling out of the nozzle when it is not printing. It also means that the rest
position of the meniscus in the nozzle bore is concave into the oriifce plate, where
the curvature of the meniscus and ink surface tension balance the negative head
produced by the ink delivery system.
This entire process, from the electrical pulse to stabilization of the nozzle
meniscus, takes fewer than 27 microseconds in a drop generator designed to eject
36,000 drops per second.
Important design parameters for the drop generator are the geometry of the
ink refill channels, the dimensions of the chamber and thickness of the nozzle,
and the shape of the nozzle bore. The important fluidic parameters are ink surface
tension and viscosity. During refill, the drop generator acts like a simple fluid
mechanical oscillator: the refill channel geometry, ink density, and ink viscosity
determine the dominant mass and damping characteristics; the nozzle meniscus
is the elastic (spring) element whose stiffness depends on its extension into the
drop generator chamber, the nozzle bore, and ink surface tension. If the system
is under-damped, the meniscus will oscillate in and out of the plane of the orifice
plate before settling. Proper design and tuning produce a critically damped configuration that minimizes the time for the meniscus to stabilize, and this maximizes the operating frequency of the drop generator. Ejecting a drop before the
meniscus has settled will affect the ejected volume: the quantity of ink in the
drop generator depends on the position of the meniscus, and this provides a
variable inertial load to the vapor bubble.
Ink viscosity affects both the dynamics of refill and drop ejection, and it is
important to control viscosity for consistent drop ejection. Ink viscosity depends
on ink formulation, and it varies significantly with temperature. The temperature
of the ink in the drop generators is nominally the same as the printhead substrate.
Given the temperatures required to form the vapor bubble, this is somewhat
surprising. But, the nucleation process is so rapid that heat from the resistor
penetrates only 0.1 micron into a column of ink several tens of microns thick.
So, the ink in the drop generator receives heat not from the nucleation event but
from contact with the substrate between drop ejection cycles. This has been
experimentally confirmed by simultaneous measurement of substrate and droplet
temperatures in operating printheads.
Left uncontrolled, the printhead temperature can vary by 35ºC or more
depending on room temperature, print density, and printing duty cycles. Stabilization of the operating temperature to about 60°C during printing is easily
achieved with the same heaters used for drop ejection. Active logic circuits and
temperature sensors on the substrate deliver short electrical pulses to the heaters
between ejection cycles. These do not produce vapor bubbles but deliver sufficient
energy to maintain constant printhead temperature. When high print densities
produce excess printhead heating, printing may slow down to manage printhead
Because a thin film of ink is heated to more than 300°C during drop ejection,
thermal inkjet inks must be formulated to avoid kogation on the resistor surface.*
Kogation can be eliminated by assuring purity in the ink ingredients and choosing
molecular structures whose thermal decomposition products are soluble in the
ink. Although this constraint on ink design may seem at first to be a limitation,
thermal inkjet drop generators have a practical lifetime of hundreds of millions
of cycles. The technology’s ability to use both dye and pigment inks and to
produce photographic-quality images and high-quality, durable black text on plain
paper speaks for itself.
Kogation is derived from the Japanese word koga, referring to rice burned onto the surface of a
rice-cooking pot, and this term was coined by Canon.
Color Desktop Printer Technology
Bubble Nucleation
<3 µs
A superheated vapor
explosion occurs by
heating at 100°C/µsec
Bubble Growth
3–10 µs
Bubble expands
forming a drop
Bubble Collapse
& Drop Breakoff
10–20 µs
Bubble collapses
drawing in fresh ink
<30 µs
Orifice meniscus
settles and refill
FIGURE 4.5 (See color insert following page 176.) Top view of particle-tolerant ink
supply channels.
For high reliability, it is essential to design printheads so that the ink inlets
to each ejection chamber are resistant to clogging. Printhead components are
carefully cleaned and assembled under cleanroom conditions, but the refill channels are only 10 to 20 microns wide. They can be clogged by particles or fibers
that escape cleaning and filtration or are formed during long-term exposure of
the print cartridge materials to the ink. Particle-tolerant architectures, shown by
the examples in Figure 4.5, prevent inlet clogging by acting as a filter and provide
multiple ink refill paths to each drop generator. The dark areas in these photomicrographs show the photo-imageable polymer thick-film with the orifice plate
removed. The light areas are the exposed substrate seen through open channels
in the polymer. Note the rows of pillars that are used for particle barriers.
The vertical growth of the vapor bubble in the drop generator is on the order
of tens of microns. This is quite large, considering the heater is of similar linear
dimensions. This large volumetric displacement right at the nozzle allows the
linear packing density of thermal inkjet drop generators to be very high: 600
nozzles per linear inch in a single column (42 microns center-to-center). Printheads usually have two columns of drop generators with a half-row offset between
columns. This gives a native vertical printing resolution of 1200 dpi. Drop ejection
timing allows such printheads to place drops at 4800-dpi grid points along the
horizontal (scan) axis, and very smooth lines and character edges can be produced
with a 4800- by 1200-dpi print grid. This level of fine addressability is very useful
for color halftoning in image reproduction.
A silicon substrate allows thermal inkjet printheads to be fabricated using
techniques employed in the manufacture of integrated circuits. The details of
typical thin- and thick-film layers are shown schematically and not to scale in
Figure 4.6. For reference, the thin-films are about one micron deep and the thick
films and orifice plate form a stack about 50 to 70 tall. The heaters used in drop
ejection are thin-film electrical resistors sputtered over a thermal barrier to control
heat conduction into the substrate. The heaters are covered with dielectric and
metallic layers to resist chemical attack and cavitation. These films are designed
for high fracture toughness, which is achieved by sputtering materials with very
10 Picoliter
4 Picoliter
FIGURE 4.6 Thin- and thick-film structure of a thermal inkjet printhead.
low brittleness. Low-resistance metallic conductors and active electronic elements
drive the heaters and demultiplex signals from the electrical interconnect.
After fabricating the thin-film layers, ink channels and the drop ejection
chamber are formed, applying a photo-imageable polymer thick film tens of
microns thick to the substrate and thin films. This thick film is exposed to light
through a pattern mask and then chemically developed to remove material, forming chambers, channels, and pillars. The orifice plate, containing the inkjet nozzles, is then attached to complete the structure.
Two technologies are commonly used for orifice plates: electroformed thin
metal plates and laser-ablated polymer sheets. Nozzles must be very precisely
located with respect to the heater and chamber walls; consistent in length, diameter, and bore shape; and free of defects that could perturb or deflect the droplets.
Nozzles producing 4-picoliter drops are typically less than 15 microns in diameter: about 1/3 the diameter of a human hair.
Electroforming is an electrochemical plating technique that deposits a thin
metal layer (called a foil), usually nickel, onto a substrate with a patterned metal
layer on top of an insulator. The metal pattern defines the plating areas, and
circular dots exposing the underlying insulator define the nozzles. Other geometries of exposed insulator define openings in the foil useful for alignment targets
and separating the individual orifice plates. Hundreds of orifice plates are formed
simultaneously on the foil sheet, each with hundreds of nozzles.
As the plating process adds depth to the foil, metal advances uniformly across
the insulator dots from the edges. This creates a smooth converging nozzle bore.
The final nozzle diameter depends critically on the plating thickness because the
diameter grows smaller at twice the rate of depth accumulation.
The material composition of the foil and the patterned metal are chosen such
that the foil has weak adhesion to the substrate. This property allows it to be
peeled off after plating is completed. A thin gold passivation layer is often
electroplated onto both sides of the foil to minimize corrosion from the ink, and
Color Desktop Printer Technology
the individual orifice plates are then separated from the sheet and assembled onto
the substrate’s thick-film layer.
Laser ablation uses an excimer laser delivering bursts of pulses, each of a
nanosecond duration. Each pulse vaporizes a thin, patterned region in a polyimide
(Kapton) tape to open a hole or channel. Multiple pulses are required to ablate
through 50 microns of polyimide. The vapor produced by the ablation process
carries away the ash to leave a clean hole. The effects of laser ablation are highly
localized, and there is virtually no thermal damage to the material surrounding
the nozzle. Compared to the multiple process steps and wet chemistry of electroforming, laser ablation is a dry process offering an economical, high-productivity method to produce a sharp-edged nozzle bore. Another key advantage of
this method is that the polyimide tape can function both as the orifice plate and
the electrical interconnect.
Figure 4.7 shows a 512-nozzle thermal inkjet printhead with a silicon substrate and laser-ablated orifice plate integral to a polyimide tape. This tape has
electroplated electrical conductors connecting the substrate to the 21 gold-plated
electrical interconnect pads. These pads provide power, control signals, and data
interconnection between the printhead and printer. Control, demultiplexing, and
drive electronics integrated into the printhead substantially reduce the number of
electrical interconnect pads, and this minimizes the cost and size of the printhead
and improves the reliability of electrical interconnect.
Orifice Plate
Ink Barrier
CVD SiO2 &
Field Oxide
NMOS Transistor
Field Oxide
Ink Barrier
Orifice Plate
Gold (conductor)
Aluminum (conductor)
Chemical vapor deposition
Silicon dioxide (insulator)
Photoimageable thick film
Nickel or polyimide
Silicon carbide (insulator & anticavitation barrier)
Silicon dioxide (insulator & thermal barrier)
Silicon nitrite (insulator & anticavitation barrier)
Tantalum (anticavitation barrier)
Tantalum aluminum (resistor)
FIGURE 4.7 (See color insert following page 176.) An HP thermal inkjet printhead and
electrical interconnect.
Considering the leverage from integrated circuit manufacturing and process
control, thermal inkjet offers a very cost-effective method for high-volume production of inkjet printheads.
About 25% of inkjet printers sold worldwide use piezoelectric transducers for
drop ejection. Piezoelectric materials change their dimensions in response to
applied electric fields, and the most common material used in piezo inkjet printheads is lead zirconate titanate (PZT) ceramic.
When driven by an electric pulse, the piezo element typically changes its
long dimension or bends. In either case, the displacement is typically a micron
or less. Whereas thermal inkjet can be characterized as a piston (vapor bubble)
stroking in a cylinder (drop generator chamber), the piezo drop generator operates
on the principle of a volumetric transformer: a small displacement (piezo deformation) over a large area (diaphragm) produces a large displacement (the ink
droplet) over a small area (nozzle). To work properly, the fluid in the drop
generator must be practically incompressible.
Epson’s MLP drop generator, circa 1997,1 shown in Figure 4.8 is an example
of a device operating by direct elongation of a piezo crystal. In this printhead,
each ejection chamber is formed in a silicon channel layer with a stainless steel
orifice plate on one side and a diaphragm on the other. A 5-mm-long PZT
transducer bonded to its exterior displaces the diaphragm wall of each chamber.
Polyimide Tape
Silicon Substrate (see-through tape)
Gold-Plated Electrical
Interconnect Pads
FIGURE 4.8 (See color insert following page 176.) Epson MLP piezo inkjet drop
Color Desktop Printer Technology
The diaphragm seals the chamber, allows displacement of the wall, and isolates
the PZT ceramic from the ink. The PZT transducer is fabricated from a stack of
20-micron-thick layers to reduce the required drive voltage. Dimensional changes
are small: in operation, the 5-mm-long transducer elongates by only one micron.
As a result, the drop generator chamber requires a large diaphragm to create the
required volumetric displacement, and this makes the piezo drop generator much
bulkier than a thermal inkjet producing a comparable drop volume. Nozzles are
spaced 140 microns apart, giving a linear density of 180 orifices per inch. This
design prints dot rows at 180-dpi vertical resolution, so the complete area fills at
720 dpi require 6 passes where the paper is advanced 1/720 inch between each
printhead scan.
Rather than using direct elongation, many piezo printheads employ a bendingmode bimorph sandwich using a thin piezo element bonded along its length to a
thin piece of another material. The Epson MLChips printhead,2 shown schematically in Figure 4.9 is an example of such a design with a PZT/zirconia bimorph
attached to a zirconia ink chamber. Three laminated stainless steel plates forming
the nozzle layer, ink manifold layer, and ink inlet layer are adhesively attached
to the zirconia layers to complete the structure. The bimorph structure bends
when voltage is applied to the piezo element, and this produces a reduction in
the chamber volume to eject a droplet of ink. Bimorph deflection in this design
is about 0.1 microns, so the lateral dimensions need to be larger than in the MLP
design to achieve the desired volumetric displacement. Chamber width is 340
micron, compared to 100 microns for the MLP design, and the nozzle spacing is
therefore larger. This printhead can be produced at lower cost than the MLP, and
high printing resolution is achieved with multiple passes.
Orifice Plate
Silicon Channel Layer
Ink Path
Piezo Transducers
FIGURE 4.9 Epson ML chips drop generator.
A bending-mode bimorph configuration is also used in the solid inkjet Xerox
Phaser printheads.3 These printheads are fabricated entirely from a stack of
photomachined stainless steel plates, brazed together with thin gold bonding
layers. Photochemical machining typically uses a dry-film resist and spray
etchants to produce plates with holes and channels.
A variety of electrical waveforms is used to drive piezo inkjet printheads.
This is an advantage of piezo inkjets that allows a given design to deliver the
same drop volume using inks of different viscosities and surface tensions, to
adjust for manufacturing variations, and to eject drops with different volumes
from a single orifice.
A common waveform is the bipolar drive pulse, in which a pulse of reverse
polarity is first applied to the piezo transducer to enlarge the ejection chamber.
This condition is maintained until ink has been drawn into the chamber from the
ink reservoir. Then a normal polarity pulse is applied to reduce the chamber
volume and eject the ink from the nozzle. Compared to a unipolar pulse, a bipolar
pulse with the same peak voltage can eject higher drop volumes.
More complex waveforms enable piezo inkjets to eject drops of different
volumes from an orifice, as seen schematically in Figure 4.10. This technique
uses a first pulse to excite an oscillation of the meniscus in the nozzle, followed
by a higher-amplitude pulse (or pulses) to eject the drop. Depending on the relative
timing and amplitude of these pulses, 2.5-, 5-, and 11-pl drops can be produced.
One inherent disadvantage of piezo inkjet is its high sensitivity to air bubbles.
Unlike thermal inkjet, in which the pistonlike vapor bubble produces high ink
velocities throughout the drop generator to sweep out air bubbles, the ink in a
piezo drop generator is virtually at rest except where the chamber cross-sectional
area is very small compared to the diaphragm: near the orifice. This makes it
difficult for piezo inkjets to purge bubbles in normal operation. Bubbles trapped
Metal Plates
Zirconia Ink
FIGURE 4.10 (See color insert following page 176.) Drive waveforms for variable drop
Color Desktop Printer Technology
in the drop generator can completely disable the piezo drop ejection process
because it critically depends on the incompressibility of the ink in the drop
ejection chamber. Bubbles are compliant and reduce the amplitude of the volumetric displacement at the nozzle.*
To prevent performance degradation from bubbles, piezoelectric printheads
require regular priming by pumping ink through the printhead and disposing of
it in the printhead service station. The large waste ink disposal pads found upon
disassembly of piezo inkjet printers indicate the large quantity of ink required
for this process over the lifetime of the printer..
A key difference between piezo and thermal inkjet is that the piezo drop
ejection process does not heat the ink. This allows piezo inkjets to use ink
components and colorants that would be unsuitable for use in thermal inkjet
devices. So far, this has not proven to offer any practical advantage, as inks have
been developed for both technologies that achieve similar imaging performance
in desktop color printing applications. Ink compatibility with the materials used
in the printhead and ink delivery system is by far the major determinant of the
range of solvents and other ink components that can be used by a particular
Some disadvantages of piezo technology over thermal inkjet include higher
fabrication cost, lower nozzle density, larger overall size, and greater sensitivity
to trapped bubbles.
Piezo printheads are designed to last the lifetime of the printer and are
typically not user-replaceable. Kogation and cavitation are not failure mechanisms
for piezo printheads. Piezo and thermal inkjet printheads are both subject to
degradation from long-term exposure to ink, and this can cause failure of adhesives and structural delamination, changes in nozzle shape, unrecoverable clogs,
corrosion, and material property changes.
The early history of unreliable products with messy liquid ink supplies created
significant barriers to adoption for first generations of modern inkjet printers.
Providing clean, simple, and economical replacement of consumables would seem
to be the whole story for the ink delivery system. To understand the success of
desktop color printing is to recognize the impact the ink storage and delivery
system has on consistent print quality and reliable operation and how it is one
of the most important determinants of printer manufacturing cost and the cost to
print a page.
This effect is seen in a symbolic volume conservation (continuity) equation for the drop generator:
(diaphragm stroke × diaphragm area) = (total change in trapped bubble volumes) + (ink volume
backflow into the refill channels) + (ink displacement × nozzle area). A non-zero first term on the
right-hand side of the equation obviously reduces the subsequent two terms. This is only a kinematic
equation that does not take into account viscous and inertial effects in the nozzle and refill channel,
and these set the ratio between the second and third terms.
Desktop color inkjet printers have multiple printheads mounted side by side
on a carriage that scans across the paper. The printheads print rows of dots, the
paper is advanced, and the process repeats. The print carriage must be able to
move out of the print zone on either side of the scan to allow all printheads to
print up to the left- and right-hand margins; to allow the carriage to decelerate,
stop, accelerate back across the print zone; and to park the carriage in the
printhead service station. Compact printers fulfill an important user need for the
efficient use of desktop space. The width of the printer is driven by the widest
media it can handle plus about twice the width of the print carriage. Minimizing
the width of the carriage is an important design consideration, because printer
width has a strong influence on manufacturing cost. For example, manufacturing
costs increase for a wider printer that requires a longer print carriage slider rod,
longer electrical cables to the carriage, a longer encoder strip, a longer drive belts,
and the need for stiffer plastic and sheet metal parts in the chassis.
High printing throughout depends on minimizing the time the printheads are
not printing. This means that the print carriage must stop quickly at the end of
a scan and rapidly accelerate to constant velocity (typically 0.5–1 meter/sec) back
across the print zone. To minimize printer width, an acceleration of 0.5–1 g*
requires the printer to deliver a force on the print carriage equal to 50–100% of
its weight. It is easily appreciated that minimizing carriage mass reduces the cost
of the scan axis servo motor and power supply, reduces an annoying side-to-side
shaking as the printer operates, and may even make the printer quieter.
The amount of ink onboard the carriage and principle of its containment and
pressure regulation directly affect print carriage width and mass. The amount of
ink contained in a disposable print cartridge has a direct impact on the cost per
page. But, the actual ink yield is a function not only of the initial fill quantity but
also how much ink is used to print, how much is consumed during printhead service
cycles, and how much is unusable due to ink delivery system characteristics.
In a drop-on-demand printer, hydrostatic pressure in the drop ejection chamber is maintained at a slightly negative pressure relative to the atmosphere to
produce enough suction to support a water column 2-5 inches high. This prevents
ink from flowing out of the nozzles when the printhead is idle. Air is not sucked
into the printhead because the nozzle menisci act like elastic membranes whose
deformation (into the nozzle bore) balances this negative pressure. Shock to the
printhead, such as dropping it on a hard surface, can cause the menisci to detach
from the edge of the nozzle and allow air to enter causing a deprime.
Consistent drop volumes depend on maintaining constant delivery pressure
at all flow rates over the life of the ink supply. Typical desktop printer ink supplies
hold between 7 and 69 ml. In some designs, the ink delivery rate to a single
printhead can vary between zero and a peak value of 7 ml/minute.
A number of different technologies are commonly used to store and deliver
ink to the printhead. Ink reservoirs moving with the print carriage, either integral
with a printhead or separately replaceable, are called on-axis supplies, meaning
1 g is the acceleration of gravity: 9.8 meters/second/second.
Color Desktop Printer Technology
“on the scan axis of the print carriage.” Fixed ink reservoirs separated from the
printheads are called off-axis supplies. The following discussion begins with
various types of on-axis ink supplies.
Early inkjet products used elastomeric bladders. Starting from a dimple,
bladders are designed to collapse in a manner that applies a nearly constant
(suction) pressure as ink is withdrawn. The need to ensure a repeatable collapse
limits bladder capacity to a few milliliters of ink. The bladder must provide a
barrier to gases and volatile ink compounds. Material compatibility with the ink
is a significant design issue for bladders: vulcanizing agents and mold release
compounds can contaminate the ink and dramatically affect physical properties
such as surface tension. Bladders are usually hemispherical or cylindrical, and
these bulky configurations limit the usefulness of bladders when a higher-volume,
compact ink delivery system is required.
A large percentage of disposable thermal inkjet print cartridges hold the ink
in a block of non-rigid, open-cell foam inside a rigid plastic container. Foam
provides a porous medium for ink storage and flow. The capillary action at the
ink–foam–air interface maintains suction as ink is withdrawn. Foam ink cartridges
have a small vent to allow air to replace the ink as it is consumed. The pore size
and compression of the foam must be carefully controlled to produce a nearly
constant delivery pressure over the life of the ink supply.
Given the large surface area it presents to the ink, ink–foam material compatibility presents a major issue whenever new inks are developed: the foam must
be free of compounds such as plasticizers and surfactants, that could interact with
ink components. Otherwise, significant changes in ink surface tension can occur,
leading to loss of pressure regulation and allowing ink to drool out of the nozzles.
Rapid withdrawal of ink from a foam reservoir can cause a large increase in
suction. This can deprime the printhead in severe cases and modulate the drop
volume in others. In addition, air can enter the pores and isolate ink-filled regions.
Unless these regions reconnect by capillary action, significant quantities of ink
can be stranded and the full capacity of the ink cartridge cannot be utilized.
In foam blocks with linear dimensions of more than a few inches, hydrostatic
forces can overcome capillary forces. In this case, ink can drool out of the orifice
plate in the normal printing orientation or can migrate away from (and possibly
deprime) the ink supply pipe to the printhead when the ink cartridge is handled.
Long ink paths to the supply pipe increase the possibility of stranding ink. For
these reasons, and to minimize the linear dimension of the ink cartridge, foam
systems are useful up to about 20 ml of ink per color.
Some simple mechanical systems, including bubblers, spring bags, ink bags,
and regulators, have been developed to overcome the limitations of bladders and
A bubbler uses a small orifice and a labyrinth baffle to allow air to bubble
into a rigid chamber holding liquid ink. This produces a controlled suction in the
ink supply as ink is consumed. A small spring-loaded plastic bag in the chamber,
vented to the atmosphere, acts like a bellows to balance changes in atmospheric
pressure. Bubblers can supply about 40 ml of ink, twice as much in the same
(nearly cubical) volume as a 20-ml foam-based ink cartridge.
Spring bags contain the ink in a sealed, flexible, metalized plastic bag. The
bag contains two metal plates held apart by a leaf spring. This mechanism is
immersed in the ink, and the plates press outward against the walls of the bag to
resist its collapse as ink is withdrawn. This allows a large quantity of ink, up to
about 100 ml, to be withdrawn at nearly constant pressure. The bag is contained
within a rigid cartridge body. Spring-bags allow the design of a narrow and tall
ink cartridge, and the collapse of the bag can operate a simple mechanical ink
quantity indicator on the print cartridge. Spring-bag ink cartridges have been used
in more than 100 million HP DeskJet and PhotoSmart printers sold since 1993.
Metalized plastic bags of ink contained in rigid plastic cartridge bodies are
used in off-axis ink delivery systems. Only the space available and industrial
design of the printer limit the capacity of this system, which delivers 69 ml of
ink per color in some configurations. Ink is pumped through pressurized flexible
tubes from the ink bags to the printheads. The tubes are designed to limit evaporation of volatile ink components, provide a barrier to air, and remain flexible
for the life of the printer. A pressure regulator on or near the printhead admits
pressurized ink as it is consumed and delivers it to the printhead at a constant
negative pressure. Some ink cartridges also contain a pad to absorb waste ink
produced during printhead servicing.
Desktop inkjet printers require service stations to ensure reliable operation of the
printheads. Service stations perform the essential functions of capping, spitting,
priming/purging, and wiping. These functions are controlled by printer firmware
that keeps track of drops printed by each nozzle on each printhead, the time the
printhead is exposed to air while printing, the longest time any nozzle has gone
without printing, the number of pages printed between service cycles, the time
since the last page printed, and other operational parameters.
Some sophisticated service stations use optical or electrostatic drop detectors
to determine which nozzles are operating within drop weight, drop velocity, and
trajectory specifications. This information is used in nozzle substitution algorithms to replace unfit nozzles by good ones in multiple-pass print modes until
the nozzles are either recovered by a service cycle or identified by the printer as
An elastomeric cap in the service station contacts the orifice plate of each
printhead to seal the printhead nozzles from the atmosphere when not in use. The
cap provides a humid environment that limits the evaporation of volatile ink
components, preventing the formation of viscous plugs and ink crusts.
Viscous plugs and ink crusts can make drop ejection impossible, and this
problem is more severe with piezo inkjets than thermal inkjets due to piezo’s less
energetic drop ejection process. Ink crust at a nozzle’s edge can misdirect ejected
drops. With some inks, the ink crust, once formed, is insoluble in the ink vehicle.
Color Desktop Printer Technology
Allowed to harden, it may be impossible to remove and the printhead must be
replaced. A soft crust may be mechanically removed by wiping. Capping provides
a critical function: for some inks, the nozzles can be exposed to air for only a
few seconds without forming crusts or viscous plugs.
Along with viscosity increases, the loss of volatile ink components from
the nozzles can concentrate colorants. If the nozzle is actually able to eject this
concentrated ink, which will have different fluid properties, then its printing
characteristics will be different from normal ink. The dots may be misplaced
and have higher optical density. Different spread and penetration characteristics
on the print medium will produce a different-size dot. For this reason, it is
important to keep fresh ink in the nozzles, especially when certain nozzles or
colors have not printed for a while. For example, when printing a black text
document, the color printheads are exposed to air in the scanning carriage but
may not print for several minutes. However, they must be ready to produce
drops meeting print quality specifications at any time. This is assured by a
process known as “ spitting.”
Spitting periodically ejects drops from each nozzle into the service station’s
waste ink reservoir. Spitting a few drops from each nozzle can eliminate concentrated colorant and viscous plugs developing in unused nozzles. If a printer is
observed to pause in the middle of a page to return the print carriage to the service
station for a few seconds, it is almost certainly to spit from unused nozzles.
Priming and purging are processes that remove air bubbles trapped in the
drop generators. A pump is required to create a subatmospheric pressure (or
impulse) to draw ink out of the printhead. Suction is preferred to pressurizing
the drop generators through their ink supply channels, because suction causes
trapped air bubbles to expand and be drawn out by ink flow. The amount of ink
used in servicing the printhead can be substantial, especially for piezo printheads,
which do not have an effective means for continuously purging the drop generator
of trapped gas bubbles.
The ink used in spitting, priming, and purging must be stored in a waste
receptacle designed to last the lifetime of the printer or in a waste ink container
in a disposable ink cartridge. This is typically an absorbent pad connected by
tube or capillary to the service station spittoons and priming pumps. Evaporation
of volatile ink components over time, especially water, helps limit the volume of
the waste ink storage medium.
A flexible, elastomeric wiper blade can be passed across the orifice plate to
remove ink spray, ink crust, and paper dust. In some cases, the printheads eject
a small amount of ink on the wiper to obtain a wet-wipe, and this helps to dislodge
ink crusts.
Materials compatibility between the inks and service station caps and wipers
must be assured. Otherwise, the elastomer may decompose, leaving sticky residues on the orifice plate that permanently clog nozzles. In the worst case, these
residues can affect the wetability of the orifice plate, causing a complete loss of
ink containment with the potential of irreparable damage to the printer and
its environment. Apart from image quality, the consequences of material
incompatibilities are the reason why most inkjet manufacturers recommend their
own original ink supplies and caution against ink cartridges supplied or refilled
by third parties.
Wiping multicolor printheads, where two or three different inks share the
same orifice plate, can pose problems from ink mixing. Liquid ink from one
color, or from spitting on the wiper, can be drawn by capillary action and negative
head into drop generators of another color. This can produce off-color droplets
or chemical reactions that clog the nozzles. In this case, spitting occurs during
or immediately after the wipe to ensure ink purity in each color printer.
Excessive wiping can scratch the orifice plate and produce capillary channels
drawing out ink and accelerating crusting. Wiping can damage non-wetting coatings and wrinkle a flexible plastic laser-ablated orifice plate causing it to delaminate from the underlying structure. In the worst case, wiping can exacerbate
nozzle clogging by forcing foreign materials into them. For these reasons, wiping
is done only when necessary. Some service stations keep track of the number of
wipe cycles for each printhead in order to balance the benefits and harms of
The user can initiate cleaning and priming service cycles if the print quality
becomes unsatisfactory (e.g., white streaks are seen in characters or area fills),
using commands in the printer driver or toolkit installed with the driver on the
user’s computer.
Both liquid and solid (hot-melt) inks find application in desktop color inkjet
printers, with liquid ink used in virtually all consumer and business inkjets and
solid ink in some graphic arts and proofing applications.
Whether liquid or solid, ink has two functional components: the colorant and
the vehicle. Although the role of the ink vehicle would appear simply to deliver
colorant to the paper, in fact, ink touches everything in the printing process and
it is the most functionally complex part of any inkjet device.
The journey of a liquid ink drop begins with the storage of ink over a shelf life
of 6 to 54 months and a service life of many months in the printer. During this
time, the colorant must remain dissolved (dyes) or suspended (pigments). Chemical and physical properties must be stable without chemical interactions between
ink components and materials in the ink containment, pressure regulation, and
printhead components. Ink surface tension may play a role in regulating pressure
in the ink delivery system. Service station features and functions depend on ink
volatility and crusting characteristics. While the nozzles are exposed to air, ink
* The authors acknowledge contributions to this section by Dr. Nils Miller and Dr. John Stoffel of
the Hewlett Packard Company.
Color Desktop Printer Technology
surface tension and viscosity must remain within narrow limits to ensure proper
drop volume, velocity, and trajectory. In a thermal inkjet, the ink’s vaporization
characteristics are part of the drop ejection process, and any thermal decomposition products must remain soluble in the vehicle. Once the droplet reaches the
paper, the vehicle must rapidly wet and penetrate the surface to control dot spread.
While penetrating, the vehicle separates from the colorant, leaving it near the
surface to maximize optical density. The vehicle should minimize cockling of
plain papers and evaporate, leaving behind a dot that meets print-quality objectives for size, shape, optical density, fade resistance, and waterfastness.
Typically, ink vehicle in desktop color printers is mostly water with various
additives. The colorant is only about 5 to 10% of the ink by weight. Although water
is not necessarily the ideal solvent from printing considerations, especially on plain
papers, it is the only solvent that meets worldwide health and safety regulations for
the home and office. These regulations also set guidelines for flammability and
toxicity and limit the percentages of organic compounds, such as alcohols and
glycols, that may be used to meet various performance requirements.
The formulation of inks represents significant intellectual property developed
over years of research and development, covering thousands of compounds and
combinations. Introducing a new ink can cost millions of dollars exclusive of
investment in production facilities. Although manufacturers may have hundreds
of patents issued on ink, certain properties, ingredients, and manufacturing processes remain trade secrets. Some essential components in the finished ink are
difficult to characterize, even with the most modern analytical chemistry techniques. For this reason, inkjet inks are difficult to completely reverse-engineer.
Typically, inkjet manufacturers partner with chemical companies to research
and develop new colorants and other key ink components, and the inks are
formulated by the printer manufacturer to be part of a system with the printer,
printheads, and specialty media (such as photo papers). Once the ink properties
are fixed for a specific product platform, the partner builds and operates the ink
production facilities. To give a sense of scale, ink production began at 250,000
liters a month for a new print cartridge platform recently introduced by a major
inkjet manufacturer. The ink was delivered to the print cartridge production line
in tanks each containing a cubic meter (1000 liters).
Before printheads or ink cartridges are filled with ink, the ink undergoes
hundreds of physical and chemical analyses in the manufacturing process to
ensure consistent properties and performance. These analyses include measurements of pH, viscosity, surface tension, and particle size distribution; gas and
liquid chromographic analyses of solvents and dyes; spectrographic elemental
analysis; anion analysis; testing in printheads for drop ejection consistency, storage and operational life, and materials compatibility; and print testing for
smearfastness, fade, waterfastness, color, and optical density.
The largest component of ink is typically highly purified water, and most
inks used in desktop inkjet printers have a viscosity of between 1.5 and 3
centipoise. Surfactants are added to obtain a surface tension in the range of 20
to 30 dynes/cm. Humectants, usually glycols, may be added to minimize the
effects of viscous plugs in the nozzles by maintaining equilibrium between water
vapor loss to and absorption from the atmosphere. Small amounts of alcohol
increase the rate of vehicle penetration into the paper by increasing the ink’s
ability to wet the surface. Cosolvents keep dyes in solution and pigments in
suspension. Buffering agents maintain ink pH within a narrow value. Biocides
and fungicides prevent growth of organisms in the ink during storage and use.
Agents are added to control cockle on plain papers.
The payload of the ink vehicle is the colorant. All manufacturers of desktop
color inkjet printers offer both dye and pigment colorants. Each type has its
advantages and disadvantages and is ideal for certain applications and somewhat
less suited to others.
4.5.2 DYES
Dyes are molecules designed to absorb specific frequencies of light, and they are
chemically dissolved in the ink vehicle. With dimensions on the order of a
nanometer, dye molecules are small enough to penetrate into any absorbent
medium, so the surface of the print medium determines the gloss. Because they
are too small to scatter light, dyes are typically brighter and more colorful than
pigments, and this usually allows the dye-based printer to produce the largest
color gamut.
Dyes are individual molecules, and after printing their internal chemical
bonds can break due to exposure to light, moisture, oxygen, and other environmental chemicals. The interaction of dye molecules with the print medium has
a dramatic effect on the fade resistance and waterfastness of the print. Encapsulation of dyes within a coating on the print medium mitigates all effects but light
exposure, and UV-absorbing laminates can significantly extend the time to fade.
For example, a certain set of inks printed on photo papers specifically designed
to resist fade can deliver more than an order-of-magnitude longer fade resistance
than on other media. Similarly, some dye-based inks may deliver poor waterfastness on some media, but others can provide a completely waterfast solution.*
One method for improving fade resistance is to design the primary colorants
to fade at the same rate. When one primary, usually magenta, fades faster than
the others, an objectionable hue shift will occur. This is commonly seen in
conventional photographs after long-term exposure to florescent lights or sunlight:
skin tones become greenish with the fading of the photographic magenta dye.
When all the dyes fade uniformly, the decrease in color saturation is far less
noticeable to the eye than a hue shift.
Pigments are particles about 50 to 150 nanometers in diameter composed of tens
of thousands of dye molecules bound together. Unlike dyes, pigments are not
dissolved in the ink vehicle: they are formulated with dispersing agents to create
See the discussion of porous and encapsulating coatings in the section on inkjet media.
Color Desktop Printer Technology
a stable suspension. If this suspension fails during storage or from exposure to
temperature extremes, contaminants, or printhead materials, the pigments will
come out of suspension in the ink vehicle to form sludge in the ink container or
printhead. This is called a pigment “crash.” This is a failure that can be recovered
only by replacing the ink supply, printhead, or both. It is why it is important to
observe the “Use By” date the manufacturer specifies for the ink cartridge.
Pigment ink cartridges from some manufacturers have a shelf life of only six
Pigment particles are too large to penetrate most inkjet papers or photo media.
Instead, they form a thin film on the surface, and it is difficult to achieve a high,
uniform gloss meeting photo-quality expectations: the printed regions can appear
dull when viewed at certain angles. Some manufacturers offer special media for
pigment inks that provide a uniform satin gloss.
Although pigment inks for desktop printers are water-based, pigments offer
excellent waterfastness and resistance to highlighter smear on plain papers by
forming stable chemical bonds with cellulose fibers. Waterfastness usually takes
several hours to fully develop. Pigments can offer excellent fade resistance: when
dye molecules on the surface of the particle decompose, underlying molecules
take their place. The more chromatic pigments tend to be less fade resistant.
The most recent generations of high chroma color pigments for photo applications offer fade resistance in the 100-year range according to predictions from
industry-standard test methods, such as those performed by Wilhelm Imaging
In some cases, dyes are added to pigment inks to achieve higher chroma or
to adjust the hue. Fading of these dyes over time or reduced waterfastness can
undo their benefits.
The chemistry of pigment dispersing agents can be designed to react with
dye-based color inks so that mixing the two on the print medium will cause the
pigment to crash. This achieves high edge sharpness in black text or lines printed
on a colored background, particularly a light color such as yellow, because the
feathering of black text into light colors produces an objectionable loss of sharpness.. When printing on plain paper, a process known as “black fortification”
prints a small amount of dye-based cyan and magenta inks under the areas to be
printed with black pigments. The pigment particles are quickly immobilized
before capillary forces can carry them away, feathering the dot edge. This allows
some inkjet printers to produce uniformly high quality black text on a variety of
plain papers, and this media-independent print quality meets a significant user
This type of reactive pigment/dye system is incompatible with printing continuous-tone color images. Cyan, magenta, yellow, and black (CMYK) printers
using pigment black ink print color images on photo media with only the cyan,
magenta, and yellow (CMY) dyes. Overprinting cyan, magenta, and yellow drops
at the same location produces black dots. Even disregarding chemical reactivity,
the different surface penetration characteristics of dyes and pigments would
produce unacceptable non-uniformity in gloss.
Some pigment ink printers use two black inks: one for satin or glossy photo
media and one for matte surfaces, such as fine art papers. These inks match the
gloss of the substrate, and the matte black usually delivers a higher optical density.
With a diameter between 1/10 and π the wavelength of green light, pigment
particles are large enough to scatter visible light, so they produce less saturated
colors than dyes. However, even the smallest inkjet nozzles are 100 times larger
than a pigment particle, so pigmented inks still behave like liquids at this scale.
Solid inks are also called hot-melt inks and phase-change inks. They are based
on materials that are solid at room temperature but have low-enough viscosity in
the liquid phase to be jettable. The ink vehicle is typically a natural or synthetic
wax or a mixture of both. These materials have glass transition points between
80 and 100ºC and are heated to around 130ºC to bring their viscosities below
about 20 centipoise for ejection.
Once printed, cooling of the ink droplets produces a rapid increase in viscosity
followed by a phase-change back to solid form. The ink forms a thin layer on
the surface of papers and other substrates, and it does not penetrate like liquid
ink. This minimizes the effect capillaries can have on dot size and shape, and
keeping the colorant on the surface produces high print density and color saturation. For these reasons, solid inkjets have the potential to deliver print quality
with higher media independence compared to liquid ink systems. Despite these
advantages, the mechanical properties of solid inks at room temperature present
a number of practical challenges. If the layer is too brittle, bending or creasing
the paper will cause it to crack. If the ink is made ductile to minimize cracking,
it may be too soft and susceptible to scratching. Scratches dull the surface gloss
of printed regions and can occur when printed sheets rub against each other in a
Mechanical durability of printed text and images improves with higher glass
transition temperature, Tg. A higher Tg protects the printed dots from smearing
if they are exposed to high environmental temperatures (such as in a closed vehicle
parked in the sun). However, the temperature at which the melt has an ejectable
viscosity is typically 20 to 30ºC above Tg, and a high Tg presents a problem for
using piezo materials. Piezo devices must be operated below their Curie Point,**
and this presents a constraint of about 130ºC on the melt-temperature for solid
ink printers.
Solid ink is usually supplied in the form of a solid rod or block, and this is
inserted into the printer and melted in a chamber close to the printhead. Xerox
keys the cross-sectional shape of each color ink block to that color’s re-supply
slot thus ensuring that inks are always loaded properly. Some solid ink printers
The authors acknowledge contributions to this section by Dr. Wayne Jaeger of Xerox Corporation.
At temperatures above the Curie Point, piezoelectric transducers lose their ability to produce
dimensional changes with applied electric fields.
Color Desktop Printer Technology
allow ink blocks to be reloaded on-the-fly, and this is a useful feature that allows
print jobs to continue without interruption.
Melt-on-demand systems are preferred over systems that keep the ink in liquid
form, because keeping the ink hot for long periods can produce changes in
material properties due to polymerization, decomposition, and loss of volatile
compounds. On the other hand, repeated freeze–thaw cycles, in which the printer
is completely turned off between uses, can produce significant mechanical stresses
on the printheads due to specific volume differences between the solid and liquid
phase. Freezing and thawing can also affect the solubility of the components of
solid ink.
Because solid ink printers use piezo drop generators, removing dissolved
gases and bubbles from the liquid phase presents a significant challenge. Manufacturers such as Xerox have developed ingenious schemes to extract gases from
the melt near the printhead using, for example, a gas-permeable membrane
between the melt and a vacuum chamber.
The phase-change process is so rapid that drops form nearly hemispherical
beads on the surface. Without a post-treatment to flatten the solidified drops, these
beads could give an undesirable grainy texture to text and area fills. Acting as
tiny lenses, the beads would scatter light to make solid ink useless for printing
color on overhead transparency films. Xerox Phaser products print solid ink drops
onto a heated intermediate transfer drum. The image on the drum is transferred
under pressure to the print medium, and this produces a smooth surface. Other
solid ink printers use a cold pressure roller or re-melt the ink after printing to
form a smooth film.
The media available for desktop color inkjet printers present the user with broad
choices for meeting cost and quality objectives in document and photo formats.
Special media are available for printing banners, brochures, greeting and business
cards, iron-on transfers, and other creative projects.
Each major inkjet printer manufacturer develops its own ink formulations,
and because inkjet print quality depends on the print medium, inks and inkreceptive coatings are generally designed together for best performance. Therefore, major manufacturers offer a collection of special media optimized for use
in its own printers. Because all manufacturers endeavor to meet similar performance and quality objectives, there are enough similarities in their solutions to
support third-party suppliers of inkjet papers and specialty media in markets
where low-cost and unique features are more important than ultimate image
quality or permanence.
The print medium has many requirements for mechanical and imaging performance. Mechanical performance relates to handling and reliability in the
* The authors acknowledge contributions to this section by Dr. Nils Miller of the Hewlett Packard
printer paper path and involves material properties of the surface and substrate.
Imaging performance relates to optical properties of the surface and substrate
and to physical properties that control the spread, penetration, and absorption of
Surface finish and coatings, which control the coefficient of friction with the
printer’s feed and drive rollers, and the medium’s bending stiffness are the most
mechanical properties of print media. These determine the reliability of picking
single sheets from the input tray and feeding them without jamming or skew
through the printer paper path.
Cockle is an uneven, bumpy texture across areas of high print density. Plain
papers can cockle in wet regions, because water causes cellulose fibers to swell
and water breaks the hydrogen bonds between fibers. This allows fibers to move
relative to one another, relieving mechanical stresses built-in during paper manufacture. Cockle can form immediately in the print zones (wet cockle) or in the
output tray (stacker cockle), where moisture lost by neighbor sheets causes
unprinted regions to cockle. Cockle can partially or completely disappear as the
sheet dries, but in the worst case leaves permanent wrinkles that can render the
print unusable. Water-based inks usually contain anti-cockle agents to minimize
this effect. Severe wet cockle produces wrinkles out of the print plane that can
cause printheads to crash against the paper. The printhead-paper distance is
typically 1–2 mm, and wet cockle can reach this height. A printhead crash
smears ink across the sheet and can damage the printhead’s orifice plate. Cockled
sheets can cause feed problems during duplex printing. Curl is another moisturerelated phenomenon, in which differential expansion (or contraction) of paper
layers cause the paper to warp or even form a cylinder.
The feel of inkjet papers and photo media meets important user needs.
Documents printed by inkjet should have the same surface texture, stiffness, and
weight as those produced by laser printers and copiers. For consumer acceptance
of photos printed by inkjet, it is important not only to deliver the image quality
of conventional photographs but also to reproduce the tactile properties of the
photographic surface and substrate.
Surface treatments and coatings, ink-receptive layer(s), and the base color of the
print medium play an important role in image quality. These determine the surface
gloss and finish; size, dot shape, and optical density; overall image brightness
and hue; ink capacity; and dry time. Ink-receptive layers play a significant role
in print fade resistance and waterfastness. The opacity and ink permeability of
the substrate is important to minimize or eliminate strikethrough. Strikethrough
is an undesirable property where an image printed on one side of the sheet is
visible (to some extent) on the other side. This is often a problem when printing
Color Desktop Printer Technology
images or color graphics on plain papers, especially thinner (e.g., 16 pound)
papers. Strikethrough limits the suitability of paper for duplex printing in all but
the lowest quality applications. Typically, strikethrough can be minimized or
eliminated by using a heavier grade of plain paper or an inkjet paper that is coated
on both sides.
Substrates come in two general forms: absorbent and nonabsorbent. Paper is
the most common example of an absorbent substrate; the print medium itself
absorbs the ink. Because untreated paper can allow excessive penetration of ink,
coatings are applied to hold the colorant near the surface to achieve high color
saturation and minimize strikethrough.
Nonabsorbent substrates are used in overhead transparency (OHT) films and
some photo papers to give a smooth base for transmission or reflection of light.
PET (i.e., polyethylene terephthalate) is commonly used for OHTs; some photo
papers use a polyester base or highly sized* papers made waterproof by laminating
a polyethylene film on both sides. For these substrates, the ink receptive coating
must hold the entire quantity of ink deposited during the printing process, and
this can be as much as 35 ml/m2. Although some ink vehicle may escape by
evaporation over time, failure to absorb all the ink results in ink pooling on the
surface, wasting the print and contamination of the printer paper path from liquid
During printing, inkjet-receptive coatings control the movement of the ink
vehicle on the surface of the print medium affecting color bleed, area-fill uniformity, and dot gain. Coatings remove the vehicle from the surface by absorption,
and this controls short-term drying that affects smearfastness (for duplex printing)
and the elimination of blocking. Blocking occurs when a sheet is stacked on top
of a freshly printed sheet in the printer output tray. The printed image of the
damp sheet may transfer colorant to the bottom side of the upper sheet.
Most ink-receptive coatings are composed primarily of pigments and polymers with small amounts of additives. Pigments used in coatings may be particles
of silica, alumina, clay, titania, and various carbonates that are mostly non-soluble
in the ink vehicle. Irregularly shaped, their dimensions can range from tens of
nanometers to tens of microns, depending on the properties desired. Pigments
may function as inert fillers or provide most of the ink receptivity and absorption
capacity in the void space between particles.
There are two primary types of coatings: porous and encapsulating (often
called “swellable”).
Porous coatings are mostly silica or alumina particles held together by a
polymer binder. Pore sizes range between ten and several hundreds of nanometers.
The ink quickly penetrates into the voids and is captured to produce very rapid
dry-to-touch times (typically less than 1 second). The void volume strictly limits
the ink capacity: exceeding the ink capacity will cause liquid ink to flood the
Sizing is added during paper manufacture to increase strength, water resistance, abrasion, opacity,
smoothness, and optical brightness. Sizing agents include natural resins, alum, starch, waxes, and
other materials.
surface. Porous coatings can deliver very high and uniform gloss, high scratch
resistance, and excellent waterfastness. Because dyes are not protected as in
encapsulating coatings, oxygen and other radicals can diffuse into the pores
causing dyes to decompose and fade. Fade resistance is one of the major challenges in the development of porous coatings.
Encapsulating coatings use water-soluble polymers such as polyvinyl alcohol,
polyvinyl pyrollidone, or gelatin. They offer complementary attributes to porous
coatings: high ink capacity and good-to-excellent fade resistance but longer dry
times, marginal waterfastness, and less scratch resistance. The extremely small
size of dye molecules enables them to penetrate the partially solvated polymer
coating. Most of the ink vehicle eventually evaporates, leaving the dyes absorbed
into the polymer, which acts as an oxygen and airborne pollutant barrier. Encapsulating coatings cannot effectively encapsulate pigment-based colorants because
of their larger size, and the pigments remain on the (smooth) surface forming a
high-viscosity film that produces variations in gloss and often has poor smearfastness and abrasion resistance.
Polymers have a great influence on the performance and function of ink
receptive coatings. For example, water-soluble or swellable polymers can be used
to absorb the ink vehicle, rigid polymers can add mechanical strength, filmforming polymers can provide surface gloss, cross-linking polymers can impart
durability, charged polymers can act as mordants,* and hydrophobic polymers
can reduce surface tack.
Performance additives are used in small quantities (usually less than 5% by
weight) to control the properties of coating materials during manufacture and
adjust the performance of the finished sheet. There are numerous reasons for
using additives, including defoaming, viscosity modification, leveling, surface
energy modification, and control of curl.
The backside of imaging media has its own functional requirements. In the
input tray, the frictional characteristics of the backside of a sheet being picked
help separate it from the imaging side of the sheet below. Backside layers provide
curl control so that the paper lies flat under a range of temperatures and humidities.
In the output tray, the backside coating prevents printed sheets from sticking
together and blocking. These characteristics are often obtained using a stacking
layer formed by a water-resistant polymer with a surface textured by small plastic
An off-white color and markings (e.g., manufacturer’s logo, machine-readable
codes, arrows, etc.) on the backside help the user load photo papers with proper
side up for printing. Recently, codes have been introduced that can be read by
the printer to give information about paper size, coating type, and whether the
paper is correctly loaded. This improves overall ease of use by ensuring that the
print mode is automatically matched to the paper and the right side is up for
One function of a mordant is to bind the colorant (dye) to the ink receptive layer. Inks can also
contain mordants.
Color Desktop Printer Technology Plain Papers
About 98% of all inkjet printing is done on plain paper. This is often a multipurpose paper designed for copiers, laser printers, and inkjets. The surface of plain
paper is usually smooth but porous. This property ensures that a multipurpose
paper has good toner transfer characteristics for electrophotographic printers and
copiers. Dyes can penetrate deeply into plain papers reducing color saturation
and strikethrough. Because pigment particles are too large to penetrate into the
paper, they stay on the surface to give the dark blacks important for printing text.
Pigments typically produce duller colors than dyes on plain papers.
Plain papers are composed of fibers from hardwoods, softwoods, and recycled
materials. Some have outer layers of high-quality fibers designed for surface and
imaging properties. These surround a thick core of inexpensive, often recycled
materials, providing bulk and strength. Sources of fiber are highly variable with
season, location, and manufacturer, so additives and sizing are added to achieve
consistent levels of performance for a particular brand. However, the worldwide
variation in the properties of plain papers offers challenges to inkjets for the
reproduction of color and sharp, black text.
Most plain papers are engineered to meet a user’s primary printing needs at
the lowest possible cost, about $0.01 per A/A4 sheet. These needs include good
color gamut, a white background, minimal feathering, minimal cockle, reliable
paper pick and minimal jamming, acceptable dry time, and good smearfastness. Coated Inkjet Papers
Coated inkjet papers provide higher image quality while retaining some of the
desirable tactile and other physical characteristics of plain paper. Papers may be
coated on one or both sides. An ink-receptive coating tightly controls dot spread
and dot shape and offers a bright, white background for contrast and color balance.
Some coatings use fluorescent dyes to make the paper appear whiter, but these
papers can have a bluish cast viewed alongside plain or bond papers in a document. To obtain high reliability in paper pick and feed, additives are used to
control the surface friction characteristics. Photo Papers
Each major inkjet printer manufacturer offers a line of photo papers to ensure
the highest image quality. These use porous or encapsulating coatings over a
bright-white photobase or white polymer film. Photo papers provide the greatest
control over dot spread and ink penetration.
Glossy photo papers offer higher color saturation than matte papers. Glossy
surfaces provide specular reflection: the light absorbed by a printed dot is reflected
back at (nearly) its angle of incidence. Matte surfaces reflect light received over
a wider solid angle, including room light, and colors may appear duller and less
saturated because light absorbed by a dot is mixed with diffusely reflected light
before reaching the eye.
143 Overhead Transparency Films
Transparency materials are usually coated on a single side with a swellable inkreceptive coating. To produce a reliable paper pick and ensure the ink-receptive
coating is properly oriented for printing, the leading edge of the coated side of
OHTs often has an adhesively backed, removable paper strip. Because the polymers used in OHT ink receptive coating have a low melting point, OHTs designed
only for inkjet can melt and severely damage fuser rollers in electrophotographic
(i.e., laser) printers. In recent years, the widespread adoption of PC projectors
has dramatically reduced the demand for OHT films. Specialty Media
Inkjet printer manufacturers and a large number of third-party suppliers offer
specialty media for business and creative applications. These include business
cards, brochure paper, textured papers, heavyweight papers for cover stock, banners, greeting cards, CD labels, and iron-on transfers.
Inkjets support a wide range of print media, and the imaging properties of different
media vary dramatically in terms of dot gain, feathering, color-to-color bleed,
dry time, and ink capacity. In addition, the user may at times require different
levels of speed, quality, and cost per page. Plain paper, economy, and speed are
important attributes for documents printed for quick review and markup. Photos
and documents for external distribution usually require the highest levels of print
quality and materials. To meet these needs, inkjet printers offer flexible solutions
called print modes.
Inkjet printers almost never scan the printheads across the page and print a
dot at every required location in one pass. Instead, they can print in single,
multiple, unidirectional, and bi-directional passes over the paper depositing ink
drops in patterns and combinations of bewildering complexity that are, fortunately, controlled by printer firmware from simple choices made by the user.
Selection of the print mode and paper type is made in the print driver, which
pops up whenever the user prints a page or document. Figure 4.11 shows an
example of quality choices among draft, normal, and best print modes and media
choices among a variety of paper types.
Draft mode uses the least amount of ink and is the fastest and most economical
print mode. It is usually reserved for plain paper and locked out when the user
selects expensive media, such as photo papers. In draft mode, ink use is reduced
by not printing every dot. This is called “dot depletion,” and it helps the sheet
to dry quickly and prevents dots from merging when printing in a single pass.
Solid colors will appear de-saturated because the paper background is not completely covered with dots. To obtain the highest possible print speed, the print
carriage scans only once over the print zone and the paper is advanced the full
Color Desktop Printer Technology
FIGURE 4.11 Magenta density ramp using light and dark inks.
swath height of the printheads. The black printhead often has a wider swath (and
more nozzles) than the color printheads on low-cost printers because of the need
for fast text printing. In this case, black and color scans may be made separately.
One of the common issues in draft mode is that nozzles that are clogged or
produce misdirected drops will leave a white (or uncolored) band that the eye is
very adept at seeing.
In some printers, draft mode is printed bi-directionally producing color shifts
between bands that depend on the direction of carriage travel. When multiple
color dots overprint, the hue depends on drop order. For example, a different blue
is obtained by printing cyan first and then magenta compared to magenta followed
by cyan. This effect is called bi-di hue shift. Printing bi-di can also cause vertical
lines to be misaligned between swaths when drop placement characteristics
depend on the direction of printhead travel and cannot be eliminated by compensating the timing of drop ejection. Some printers can print and scan a calibration
page with test patterns to align colors, minimize bi-di misalignment (usually only
for normal and best modes), and minimize paper advance banding.
Normal mode provides a good balance between speed and quality for most
applications. Here, solid colors are printed with full coverage and text-enhancing
processes, such as black fortification, are used. Partial swath advance of the paper
allows multiple scans of the printheads over the same area so that good nozzles
substitute for unfit nozzles to print all pixel rows. Small errors in drop trajectories,
nozzle-to-nozzle drop volume variations, and paper advance are effectively hidden
by multipass printing in normal mode.
Normal mode is almost always printed uni-directionally to eliminate bi-di
hue shift and drop placement errors. However, some high-speed workgroup inkjet
printers offer fast normal print modes with correction algorithms that minimize
bi-di color shifts.
Best mode is the slowest and highest quality mode because more passes are
used over a given area to average out nozzle-to-nozzle variations, and ink has the
longest time to be absorbed by the print medium, minimizing color-to-color bleed.
In best mode, every neighbor to any given dot will be printed by a different nozzle
on a different pass. Best mode is typically the default for printing photos on photo
media, but recent advances in media and printhead design allow some printers
to produce high-quality 4” × 6” photos in about 15 seconds in a fast draft mode
that are nearly indistinguishable from normal and best modes.
The number of ink droplets printed into each pixel depends on the quality
mode and the ink capacity of the print medium. Proper color rendering depends
on how the print medium’s ink capacity and dot formation processes affect tone
reproduction characteristics. This is why selecting the correct medium is important to meeting print quality expectations. For example, selecting best mode and
glossy photo paper but loading plain paper in the input tray will put too much
ink on the paper, giving a muddy, dark image with poor color fidelity.
Selecting the proper print mode and print medium has been a source of user
confusion with desktop inkjet color printers. In recognition of meeting a user
need for reliable imaging, some manufacturers have added an automatic media
type sensor to the printer. This sensor matches the print mode to the type of
medium sensed in the print zone. These sensors use light emitting diodes (LED)
and a photodetector to measure the difference between specular and diffuse
reflection from the surface of the print medium, and they can distinguish between
plain paper, coated paper, photo paper, and overhead transparency film. Printed,
machine-readable marks on the backside of photo papers can further refine the
selection to optimize photo print modes to different photo coatings.
Desktop color inkjet printers place individual colored dots a few tens of microns
in diameter on addressable grids as fine as 5760 points per inch. A four-color
CMYK* printer uses cyan, magenta, yellow, and black inks chosen to deliver a
large color gamut with highly saturated colors, dense shadows in images, and
high-contrast black text. The yellow ink adds little optical density and is used to
This is the common designation for a cyan (C), magenta (M), yellow (Y), and black (K) ink system.
Color Desktop Printer Technology
control the hue of the printed pixel. Patterns of isolated cyan, magenta, and black
dots in image highlights and midtones can have high contrast against the bright
white background of inkjet photo papers. This produces a variation in lightness
perceived as an undesirable, random texture called grain. Minimizing the visible
grain in an image is a key objective for achieving high image quality.
Although individual printed dots span a very small visual angle at normal
viewing distances, the contrast sensitivity of the human visual system allows
lightness variations to be perceived.4 However, the eye is less sensitive to color
variations under these conditions. The first implication of all this is that very tiny
dots of full-density cyan, magenta, and black inks are required before they can
produce images free of noticeable grain. This requires small drop volumes, usually
about 2 pl or less. The second implication is that combinations of cyan, magenta,
yellow, and black dots can produce the perception of millions of distinct colors,
and this principle forms the basis for all color halftoning algorithms.
A binary pixel is one in which a single drop of ink fills the pixel area. In a
CMYK system, binary pixels offer only eight addressable colors: white (background), the four primaries (cyan, magenta, yellow, and black), and the three
secondaries (red, green, and blue). Even at 600 pixels per inch, the granularity
of binary pixels printed with CMYK inks is too high for true photographic quality
at the normal viewing distance of a handheld print. Some large-format CMYK
inkjets use a 600 dpi binary print system. Unlike typical photographs, these large
(poster-size) prints are viewed from a distance of several meters where the visual
angle subtended by the pixels is equivalent to more than 2400 dpi in a handheld
Reductions in granularity have historically been achieved with smaller dots.
However, the trend in Figure 4.2 shows diminishing reductions in dot size with
drop volume below about 4 pl. Also, 2 pl may be close to a practical manufacturing
limit for drop-on-demand drop generators because features with very small physical dimensions are required. Small features will be more susceptible to clogging
and sensitive to manufacturing tolerances. Most manufacturers are now producing
the smallest drops (2–5 pl) that can be reliably ejected for their dark inks.
The most effective and practical means to reduce granularity uses a halftone
pixel. In a halftone pixel, the printer has control of the dot size, number and
placement of dots within a single pixel, dot optical density, or all three. The
method used by most desktop color printers is a combination of small drop
volumes and “light” primaries, typically light cyan (c) and light magenta (m).
Such a six-color system is called a CcMmYK printer, and processes using more
than four primary colors are known generically as high-fidelity color systems.
This section will show how such color systems can reduce image granularity and
improve the printer’s color gamut.
Recently, desktop color printers have introduced systems with one or more
neutral gray inks. This offers several benefits: a reduction in image grain,
improved dark colors and shadow detail, better and more stable neutral tones,
and reduced metamerism. When composite grays are printed from combinations
of cyan, magenta, and yellow inks, any variation in drop volume from one of the
primaries can cause the gray to shift toward red, green, or blue. Printing neutral
tones directly with gray inks avoids this problem.
Color constancy effects are particularly noticeable in gray tones and in black
and white prints. Composite grays usually exhibit some degree of metamerism:
a gray viewed under one illumination (for example, tungsten light) can appear
greenish or reddish under daylight or florescent light. Gray inks with uniform
spectral reflection can significantly reduce metamerism.
Multi-ink strategies can be used to increase gamut size by adding primaries
around the gamut equator: highly saturated oranges, greens, or blues can significantly increase gamut volume. Use of additional colors has also been considered
to improve the spectral match for reproduction of fine art prints with less metamerism, and can bring more Pantone® colors in-gamut for graphic design applications.
In CMYK printing, it is well known that many colors can be produced from
different ink combinations, especially different amounts of black relative to cyan,
magenta, and yellow. In mathematical terms, this is an under-constrained situation: tristimulus human color perception is well modeled with three dimensions
while four inks are often available to reproduce color. There are a wide variety
of undercolor removal (UCR) and gray component replacement (GCR) strategies
to determine how much black ink to use in the color separation. This produces
a more neutral, higher-density black and reduces the amount of ink used compared
to composite (CMY) black.5 Factors to be considered include image grain, ink
usage, ink capacity of the print medium, dry time, and color constancy.
Adding colors, such as light primaries and grays to a CMY system permits
printing more colors per pixel. With CcMmY inks and three gray inks (light,
medium, and dark), the number of theoretical dot combinations per pixel gives
over 72 million addressable colors.* While all these colors may not be practically
distinguishable, and printer color tables will use a subset of all the possible
combinations, this is continuous-tone printing for all practical purposes.
In subtractive color printing, the base color is the white paper and care is
taken in photographic media to provide a bright, neutral background. In the light
tones of image highlights, light cyan and light magenta inks are printed with
yellow on this white background.** There is low contrast dot-to-dot and for the
dots against the background. With increasing print density, the light inks are
printed, with area coverage increasing up to 100%. Because this produces only
a midtone color, higher density is achieved by printing full-density (i.e., dark)
cyan, magenta, and black inks on a background of light inks. This serves to reduce
This is HP’s claim for its PhotoREt Pro color layering technology, where up to 29 drops of ink can
be printed in a single 300 dpi pixel.
** Because the full-density yellow ink has very low contrast on the white background, there is little
justification from enhanced image quality for adding a light yellow printer. The additional nozzles
and drive electronics, which add to printer cost, are better distributed among the CcMmYK primaries
to obtain higher color throughput.
Color Desktop Printer Technology
the contrast between adjacent dots for lower perceived image grain. This process
is shown schematically in Figure 4.12 for a magenta density ramp. Similar ramps
are used for light and dark cyan and combinations of gray inks.
In this figure, an output level of magenta is produced by combination of light
and dark components. For low magenta densities, only the light ink is used. At
the midtones, dark magenta is introduced on the light magenta background. At
high magenta densities, mostly dark magenta is printed and the amount of light
magenta is reduced to avoid exceeding the ink capacity of the print medium. As
is typical of all color separation schemes, the quantity of each ink is varied
continuously to produce a smooth density ramp.
LIGHT INKS Light Ink Strategies: Halftoning
The digital halftoning process plays an important role in reducing image grain
in CcMmYK printing because it accounts for dot-to-dot contrast information.
The image processing pipeline for digital color printers, including those that use
light inks, is shown schematically in Figure 4.13.
The continuous tone image data from the source image in its color space
(e.g., RGB, sRGB, and CIELAB) is first transformed into printer CMY densities,
considering the spectral characteristics of its inks. This is called the device (color)
space. At this point, input colors that are outside the printer’s color gamut are
brought into the gamut using various strategies. Next, the printer CMY densities
are transformed into ink amounts, taking into account drop volumes, dot size on
a specific print medium, the type of print medium and its ink capacity, and the
number and spectral absorption characteristics of the inks. Additional factors may
be considered; for example, dye-based black ink can be mixed with dye-based
colored inks, but pigment black inks are not generally mixed with dyes in images
on photo media. Given the amount of data to be processed in real-time while
printing, these color transformations are usually done in a lookup table (LUT) in
the printer’s firmware.
Finally, halftone processing determines the number and location in each pixel
for dots of each ink color. The halftoning process typically takes CMYK densities
for each pixel and computes the number and location of individual ink dots. At the
level of individual pixels, halftoning may introduce some error between the desired
and printed color, particularly when only three or four inks are available. Some
algorithms, such as error diffusion, operate by distributing this error to neighboring
pixels and the spatial response function of the eye produces the perception of a
uniform, desired value. Error diffusion becomes less important when more addressable colors are available (e.g., with six- and eight-ink systems) because the color
errors are much smaller.
A color plane is an array of density values for a single color at each pixel in the
image. In a CMYK printer, there will be three color planes: C, M, and Y, because
black pixels can be represented by C + M + Y. The halftoning process converts the
Part 1
Part 2
Part 3
FIGURE 4.12 An inkjet printer configuration menu.
Color Desktop Printer Technology
FIGURE 4.13 (See color insert following page 176.) Image-processing pipeline.
three color planes into four dot masks, which are instructions to the printheads to
print or not to print a dot at each possible location on the image. There will be six
dot masks in a CcMmYK printer. Halftoning can operate on each color plane
independently or all color planes together. If the colors are halftoned independently
of one another, there is no control over how dots of one color will spatially interact
with dots of another color; they may partially or completely overlap, producing
unintended results such as dark dots. If the color planes are halftoned together, then
the dot patterns for each color can be interdependent and control can be exercised
over dot overlap and adjacency to produce optimal results.
There is a significant advantage to exercising interdependent color dot placement.
In many cases, image grain can be reduced by preventing dot overlap until midtones
or darker colors are printed. In the case of independent halftoning, light cyan and
light magenta dots may overlap, producing a bluish dot that is darker than either the
light magenta or the light cyan individually. In the case of a 20% fill factor for both
inks, c and m will overlap at about 0.2 × 0.2 = 0.04 or 4% of the print locations.
Yellow and the light inks interact similarly. Statistically, all three inks will overlap
at some points, producing a gray print spot that is darker than any of the three primary
inks used. The overlapping dots will result in more contrast between adjacent print
locations (bluish, greenish, reddish, or gray against the white background) than any
single dot alone. The resulting print appears grainier than when the dots can be
distributed to minimize overlapping (i.e., interdependent halftoning). Of course, once
the required density from one or more colors exceeds 100%, some locations must
be overprinted. However, even for darker colors, some reduction in grain can be
realized by preventing yellow and dark cyan and dark magenta from printing in the
same location (making the blackest possible dot) until necessary. Light Ink Strategies: Light Ink Color
An obvious method for producing the light inks is to simply dilute the colorant
used in the dark inks with more (clear) ink vehicle. This produces a less chromatic
ink of the same hue as the dark ink, and light inks are sometimes made this way.
An advantage of dilution is that the light and dark inks will have similar, but
scaled, reflectance spectra. This means that they will change color similarly under
different light sources; they exhibit similar color constancy. Thus, a range of gray
values from white to black (i.e., a neutral ramp) produced with light inks for
highlights, a mixture of light and dark inks for the midtones, and dark inks for
shadows will shift in the same way as the image is viewed under different
illumination sources. An additional benefit of hue matching between light and
dark primaries is hue constancy in the midtones as dark inks replace light inks.
For example, with similar color constancy between light and dark primaries,
if the hue shift observed between two different illuminants is toward red, then
all tones from light to dark along the neutral ramp will red shift. If the light inks
and dark inks are formulated without similar color constancy, then a density ramp
appearing neutral when viewed under one illuminant may shift to green for some
gray values and to red for others when viewed under another illuminant. This
wandering neutral poses significant image quality problems and can cause a color
printer to be essentially useless for printing high-quality black and white photos.
A potential problem in making light primaries by dilution is unacceptable
fade resistance for the light inks. Thus, manufacturers may use more fade-resistant
colorants for the light inks with an effort to ensure that the light and dark inks
exhibit similar hues viewed under the most important illumination sources.
The ink capacity of the print medium must also be considered when designing
the light inks. If the optical density is too low, excessive amounts of the light ink
may be required to achieve midtone densities, and this could exceed the ink
capacity of the print medium, especially for porous coatings.
Printing with more than four color primaries is common in commercial offset
lithography, digital presses based on liquid toner electrophotography, and large
format inkjet printers. Additional inks permit direct printing of spot, logo, and
Pantone® colors used for corporate and brand identities. They also increase the
gamut size, improve color constancy, reduce ink usage, and create smooth, grainfree gradations for certain critical colors such as skin tones. As of this writing,
desktop color printers offer up to nine inks adding to CcMmY various combinations of gray, blue, red, and green inks.
Achieving a larger printed color gamut is the usual reason for adding primary
colors. Because of the large installed base of six-color lithographic presses, much of
the effort in this area has been on CMYK systems with two additional primary inks. Dark Ink Strategies: Selecting Additional Primaries
Consider now how green and blue inks can expand the gamut of a printer that
complies with the SWOP* color standard. Being able to print the full SWOP
gamut (as well as other standards for commercial color printing used in Europe
and Japan) is important for accurate prepress proofing on inkjet printers. This
For example, Standard Web Offset Press specification: ISO 2846-1, Graphic Technology - Color
and Transparency of Ink Sets for Four-Color Printing - Part 1: Sheet-fed and Heat-set Web Offset
Lithographic Printing (1997).
Color Desktop Printer Technology
Dark Magenta
Output Magenta
& Light Magenta
Input Magenta Value
FIGURE 4.14 Reflectance of SWOP primaries.
example is only for tutorial purposes and is not provided as an ideal implementation. It has been highly simplified to reduce the complexity of the analysis while
preserving an indication of the potential benefits.
Figure 4.14 shows measured reflectance curves for a SWOP press. Individual
reflectance curves are presented for the substrate (i.e., white paper) and the cyan,
magenta and yellow colorants.
Figure 4.15 shows the measured reflectance of two secondary colors, green,
and blue, produced by overprinting the primaries. The figure also presents scaled
curves produced by hypothetical green (G) and blue (B) primaries for use with
the substrate and CMY colorants of Figure 4.14. The curves for G and B are not
based on any specific colorants; they have been chosen to yield higher chroma
at roughly the same hue angles of actual SWOP green and blue secondaries. This
also reduces the amount of cyan and yellow inks to print green, cyan, and magenta
inks to print blue.
Figure 4.16 presents gamut equators in CIELAB color space, where the xaxis is a* (red-green) and the y-axis is b* (yellow-blue). The gamut of the CMY
inks of Figure 4.14 is shown as a solid line. The gamut of the CMYGB system
Typically One LUT
Convert to
Printer CMY,
including Gamut
Color Separation
to Ink Amounts
Contone Ink
FIGURE 4.15 Reflectance of overprinted secondaries and G, B inks.
FIGURE 4.16 (See color insert following page 176.) Gamut of CMY and CMYGB
is shown by the dotted line. Here, G and B values from Figure 4.16 have simply
been substituted for the composite green and blue, although Viggiano6 describes
how secondaries might be predicted from a given set of primaries.
The original CMYK printer had a gamut volume of roughly 415,000 in
CIELAB color space. Adding the green printer (G) increases the CMYG gamut
volume to 498,000; adding the blue printer increases the CMYB gamut volume
to 518,000. The CMYGB printer delivers a gamut volume of 601,000.
The hypothetical CMYKGB printer produces a gamut volume about 40%
larger than obtainable with CMYK primaries. Because the blue and green colorants are hypothetical, this gamut expansion is not necessarily typical. Inkjet
printers using cyan, magenta, and yellow inks with higher chroma compared to
SWOP would not realize such large increases in gamut volume by the addition
of blue or green printers. In this case, additional primaries such as violet or orange
are often used to achieve gamut expansion. Dark Ink Strategies: Image Processing
The image processing used with more than four full-density primaries is similar
to the flowchart shown in Figure 4.13. In this case, the color separation process
between printer CMY conversion and halftoning involves the custom colorants
instead of light cyan and magenta. There are many to perform this separation,
but one example can be demonstrated that is similar to the gray component
replacement process.
For a six-color printer with additional green and blue primaries, a hue
sequence determines the nearest color neighbors. For example, printing cyan and
yellow primaries produces green, and printing cyan and magenta produces blue.
The hue of the secondary color (i.e., green and blue) will not necessarily be a
simple linear mixture of the two, but it generally falls somewhere between the
hues of the two primaries.
Assuming a gray component replacement (GCR) algorithm has been applied
to the color data to control the black printer, the values cyan′, magenta′, and
yellow′ in the following example are the remaining colors required to produce
pixel chroma. For 8-bit color planes, these values are typically expressed as
integers ranging between 0 and 255.
Color Desktop Printer Technology
A supplementary color replacement (SCR) algorithm can be applied to the
color data after GCR. For green, a simple SCR process is:
green = min (cyan′, yellow′)
cyan′′ = cyan′ – green
yellow′′ = yellow′ – green
where cyan′′ and yellow′′ are the new printer cyan and yellow values for the
printer given any value of green produced by Equation 4.1.
In the same manner, a simple blue SCR is:
blue = min (cyan′, magenta′)
cyan′′ = cyan′ – blue
magenta′′ = magenta′ – blue
where cyan′′ and magenta′′ are the cyan and magenta values for the printer given
any value of blue from Equation 4.4. It is assumed, of course, that blue and green
do not overprint in any pixel.
Note that Equation 4.1 through Equation 4.6 are highly simplified and could
add granularity where the minimum of cyan, magenta, and yellow is not zero.
This recalls the earlier discussion of dot visibility as a function of color and
density. However, these equations provide an example demonstrating how deviceto-device conversions might be performed during the color separation process
with four full-density color primaries. Additional and more complex separation
schemes may be found in the literature.7
Like the printed word centuries ago, the production of color images and documents was once the exclusive domain of a small number of highly trained and
skilled people working within narrow professional communities. The past 2
decades of extraordinary developments in desktop color inkjet technology and
inkjet printer features have enabled a level of personal expression through word
and image scarcely imaginable by either Gutenberg or Lord Rayleigh.
1. T. Kitahara, Ink-jet head with multi-layer piezoelectric actuator, Proc. IS&T 11th
Int’l Congress on Advances in Non-Impact Printing Technologies, IS&T, Springfield, VA, 1997, pp. 346–349.
2. M. Usui, Development of the New MACH (MACH with MLChips), Proc. IS&T
12th Int’l Congress on Advances in Non-Impact Printing Technologies, IS&T,
Springfield, VA, 1998, pp. 50–53.
3. H. P. Le, Progress and trends in ink-jet printing technology, Journal of Imaging
Science and Technology, 42(1), pp. 49–62 (1998), and on the IS&T website:
4. Lau and Arce, Modern Digital Halftoning, p. 42 (2001).
5. M. Southworth, Color Separation on the Desktop, Graphic Arts Publishing, Livonia, NY (1993).
6. Viggiano, Colorant Selection for Six-Color Lithographic Print, 6th CIC (1998).
7. E. Stollnitz, Reproducing Color Images Using Custom Inks, Ph.D. Thesis, 1998.
Laser Printer
Fumio Nakaya and Yasuji Fukase
History ....................................................................................................158
Marking Technology................................................................................159
5.2.1 Electrophotography......................................................................159
5.2.2 Marking Process ..........................................................................160
5.2.3 Technology Elements ..................................................................162 Photoreceptor ................................................................162 Charging ........................................................................164 Laser ROS Exposure.....................................................164 Development and Developer .......................................165 Transfer .........................................................................169 Fusing ............................................................................171 Cleaning ........................................................................172 Process Control .............................................................173
5.2.4 Image Quality ..............................................................................173 Color Fidelity/Color Stability .......................................174 Color Uniformity...........................................................175 Tone ...............................................................................175 Reproduction of Fonts ..................................................177 Defects...........................................................................178 Graininess......................................................................178 Banding, Colors to Colors Miss-Registration ..............178 Color Gamut..................................................................179
5.3 Toner ........................................................................................................180
5.3.1 Types of Developer (Toner and Carrier).....................................180
5.3.2 Ingredients of Toner and Carrier.................................................181 Toner..............................................................................182 Carrier............................................................................185
5.3.3 Toner Manufacturing Process......................................................187
5.4 Media and Consumables .........................................................................189
References ........................................................................................................192
Color Desktop Printer Technology
Electronic image information is converted into the hard print that human eyes
can see; the equipment using an electrophotographic method with a laser beam
as a light source is called a laser printer. An electrostatic latent image is formed
on a photoreceptor by sequential laser spots that are modulated based on the
image signal with a rotated polygon mirror, and a visualized image is formed by
the electrophotographic method.
Electrophotography started from the copying machine application; it has
evolved for years and now can handle digital signal information by using a laser
beam source. The laser beam printer with a He–Ne laser was put into practical
use for the first time as a high-speed output terminal of a computer at IBM
Corporation. Now it serves an important role as an output device for workstations
and personal computers. Furthermore, color copying machines and printers are
advancing by using the electrophotograph method.
As another effect of being able to deal with a digital image signal, electrophotography has affected progress of this colorization technology. The raw characteristic of electrophotography is very high contrast reproduction, and smooth
gradation is difficult to reproduce. The digital image signal is compensated on
image-processing technology, and reproduction characteristic of an electronic
photograph has become possible with the linear input/output characteristic of
image density.
Moreover, rapid development of electronic information machines and equipment has a great influence in all areas, including intellectual activity and production activity, and recombination of industrial structure is progressing steadily.
The conventional division-of-work organization in printing from individual work
to the printing industry is about to collapse.
Enhancement of network infrastructures and improvement in performance
has developed a new print system called Over the Internet Printing or Distribute
and Print. Under this new infrastructure, it has researched what is the most
balanced (quality/cost) total system of printing through printing production work
to delivery to the customer.
Laser printers have extended the range of print speed from 1 ppm (page per
minute) for personal usage to 180 ppm for on-demand printing. Moreover, a color
laser printer also has the capability of producing several ppm to over 110 ppm
and has a key position in the print market.
The improvement in image quality by digital image data processing is remarkable, not only with electrophotography but also with various marking technologies, such as inkjet, dye sublimation, silver halide, and conventional lithographic
printing. The laser printer with the electrophotographic system is in a middle
position for quality of image and productivity in these hard copy printing systems.
In the future, the laser printer will possibly extend the application range to both
ends, but it also runs the risk of shrinking the range by the progress of other
marking technologies.
Laser Printer
The printer market overall had about 90 million unit shipments annually in
fiscal year 2001, and inkjet systems, which show remarkable growth in home use
applications, constituted 70% of this market; whereas electrophotographic printers constituted only about 12%. However, from a revenue perspective, inkjet
systems constituted about $10 billion and electrophotographic laser printers,
including monochrome and color printers, constituted $13 billion. Moreover,
consumable sales, such as ink for inkjet printers or toner for laser printers,
matched hardware sales, with inkjet cartridges at $8.5 billion and electrophotographic cartridges at $10 billion in 1998.1
Regarding color printer products, most inkjet printers are color printers.
However, shipment of electrophotographic color printers is still small, with
166,000 shipments in 2001 in the Japanese market. (This is equivalent to 14%
of the number of monochrome page printers shipped in Japan.)
1996 print volume data show 210 billion prints in Japan and 600 billion prints
in United States. Print volume is increasing rapidly vs. copy volume. Now the
number of prints is close to or exceeds the number of copies, and the trend is
expected to continue.
The Carlson process, developed in 1938, is one of the most common electrophotographic methods. The basic principle of the Carlson process is charging photoconductive material uniformly, exposing an optical image to form an electrostatic
latent image, developing a latent image with particles that have an electrostatic
attraction and visualizing the image, transferring the particles onto media, and
fusing the particles permanently onto the media.
A detailed explanation of the electrophotographic process is available in
technical books.2,3 The electrophotographic process, as shown in Figure 5.1,
iterates the seven steps of charging, exposing, developing, transferring, fusing,
discharging, and cleaning consecutively to get multiple sheets of images. The
electrophotographic process forms a latent image each time, so it is also called
no-plate printing.
For color print, the above electrophotographic process is iterated four times
for cyan, magenta, yellow, and black images, and then the colors are overlapped.
Color electrophotographic system architectures are categorized into three types:
(1) overprint on photoreceptors, (2) overprint on intermediate media, and (3)
overprint on paper4 (see Figure 5.2).
Other categorizations can be done to count the number of photoreceptors
in a system. A single photoreceptor type with multiple development housing is
called a multi-path type, and multiple photoreceptors, each with one development housing, is called a tandem type. Figure 5.3 shows a typical tandem
Color Desktop Printer Technology
5. Fusing
2. Exposure
6. Discharging
7. Cleaning
1. Charging
3. Developing
4. Transfer
FIGURE 5.1 Electrophotography process.
Table 5.1 shows a comparison of various types of electrophotographic system architectures. Tandem type is theoretically four times faster than multi-path
type; however, mechanical architecture is more complicated and a sophisticated
color-registration mechanism is required. The tandem type system was first
introduced in the high-end and high-price market. Another feature of the tandem
type system architecture is to handle a wider range of media due to its simple
media path, and it can be dominant with an evolution of a low-cost colorregistration mechanism.
A schematic diagram of a multi-path intermediate belt transfer color laser printer
is shown in Figure 5.4.5 Yellow, magenta, cyan, and black image signals are
converted from device-independent signals to device-dependent signals, in terms
of color rendering, Modulation Transfer Function (MTF), Tone Reproduction
Curve (TRC), halftone screening, etc., and sent to a laser Raster Output Scanner
(ROS) driver. The laser ROS exposes the photoreceptor to discharge the photoreceptor electrostatic charge and form a latent image of the original. The development device, in this case a rotary development housing, develops the latent
image on the photoreceptor with a toner, by electrostatic force. Each color is
transferred to the intermediate transfer belt, color by color, and four toner layers
are transferred to media. Finally, the four-color image is melted onto the media
by a fusing device and fixed at room temperature. The rest of the toner on the
photoreceptor is removed with an elastomer blade to be ready for the next cycle.
In the case of black, the printer development device is stationary and the developed toner is directly transferred from the photoreceptor to the media.
A schematic diagram of a tandem intermediate belt transfer color laser printer
is shown in Figure 5.5.6
Transfer Corotron
FIGURE 5.2 Multi-path color laser printer.
(a) Overprint on Photoreceptor
(b) Overprint on Intermediate Medium
2nd Transfer Corotron
Intermediate 1st
Transfer Transfer
Medium Corotron
Paper Holding
(c) Overprint on Paper
Laser Printer
Color Desktop Printer Technology
Charge Corotron
Laser Scanning Optical System
Transfer Corotron
Paper Holding & Transport Belt
FIGURE 5.3 Tandem color laser printer (overprint on paper type).
Comparison of Various Types of Color Electrophotography Architecture
System Architecture
Overprint on
Overprint on intermediate
Overprint on paper
Few composition element parts of a mechanism.
Few color-to-color registration errors.
Latent image is influenced by the last development
Possible to handle wide range of media.
Image noise is accumulated through twice-transfer
process (toner disturbance at transfer area).
Stable latent and developed image
The kind of applicable media is restricted, for it needs
to wrap onto transfer drum.
Same productivity between mono color and color
Factors of color registration errors increase.
Increase composition element parts and unit cost.
5.2.3 TECHNOLOGY ELEMENTS Photoreceptor
An organic photo conductor (OPC) is commonly used as a photoreceptor. The
OPC has a layered structure consisting of an aluminum base plate, an electrical
charge blocking layer (BL), an electrical charge generation layer (CGL)m and
an electrical charge transport layer (CTL). Key factors in the OPC are
Interference of incident light
Electrical charge uniformity along its surface
Electrical charge cyclic stability
Photoreceptor drum eccentricity
Laser Printer
Development Unit
Laser ROS
Transfer Unit
Paper Stack
FIGURE 5.4 (See color insert following page 176.) A schematic diagram of multi-path
intermediate belt transfer color laser printer.
Transfer Belt
1st Transfer Roller
Developing Unit
Laser ROS
Paper Stack
FIGURE 5.5 (See color insert following page 176.) A schematic diagram of tandem
intermediate belt transfer color laser printer.
Color Desktop Printer Technology
The photoreceptor drum eccentricity, electrical charge uniformity, and cyclic
stability normally should be less than 50 microns and 20V. Charging
A charging device puts an electrical charge uniformly onto the photoreceptor.
The charging sign depends on the photoreceptor and toner material. Various types
of charging devices are available, such as wire type, pin type, and roller type.
The wire type is called a corotron. It has a fine wire (diameter: 50–60 micron
meters) and a nearby grounded shield. When a high voltage (5–10 kV) is applied
between the wire and the shield, the air near the shield will become ionized. The
pin and roller types are almost the same as wire except for their shapes and
distances. Key factors in the charge device are geometrically skewed against the
photoreceptor and dirt management to prevent its contamination. Laser ROS Exposure
The laser ROS exposes the photoreceptor to discharge the photoreceptor electrostatic charge and form a latent image of the original. Figure 5.6 shows an example
of laser ROS optics. Quantized image signals are modulated for tone reproduction
and sent to the laser driver circuit; then a laser beam is emitted, with the beam
hitting a polygon mirror, reflecting from it, focusing on a photoreceptor surface
as a spot, and scanning the photoreceptor surface diagonal to the process direction.
There are two types of writing methods. One is to write white, which exposes
the background area of the original (similar to negative in photography), and
another is to write black, which exposes the image area of the original (similar
to positive in photography). The latter case is more popular. Key factors in laser
ROS optics are to have suitable laser power with photoreceptor photosensitivity,
laser spot size, and beam positioning along the process direction. Only a few
microns of beam disposition may cause banding defects such as a dark and light
stripe along the perpendicular to the process direction. Also, laser ROS jitter and
wobble may cause graininess, so those performances should be better for color
printers than for black and white.
A laser diode emission wavelength of about 760 nm–830 nm, which is similar
to application for a compact disc reader, is normally used.
A smaller laser ROS beam spot size gives higher image quality. The laser
ROS beam spot size is proportional to the laser diode emission wavelength.7
Therefore, 670 nm (GaAIP; used in DVD) or 400 nm (GaN) laser diodes are
candidates for the next laser ROS.
There are two methods modulating laser ROS intensity. One is emission
intensity modulation, and the other is emission pulse width modulation. In
either case, it is important to keep the linearity between the image signal level
and the laser beam intensity. Basically, some compensating circuit for temperature dependency of the laser light source is needed. Additionally, it must treat
carefully any response delay in laser pulse rise time. Figure 5.7 shows an
Laser Printer
FΘ lens
Polygon Mirror
Collimating Lens
Polygon Motor
Diode Laser
FIGURE 5.6 Laser ROS optics.
example of nonlinearity in pulse width modulation.8 It needs some artifice to
compensate for it. For example, in Figure 5.7, halftone image density becomes
stable when active pixels are set side by side, compared with dispersed layout
of the active pixel case in the ordered dither screen method. This filling order
helps to stabilize electrostatic contrast in electrophotography.
For faster laser printers, various multiple laser beam approaches have been
commercialized or are under development as shown in Figure 5.8.9 All
approaches have a multiple laser beam along the process direction to gain higher
writing speed. A two-dimensional laser array has been proposed as a static
scanning device to eliminate mechanical scanning devices such as polygon
mirror systems. Development and Developer
The development device develops a latent image on the photoreceptor with
powder ink, called toner, by electrostatic force. The toner is carried to the photoreceptor by a carrier, which is oppositely charged during the mixing process in
the development housing. The toner and carrier mixed together is called the
Color Desktop Printer Technology
Light Energy (mJ/cm2) × 102
Xp: Pixel Size
x: Laser spot scan length
a: 2-pixel width
b: 1-pixel width
c: 2/3 pixel width
d: 1/3 pixel width
FIGURE 5.7 Nonlinearity of light energy in pulse width modulation.
developer. This process is done four times before proceeding to the next process
or one color at a time.
The primary subjects for designing the developing unit for color printers are
the layout of four or more color development devices along the photoreceptor,
and the method of supplying a significantly higher amount of toner into the
development device.
Commercialized development device layouts, for example, include rotary
(CLC-1, CLC-500, A-color), horizontal (CLC-200), and vertical (CF-70). All of
these layouts move each color development device to switch colors during the
marking process. Thus, special care is needed to minimize mechanical vibration
caused by this movement to prevent the banding defect.
The colored original image area is typically five times greater than black and
white originals, and toner consumption is higher and faster. Thus, better toner
and carrier mixing characteristics, called admix characteristics, are required.
Laser Printer
Laser Diode
Galvanometer Mirror
Half Mirror
Argon Grating
Position Sensor
(a) Beam synthetized by optics
Multi Channel
(b) Diffraction grating method
(c) Beam emission diode array method
Active Layer
FIGURE 5.8 Multiple laser beam approaches.
The most significant aspect of achieving higher image quality in electrophotography using powder toner is to make the toner particle size as small as possible.
For color toner, good pigment dispersion in resin provides a more desirable
Development methods are generally classified into two types. One is powder
toner development, as shown in Table 5.2, and the other is liquid toner development, such as pigment-dispersed iso-paraffin. The typical two-component magnetic brush development method and magnetic single-component development
method are shown in Figure 5.9.10
In the two-component magnetic brush development method, the carrier has
three functions: put an adequate electric charge quantity to the toner by friction
charging, carry the toner to the development zone, and remove any unwanted
toner from the photoreceptor. However, due to life issues (the deterioration of
triboelectric charging performance with toner contamination on carrier surface)
and complicated mechanisms, magnetic single-component development is dominant, especially in black and white printers. For color, due to magnetized toner
opaqueness, magnetic single-component development is not suitable and twocomponent magnetic brush development method is used. The non-magnetic
toner development method has good color reproduction characteristics and is
Color Desktop Printer Technology
Electrophotographic Development Methods
Colored pigment main
material: thermoplastic
Glass bead, steel
ball, etc.
Two component
magnetic brush
Colored pigment main
material: thermoplastic
particle: steel
powder, ferrite,
single component
Colored pigment main
material: thermoplastic
No need
Magnetic single
Colored pigment
thermoplastic resin and
magnetic particle
No need
Developer (mixture of toner
and carrier) is sprinkled
over electrostatic latent
image using gravity, and
only toner is made to stick
to photoreceptor.
Developer is adhered to a
cylinder surface with
magnetic force. Developer
is made into the shape of a
brush. Electrostatic latent
image is rubbed by the
brush, and toner is
transferred to a latent image
on a photoreceptor.
Toner is contacted with the
cylinder sleeve (conductive
rubber or resin tube),
conveyed by image force,
and transferred to latent
image on photoreceptor.
Toner is held on the cylinder
surface with the magnet
inside it, and transferred to
latent image on
Development Sleeve
Toner Dispenser
N Two Component
Development Sleeve
Magnetic Toner
Paddle Developer Mixing Roller
(a) Two component magnetic brush development
FIGURE 5.9 Typical development method.
Toner Feeder
Trimmer Blade
(b) Magnetic single component development
Laser Printer
smaller and simpler than the two-component magnetic brush development
method. However, it has several issues such as the stability of frictional electrification, the uniformity of the thin toner layer, and the maintenance of the
thin gap (within about 200 microns) between the photoreceptor and the development sleeve. Transfer
Developed toner on a photoreceptor is transferred to either the final media to
form a hard copy of the original image or to an intermediate transfer material
once and then transferred to the final media. Generally, in this process, an electrostatic force is applied. Corona transfer and roller transfer are popularly used. For
color print, this process is done either once (four colors at once) or four times
(one color at a time).
Difficulties in this process are to achieve a good color-registration level and
to achieve a good transfer ratio through the four-toner layer. The requirement of
the color-registration error level depends on the original and the type of halftone
screen used. Normally about 100–150 microns is good enough. The stress case
is a uniform gray with the same halftone screen angle. For example, a 200-lineper-inch screen case, 64 microns out of the color-registration error causes severe
color shift. It can be resolved by adopting a different halftone screen angle with
a limited choice. Transfer non-uniformity may cause bad graininess and an
unsaturated color image. Geometric settings of the photoreceptor and transfer
device and differentiation of electrostatic force applied to each color are effective
to improve it. Hollow character is one of the defects in roller transfer. Toner
image is compressed by mechanical pressure between a photoreceptor and a
transfer roll in the transfer nip area. After the nip, a part of the toner is taken off
by the cohesive force between toner particles and adhesive force with the photoreceptor surface (Figure 5.10 and Figure 5.11).10
Hollow Defects
Line Image
FIGURE 5.10 Hollow character.
100 µm
Color Desktop Printer Technology
Photoreceptor Surface
22 mm
Intermediate Belt Surface
189 mm
(a) Pressure 0.0 kPa
(b) Pressure 0.7 kPa
(h) Transferred from (b)
(c) Pressure 22.1 kPa
(i) Transferred from (c)
(d) Pressure 31.7 kPa
(j) Transferred from (d)
(e) Pressure 371.5 kPa
(k) Transferred from (e)
(f ) Pressure 386.9 kPa
(l) Transferred from (f )
(g) Pressure 168.8 kPa
(m) Transferred from (g)
FIGURE 5.11 Adhesion phenomena between toner image and photoreceptor (physical
Effective methods to avoid this problem include driving the photoreceptor
and the transfer roll at different surface speeds and using some additive on the
toner surface to improve its fluidity. The intermediate transfer method is good at
handling various papers in color printers, because of its weight range of 60–220
g/m2 and because it reduces transfer action to the media to one time. The intermediate transfer method is becoming mainstream in color printers. The material
for the intermediate transfer drum is elastic rubber, and the material for the belt
type is a thin plastic film that is flexible enough to track photoreceptor shape.
Both types have an electric resistance range of 108 1010 Ω cm semi-conductive
to reserve appropriate transfer voltages (Figure 5.12).12
Laser Printer
1st Transfer
Cleaning Blade
Film Seal
Transfer Belt
2nd Transfer
Cleaning Blade
FIGURE 5.12 Intermediate belt transfer method.
Recent technology progress in this transfer process is remarkable for digital
color image reproduction. However, there are still many issues to resolve, for
example, its mechanical simplicity, stability, etc. Fusing
The fusing process is to melt and fix a toner layer on a medium. Pairs of roll
fusers are normally used. Rolls, such as fuser rolls and pressure rolls, contact
each other to form a nip. Toner image on a medium passes through the nip.
Thermal energy and mechanical pressure are impressed on the toner image in the
nip. After the nip it is cooled to room temperature and the toner image is fixed.
Therefore, an electronograph consists mainly of thermoplastic materials.
To achieve a maximum color gamut with a given toner set, the fused toner
surface should be as smooth as possible. A smooth and soft surface fuser roll is
suitable for this purpose.
On the other hand, this type of fuser roll fades quickly. Also, the peeling
force is higher than on a rough surface, so toner offset from the media to the
fuser roll. To achieve both longer life and a lower peeling force, more fuser oil
should be applied or more wax should be added to the toner. The former case
produces many side effects, so the latter case is better.
Color Desktop Printer Technology
Microscopic random agglomeration of adjacent toner in the fusing process
makes the graininess level worse. Agglomeration of the toner occurs if the viscosity of the toner at the fusing temperature is too low, so the adjacent toners
literally stick to each other.
The binder resin’s glass transition temperature Tg should be high enough for
storage and the melting point Tm should be as low as possible for conservation
of energy. Normally, Tg is about 50–70ºC. Polyester and polystyrene are popular
for binder resin. Viscosity and elasticity are both important for avoiding toner
offset on the fuser roll. The dynamic shear elastic modulus range should be 2 ×
106 to 5 × 105 dyn/cm2 under a shear velocity of 102s–1 corresponding to the
effective fusing time.13 Usually, toner material is designed to have two peaks of
molecular weight distribution; the lower peak is for bonding between the toner
and the medium, and the higher peak is for avoidance of hot offset of the toner
layer. Polypropylene wax is sometimes used as a release agent.
Conservation of energy is now a big challenge for fusing; 60–80% of the
total fusing energy consumption is spent just for standby. Thus, challenging
points include a lower standby temperature and a faster revival time. Double
roll-fusing and belt-fusing system geometries are shown in Figure 5.13, and a
comparison of fusing technology in terms of energy consumption is shown in
Table 5.3.14 Cleaning
After the transfer process, the cleaning device cleans the remaining toner on the
photoreceptor surface for the successive electrophotographic cycle. Blades and
brushes are popular, and a blade is usually used. The blade material is polyurethane resin, which is mechanically and chemically (ozone) strong. The cleaning
process has great effect on image quality, such as background image density and
image defects. However, the toner particle becomes smaller in size and spherical
Fuser Roll
Heat-Resisting Release Layer
(silicone rubber, fluorocarbon resin)
Fuser Roll
Stripper Finger
Core (aluminum)
Inlet Chute
Core (steel)
Elastic Layer
Pressure Roll
(silicone rubber)
(a) Double roll fusing system
FIGURE 5.13 Typical fusing systems.
P2 P1
Endless Belt
(b) Belt fusing system
Pressure Pad
Soft Pad: P1
Hard Pad: P2
Laser Printer
Comparison of Fusing Technology in Terms of Energy Consumption
Warm Up
Heat roller
Heat belt
No need
Neither good nor bad
Belt drive stability
Lowest thermal efficiency
Big and heavy device
High power consumption
Low fixing quality
Heavy device weight
High torque to drive
form to get higher print quality, and the smaller and smaller, and cleaning becomes
more and more difficult in terms of photoreceptor life and overload, resulting in
a banding defect. Process Control
The electrophotographic process needs a certain precision of process control
devices to stabilize its temperature/humidity dependency and cyclic instability.
Electrostatic charge on the photoreceptor is discharged according to its resistance,
and the resistance is a function of temperature. Powder toner holds an electrical
charge and is discharged due to the surrounding humidity. So, temperature and
humidity variation strongly affects the electrophotographic process. Control targets include laser ROS exposure intensity, electrostatic voltage on the photoreceptor, developed/transfer toner amount, toner concentration in development
housing, and fused image density. The required precision depends on the application. One stressful application, for example, is graphic arts, because it requires
higher color stability and faster speed, which consume toner rapidly. Figure 5.14
shows an example of a process control system.15
Optical density data of developed toner on the photoreceptor and a pixel count
of original images feed back to the toner dispense motor to add an appropriate amount
of toner supplied into the development housing. The optical density information is
used to control applied voltages of electrostatic charges or to vary exposure intensity
of laser ROS.
Due to the increase in electronic originals, the relative amounts of printer output
of all hard copies have been increasing. Since images created by PCs are noiseless,
defects or non-uniformity standards required for printers are higher than those
for copiers.
Color Desktop Printer Technology
Computing Unit
High Voltage
Power Supply
Laser Scanner
Toner Dispense
ADC Sensor
Computing Unit
Video Signal
Pixel Counter
FIGURE 5.14 An example of process control systems.
With technological progress of high-definition displays and ecological trends,
temporary information is sometimes processed only on displays, with no hard
copies being made. For these reasons, higher image qualities are demanded for
printers. Needless to say, preventive measures for negotiable securities, counterfeits, and copyright infringement will become vital.
To combine various imaging devices for creation and transmission of highquality images, a method for reducing machine-to-machine color difference, dayto-day color instability, and also image deterioration at the time of compression/expansion will be required. Also, the color reproduction range needs to be
expanded, and image/color management with enhanced accuracy is required.
Color management includes gamut mapping, the color appearance model, measures on flare influence to display, and profile connection space (due to the
differences in paper, UCR method, and colorant). Color Fidelity/Color Stability
For color management in the office, it is desirable to achieve color fidelity
regardless of which device to use (without designating the specific device).
Required color fidelity levels vary according to image types. Examples show that
the color tolerance range of skin color for natural images is CIELAB ∆E ≤ 2 for
the green–red direction and CIELAB ∆E ≤ 3 for the blue–yellow direction.16
Because characters have been output mostly in black and white, density stability
in a high-density area has been critical. On the other hand, because pictures and
diagrams, rather than characters, are output in color, it is critical to ensure density
Laser Printer
Comparison of Color Stability in Various Marking
Technologies (Day-to-Day Color Stability: CIELAB ∆E)
Values are measured on mid-level density gray path.
* CIELAB ∆E of color variation caused by donor roll replacement is
about 4–5.
** Value when ink tank and marking head are not replaced.
stability ranging from low-density areas to high-density areas. Table 5.4 shows
the comparison of color stability with various marking technologies.17
Table 5.4 indicates that currently, electrophotography needs measures to reduce
day-to-day color instability. Table 5.5 shows examples of variation factors and
amounts of each subsystem at a low-density area in electrophotography.18
The illuminated light intensity variation of the image input terminal (IIT) and
the screen generator (SG) variation of the image output terminal (IOT) have a
high contribution ratio to density variation at low-density areas. Table 5.6 shows
a comparison of analog black-and-white, digital black-and-white, and digital color
Due to digitization, the number of variable elements increases. For example,
ROS and SG are the new variables. For color output, three signals (red, green,
blue (RGB)) are necessary and three to four colors are required for developer
units. For color, because the variation tolerance range will be twice as narrow,
the difficulty of achieving the image quality target is estimated to be more difficult. Color Uniformity
In electrophotography, color uniformity as well as color stability needs further
improvements. As shown in Table 5.5, non-uniformity of illuminated light intensity, drum to wire space (DWS), electronic potential of photoreceptor, drum to
roll space (DRS), and mass on sleeve (MOS) (amount of developer on developer
roll) has a high contribution ratio to density non-uniformity for one copy. Measures to prevent cost increases while increasing accuracy need to be taken. Tone
To reproduce high image quality, tones should be reproduced from 10% area
coverage in the highlighted area, and the maximum density should be ≥ 1.8.
Because defects such as rosette and moiré occur when reducing or enlarging
Color Desktop Printer Technology
Variation Factors and Amounts of Each Subsytem at Low-Density
Area (10% Area Coverage)
IOT (charge)
IOT (develop)
Density Nonuniformity
Density Variation
Illum. light intensity
White reference plate
Light intensity
Background potential
Dark potential
Toner density
Transfer efficiency
± 0.013
± 0.010
± 0.005
± 0.008
± 0.006
± 0.001
± 0.005
± 0.005
± 0.002
± 0.024
≤ 0.016
± 0.021
± 0.020
± 0.001
± 0.008
± 0.006
± 0.002
± 0.001
± 0.055
≤ 0.028
IOT (transfer)
System total
Target level
1 SG = Halftone screen generator
2 DWS = Drum to wire space
3 GDS = Grid to drum space
4 DRS = Drum to roll space
5 MOS = Mass on sleeve
Comparison of Analog Black-and-White, Digital Black-and-White,
and Digital Color Copiers
optical system
optical system
Number of variable
Required accuracy
Digital Color
Illumination optical
RGB separation
DEVE 3–4
Approximately twice
Approximately six
Laser Printer
images, it will be critical to increase the number of lines. Inkjet manages to
improve both tone and resolution using various measures.
In current electrophotography, the typical halftone screen frequency is 200
lines/inch and the typical depth for each halftone screen cell is 8 bits. It should
be noted that if one wishes to achieve higher resolution image reproduction, it
requires not only adoption of a higher halftone screen frequency, but also improvement in microscopic image reproduction accuracy. Otherwise, it creates more
microscopic image noise and a higher gammer tone curve. Reproduction of Fonts
In lithographic printing, different fonts are used for different purposes. 1800-dotsper-inch (dpi) lithographic printers for the office, with 14 types of Japanese fonts
and 64 types of European fonts, are already available. The number of users who
select various types of fonts has gradually been increasing. Some use unique
fonts to express originality. However, it is challenging to reproduce, in one page,
high image quality text/image mixed originals in various fonts sent from various
recording media or digital cameras. It is also important to offer products of this
capability in prices available for general offices. Thus, more improvements are
vital for image formatting and speed-up of image processing. Marking technology
varies from font to font. In digital lithographic printing, marking technology
ranges from 1200 dpi to 4000 dpi (though they may vary according to text size).
This is considerably higher than current non-impact printers. Table 5.7 shows
difference of aspect ratios among various fonts.
This example shows that in order to reproduce the aspect ratio accurately, it
is necessary to adjust the line width by approximately 22 µm for 8-point characters. Font form and density reproduction must be identical when printers of
different manufacturers or types are used to print electronic originals. Thus,
technology to adjust difference in output among printers has come to be the
critical issue.
Difference of Various Fonts’ Aspect
Aspect Difference of
8-Point Character
Color Desktop Printer Technology Defects
In network systems, it is also necessary to consider elements other than marking
technology. Image deterioration (a defect which does not exist in the original
image) caused by image compression or expansion (e.g., J-PEG) becomes a
critical issue when retrieving data stored in memory and then copying (e.g.,
electronic pre-collate copy (EPC)) or printing via network. Mosquito noise, such
as random small mosquito shape artifacts, is one example of this type of image
deterioration. Graininess
Graininess characteristics vary for each marking technology. Lowest graininess
can be achieved with the sublimation thermal transfer method. Because it is easy
for inkjet to maintain the same size of ink particle, inkjet has an advantage
compared with other marking technologies. With thermal transfer printing (TTP),
blotch occurs in the secondary color caused by the boundary of primary and
secondary colors. Because graininess is high for two-color and three-color, it is
vital to make the ink donor film thinner. In electrophotography, graininess is
improved with smaller toner particles.19–21 Figure 5.15 shows an example.22
Figure 5.15 indicates that low graininess can be obtained if the toner particle
diameter is smaller than 2 µm, regardless of ROS beam diameter or number of
lines. In order to obtain an ideal particle property when the toner particle diameter
is larger than 2 µm, it is desirable to keep the ROS beam diameter small and
minimize the number of lines within the range so image structure cannot be
recognized as a visual characteristic. Banding, Colors to Colors Miss-Registration
Stripes of density non-uniformity that occur in the uniform shading area in tables
or graphs or white gaps between text and shading background caused by the
colors to colors miss-registration (Figure 5.16) greatly influence the quality of
business documents.
The striped patterns called banding are normally observed at the 0.2–2
cycle/mm spatial frequency where the human visual system has high sensitivity.
Figure 5.17 shows that banding is visually recognized most easily at about 1
cycle/mm spatial frequency.23 A report shows banding is also dependent on the
number of lines. With 200 lines 8 bits, 0.5% peak-to-peak frequency is necessary
for a spatial frequency of 0.38 cycle/mm.24 In the printing industry, 25–30 µm25
colors to colors misregistration is necessary in a high-quality image.
With inkjet, not only colors to colors miss-registration but also stitching noise,
which occurs in the process direction with the interval of the ink head, have been
the problems to be solved. However, by changing the marking order, they have
improved dramatically.
In digital electrophotographic technology, the main factors that cause banding
or colors to colors misregistration are positional inaccuracy when writing in light
Laser Printer
400 lpi
Visual Noise
200 lpi
200 lpi × 22 µm
200 lpi × 56 µm
400 lpi × 22 µm
400 lpi × 56 µm
Toner Particle Size, µm
FIGURE 5.15 Effect of toner particle size, screen frequency, laser ROS beam diameter
in graininess.
signals to photoreceptors by ROS and zigzag motions or velocity variation of
photoreceptors/transfer belt.
Factors for fast scan direction positional inaccuracy when writing in light
signals by laser ROS are considered to be rotational inconsistency of polygonal
mirror or jitter.26 The photoreceptor velocity varies according to the photoreceptor’s gear pitch in the mechanical driving system, the number of teeth, or the
source resultant pulse number for motor excitation. How to prevent transition of
unique vibrations for the driving transfer system configuration factor and how to
reduce vibration transfer sensitivity were presented in a previous study.27 Color Gamut
Figure 5.18 shows the color gamut of color copiers that are already commercialized. To incorporate a large color gamut of red and blue areas, it is effective to
make the bevel of magenta colorant on the long wavelength sharp and the reflectance maximum on the short wavelength.28 Such measures are now being discussed (Figure 5.19).
Color Desktop Printer Technology
Shaded Background
Shaded Background
Black Text
Model: Blotch around text area (cross section)
Model: Blotch on border area (plain view)
FIGURE 5.16 Examples of defects in print.
5.3.1 TYPES
Developer for laser printers is categorized in Figure 5.20. Powder and liquid are
the primary classifications. Powder developer can be categorized into two types,
single-component and two-component, by the difference in the triboelectric
charging method.
In single-component developers, the toner obtains its triboelectric charge by
interacting with the development sleeve surface and the trimming blade (see
Figure 5.9b).
In the case of magnetic single-component developer, the toner particles are
attracted onto the development sleeve by magnetic force. In the case of a nonmagnetic developer, the toner particles are transported to the development zone
by an electrostatic force.
Two-component developers consist of carrier beads and toner. These particles
are mixed together and generate a triboelectric charge. Usually, magnetic powder
is used as a carrier. The electric property of the carrier beads affects development
characteristics. Development with insulating carrier beads is good for fine-line
image reproduction, and development with conductive carrier is good for largearea solid image reproduction. The difference is because of the strength of the
Laser Printer
Photoreceptor Velocity Variation Ratio % p.p
Screen Lines/Tone Levels
= 72/32
= 98/18
= 94/32
Spatial Frequency, Cycles/mm
FIGURE 5.17 Acceptability of banding level.
effective electric field in the development area and the residual electric charge in
carrier beads after development.
Liquid developer consists of toner dispersed in insulating liquid (carrier).
Insulating carrier is organic solvent, and toner particle size is smaller, i.e., 1–2
microns, than powder developer, which gives better image quality. Thus, liquid
developer is suitable for publishing laser printers.
Characteristics of developers are shown in Table 5.8.
Developer should satisfy various functions such as storage stability, environmental protection (safely disposable and recycling use), and electrophotographic
capabilities. Environmental impact and influence become more important year
by year in toner material design.
Toner is for visualization of images, so colorant and binder resin are obligate
ingredients. Carrier is for tribo electric charging and retention, and held and feed
by magnetic force, so magnetic material and surface coating resin to control tribo
electric charging are obligate ingredients.
Figure 5.21 shows the toner surface by electron micrograph.29 A powder toner
particle is 5–12 microns in diameter, and the shape is a round or indeterminate
form, depending on its manufacturing method. Many toners consist of binder
Color Desktop Printer Technology
FIGURE 5.18 Color gamuts in color copiers.
resin, colorant, charge control agent, and release agent. In many cases, the surface
has an external additive for its fluidity and charge control.
Figure 5.22 shows a photograph of a ferrite carrier core whose surface is
coated by resin to control electric resistance.30 Toner Binder Resin
Binder resin bonds toner onto media to form visible images. This bonding and
preventing from fading are main functions of the binder resin. However, binder
resin is a fundamental material, because its characteristics affect every electrophotographic subsystem.
Resin is a polymerized monomer, and composition of monomers can control
resin characteristics. Molecular weight distribution of resin is controlled to have
an appropriate bonding and releasing characteristic.
Figure 5.23 shows a toner viscoelastic behavior conceptual diagram. Fusing
of toner image is done in the elastomeric region. If the toner temperature is lower
than the elastomeric region, the toner is not able to fix on a medium. If the
temperature is in the fluid region, it is easy to cause a hot offset problem (toner
image sticks to a fusing roll). Therefore, keeping the elastomeric region as wide
as possible is necessary to prevent unfusing and hot offset problems.
Laser Printer
Pigment Type Colorant
Dye Type Colorant
Spectral Reflectance %
More Ideal Colorant
Enlarge Red Area
Enlarge Blue Area
Wavelength, nm
FIGURE 5.19 Schematic diagram of magenta spectral reflectance curve.
Dry (powder)
Wet (liquid)
FIGURE 5.20 Classification of developer for laser printers.
Resin with multiple peaks in molecular weight distribution is a mixture of a
high-molecular-weight component and a low-molecular-weight component. The
low-molecular-weight component improves bonding characteristics, and the highmolecular-weight component prevents toner offset.
Color Desktop Printer Technology
Characteristics of Developers
Easy to control tribo electric charge
High-speed develop ability
Easy-to-use color toner
Less environment dependent
Need to control toner concentration
Complicated structure of deve. Unit
developer needs to be exchanged.
Magnetic singlecomponent
Simple structure of development unit
Free to control toner concentration
High accuracy is required
Difficult to use color toner
Difficult to control tribo electric
Environment dependent
Simple structure of development unit
free to control toner concentration
Easy-to-use color toner
Difficult to control tribo electric
Difficult to improve speed of
6.5 µm
FIGURE 5.21 Toner surface view.
Lowering fusing energy can be done by lowering molecular weight. However,
lowering molecular weight lowers the resin glass transition temperature and
causes bad storage stability. Therefore, not only lowering molecular weight but
also changing the composition of resin has been implemented.
Typical resins for the heat-conduction fusing method are styrene–acrylic
copolymer, styrene–butadiene copolymer, and polyester.
Styrene–acrylic copolymer resin is inexpensive because of its monomer composition, which has triboelectric charging control ability and ease of molecular
weight control.
Polyester resin has high mechanical strength and superb viscoelasticity characteristics, but it is more expensive than styrene–acrylic copolymer resin in general.
Laser Printer
Resin Coated
Dynamic Shear Elastic Modulus G
FIGURE 5.22 Ferrite carrier.
Glass Transition Region
Elastomeric Region
Fluid Region
Fusing Latitude
Toner Temperature
FIGURE 5.23 Toner viscoelastic behavior.
Styrene–butadiene copolymer resin is also inexpensive; however, the molten
viscosity of it is relatively low, so silicone oil and its applicator mechanism are
obligate. Other Typical Construction Materials of Toner
Toner consists of various materials besides binder resin to achieve many electrographic functions. Colorant, charge control agent, release agent, and external
additive are typical secondary toner materials. The main compositions are shown
in Table 5.9. Carrier
The basic functions of the carrier are to charge toner particles with the triboelectric
charging process and to hold toner particles on its surface until it reaches the
development zone.
The magnetic developer feeding mechanism and magnetic carrier are commonly used. The magnetic carrier is categorized into two types: magnetic material
For fluidity improvement
For cleanability
Release agent Lowering critical surface
tension of toner
Minus charging
Plus charging
control agent
Magnetite (100–500 nm) cubic,
spherical, polyhedral structure
C.I. Pigment Yellow 17
Non-benzene yellow pigments
Quinacridone carmine 6B
Magnetic particle
(magnetic toner)
Nigrosin (for black toner)
4 class ammonium salt series (color
Halogenated compound Azo series or salicylic series metallic
and metal complex
Olefin series wax
Polypropylene and polyethylene
Ester series natural
Carnuba wax
Fine inorganic particle Silica (about 7–15 nm)
Higher fatty acid
Zinc stearate
metallic salt
Nitrogen compound
Furnase black, channel black
Typical Materials
Carbon black (nonmagnetic toner)
Typical Construction Materials of Toner
Carbon black is sometimes used for
magnetic toner as an electric resistance
20–80 wt% Spherical shape looks red-tinged black
and cubic shape has a bluish tinge.
5–10 wt% To achieve a good transparency, colorant
should be dispersed evenly and finely in
toner resin. Surface treatment or
processed pigment to meet dispersion
1–5 wt%
4 class ammonium salt series charge
control agents have no color and are
Non-metallic complexes are approached
for heavy metal free movement.
0–5 wt%
Low critical surface tension, melting at
low temperature sharply and high
hardness under room temperature.
0–5 wt%
Other function of external additive is to
improve tribo electric charging
characteristics of toner.
5–10 wt%
Color Desktop Printer Technology
Laser Printer
coated with resin and magnetic material dispersed in resin. Magnetic materials
for the carrier are iron, magnetite, copper–zinc ferrite, manganese ferrite, and
light metal ferrite.
Resin coating provides an electrostatic property and electric resistance adjustment. It prevents toner contamination onto the carrier surface as well. Resin
coating materials include styrene–acrylic copolymer, fluorine series such as
vinylidene fluoride, polyethylene, and silicon resin.
The magnetic material dispersed resin type carrier is manufactured like magnetic toner; the resin and magnetic material melt together, knead well, cool down,
and then smash up. The smaller particle size and lighter weight are selling points
for this type of carrier.
The schematic flow of a conventional pulverizing process is shown in Figure
5.24.31 First, the toner components of resin, colorant, and other materials are
mixed in a dry process (pre-mixture process). Subsequently, the resin is softened
by applying heat and kneading it with strong shearing stress, and the resin bulk
or pellet is obtained that has a uniform distribution of the component, such as
colorant (knead process). Then it is pulverized to several micron meter-sized
particles (pulverization process). The air jet pulverizing method uses high-speed
airflow, and the mechanical pulverizing method uses shear stress. Finally, toner
is obtained through a classification process that takes out the desired particle
diameter from the obtained pulverized material (classification process). The rough
Charge Control Agent
Surface Treatment
FIGURE 5.24 Process flow of pulverized toner.
powder and fine powder that are generated in this process are collected and reused
Color Desktop Printer Technology
as raw material. After that, some external additives are mixed and put into a
Polymerization is a method of manufacturing, including the stage of synthesizing resin that is different from the conventional pulverizing process. This
process can be roughly classified into suspension polymerization and emulsion
aggregation. In the suspension polymerization process, the contents of the toner
component such as colorant are dispersed in a monomer. The droplet of a predetermined size as a toner is made from the monomer composite that is suspension
dispersed in water using a dispersion stabilization agent.
Subsequently, the coloring particle that dispersed the content of the toner
composition ingredient is formed by suspension polymerization. After removing
the dispersion stabilization agent, filtration, washing, drying, and mixing with
external additives, toner for the laser printer is obtained.
The schematic flow of this manufacturing process is shown in Figure 5.25.32
Because a droplet is formed in water by the suspension polymerization method,
the droplet (toner) becomes a spherical shape.
Considering the principle of electrophotography, the spherical shape toner
has an ideal performance. However, it has problems such as poor cleaning of the
remaining toner because of the larger adhesion force to a photoreceptor.
The emulsion aggregation manufacture process is also put into practical use
as a method to form a semi-spherical shape toner by the polymerization method.
Figure 5.26 shows the schematic flow of toner preparation.33
Emulsion aggregation toner may be prepared by using the following steps.
First, there is pigment/wax dispersion preparation (dispersed in de-ionized water),
and emulsion polymerization (obtained latex is about 200 nm in size). Latex size
can be controlled by surfactant concentration, homogenization (the mixture is
homogenized with a high shear mixer; average size is about 2.5 microns after
Charge Control Agent
Dispersion Stabilization Agent
Filtration, Washing
FIGURE 5.25 Process flow of suspension polymerization toner.
Laser Printer
Latex Addition
FIGURE 5.26 Process flow of emulsion aggregation toner.
homogenization), aggregation (the homogenized mixture is heated with continuous mixing in the reactor and with temperature ramping), and coalescence (the
pH of the slurry is controlled to freeze the aggregated size and then the mixture
is ramped over the glass transition temperature and held for several hours with
mixing). After cool down, the filtered coalesced particle is washed with de-ionized
water and dried. Narrow size distribution can be obtained by optimization of the
latex size and Zeta potential, controlled in the emulsion polymerization step
(Figure 5.27).34
High color image quality may be achieved by fine image reproduction and
wide color gamut based on small particle size, narrow distribution, and the
optimization of fusing properties with low-molecular-weight polyethylene wax
Although it is not a polymerization method, another toner-manufacturing
process, without passing through pulverization process, has also been proposed.
Resin, colorants, and other materials are dissolved and dispersed in solvent that
is insoluble in water. The solvent is dispersed in water to form oil drops of toner
size. Removing a solvent forms a toner particle. In this case, the resin can use
various kinds (i.e., polyester), because this process can use materials that polymerized beforehand. Figure 5.28 shows three kinds of toner surfaces manufactured by various processes
Media, toners/developers (see Section 5.3), and photoreceptors (see Section are the major consumables for color laser printers. Laser color printers
should be able to handle any kind of medium. However, there are several aspects
that are suitable to color laser printers. These include electrostatic stability with
temperature and humidity, appropriate elasticity for feeding, and resistance to
Color Desktop Printer Technology
Aggregation Toner
without Classification
Volume Rate %
Pulverizing Processed
Toner with Classification
Particle Size (diameter) Micron Meter
FIGURE 5.27 Particle size distribution of emulsion aggregation toner and conventional
FIGURE 5.28 Comparison of toner surface.
curl with toner and media thermal contraction after fusing. Unique aspects for
transparency media are false contour and transparency projection quality. If the
transparency media surface is not soft enough to penetrate the toner image, the
edge of the toner image will turn out to be a bump, reflecting incident light and
becoming a dark band looking like a false contour. Transparency media projection
images sometimes look dull, dark, and less saturated. This occurs when either
the toner is not well melted in the fusing process or the transparency medium
itself has low transparency.
Laser Printer
Grade of Graininess
Bekk Surface Smoothness (sec.)
FIGURE 5.29 The relationship between paper surface smoothness and halftone graininess.
An image quality relating aspect is paper surface roughness, as shown in
Figure 5.29, where smoother paper gives less graininess, which is an important
attribute for color.35 However, smoother paper has high material density that is
too soft for feeding and too thin (with the same paper weight) to obtain enough
opacity for showthrough. Thus, for colored paper, heavier weight with more filler,
such as potassium carbonate added to it, is preferred. Smoother paper is also easy
to curl. Additional tweaking such as optimizing paper fiber alignment, reduces
paper curl. Figure 5.30 shows the relationship of the paper fiber orientation ratio
Curl Radius (1/m)
Sonic Velocity Ratio, MD/CD
FIGURE 5.30 The relationship between fiber orientation ration and radius of post-fusercurl.
Color Desktop Printer Technology
(MD [machine direction]/CD [cross direction] of a paper pressing machine and
paper curl under stress conditions (high image area coverage).36
The customer replaceable unit (CRU) and service engineer replaceable unit
(ERU) are other consumables for color laser printers. CRU and ERU are cartridges
usually consisting of a toner bottle, photoreceptor, development housing, and
cleaner that give better maintainability.
1. JEITA research report, JEITA, 04-P-2 (2004).
2. Dessauer, Clark, Xerography and related processes, The Focal Press, Burlington,
MA (1968).
3. Schaffert, R.M., Electrophotography, The Focal Press, Burlington, MA (1975).
4. Fukase, Yasuji, Journal of Printing Science and Technology, 33, 2, pp. 29–35
5. Kimura, Kiyoshi et al., Color Laser Wind 3310, Fuji XEROX Technical Report
No. 12, p. 161 (1998).
6. Tamura, Kazuo et al., DocuPrint C2220/C2221, Fuji XEROX Technical Report
No. 14, pp. 104–115 (2002).
7. Uchida, Teiji, Optics for engineers, The Journal of the Institute of Telecommunications Engineers, 62, 5, pp. 538–545 (1979).
8. Kawamura, Naoto, Digital marking technology, Journal of The Society of Electrophotography of Japan, 26, 1, pp. 75–82 (1987).
9. Kataoka, Keizo, High speed and high resolution laser scanning optics using multi
beam laser, Journal of The Imaging Society of Japan, 37, 1, pp. 91–98 (1998).
10. Fukase, Yasuji et al., Digital Hardcopy Technology, Kyoritsu-Shuppan Co., Ltd.,
p. 199 (2000).
11. Nakayama, Nobuyuki et al., Numerical Simulation of Electrostatic Transfer Process Using Discrete Element Method, PPIC/Japan Hardcopy Conference ’98 Proceedings, pp. 261–264 (1998).
12. Okamoto, Yoshikazu et al., Intermediate Transfer Belt System for Color Xerography, Fuji XEROX Technical Report No. 12, pp. 22–31 (1998).
13. The Imaging Society of Japan, Basics and Application of Electrophotography,
Corona Publishing Co., Ltd., p. 197 (1988).
14. Matsumoto, Shinji, Energy Saved Fusing System, Technology Seminar of Fall
Meeting, The Imaging Society of Japan (1999).
15. Kimura, Kiyoshi et al., Color Laser Wind 3310, Fuji XEROX Technical Report
No. 12, p. 161 (1998).
16. Nakaya, Fumio, The 16th Fall Seminar of The Institute of Image Electronics
Engineers of Japan Proceedings (1992).
17. Bolte, Steve, SPIE, 1670 Color Hard Copy and Graphic Arts, pp. 3 (1992).
18. Nakaya, Fumio, The 8th Conference of the Color Technology ’91 Proceedings,
pp. 141–148 (1991).
19. Shaw, R. P. Dooley, Noise perception in electrophotography, Applied Photographic
Engineering, 5(4), pp. 190–196 (1979).
20. Yamazaki, T., et al., The Journal of The Institute of Image Electronics Engineers
of Japan, 21, 2, pp. 106 (1992).
Laser Printer
21. Chiba, T., The 4th NIP Symposium Proceedings, pp. 129 (1988).
22. Shigehiro, Kiyoshi, The Effects of Toner Particle Size and Image Structure on the
Image Quality in Electrophotography, The 9th NIP Symposium Proceedings, pp.
97 (1993).
23. Ishikawa, Hiroshi et al., Effect of Photoreceptor Velocity Variation on Halftone
Image Reproduction, The 3rd NIP Symposium Proceedings, pp. 133 (1986).
24. Hirakura, Koji, Drum Digital Color Electo Photographic System, Japan Hard Copy
’91 Proceedings, pp. 101 (1991).
25. Kume, Masatsugu, Insatsu Joho, pp. 23 (1992).
26. Sakaue, Eiichi, Brief Summary and Outlook of Laser Direct Prepress, The Journal
of The Society of Electrophotography of Japan, 33, 2, pp. 170 (1994).
27. Miwa, Tadashi et al., Konica Technical Report, 6, pp. 29 (1993).
28. Tsuda, Shinich et al., The 5th Information and Image Processing Conference of
The Japan Society of Colour Material Proceedings, pp. 26–33 (1994).
29. Ichimura, Masanori et al., The Smallest Particle Size Full Color Toner, Fuji
XEROX Technical Report No. 13, p. 170 (2000).
30. Yamazaki, Hiroshi, Basics of Developer Technology and Its Future Trend, 53rd
Technology Seminar: The Imaging Society of Japan, p. 109 (2002).
31. The Imaging Society of Japan, Basics and Application of Electrophotography,
Corona-Pub., pp. 482–485 (1988).
32. Yamazaki, Hiroshi, Basics of Developer Technology and Its Future Trend, 53rd
Technology Seminar: The Imaging Society of Japan, pp. 118–120 (2002).
33. Matsumura, Yasuo et al., Encapsulated Emulsion Aggregation Toner for High
Quality Color Printing, IS&T NIP17: International Conference on Digital Printing
Technologies, pp. 341–344 (2001).
34. Matsumura, Yasuo et al., Development of EA Toner (Emulsion Aggregation Toner)
for High Quality and Oil-Less Printing, Fuji XEROX Technical Report No. 14,
p. 98 (2002).
35. Matsuda, Tsukasa, Fuji Xerox J Paper, Fuji XEROX Technical Report No. 7, p.
56 (1992).
Dye Thermal-Transfer
Nobuhito Matsushiro
Driving Mechanism .................................................................................197
6.2.1 Line Sequential Methods.............................................................197
6.2.2 Area Sequential Method..............................................................197
6.2.3 Typical Configuration ..................................................................198
Details of Dye Sublimation Printer.........................................................199
6.3.1 Basic Structure.............................................................................199
6.3.2 Sublimation Dye and Sheets .......................................................199
Details of Wax Melt Printers...................................................................200
6.4.1 Basic Structure.............................................................................200
6.4.2 Melt Wax and Sheets...................................................................200
Thermal Head ..........................................................................................200
6.5.1 Requirements ...............................................................................200
6.5.2 Structures and Features ...............................................................200
6.5.3 Temperature Control....................................................................201
6.5.4 Concentrated Thermal-Transfer Head.........................................202
Various Improved Printer Engines ..........................................................203
6.6.1 Improvement in Processing Speed ..............................................203
6.6.2 Improvement of Durability of Thermal-Transfer Printer ...........203
Other Printers Based on Thermal Transfer .............................................203
6.7.1 Thermal Rheography ...................................................................203
6.7.2 Electrosensitive Transfer Printer .................................................204
6.7.3 Light-Sensitive Microcapsule Printer..........................................204
6.7.4 Laser Thermal-Transfer Printer...................................................204
6.7.5 Direct Thermal Printer ................................................................205
Controller Aspect.....................................................................................206
6.8.1 Resolution ....................................................................................206
6.8.2 Tone Reproduction ......................................................................206
6.8.3 Color Gamut ................................................................................207
User Aspects ............................................................................................207
6.9.1 Characteristics for Application....................................................207
Color Desktop Printer Technology
6.9.2 Product Range .............................................................................207
6.9.3 Running Cost ...............................................................................207
6.10 Stability Issues.........................................................................................208
6.10.1 Recording Density .......................................................................208
6.10.2 Thermal Issues.............................................................................208
6.10.3 Wear Issue ...................................................................................208
6.11 Conclusions..............................................................................................208
The principle of image formation by thermal-transfer printers is depicted in Figure
6.1. Figure 6.1 relates to the dye sublimation and the wax melt printer as most
representative of thermal-transfer printers. The principle of thermal-transfer printers is that using a heating element, physical or chemical reactions of solid-state
ink form images. An input voltage heats the thermal-generating elements. Onto
the recording sheet, the base material sublimes or melts from the heated area of
the ink.
In a thermal-transfer printer, an ink supply and ink recovery mechanism or
clog recovery system are not required as in inkjet printers, and these are advantages of the thermal-transfer printer. The only mechanical parts required are the
driving mechanism of ink sheets and coloring sheets, making the recording and
printing mechanisms relatively simple. The product can be compact, lightweight,
and inexpensive because the structure of the printer engine is so simple.
Thermal Head
Thermal Head
Base Film
Ink Layer (dye)
Base Film
Ink Layer (pigment)
Accepting Layer
Recording Paper
Paper Base
Platen Drum
Platen Drum
Dye Sublimation
Thermal Transfer
Wax Melt
Thermal Transfer
FIGURE 6.1 Dye thermal transfer and wax melt thermal transfer.
Dye Thermal-Transfer Printer
The serial line sequential method and the parallel line sequential method are two
driving methods (Figure 6.2). In the serial line sequential method, the recording
sheet moves forward gradually, and the thermal head shifts for each color. In the
parallel line sequential method, each color is aligned in parallel, the recording
sheet is forwarded, and the thermal head is shifted in parallel.
The processing speed by this method is high. Currently, most printers adopt this
method. As shown in Figure 6.3, transfer sheets of the same size as the recording
sheets are continuously aligned, and transferring is carried out onto the entire
area of the sheet.
Yellow (Y)
Magenta (M)
Cyan (C)
FIGURE 6.2 (a) Line sequential method (parallel sheet); (b) line sequential method
(sequential sheet).
Color Desktop Printer Technology
Recording Paper
FIGURE 6.3 Area sequential method.
A typical driving configuration of a thermal-transfer printer is depicted in Figure
6.4. The engine shown is the swing type. The recording paper returns to its initial
position when one color is completely applied to the paper. Low expense and
compactness are advantages of this type of printer. A drum-winding system and
a three-head system also exist. In the drum-winding system, registration is relatively simple and the return of the recording sheet is not required; this contributes
to a reduction in the recording time. The three-head system is suitable for highspeed recording.
Platen Roller
Ink Sheet
Thermal Head
FIGURE 6.4 Typical configuration of a thermal-transfer printer.
Dye Thermal-Transfer Printer
The dye sublimination printer is one of the most promising printers of recent
years. Sublimation is the cycle that forms the basis of dye sublimation printers.
Sublimation is a process by which solids are transformed directly between solid
and vapor without passing through a liquid phase. The outline structure of the
dye sublimation printer engine is shown in Figure 6.1. Ink sheets are media onto
which dye has been dispersed for subsequent transfer to the recording sheet.
In a sublimation dye system, the energy required for transfer is greater than that
used for a wax melt printer. The color carrier in the sublimation dyes must by
design be unstable but, at the same time, recorded materials must be stable. This
is contradictory to the nature of sublimable dyes. Accordingly, the selection of
dyes is the most important factor in determining the success or failure of ink
sheet manufacturing.
Reproduction of colors is realized by subtractive color mixture. Ink sheets
used in the system have the three subtractive primary colors of cyan, magenta,
and yellow, and often black is also used. To satisfy requirements such as light
resistance and color reproducibility, special dyes were developed from the middle
of the 1980s. At the beginning of the 1990s, new dyes with the same molecular
structure as pigments used for color photosensitive materials were developed.
Progress has been made in the area of materials development of dyes with
appropriate subliming characteristics — characteristics for dispersing by diffusion, high absorption coefficient, severe weather tolerance, and good saturation.
The requirements for the dyes applied onto the recording sheet include high
linearity with respect to the applied heat and many levels of gradation. The ink
sheets should not degrade the thermal head, the ink sheets should not adhere to
the thermal head, and no ink should remain on the thermal head. The thermalresistant layer should not affect the properties of the dye layer. In the recording
sheets, chemical compounds with high hardness are included.
It is essential that the ink sheets be made of thermal-resistant materials
because the the thermal head reaches 280–340°C momentarily. Only polyester
films are available as low-cost sheets with thermal-resistant properties and
strength. However, until recently, polyester films have had short useful lives under
these conditions. In recent years, methods have been developed to incorporate a
layer with improved longevity.
Recording sheets should not degrade the thermal head with the compounds
when they are in contact. Recording sheets for dye sublimation printers are special
media. There is no path currently foreseen to support plain paper.
Color Desktop Printer Technology
The mechanisms of two thermal-transfer printers are compared in Figure 6.1.
The major difference in terms of the thermal-transfer mechanisms of the two
systems is that, compared with the wax melt printer, the dye sublimation printer
requires higher thermal energy.
6.4.2 MELT WAX
In the wax melt printer, a wax containing pigments is used as ink. Typical wax
media such as paraffin are used. Coloring materials such as pigments are dispersed, and coated sheets of inks are produced. Pigments are used as the colorants.
Pigments whose characteristics are very close to those of printing ink can be
used. The wax functions as a carrier for the pigments during heating, melting,
and transferring by the thermal head. When cooled, the dried wax fixes the
pigments as a binder onto the recording sheets. As vehicles, paraffin wax, carnauba wax, and polyethylene wax are used.
The wax melt printer employs more stable dyes and pigments than does the
dye sublimation printer for high-quality images. Many media, including plain
paper, may be used for recording.
The thermal head must be durable throughout continuous rapid heating and
cooling cycles. For high-speed printing, the thermal head must have thermal
properties that allow rapid heating and cooling. Thus, the thermal head must have
low heat retention capacity. To reduce power consumption, the heat energy supply
must be highly efficient. For high-resolution printing, high-density resistor patterns must be realized.
There are three types of typical thermal heads: thin-film, thick-film, and semiconductor. The thin-film head is suitable for high-speed printing because of its
excellent thermal response. The structure of the head is shown in Figure 6.5(a).
However, the production of large-sized thin-film heads is difficult, and the manufacturing process is complex. This head is frequently used in both sublimation
and melt printers. The heating element is formed using screen-printing and sintering technology. The simple fabrication process is an advantage, and thus it is
suitable for mass production.
A thick-film head is most suitable for large printer engines and used for largesized sheet printers. In the case of sublimation transferring, the conduction time
of the current to the thermal head is 2–20 ms, and that in melt transferring is
Dye Thermal-Transfer Printer
Glass Layer
Lead Line
Glass Layer
Thin-Film Head
Thick-Film Head
FIGURE 6.5 Thin-film head and thick-film head.
approximately 1 ms. The difference in duration arises from the difference in
recording processes, reflected in the recording speed. Figure 6.5(b) shows the
structure of thick-film heads.
A semiconductor head has not yet been practically applied due to the disadvantage of its slow speed from poor thermal response.
Head Temperature
An important problem related to temperature control of the thermal head is heat
hysteresis (Figure 6.6). For ideal temperature control, the thermal head temperature is controlled using a reference temperature when an input voltage signal is
applied, and the thermal head temperature returns to its original value immediately
when the input signal is set to zero. Due to heat hysteresis, this kind of control
is not easily realized. However, several methods that make use of special arrangements have been developed.
FIGURE 6.6 Transition of the surface temperature of a thermal head.
Color Desktop Printer Technology
Figure 6.7 shows a concentrated thermal head in which the pattern of the heating
elements is used. The thermal head has many narrow regions, as shown in Figure
6.8. With a heating element of this type, high-temperature sections are created
in the high-resistance section. Then at areas centered on these high-temperature
sections, heat transfer starts. The transferred areas expand around these centering
points as the applied energy is increased. The above is the major feature of this
method. This method can also be used with melt ink sheets.
Recording Voltage Pole
Thermal Spot
FIGURE 6.7 Concentrated thermal-transfer head.
Thermal-Transfer Head
FIGURE 6.8 Heating elements of conventional and concentrated thermal heads.
Dye Thermal-Transfer Printer
A one-pass full-color printer engine has been developed because thermal-transfer
color printers are inferior to the electrophotographic printing system in relation
to processing speed. Figure 6.9 shows the structure of the engine. The one-pass
system has four independent recording sections: yellow (Y), magenta (M), cyan
(C), and black (K). The recording sheet is passed to the four sections in order.
The disadvantage of this system is that the sheet-forwarding section is complex
and, therefore, color deviations tend to manifest easily. The speed of recording
sheet varies due to the load and back tension during forwarding. This results in
color deviations, even if the rotation speed of the drive roller is constant. By
detecting the speed of the recording sheet directly using a detection roller and
by controlling the drive roller to maintain a constant recording sheet speed, color
deviations are prevented.
In dye thermal-transfer, durability characteristics of the images are poor due to
the inherent nature of the dyes themselves. For example, discoloration due to
light exposure is a problem. To improve this problem, the following measures
have been developed. In the recording layer, a compound is placed that reacts
with a transferred dye and improves its stability. Also in the recording layer, an
ultraviolet absorbing material or similar substance has been added. A transparent
protective layer is sometimes applied on top for added stability.
With this printer, a small hole is made at the center of the resistor in the thermal
head and melted solid ink is transferred onto the recording sheets through the
hole. This type of printer is based on the idea of inkjet recording.
FIGURE 6.9 One-pass printer engine.
Ribbon Ribbon
Color Desktop Printer Technology
Ink Layer
Electric Conduction
FIGURE 6.10 Structure of electrosensitive transfer.
Colored materials are transferred using joule heat from electric conduction
through a coated layer of high electric resistance placed under the sublimable
dye or thermal-transfer ink layer (Figure 6.10). This printer modulates the
amount of ink transferred at each pixel in accordance with the duration of
electric conduction. This printer is not popular because the production cost of
recording materials for this system is high compared with that of other printer
In place of an ink sheet, this system uses media with dispersed microcapsules.
Capsule-containing sheets are exposed to ultraviolet light in an imagewise manner, creating a latent image. The ultraviolet exposure hardens capsules. The sheets
are superimposed with recording sheets and put through a pressure roller. The
pressure roller is able to crush only the unexposed capsules. The pigments in the
crushed capsules are released and transferred onto the recording sheet.
The capsule wall is a polymer of urea and formaldehyde. Acryl monomers,
polymerization initiator, spectral sensitizers, and leuco dyes are contained in
the capsules. The recording sheet contains a developer. When released leuco
dyes come into contact with the developer in the recording sheet, colors are
When higher resolution is desired, limitations in the thermal-head fabrication
process become important because image resolution depends on the level of
integration of heating elements in the thermal head. A new engine is being pursued
to resolve this problem, i.e., a dye transfer printer using a laser as a heat source.
Dye Thermal-Transfer Printer
Laser Beam
Base Film
Laser Beam
Base Film
Dye Receiving Layer
Dye Donor Layer
+IR Absorbent
Dye Receiving Layer
Base Sheet
Base Sheet
(a) Surface-Absorbing Type
(b) Dye-Donor-Layer
Absorbing Type
Dye Donor Layer
FIGURE 6.11 (a) Surface-absorbing printing media type and (b) dye-donor-layer-absorbing printing media type.
The dye sheets consist of a base film and a dye layer, and the recording sheets
consist of a base sheet and a dye-receiving layer. This is basically the same as
that used in conventional dye transfer printers. In addition, it is necessary to
incorporate a layer that converts laser light to heat in the dye sheet. Figure 6.11
shows two types of printing media used for this printer. In contrast to the situation
in Figure 6.11(a), in the construction of Figure 6.11(b), a layer that is infraredray absorbent and selectively absorbs semiconductor laser waves is included, and
part of the laser beam is absorbed in the dye layer. By appropriately setting the
parameters related to absorption, the amount of dye transferred can be increased.
A relatively high-power laser is required because the system is a heat mode.
Previous research used a gas laser. In recent years, a high-power semiconductor
laser has been developed. The semiconductor laser came from compact optical
disc recording development.
In laser drawing, there are two types of image creation printer engines with
respect to the recording surface on a drum. One is a printer engine in which the
laser moves, and the other is a printer engine in which the laser from a fixed
source is scanned using a mirror. For thermal-transfer recording with low recording sensitivity, the former systems are popular. Resolutions of 2540 dots per inch
(dpi) or higher have been realized. As a countermeasure for the slow writing
speed of the semiconductor laser system, multihead printer engines are being
Coloring agents are applied in advance to the recording sheet. Color is formed
when heat is applied locally onto the sheet using a thermal head. This system has
advantages over others because of the possibility of high-speed printing.
Color Desktop Printer Technology
In regard to image resolution, 150 dpi (1980), 200 dpi (1987), 400 dpi (1996),
and 600 dpi (1998) have been introduced. Greater than 1200 dpi has been available, but for limited purposes, such as color proofing.
The essence of sublimation thermal-transfer is that density modulation is possible
for each dot. By an increase or decrease in joule heat, the density of each dot
can be controlled. Heat levels are controlled by the pulse width of input signals.
This control is an analog process.
The melt-type thermal-transfer printer produces solid and clear images. Thus,
the printer is used in the preparation of bar codes. To achieve gradation, the dither
method or another similar method is used, because the density characteristic of
this recording method is binary. Thus, the resolution of this method tends to be
From the direct thermal printer, excellent coloring and image quality can be
obtained, because it utilizes special sheets on which a coloring agent is applied.
As illustrated in Figure 6.12, the reproduction of color density by various
printers can be classified into three methods. In the first method (bi-level pseudo
density), the area of ink-covered dots changes; thus, when an image produced by
this method is observed from a distance, a coherent density change can be
observed. In the second method (multi-level pseudo density), an area is covered
by a fixed number of dots whose color density is constant and whose dot size
changes in relation to the density. The number of dots whose dot density and size
are constant is changed in an area; this is called the bi-level area method. In the
Pseudo Density Method (bi-level pseudo)
Pseudo Density Method (multi-level pseudo)
Density Modulation Method
FIGURE 6.12 Methods of the reproduction of color density.
Dye Thermal-Transfer Printer
third method, the density of each pixel changes. In this processing, each pixel
can be given a continuous change of color density, and near-perfect gradation
can be reproduced over the entire gradation range. In the first method, by incorporating a sharp heat-generating distribution in the narrow regions, area-based
gradation, in which the transfer recording area within one pixel of the heating
element is altered, has been realized. This combination method is called the dotsize variable multivalue area pseudo density reproduction method.
Sublimable dyes are suitable for use in full-color printer recording systems. The
three primary colors, Y (yellow), M (magenta), C (cyan), of sublimable dyes have
almost the same color gamut range as that realized by color offset printing.
The dye sublimation printer is capable of producing color pictures whose image
quality is comparable to that obtained by the photographic systems. Thus, when
printing quality close to that of photographic images is required, such as in color
proofing and production of pictorial color copies, this is suitable. The characteristics of the melt thermal-transfer printer offer sharp dot matrix images.
As for disadvantages, there are problems of poor durability, retransferring of
transferred dyes, and early degradation of sections touched by fingers. Various
methods to alleviate these problems have been developed for practical use. For
example, sometimes the entire recording sheet is covered with a protective layer
containing ultraviolet absorbents.
There is a wide range of products using thermal-transfer technologies, from lowcost printers to high-end printers, from personal use to office use, from desktop
publishing (DTP) use to color proofing use.
Small mobile printers for in-the-field use have been developed using thermal
transfer technologies.
Instant photo printing at stores and home often depends on thermal-transfer
technologies. The direct thermal printer is widely used in simple printers such as
facsimiles and personal computers. In addition, this system is used in printers
incorporated with measuring instruments, and high-end thermal-transfer printers
are also used for color proofing in printing processes.
Although the running cost is high, many features make the printers cost effective
for some applications. Because of the excellent color reproducibility, high density
Color Desktop Printer Technology
level (optical density value: 2.0), and high tone image (gradation: 64 or higher),
they have been incorporated into products requiring high image quality, such as
digital color proofing systems, video printers, DTP printers, and card-transferring
For sublimation printers, multiple transferable ink sheets have been developed
to reduce running cost. With these sheets, up to ten printing cycles with the same
sheet are possible with no loss of density. In these ink sheets, large quantities of
sublimable dyes are included in a thermoplastic resin layer.
Ambient temperature and heat accumulation in the thermal head can cause the
thermal balance to be misdistributed, resulting in errors in recording density.
These factors can cause problems of non-uniformity and problems in color reproducibility and resolution.
There are two main thermal issues related to the thermal head. The first issue
arises from the temperature increase of the thermal head. To maintain the operating temperature of the thermal head within limits that prevent destruction of
either the head or the heat-resistant layer on the ink sheet, the head radiation fin
is carefully designed. The second issue is related to the problem of constant
changes in the thermal head temperature during printing. Measures counteracting
transient temperature changes are required. These include optimum control by
temperature detection using thermistors and temperature prediction using hysteresis data from gradation printing.
The thermal head wears due to friction because the thermal head maintains contact
with the thermal-sensitive recording sheet as it moves. Head wear advances
rapidly, in particular, if the thermal-sensitive recording sheet contains chemical
compounds with high hardness. Chemical wear occurs in addition to this type of
mechanical wear. It is corroded by alkali ions and other substances contained in
the thermal sensitive recording sheets, because the surface of the thermal head
is glassy.
There is a very wide range of products using thermal-transfer technologies, from
low-cost printers to high-class printers, from personal use to office use, and from
DTP use to color proofing use. The system has now been improved in terms of
Dye Thermal-Transfer Printer
performance. The resolution of the thermal head has been improved from 600 to
1200 dpi. Further effort is being put forth toward high precision and low cost.
The problem of image storage properties, which was the crucial issue for dye
heat transfer, has been solved substantially by the stabilization of the dye through
chemical reactions. Research continues in the search for better recording materials.
As described, there will be many improvements from the user’s point of view
and there are many expectations for dye transfer printers.
T. Abe, Trends of thermography, ITE Technical Report, 11(26), 7–12, 1987.
S. Ando et al., A basic study of thermal transfer printing for improvement of print quality,
SID 85 Digest, 160–163, 1985.
C.A. Bruce and J.T. Jacobs, Laser transfer of volatile dye, J. App. Photo Eng., 3(1), 40–43,
H. Genno et al., Correction method of printed density with the sublimation dye transfer
process, SID Int. Symposium Digest, 284–287, 1990.
T. Goto, Color reproduction of video printer, ITE Technical Report, 47(10), 1397–1400,
W. Grooks et al., Ribbon thermal printing, ribbon and head requirements, IS&T, The 2nd
International Congress on Advances in Non-Impact Printing Technologies, p. 237,
K. Hanma et al., A color video printer with sublimation dye transfer method, IEEE Trans.,
CE-31, 1985.
Y. Hori et al., Development of high definition video copy equipment, IEEE Trans., CE32, 1985.
M. Irie and T. Kitamura, High-definition thermal transfer printing using laser heating,
Journal of Imaging Science and Technology, 37(3), 1993.
A. Iwamoto, Thermal printing technology, EID88–29, Japan, 1988.
T. Kanai, S. Hirahara, T. Ohno, K. Yamada, H. Nagato, and K. Higuchi, Digital–analog
halftone rendition using ink-transfer thermal printer, 6th International Display
Research Conference (Japan Display 86), 3, 1986.
S. Masuda et al., Color video printer, IEEE, Trans. on Consumer Electronics, CE-28 (3),
226–232, 1982.
N. Matsushiro, Dye transfer printing technology, Encyclopedia of Imaging Science and
Technology, vol. 1, pp. 189–197, John Wiley, New York, 2002.
M. Mizutani and S. Ito, Thermal transfer printer, Oki Technical Journal, 51(2), 1984.
S. Nakaya, K. Murasugi, M. Kazama, and Y. Sekido, New thermal ink-transfer printing,
Proc. SID, 23(1), 51–56, 1982.
I. Nose et al., A color thermal transfer printer with recoating mechanism, SID 85 Digest,
143–144, 1985.
H. Ohnishi et al., Thermal dissolution ink transfer for full-color printing, IEEE Trans. on
Electron Devices, 40(1), 69–74, 1993.
K.S. Pennington and W. Crooks, Resistive ribbon thermal transfer printing, SPSE Proc.
of 2nd Int. Congress on Advances in NIP Technology, 236, 1984.
Color Desktop Printer Technology
O. Sahni et al., Thermal characterization of resistive ribbon printing, SID 85 Digest, 152,
M. Shiraish et al., Development of and A4-size color video printer, SID 87, Digest,
424–427, 1987.
N. Taguchi, H. Matsuda, T. Kawakami, and A. Imai, Dye transfer resistive sheet printing,
SPSE Proc. of the 4th Int. Congress on Advances in NIP Technology, Thermograph
Session, 532–543.
H. Tanaka, Multi contrast steps record of the thermal transfer printing method by wax ink,
ITE Technical Report, 17(27), 19–24, 1993.
A. Tomotake et al., Structure/activity relationship of post-chelating azo dyes in thermal
dye transfer printing system, IS&T, The 14th International Congress on Advances
in Non-Impact Printing Technologies, 269–272, 1998.
K. Ttsuji, The trend of the research in thermal printing technology, EID89–38, Japan, 1989.
Film-Based Printers
Tsutomu Kimura, Atsuhiro Doi, Toshiya Kojima,
Masahiro Kubo, and Akira Igarashi
History .....................................................................................................211
Photo Printers ..........................................................................................216
7.2.1 Film Scanning Technology..........................................................216 Line CCD Image Reading Technology ........................216 Area CCD Image Reading Technology........................217
7.2.2 Marking Technology....................................................................220 Exposure Technology....................................................220
7.3 Emerging Technology..............................................................................221
7.3.1 Pictrography.................................................................................221 Principle of Recording ..................................................222 Material Composition and Image Forming Process.....222 PG Printer......................................................................224 High-Quality Imaging Technology ...............................225
7.3.2 Thermo-Autochrome Method......................................................226 Recording Principle.......................................................226 TA Paper........................................................................228 Heat-Responsive Microcapsule.....................................229 TA Printer......................................................................230 Summary .......................................................................232
References .........................................................................................................232
Traditional photography is based on silver-halide chemistry. Historically, the
ability to capture and keep photographic images became a reality in the 1830s
when William Fox Talbot discovered a technique for preventing images from
fading over short periods of time, and Louis Daguerre published his method for
making daguerreotypes. Color film for consumers became available when
Kodak’s Kodachrome entered the market in 1935. As for instant photography,
Polaroid Corporation put the first color film on the market in 1963, which was
followed by the development of other types of instant photography. Throughout
Color Desktop Printer Technology
the years, photography has continuously improved. One important innovation
was the processing method invented by 3M in 1964 that also led to the development of dry-silver color technology by 3M in 1986. Also in 1986, Fuji Photo
Film announced the development of diffusion-transfer-type color thermal development silver-halide material technology, putting it on the market the next year.
Television technology made rapid progress following the beginning of practical broadcasting in 1935 (in Germany), when the technologies for handling an
image as electronic signals and the displaying them on the cathode ray tube (CRT)
advanced. In the late 1950s, during the space race, technologies for transmitting
images, digital processing, and receiving their outputs developed. With such
technological development, naturally, there was a growing need for printing
images from electronic signals and digital data. Furthermore, when Sony introduced the Mavic digital camera in 1981, development of color printers for general
use was accelerated remarkably.
In the course of this development, silver-halide photosensitive materials were
naturally regarded as an advantageous means for making a hard copy of color
electronic signals and digital images because of their high sensitivity and high
image quality and also because of their diversified characteristics. There has been
much technological research for the development of photographic materials
designed for compatibility with a host of new recording systems. Various kinds
of technological developments have been made for getting new materials for
various recording systems matched with the exposure to light.
Exposing silver-halide photosensitive material with an image is known as
recording the image. One common recording device is the CRT because of its
relative compactness, and low cost CRTs are traditionally used to expose the
entire image at once.
As a point light-source system, the drum-scanning system (halogen lamp,
light emitting diode (LED), etc. as a light source) was developed early. In recent
years a laser-scanning system has been developed, in which a laser beam is
modulated to scan the exposure using a polygon mirror. As one-dimensional array
exposure systems, there are LED array systems in which multiple LEDs are
aligned, an electroluminescent display tube array system in which multiple electroluminescent tubes are aligned, and a light shutter array system in which PLZT
(ceramic comprising Pb, La, Zr, and Ti) controls the polarized direction of
penetrating light by applying a voltage. Another practical image recording system
uses a two-dimensional array of digital micro-mirror devices (DMD) for controlling exposure (see Figure 7.1).
Corresponding silver-halide photosensitive materials are instant films, reversal films, color papers for direct recording, and positive prints after recording on
negatives. It is possible to use non-specialized films that are used in standard
photographic cameras, although in many cases their sensitivity to light is not
appropriate for the exposure means of various systems. Technical developments
have continued, including improvements in sensitivity, gradation, color-reproducibility range, and resolution of materials. Likewise, developments in exposure
Film-Based Printers
Color Separation
FIGURE 7.1 DMD exposure system.
means have contributed to changes in the light-emission spectrum, quantity of
light, addressability, gradation control, and color-reproducibility.
In the following, we describe the exposure and printing technologies used
widely by CRT and by fiber optical tube (FOT) as its modification. CRT exposure
is a system in which a two-dimensional image displayed on the CRT screen is
projected and exposed on a photosensitive material by optical means (Figure
7.2a). FOT exposure is a system in which the display area is a one-dimensional
device of a flattened CRT and the image on the display and the photosensitive
material are simultaneously moved for exposure while forming the image on the
photosensitive material by optical means (Figure 7.2b and c).
When exposing a color image using a CRT, one usually exposes three times
by applying three color filters to a black-and-white CRT successively (Figure
7.2a). Because no mechanical scanning system is necessary, the device is
simple. Many video-printer products are on the market that are specially combined with instant photographic materials. Printers combined with color-paper
systems were put on the market in the late 1980s. The FVP600 of Fuji Photo
Film is an example (Figure 7.3).1 The maximum recording width is 102 mm,
and multiple exposures are executed using a high-brightness 7-in. CRT with
525 scanning lines, being able to produce an E-size print in an exposure time
of approximately 3 seconds. The system used is shown in Figure 7.2, and the
spectral sensitivities of the color-separation filter and color paper are shown in
Figure 7.4. In the 1990s, large-size, high-performance devices were developed,
and Agfa commercialized the AGFA DSP, which is a printer with a maximum
width of 203 mm, utilizing a 9-in. flat-type CRT with 1024 scanning lines. It
had a capacity of 1400 sheets per hour for L-size media and about 400 sheets
per hour for a maximum size of 89 mm × 127 mm.
In color exposure systems using FOT, there are systems in which an image
is created by applying three color filters to each of three sets of FOT and focusing
them on a single point of the material using a mirror-lens system (Figure 7.2b),
Color Desktop Printer Technology
B- T
Color Separation
(a) CRT Exposure System
(b) RGB Monochromatic
FOT Exposure System
R Fluorescent
G Fluorescent
B Fluorescent
ThreeColor (RGB)
(c) Three Color (R,G,B) Divided
FOT (fiber optical tube) System
FIGURE 7.2 Three-color (RGB) divided FOT system.
Video Image
CG Image Input
Color Monitor
Character Image
TV Camera
Image Input
Frame Nega/Posi
Memory Conversion
Image Input
Processing Unit
FIGURE 7.3 Diagram of an FVP600.
CRT Driver
Film-Based Printers
Spectral energy distribution of CRT
Relative Intensity
Wavelength (nm)
Spectral transmissivity of color separation filters
Wavelength (nm)
Relative Intensity
Spectral sensitivity of color paper
Wavelength (nm)
FIGURE 7.4 Spectral characteristics of FVP600 components.
or in which one FOT coated with three colors of electroluminescent materials
(Figure 7.2c) is used. Both systems need an auxiliary scanning system to transport
the material at a constant speed. The single FOT system is capable of exposing
with the light source and the photosensitive material in contact so that the device
can be downsized. The Konica VP100 was a commercialized example in which
a single FOT system records on color paper. The first copy was obtained in 7
minutes and 30 seconds, and its capacity was 60 sheets per hour for the size of
130 × 180 mm.
Color Desktop Printer Technology
Conventional mini-labs produce the image of a negative film on a silver-halide
color paper using an illumination system, an optical system that focuses the image
on color paper. After a certain exposure time, these are followed by development
processing to get a print. At digital mini-labs, existing slides and prints may be
copied. They are converted to digital data by scanning the originals using a CCD
scanner. The digital data undergoes image processing, and the image is exposed
on a silver-halide color paper, followed by development processing to get a print.
Also, mini-labs record onto print paper digital data that come directly from digital
still cameras (DSCs).
As for the CCD scanners used for reading images, there are line and area types,
either of which may be selected on the basis of the required qualities for film
scanning, mainly the number of pixels for reading and the reading speed. Line CCD Image Reading Technology
Figure 7.5 shows the structure of the Fuji Film Frontier 350 film scanner using
a line CCD. From the bottom, it consists of a light source for illumination, a film
carrier, an image-forming optical system, and a CCD sensor for reading.
For illumination, a slit illumination optical system with a high light intensity
has been developed to read a high-density negative film at a sufficient level of
signal-to-noise ratio (S/N). As the light source, a halogen lamp with high stability
was adopted, illuminating the slit illumination part in the film carrier through the
cold filter, the light source iris mechanism, the negative and positive spectrumcompensating filters, and the diffusion box. Here, for illuminating a film with the
high light intensity, a film-cooling mechanism has been adopted to prevent a rise
in film temperature, making it a highly reliable and safe system.
The film carrier utilizes a line CCD to provide a mechanism for transporting
film at a constant speed with high precision. A number of types are available for
various film sizes.
Lenses for the image-forming optical system are sufficiently compensated
for aberrations such as magnification chromatic errors, axial chromatic problems,
and image distortion, and the optical magnification is adjustable in accordance
with the necessary pixels for the output. Because of this, even for a large-size
print over 10 in. × 12 in., it can secure sufficient reading resolution so that it may
realize a sharpness equal to or better than that of analog printers and also provide
optimal reading resolution for a wide range of negative sizes. Also, it can accurately read slide films that have a differential thickness or an uneven shape due
to mount curling.
Film-Based Printers
3 Line C C D
Variable Iris
Variable Position
(magnification, auto focus)
Up–down of Lens for Change of Magnification
Film Carrier
Light Diffusion Box
Nega–Posi Balance Filter
Iris for Light
Intake Fan
Exhaust Fan
(halogen 400 W)
FIGURE 7.5 Film scanner unit of a Frontier 350 (line CCD). Area CCD Image Reading Technology
An area sensor has higher light utilization efficiency compared with a line sensor,
and the film transport mechanism becomes simpler, which greatly contributes to
miniaturization of the device. Figure 7.6 shows the structure of a film scanner
for the Frontier 330, Fuji Photo Film, using the area CCD.3
The light source for the film scanner is LED (see Figure 7.7). LED is installed
on a ceramic base with high thermal conductivity so that it has a high level of
safety, an excellent characteristic of a film scanner. LED has excellent features
such as lower power consumption than a halogen lamp, long life, its space-saving
property, and negative film cooling is not needed. In particular, power consumption is a few watts, which is less than 1/100 of a halogen lamp, and, further, it
lights up only at the time of scanning, contributing greatly to reduction in power
consumption of the device. Problems with LED as the light source of a film
scanner include securing a sufficiently high light intensity and controlling the
Color Desktop Printer Technology
ADC Board
BLK Shutter
Mirror Box
Peltier Thermostat
R,G,B,IR Array
FIGURE 7.6 Film scanner unit of a Frontier 330 (area CCD).
FIGURE 7.7 (See color insert following page 176.) LED light source for a film scanner.
variations in light intensity and wavelength. The light intensity problem could be
solved by using a honeycomb-type CCD, which will be described later. The
variations in light intensity and wavelength could be suppressed by the use of a
ceramic base and the temperature control by a Peltier element.
Figure 7.6 shows a diagram of a film scanner unit in which a Peltier element
is used to stabilize the temperature of the light source. Figure 7.8 shows the light
intensity of the light source before stabilization and after stabilization.
Film-Based Printers
− 0.2
High Temp
Low Temp
Light Intensity
High Temp
Low Temp
Light Intensity
− 0.2
Wave Length (nm)
Before Stabilization
Wave Length (nm)
After Stabilization
FIGURE 7.8 Stability characteristics of light sources.
FIGURE 7.9 Honeycomb type of CCD.
Figure 7.9 shows a CCD developed exclusively for the Frontier 330. The
effective pixel number is 3.2 million and it forms a honeycomb structure in which
the pixels are arranged in a triangular arrangement. The opening shape is almost
circular, and the area of the photoreceiver can be made large enough so that
modulation transfer function (MTF) may have high isotropy and sensitivity to allow
a wide dynamic range, an excellent characteristic for a film scanner. Also, to secure
the necessary pixels for a large-size print, the Frontier 330 has two axes of a highprecision pixel-shifting mechanism using piezo. As shown in Figure 7.10, it can
pick up an image, 1448 × 2172 pixels for each color with direct reading, 2048 ×
3072 pixels with two readings having one shift between the readings, and 2896 ×
4344 pixels for each color with four readings having three shifts between the
readings. The maximum delivers more than four pages of about 12 million pixels.
Further, infrared radiation (IR) is added to the illumination source of this device,
allowing for a special channel that detects defects (Figure 7.11).
Color Desktop Printer Technology
FIGURE 7.10 (See color insert following page 176.) Automatic scratch and dust restoration function.
Without pixel shifting
(1448 × 2172 pixels)
With pixel shifting done once,
scans twice (2048 × 3072 pixels)
With pixel shifting done three times,
scanning four times (2896 × 4344 pixel
FIGURE 7.11 (See color insert following page 176.) Pixel-shifting method.
7.2.2 MARKING TECHNOLOGY Exposure Technology
Among various exposure systems, several have been commercialized for digital
mini-labs, and laser-scanning systems have become the mainstream because of
their advantage in image quality. Those systems using a solid laser have the
excellent characteristics of high image quality and the capability of miniaturizing
the device. Figure 7.12 shows the printer structure of a Frontier 350 using a solid
laser. Silver-halide paper is set in the form of a roll (two rolls in this device) and
sheets cut to a specified length are transported upward, and after changing their
courses to the horizontal direction, they are conveyed to the laser exposure part.
The laser exposure part is transported at a high-precision speed.
The laser-scanning unit modulates each laser beam of RGB at a high speed
with light-modulation elements, and the color paper being transported at high
Film-Based Printers
FIGURE 7.12 Laser exposure unit of Frontier 350.
speed is irradiated by these three laser beams for a constant-speed scanning
exposure using a polygon mirror and fθ lens (Figure 7.13).
Photographic materials using silver-halides are superior to other color hardcopying materials in sensitivity and image quality. However, conventional silverhalide color paper has problems in the control of processing liquids used for
development or processing speed. The diffusion-transfer-type color thermal development material technology (Pictrocolor System) is known by the name of Pictrography (PG).4–5 It was introduced in 1987, and uses a small amount of water
and heat, but no processing liquid.
Color Desktop Printer Technology
Photosensitive Material
F θ Lens
Collimeter Lens
B-SHG Laser
R Semiconductor
G-SHG Laser
FIGURE 7.13 (See color insert following page 176.) Optical system for laser exposure. Principle of Recording
A pictrography system is one kind of diffusion-transfer system, characterized by
having all necessary development agents built in. Image formation in the Pictrocolor system proceeds as follows (see Figure 7.14):
1. Expose a donor film (thermal development photosensitive material)
2. Supply a small amount of water to the donor (approx. 0.7 cc/A4 or 10
3. Contact an image-receiving paper with the donor
4. Thermal development and dye transfer
5. Peel off the image-receiving paper from the donor
By exposing to light, a latent image (metallic silver nucleus comprising three
to four silver atoms) is formed in the exposed silver-halide of the donor. Water
supply, contacting, and heating generate an alkali, and development proceeds to
release the water-soluble dyes. The released dyes diffuse and transfer to the
image-receiving paper and are fixed by the mordant in the image-receiving paper.
By peeling off the image-receiving paper from the donor, the residual water
quickly evaporates due to the remaining heat, providing a print (Figure 7.15).
The image-receiving paper thus obtained after printing contains only dyes and
no unfixed silver-halide that would cause deterioration of the image quality. Material Composition and Image Forming Process
The main composition of the donor is composed of hydrophilic binders such as
gelatin, spectrally sensitized photosensitive silver-halides, dye-releasing redox
Film-Based Printers
Finished copy
Light Source
(3) Thermal
(1) Exposure
Piatro film
(2) Moistened with water
(4) Peel-off
Used film
(to be discarded)
FIGURE 7.14 (See color insert following page 176.) Conceptual diagram of pictrography.
Silver Halide
color fixative
latent image
(2) wetting/
putting two sheets together
(3) thermal development/
image transfer
(4) peel apart
dye DRR compound
(1) exposure
FIGURE 7.15 Process diagram of pictrography.
compounds (DRR compounds) known as dye material, and precursor basic metal
compounds that generate alkali. The image-receiving paper consists of hydrophilic binders such as gelatin or various polymers, cationic mordant polymers
for fixing dyes, and basic precursors that react with the basic metal compounds
in the donor to form chelate compounds, generating alkali.
In the Pictrocolor system, the alkali necessary for advancing development
is not generated until the donor and the image-receiving paper are brought into
contact so as to better preserve the materials before use. For this reason, it is
Color Desktop Printer Technology
necessary for the donor and the image-receiving paper to be stable before
contact. After contact, therefore, water-soluble agents contained in the imagereceiving paper and the sparingly soluble basic metal compound react in a small
amount of water to generate alkali in 2 to 3 seconds, even at room temperature.
The alkali help to catalyze reactions in the DDR compound that releases
diffusible dye (Figure 7.15). PG Printer
Pictrography is the digital printer of the Pictrocolor system. The first-generation
machine was developed in 1987, based on a drum exposure type with an LED
light source. LED has excellent characteristics such as size, cost, dynamic range,
light intensity, and the linearity between an electric current and a light intensity.
However, the exposure system with LED has the following problems:
1. LED has a large luminous point so that an image-forming beam may
not be narrowed down on a donor, which limits improvement in the
recording density.
2. The drum exposure system has a limit in the revolution speed of the
drum, hindering improvement in the recording speed.
Thus, it has been difficult to attain high resolution and high-speed recording.
With a change in the exposure system to LD in 1993, the donor was also
improved to get much better image quality. As compared with LED, LD is easy
for narrowing the beam diameter, being suitable for high-density recording, and
replacement of the exposure system from LED to LD has improved resolution
from 284 to 400 dpi. Also, the convergence efficiency of the collimeter lens is
so high that exposure light intensity may be enough for high-speed recording.
Furthermore, by combining LD with a polarizer consisting of a polygon and an
fθ lens, photosensitive material may be exposed in the course of transport in a
plane, which makes it possible to miniaturize the device (Figure 7.16).
On the other hand, the LD exposure system has the following defects:
1. The practical range of LD wavelength is limited.
2. The electric current and optical output characteristics of LD vary significantly with temperature change.
3. When LD light beams of three wavelengths form images with the same
lens, the wavelength dependence of image-forming characteristics such
as color aberration appears as a color shift, variations in the beam
diameter, etc.
4. When the optical parts have transmittance and reflectance distributions
in the main scanning direction, it causes shading. Particularly when
the shading characteristics are different among colors, differences
become discernible as color shading.
Film-Based Printers
laser exposure unit
used donor sheet
wetting tray
donor sheet outter
magazine Receiv.
print tray
receiving sheet
FIGURE 7.16 Internal diagram of a Pictrography 3000. High-Quality Imaging Technology
LD wavelengths used in pictrography are 680, 759, 810 nm. To prevent unintentional color mixing even in this narrow wavelength range, it is necessary to
design the donor spectral sensitivity distribution so that there would be sharp,
non-overlapping peaks in this wavelength range. On the material side, it has
become practical for the first time because of J-band sensitization technology
that regularly arranges molecules of sensitizing dyes on a surface of the silverhalide crystals for a photosensitive layer to have a peak at 750 nm. For the
photosensitive layer having a peak at 810 nm, multi-functional ultra-red dye
was introduced with a sharp absorption at 750 nm. With the introduction of
these technologies, spectral sensitivities with excellent color separation were
realized (Figure 7.17).
LD exposure is a pulse-width modulation system with a fixed light intensity.
The modulation is controlled by 12 bits having a sufficient density resolution.
For variations due to differences in LD exposure parts and sensitivity variations
of photosensitive materials, the printer is provided with a calibration system in
which the built-in densitometer measures the relation between the exposure
control signal and the print density to allocate the optimal 12-bit signal to an 8bit image signal.
Color Desktop Printer Technology
wavelength (nm)
FIGURE 7.17 Spectral sensitivity of pictrography media for LD exposure.
Heat-resistant protective layer
Yellow color forming layer
High thermosensitivity
419 nm Light sensitivity
Magenta color forming layer
Mid thermosensitivity
365 nm Light sensitivity
Cyan color forming layer
Low thermosensitivity
FIGURE 7.18 (See color insert following page 176.) Basic structure of TA paper.
7.3.2 THERMO-AUTOCHROME METHOD Recording Principle
Until the 1980s, it was believed that full-color printing by a direct thermal print
system was possible in principle but not in practice. In 1994, however, Fuji Photo
Film commercialized a direct thermal printer based on a thermo autochrome (TA)
system.6–8 Thermo autochrome is a coined term meaning a system in which all
the necessary mechanisms for color printing are incorporated into recording
paper, and by repeat heating and light exposure (automatically), one may get a
color print.
Figure 7.18 shows the cross-sectional structure of TA paper. On a substrate,
each thermo-sensitive layer that renders cyan, magenta, or yellow color is
successively coated, and a heat-resistant protective layer is provided as the top
Film-Based Printers
Developed density
Recording energy (mJ/mm2)
FIGURE 7.19 Thermal recording characteristics of TA paper. (The density is measured
with an X-rite reflective densitometer. The recording energy is calculated from the electric
energy supplied to the thermal head.)
layer. Each layer reacts with a different thermal energy to develop the color. The
uppermost yellow layer reacts with the lowest thermal energy, and the lowermost
cyan layer reacts with the highest. Figure 7.19 shows the thermo-color development characteristics. The two upper layers of magenta and yellow have thermosensitivities and photosensitivities as well. In the magenta-developing layer, the
incorporated color-developing component is decomposed by 365-nm ultraviolet
rays and loses color-developing capacity by heating. Accordingly, after forming
an image by heating, the image may be fixed by exposing the total paper to 365nm ultraviolet rays. Similarly, the yellow-color developing layer loses colordeveloping capacity by 419-nm ultraviolet rays.
By building such a mechanism into the recording paper, full-color printing
becomes possible by the following simple procedures:
1. Apply low thermal energy corresponding to yellow image information
to a TA paper to record a yellow image
2. Expose it to 419-nm ultraviolet rays to fix the yellow image
3. Apply an intermediate thermal energy corresponding to magenta image
information to record a magenta image
4. Expose it to 365-nm ultraviolet rays to fix the magenta image
5. Apply high thermal energy corresponding to cyan image information
to record the cyan image
Color Desktop Printer Technology
Thermal head
Power supply
Fixing lamp
Yellow print
TA paper
Magenta print
Cyan print
FIGURE 7.20 Basic configuration of a TA printer.
Figure 7.20 shows the basic configuration of a TA printer. The printer is
composed of a thermal head for thermal recording, one ultraviolet-ray fluorescent
lamps of 419 nm and one of 365 nm, and platen rolls for driving. The TA paper
is fed from the left side of the thermal head. In the first process, the yellow image
is simultaneously recorded and fixed. That is, the yellow image is recorded by
the thermal head and, at the same time, the image is fixed by the 419-nm
ultraviolet-ray lamp placed downstream. Next, the paper is returned to its original
position to record and fix the magenta image. Last, the cyan image is recorded.
Thus, by having the paper move back and forth two and a half times, one can
complete full-color printing.
The merit of the TA system is that TA paper is the only material necessary
for printing, and it does not require any consumables such as ink or toner at all;
accordingly, there is no waste disposal associated with printing. Further, the drive
is only for transporting the paper, so one can build a simple and highly reliable
system. TA Paper
Because a yellow image layer recorded by low thermal energy may be developed
by an intermediate energy for magenta image recording if left as it is, it is
necessary to fix the yellow image (lose its thermal sensitivity) by some means.
In TA paper, this problem has been solved by introducing a diazonium-salt
compound as a color-developing material.
The diazonium salt compound reacts with a coupler to form a dye. It is
decomposed by the light corresponding to its absorption wavelength, losing its
reactivity with the coupler. In the recording layer, a diazonium salt compound
and the coupler are dispersed. They do not interact until they are heated. When
the layer is exposed to light, the thermally recorded image is fixed. Figure 7.21
shows the absorption spectrum of the diazonium salt compounds used for the
magenta color-developing layer and the yellow color-development layer. By making them have different spectral absorption characteristics, only the yellow image
may be fixed selectively when one exposes it with 419-nm ultraviolet rays in
Film-Based Printers
380 400 420
Wave length (nm)
FIGURE 7.21 Spectroscopic photosensitivity of diazonium salt compounds. (Measurements were made on methanol solutions in concentrations of 25 mg/L.)
FIGURE 7.22 Color formation reaction of a typical leuco dye.
advance. The magenta image is fixed by 365-nm ultraviolet rays because the
yellow image has already been fixed.
Because the cyan color-developing layer does not require fixation, a leuco
dye and an organic acid used for conventional heat-sensitive recording paper are
used as color-developing materials. The leuco dye is a kind of pH-indicating
agent that is colorless in the basic and neutral ranges but forms a dye under acidic
conditions (Figure 7.22). This reaction is reversible. Heat-Responsive Microcapsule
In the color-developing layers of TA paper, two colorless compounds that form
a dye through a reaction are included separately in the recording layer, and they
react by heating to form a dye image. Thus, it is important to have a system in
Color Desktop Printer Technology
which the two compounds are stably separated at room temperature and both
react rapidly by heating.
Heat-responsive microcapsules have been introduced into TA paper. Microcapsules are minute vessels of core–shell structure. The core components are
protected from the surroundings by the shell, and with a heat-responsive microcapsule, permeability of substances through a polymer membrane constituting
the shell varies significantly with the surrounding temperature. More concretely,
the shell comprises a poly(urea-urethane) membrane. Below glass transition
temperature (Tg), it has a very low permeability of substances, and above Tg, the
permeability increases by severalfold. This change is probably due to a drastic
change in the force of hydrogen bonding within molecules or among molecules
of the membrane-forming polymer around Tg.
In the magenta and yellow color-developing layers of TA paper, microcapsules consist of a core where a hydrophobic diazonium salt compound is dissolved
in a hydrophobic and high-boiling-point solvent and a shell of poly(urea-urethane)
membrane coexists with a coupler and an organic base. At normal temperature
(below the Tg of the shell membrane), the diazonium salt compound is insulated
from outside under hydrophobic conditions so that it is quite stable despite its
high activity. However, when heated, the three components are mixed to generate
a dye-forming reaction.
With the cyan color-developing layer, Leuco dye is similarly included in the
heat-responsive microcapsule. TA Printer
As shown in Figure 7.20, key parts of a TA printer consist of a thermal head, a
set of ultraviolet fluorescent lamps, and mechanical parts for paper transport.
The thermal head consists of minute heating elements arranged on a ceramic
base, which can thermally record 300 to 600 dpi. Thermal energy applied to TA
paper may be controlled by the magnitude of the electrical energy to the heating
elements; however, the heating temperature varies, depending on the surrounding
temperature or the base temperature of the thermal head, even if the same electrical energy is applied. Thus various controls are employed so that stable recording density may be reproduced, irrespective of the printing environment. Further,
as is apparent from the thermal color-development characteristics in Figure 7.19,
the magenta color development starts in the saturated density region of yellow,
and the cyan color development starts in the saturated density region of magenta.
Thus, a high-density yellow or magenta is likely to cause mixed colors. To prevent
this, a three-dimensional look-up table (LUT) is installed in the printer so optimal
heating conditions of yellow, magenta, and cyan (YMC) may be chosen for RGB
image information. More concretely, when saturation is demanded, the density
to prevent mixed colors is suppressed, and when more density than saturation is
required, the color is heated sufficiently to get a high density.
Figure 7.23 shows the emission spectra of ultraviolet fluorescent lamps for
fixation. The ultraviolet fluorescent lamps are inexpensive and highly efficient;
Film-Based Printers
0 .9
Magent a
0 .8
0 .7
0 .6
0 .5
0 .4
0 .3
0 .2
0 .1
300 320 340 360 380 400 420 440 460 480 500
Wave length (nm)
FIGURE 7.23 Emission spectra of the fixing lamps.
FIGURE 7.24 Digital color home printer using a TA system (Fuji CX-400).
however, they have a defect: illumination energy changes easily with the surrounding temperature or deterioration. A TA printer is designed so that printing
may be carried out under optimal conditions by always monitoring the light
intensities. Figure 7.24 shows a digital home printer for directly obtaining fullcolor prints from a digital camera output. At a speed of 20 sheets per hour, printing
of L-size (89 × 127 mm) without edge margins is possible, with the equivalent
image quality of a conventional silver-halide photo. Figure 7.25 shows a TA
printer in which thermal heads for YMC recording are separated individually.
Color Desktop Printer Technology
Cyan recording head
TA paper
Magenta fixing lamp
(365 nm)
Magenta recording head
Yellow fixing lamp
(419 nm)
Yellow recording head
FIGURE 7.25 (See color insert following page 176.) Configuration of high-speed threehead tandem digital color printer. Summary
TA is a simple and highly reliable full-color printing system. It provides prints
equivalent to silver-halide photos in a completely dry system, without generating
any waste materials. For the present, it is used for printing from digital cameras
that require instant output, but in the future its use may be expanded to include
applications in the medical field and various other markets.
Technically, because all the necessary components for color development are
included in the paper, it is necessary to improve the whiteness at the image base;
however, the storage level of whiteness equivalent to that of silver-halide photos
has been attained. As for the durability of printed images, it is still inferior to
that of silver-halide photos; however, it has reached the level of several tens of
years in dark storage and is being improved year by year.9
As for the printer, performance improvements in ultraviolet-ray LEDs will
allow its use as an illumination light source for fixation in the near future. The
illumination energy of LEDs is far more stable than that of fluorescent lamps,
and it has a longer life and compactness. It will enable a system design in which
ease, simplicity, and high reliability — merits of TA — will become more
1. K. Shiota, Optical printers, Journal of Institute of Image Information and Television Engineers, 43, 1223–1229, 1989.
2. Y. Ozawa et al., Development of Digital Minilab System Frontier 350/370, Fuji
Film Research & Development, 45, 45–41, 2000.
3. Y. Nakamura et al., Development of Digital Minilab System Frontier 330, Fuji
Film Research & Develoment, 48, 15–21, 2003.
Film-Based Printers
4. M. Kubo, A. Ueshima, and M. Okino, Full-Color Laser Printer, Pictrography
3000(1), Hardware, Technical Research Report, Institute of Electronic Information
Communication, Vol. 93, No. 274, Oct. 1993.
5. Kamosaki, Yokokawa, and Inagaki, Full-Color Laser Printer, Pictrography
3000(2), Photosensitive Materials, Technical Research Report, Institute of Electronic Information Communication, Vol. 93, No. 274, Oct. 1993.
6. A. Igarashi, T. Usami, and S. Ishige, IS&T’s 10th International Congress on
Advances in NIP Tech., pp. 323–326, 1994.
7. M. Sato, M. Takayama, and M. Tsugita, IS&T’s 10th International Congress on
Advances in NIP Tech., pp. 326–329, 1994.
8. A. Igarashi and T. Usami, J. Inf. Recording, 22, 347–357, 1996.
9. K. Minami, S. Sano, A. Igarashi, Japan Hard Copy, 98, 91–94, 1998.
Part III
The Management of Color
Color Management
Mitchell R. Rosen
8.1 Introduction..............................................................................................237
8.2 ICC Color Management General Approach ...........................................240
8.3 ICC Color Management ..........................................................................242
8.4 History of ICC ........................................................................................245
References ........................................................................................................247
Color reproduction workflow involves a pair of color devices: a source device
for capturing or creating a complex original stimulus and a destination device
for re-creating the original’s appearance. Image processing steps enhance, maintain, or degrade the appearance of the image or prepare it for rendering on the
destination device. In traditional analog workflows such as film photography, the
processing steps are usually chemically based. Modern workflows such as digital
photography often have processing steps implemented in software or firmware
Color management is designed for digital workflows to provide specialized
color processing between the capture and rendering stages with the goal of maintaining color appearance. The original can be found in the real world or it can
derive from a color device. The original sometimes is of unclear pedigree. Table
8.1 describes all the possible source/destination color reproduction combinations.
Table 8.1 can be deceptive, as it makes the problem of color reproduction
seem easy. The source column of the table lists only cameras, scanners, displays,
printers, and files. The destination column mentions only displays and printers.
If the problem truly boiled down to only five possible inputs and two possible
outputs, as implied by the table, then color management should be uncomplicated.
It turns out that within each category of device, there are many different
technologies that could potentially be involved. Each technology introduces different color reproduction characteristics. Also, within a family of similar technologies, there are many design choices that go into differentiating the various
devices. These, too, have a large impact on how color devices respond. Color
management, when practiced well, makes it possible for a single reproduction
system to allow any number of different device types to participate.
Color Desktop Printer Technology
Color Reproduction Input/Output Device Combinations
Where Original
Is Found
How Original Is Brought into the
Color Reproduction Workflow
Where Reproduction
Is Rendered
In the real world
On a color device
For color management to work, the stimulus and response character of a
digital device must be characterized. Digital cameras are stimulated by color and
respond with digital values. Scanners are similar. Displays and printers are the
opposite: they are stimulated by digital values and respond with colors. Table 8.2
summarizes this.
There are two basic approaches to color management. Each requires knowledge of the stimulus/response character of color devices. The two approaches
differ in how they take advantage of device characterizations. One approach to
color management calibrates every device to act like a particular, standard pseudodevice. One such popular pseudo-device is described by the sRGB standard (see
Figure 8.1).
Stimulus/Response Pairings for Digital
Digital Device
Digital Values
Digital Values
Digital Values
Digital Values
Color Management
FIGURE 8.1 sRGB calibration-based approach to color management. Source devices are
calibrated to deliver standard sRGB digital values as their “response” and destination
devices are calibrated to accept sRGB digital values as their “stimulus.”
The second color management approach is profile based. Here device characterizations are packaged into standard data structures known as profiles. Two
profiles will be used at the time of color reproduction. The first profile describes
the source device and the second profile describes the destination device. Color
processing will use these profiles to attempt to maintain the color appearance of
the original.
Each device has its own profile. The profile teaches the color management
system how to relate the device’s digital values to colorimetry. Instead of emulating a standard pseudo-device, as in the sRGB case, here each device is free
to act in its native way. Figure 8.2 illustrates this approach. The International
Color Consoritum (ICC) defines the industry standard for profile-based color
This chapter concentrates on the profile-based approach to color management
as defined by the ICC. Because this approach allows devices to retain their
intrinsic stimulus and response character, it has the potential of maximizing
The ICC profile specification determines the format of an ICC profile. Basically, the profile file is a tagged structure much as the TIFF image file format is
a tagged structure. Such a format includes a standard header, a table of contents,
FIGURE 8.2 ICC profile-based approach to color management. Every device has its own
profile that describes that device’s relationships among digital values and color.
Color Desktop Printer Technology
and a series of data structures called tags that are enumerated in the table of
An ICC source device profile contains tags that describe the relationship
between input digits and colorimetry. Tags in an ICC destination profile describe
the relationship between colorimetry and output digital values. The most popular
operating systems have implemented support for ICC profiles. These technologies
have been implemented in Windows as image color management (ICM) and in
Macintosh as ColorSync.
Figure 8.3 describes a standard profile-based color management approach. An
image is captured by or created on a source device. The profile describing the
source device and the profile describing the destination device are used to guide
image processing. Afterward, the processed image is rendered on the destination
Equation 8.1 and Equation 8.2 illustrate one of the typical transformations
from device digit to device independent color that can be encoded within an
source profile. Equation 8.1 uses one-dimensional look-up tables (1D LUTs) that
linearize an input RGB signal. Equation 8.2 follows with a 3 × 3 matrix that
transforms the linear values to XYZ.
Input Image
by Source
Source Device
Destination Device
Output Image
by Destination
FIGURE 8.3 General profile-based color management block diagram.
Color Management
[ R ′ ] = LUTR→R′ [R ]
[G ′ ] = LUTG→G′ [G]
[ B′ ] = LUTB→B′ [BB]
 X   a11
  
 Y  =  a 21
 Z   a 31
a 22
a 23
a13   R ′ 
 
a 23  G ′ 
a 33   B′ 
where RGB are the digital values of the source device, aji are the matrix coefficients to estimate colorimetry, and XYZ are the tristimulus values for an object
under a given illuminant.
For a common transformation chain often parameterized by a destination
profile, Equation 8.3 through Equation 8.5 are presented below. In Equation 8.3,
L*a*b* values are modified through individual 1D LUTs. A multidimensional
LUT in Equation 8.4 transforms the modified L*a*b* values to linearized cyan,
magenta, yellow, black (CMYK). A final set of 1D LUTS in Equation 8.5 is
applied to appropriately shape the CMYK signals for the destination printer.
 L*′  = LUTL→L′  L* 
 
 
 a *′  = LUTa→a′  a * 
 
 
 b*′  = LUTb→b′  b* 
 
 
 C′ 
M′ 
  = LUTLab′→CMYK′  L*′ , a *′ , b*′ 
 Y′ 
 
 K′ 
[C] = LUTC′→C [C′ ]
[ M ] = LUTM′→M [ M ′ ]
[ Y] = LUTY′→Y [ Y′ ]
[ K ] = LUTK′→K [ K ′ ]
To use profiles, the image-processing stage of the color reproduction block
diagram must be able to make appropriate use of the transformation coefficients
found in the tags of the input and output profiles. The actual transformation chains
Color Desktop Printer Technology
implied by the profiles may be directly applied to images or they may be concatenated to form more efficient processes. In addition to being able to perform
the color transformations described by the profiles, an image-processing engine
must also be able to convert among ICC’s supported device-independent colorspaces known as profile connection spaces (PCS). Currently supported PCSs are
variants of XYZ and L*a*b*.
If a color reproduction system were being used to convert an image from an
input device with a profile utilizing Equations 8.1 and 8.2 to an output device
with a profile utilizing Equation 8.3 through Equation 8.5, then the imageprocessing stage would need to provide an additional transform from XYZ to
L*a*b* between Equation 8.2 and Equation 8.3.
The ICC requires that tags describe device characteristics in terms of D50 colorimetry. Chromatic adaptation must be applied for measurements taken under
different conditions. Also, a special normalization of the colorimetric values is
specified. This normalization is convenient for processing images to be rendered
as reflection prints because it guarantees that very bright whites on input will be
mapped to paper white on output. In ICC terminology, this manipulated colorimetry is called profile connection space (PCS).
For most color capture devices, the mapping from its three channels and to
three dimensions of colorimetry is overall one-to-one. Where the actual mapping
is many-to-one or one-to-many, this is due to differences between the instrumental
metamerism of the input device and the metameric characteristics of the standard
observer. For most printing devices, there is much redundancy between colorimetry and output digits due to the typical presence of a fourth colorant. Makers of
ICC profiles must deal with these conflicts and provide color transformations
with single mappings in each instance.
Gamut mapping is an important concept in color reproduction. The goal of
color reproduction is to maintain the color appearance of an original. Sometimes
this means attempting to match exactly the colorimetry of the original. Other
times, color appearance actually is improved when moving between two different
media by using different colorimetry on the reproduction. Either way, an algorithm determines what color should be printed or displayed. If that color is not
achievable on the output device, gamut mapping is necessary.
Figure 8.4 is a gamut-mapping cartoon. The contour indicates the gamut limit
of a device and the star is considered an out-of-gamut color that needs to be reproduced. There are many choices available to the profile-builder as to how to map
colors into the gamut. One popular method is to choose a particular point on the L*
axis (a centroid) and map all colors toward it. Figure 8.4b illustrates such an
Regardless of how the out-of-gamut color is brought into the device gamut,
a second question is whether the point should be mapped to only the gamut
surface or brought further into the gamut. One reason to bring colors further
Color Management
FIGURE 8.4 Gamut-mapping cartoon. Contour indicates gamut limit of a device; a star
indicates out-of-gamut value to be mapped. For this particular gamut-mapping scheme, a
second star on the L* axis indicates a centroid value toward which out-of-gamut colors
are mapped.
in-gamut is to allow for differentiation among out-of-gamut colors. Bringing some
out-of-gamut colors into the bulk of the gamut is known as creating a soft-tuck
vs. the idea of a hard-tuck to the gamut surface.
Creating a soft-tuck becomes a very complicated process because in-gamut
colors must also be compressed toward the center of the gamut. Figure 8.5
illustrates this point. Figure 8.5a shows the compression of the in-gamut colors
to make room for some out-of-gamut colors.
Figure 8.5b shows four gamut-related regions that are sometimes defined.
The central core near the neutral axis is sometimes left alone without any compression. The soft region between the central core and the gamut surface is often
compressed, moving in-gamut colors and making room for some out-of-gamut
colors. Then there is an out-of-gamut region that defines which colors are brought
into the gamut bulk. Finally, there is the outermost area from which all colors
are mapped directly onto the gamut surface.
An entity that applies ICC-compliant image processing is called a color
management module (CMM). The CMM processing typically involves a pair of
profiles, one referring to the image-capture device and the other referring to the
image-rendering device, as shown in Figure 8.6a. Common image transformation
chains that may be encoded in source and destination profiles were discussed in
Section 8.2. Equation 8.1 and Equation 8.2 describe an image transformation
chain typically encoded in the tags of a source profile and Equation 8.3 through
Equation 8.5 are typically parameterized within the tags of a destination profile.
Before processing images, the CMM will often apply data processing to the
profile data for the sake of efficiency. The fact that each ICC profile is tied to the
common PCS makes the concatenation of a series of transforms fairly straightforward and low-cost computationally. Equation 8.6 through Equation 8.8 show
one possible transformation chain that could result from collapsing portions of
the full set of Equation 8.1 through Equation 8.5. These equations demonstrate
a popular image-processing chain that starts with one-dimensional LUTs applied
Color Desktop Printer Technology
FIGURE 8.5 Soft-tuck cartoon. Out-of-gamut colors are brought into the bulk of the
gamut. Some in-gamut colors need to be compressed to make room.
to three input channels followed by a three-dimensional look-up and possibly a
final one-dimensional look-up applied to each output channel (see Figure 8.6b).
Computationally and with respect to memory requirements, the image-processing
chain described here is very feasible.
Equation 8.6 is the same as Equation 8.1. Equation 8.2 through Equation 8.4
are concatenated into the multi-dimensional LUT of Equation 8.7. Equation 8.8
is no different from Equation 8.5. Note that in building this more efficient colorprocessing chain, it is still necessary for the concatenation engine to plug in an
extra transformation that resolves the XYZ and L*a*b* discontinuity found
between Equation 8.2 and Equation 8.3. That additional step is also captured in
the LUT of Equation 8.7.
[ R ′ ] = LUTR→R′ [R ]
[G ′ ] = LUTG→G′ [G]
[ B′ ] = LUTB→B′ [BB]
 C′ 
M′ 
  = LUTRGB′→CMYK′ [ R ′, G ′, B′ ]
 Y′ 
 
 K′ 
Color Management
Transform from
Input Digital Values
to Colorimetry
Transform from
Colorimetry to
Output Digital Values
Create New
of Input and
Transform from
Input Device
Values to
Output Device
FIGURE 8.6 (a) Logical ICC image-processing block diagram. (b) Typical CMM image
processing with efficiencies. Transform box may include 1-D and 3-D look-ups, matrices,
and other functional forms.
[C] = LUTC′→C [C′ ]
[ M ] = LUTM′→M [ M ′ ]
[ Y] = LUTY′→Y [ Y′ ]
[ K ] = LUTK′→K [ K ′ ]
For a number of years prior to the first ICC specification in 1994,3 a few companies
tried to promote and sell their own color management solutions including Kodak,
Tektronix, and EFI. These tended to involve special software that a user would
manually invoke to process images. The packages were commercial flops.
One company, Adobe, did have a successful color management launch, which
was implemented in every rasterizing image processor (RIP) that supported the
PostScript Level 2 Language.4 Today’s ICC approach is remarkably similar to
what was brought to the masses by Adobe in 1990. However, Adobe’s color
management was limited to only Postscript printers and, thus, did not offer a
general solution.
Color Desktop Printer Technology
Color Management Framework Interface
3rd Party
3rd Party
FIGURE 8.7 An operating system’s view of color management showing how color-naive
applications and libraries can use color management, and how those with added-colorvalue can plug-in. (Adapted from ICC specification.2)
At least one ad hoc industry group, the Association of Color Developers
(ACD), was formed during those pre-ICC days to come up with an open
standard to bootstrap the acceptance of universal color management. The effort
went nowhere. The ACD did create one lasting legacy, however. It put together
a schematic that looks much like the one in Figure 8.7. The group realized that
selling color management was going to succeed only when the operating system
vendors understood how they could facilitate better color reproduction and
when applications vendors without color expertise were supplied with an
approach that would not burden them with the development costs of creating
color smarts.
Illustrations like the one in Figure 8.7 were designed to speak to the operating
system and application provider community. To a significant extent, this sketch
was effective in selling the idea. What was needed at that point was a leader to
take the plunge and implement the infrastructure. That leader came in the form
of Apple Computer. Apple made the investment and in January 1993 introduced
an extension to its operating system toolbox: ColorSync.
The original ColorSync printer model had limited value. It was based on a
transform that was only accurate for perfect dot-on-dot printers.5 Apple realized
early on that ColorSync was going to need significant industry help and buy-in
to become a success. Gerry Murch, who led Tektronix’s early color management
efforts, was then at Apple and headed the drive to create industry consensus. He
brought together a group that, at first, had the name ColorSync 2 Consortium. At
that time it was heavily centered on helping Apple define the next level of
ColorSync. Kodak, Adobe, and others already had fully functioning color
management systems that were incompatible with the original ColorSync. Those
companies had motivation to see a more mature ColorSync that supported their
approaches. Device manufacturers and graphics arts applications providers were
also motivated to see Apple succeed so that they did not have to implement or
buy their own color management technology. Hewlett Packard, with extreme
Color Management
dominance in the desktop- and office-printer market at the time, did not join the
consortium until close to its roll-out and could easily have derailed it if the
company had been motivated.
As the movement proceeded, the name and mandate of the consortium
changed in order to encourage companies with no reason to help Apple to
consider joining. From ColorSync 2 Consortium, the next name was InterColor
and eventually the International Color Consortium (ICC). As evidence of the
successful detachment of the consortium from Apple, even Microsoft eventually joined the consortium.
In the early formation of ColorSync 2, Adobe was very successful in helping
steer the consortium in several important ways. As already mentioned, Adobe
was already a color management vendor and Adobe’s experience became the
ICC’s wisdom. Adobe’s main motivation was to support the PostScript concept
that the same document printed at any time on any printer should always look
the same. To make that happen, as much information as possible should be packed
into the profile’s required tags, with relatively little room for creativity in the
image processing. This is consistent with a “smart profile” and a “dumb CMM,”
and this is how an ICC color management system currently works.
In June 1994 the first ICC profile format specification, version 3.0, was
published.5 Currently, version 4.2 is available. ICC profile format specifications
are made available at the consortium website:
1. IEC 61966-2-1 Amendment 1: 2003, Multimedia systems and equipment —
Colour measurement and management — Part 2-1: Colour management —
Default RGB colour space — sRGB, 2003.
2. ICC Specification ICC.1:2003-09 (Profile version Image technology color
management — Architecture, profile format, and data structure, 2004.
3. InterColor Consortium, InterColor Profile Format, Version 3.0, 1994.
4. Adobe Systems, Inc., Postscript® Language Reference Manual (2nd ed.), Addison-Wesley Professional, 1990.
5. J.E. Thornton, Y.J. Lee, and M.R. Balonon-Rosen, The Apple ColorSync Printer
Profile Model and Its Optimization, Proc. of IS&T’s 46th Annual Conference, pp.
147–150, 1993.
Desktop Spectral-Based
Mitchell R. Rosen, Francisco H. Imai,
Yongda Chen, Lawrence A. Taplin, and
Roy S. Berns
9.1 Current Metameric Systems ....................................................................249
9.2 Traditional vs. Spectral-Based Systems..................................................250
9.3 Spectral Image Acquisition System ........................................................251
9.4 Spectral Color Management....................................................................254
9.5 Spectral Model for Printers .....................................................................260
9.6 Conclusions..............................................................................................263
References and Bibliography............................................................................263
Although not ready for commercial use, an active area of research with significant
potential impact on desktop printers is the innovation of spectral reproduction
systems. Output from such systems will have more stable color relative to an
original than is currently achieved. Spectral systems do not rely on colorimetry.
Instead, they attempt to reproduce actual spectral reflectance or spectral transmittance of an original or its spectral radiance.
Advantages to these new systems will be many. A printed reproduction that
matches the reflectance spectra of an original will preserve color matches over
the range of all illuminants and for all observers. Reflectance-based rendering
would be less sensitive to small characterization and calibration errors and to
rendering noise. Radiance matching could be very powerful when reconstructing
an object’s appearance for viewing under new conditions. Proofing systems will
improve as well as systems that make spot-color and specialty ink choices.
In 1802, Young presented to the Royal Society his conclusion that human perception of color is based on three primaries. Young was correct. This phenomenon
Color Desktop Printer Technology
Wavelength (nm)
FIGURE 9.1 Spectral sensitivity of the three cone types.
is due to the typical character of sensors known as cones that populate the human
retina. Cones are useful for seeing color under normal lighting conditions. There
are three types of cones. S-cones have sensitivity primarily to the relatively short
wavelengths, L-cones are biased toward relatively long wavelengths, and M-cones
fall between them (see Figure 9.1). As long as the three cone types have the same
responses when looking at two objects, the two objects will have the same color.
Thus, three well-chosen primaries can stimulate the cones to see any color. The
phosphors on a TV work in this way, as do cyan, magenta, and yellow inks in a
Pairs of objects with distinct spectral reflectance properties but that arouse
the same psychological color response when viewed under a single light source
are known as metameric pairs. Metamerism has been important for color reproduction from the time of cave drawings to modern digital systems. In fact, until
recently, almost all innovation in color imaging has been based on metamerism.
Maxwell depended upon it for his 1861 demonstration of three-channel full-color
photography and manufacturers of modern digital cameras, scanners, and printers
still rely on it. Looking at the four major desktop printer technologies discussed
in Part II of this book, it is clear that they, as currently used, create color through
the use of metamerism. They depend primarily on the use of the standard process
inks of cyan, magenta, and yellow to create the appearance of all colors.
In spite of its wide use, metameric reproduction has a number of drawbacks.
Many are addressed by spectral color reproduction [Hunt, 1995; Berns, 1999;
Hill, 2002]. Color from a printed copy will change its appearance under different
illumination. If the original color is metameric with an original, it is likely that
Desktop Spectral-Based Printing
FIGURE 9.2 (See color insert following page 176.) Garrett Johnson’s “Metameric
Cows” [Johnson, 1998]. This demonstration simulates how the same reflectance properties
under a single light source can appear different to different observers. The front half and
the back half of the cow are different reflectances. To the two-degree observer (left), the
front and the back have the same color. To the ten-degree observer (right), the front and
the back have different colors.
under a different light source, the match between original and reproduction will
break. There is also a surprisingly wide range of sensitivity differences among
people. Thus, a pair of colors may appear metameric to one observer but not to
another (see Figure 9.2).
Although the data flow diagram does not need to change between spectral
and metameric systems, the details concerning each stage and the connections
between stages are very different. Figure 9.3a shows a typical, contemporary,
completely metameric system. In such a system, source and destination devices
are characterized relative to colorimetry. This type of system was discussed in
the previous chapter on color management. The input device captures three
wideband red, green, blue (RGB) channels. The image is processed anticipating
the colorimetric rendering capabilities of a cyan, magenta, yellow, black (CMYK)
printer. Figure 9.3b shows the flow diagram for a spectral system. Here, spectral
response and spectral rendering are characterized. The source device captures
many narrowband channels. The image is processed minimizing spectral error
for a many-colorant printer.
The remainder of the chapter is devoted to detailing spectral systems, showing
their potential and limitations. Although spectral image acquisition and estimation
are not the main emphases here, spectral printing systems will be put into the
context of end-to-end scene-to-hardcopy reproduction.
As shown in Figure 9.3b, the spectral color printing system receives data captured
by a multi-channel imaging system. The overall performance of the system will
depend highly on the quality of the spectral information generated by the spectral
image acquisition system. A variety of camera approaches is available for spectral
Color Desktop Printer Technology
Input Image
by 3-Channel
Source Device
Input Image
Captured by
Source Device
Source Device
Destination Image
by 4-Channel
Destination Device
Source Device
Destination Image
Rendered by
Destination Device
FIGURE 9.3 (a) Imaging flow diagram for a metameric color reproduction system. (b)
Imaging flow diagram for a spectral color reproduction system.
image acquisition systems [Bacci et al., 1992; Bacci, 1995; Burns, 1997;
König and Praefcke, 1998; Tominaga, 1999; Rosen and Jiang, 1999; 2002;
Hauta-Kasari et al., 1999; Haneishi et al., 2000; Imai, Berns, and Tzeng, 2000a;
Imai, Rosen, and Berns, 2000b; Imai et al., 2001; Murakami et al., 2001; Hardeberg, Schmitt, and Brettel, 2002]. In the early 1990s, Bacci et al. (1992, 1995)
used a small-aperture optical fiber reflectance spectrophotomer to scan across
painting and frescos. In general, there are two different image-capturing
approaches for full-frame spectral images. The first is a narrowband approach
[König and Praefcke, 1998; Tominaga, 1999; Rosen and Jiang, 1999; Imai, Rosen,
and Berns, 2000b; Hardeberg, Schmitt, and Brettel, 2002]. Results have been
reported on using a set of interference filters [König and Praefcke, 1998] or a
narrow-bandpass tunable filter [Slawson, Ninkov, and Horch, 1999] in front of a
monochrome sensor [Tominaga, 1999; Rosen and Jiang, 1999; Imai, Rosen, and
Berns, 2000b; Hardeberg, Schmitt, and Brettel, 2002]. Both charge-coupled
devices (CCDs) [Tominaga, 1999; Imai, Rosen and Berns, 2000b; Hardeberg,
Schmitt, and Brettel, 2002] and panchromatic, black-and-white film [Rosen and
Jianag, 1999] have been utilized as sensors for narrow-bandpass capture. Alternatively, Hauta-Kasari et al. used an optical system with a diffraction grating to
produce spectrally adjustable illumination [Hauta-Kasari et al., 1999]. These
systems are analogous to using a spectrophotometer, sampling the visible spectrum at known bandpass and wavelength intervals and, because they are spectral
measurement methods, simple transformation can be generated from captured
camera signals to reflectance spectra.
Desktop Spectral-Based Printing
The second approach is an abridged technique that uses five or more absorption [Haneishi, et al., 2000; Imai, Rosen, and Berns, 2000b; Hardeberg, Schmitt,
and Brettel, 2002] or interference [Burns, 1997] wideband filters in front of a
monochromatic camera and requires spectral estimation [Marimont and Wandell,
1992]. The Visual Arts System for Archiving and Retrieval of Images project
(VASARI) developed a system to capture paintings at high resolution [Saunders
and Cupitt, 1993]. The VASARI scanner recorded the reflectance properties
between 400 and 700 nm using 7 channels, each with a 70-nm bandwidth.
Alternatively, it is possible to build a broadband rewritable filter [Miyazawa,
Hauta-Kasari, and Toyooka, 2001].
In another approach, a conventional trichromatic digital camera is combined
with selected absorption filters or light sources [Imai, 2000a; Sun, 2002]. The
filters or light sources are selected to optimize the performance of the spectral
estimation. In this broadband approach, the spectral reflectance of each pixel of
the original scene can be estimated using a priori spectral analysis with direct
spectrophotometric measurement and imaging of samples of the object to establish a relationship between the camera signals and spectral reflectance. The
wideband acquisition takes advantage of the possibility of decreasing the spectral
sampling increment without a significant loss of spectral information because of
the smooth absorption characteristics of both manmade and natural colorants
within the visible spectrum [Maloney, 1986; Parkkinen, 1989; Jaaskelainen, 1990;
Dannemiller 1992; Vrhel 1992, 1994]. However, this technique has poorer
performance compared to the previous methods because it is based on a color
camera limited by its inherent spectral sensitivities. However, this method has
the advantage of being more easily implemented.
The accuracy of the spectral estimation method can be evaluated by taking
the similarity of the original and the estimated spectral curves into account or
considering how the reproduced spectrum resembles the original scene when
viewed by a human observer. Therefore, both spectral curve error metrics and
colorimetric metrics have to be evaluated [Imai 2002a, 2002b, 2003a]. We recommend manually viewing spectral difference curves as an important aspect of
system evaluation along with the calculation of multiple objective metrics such
as color-difference equations, spectral curve difference metrics, and a metameric
index calculation. A metameric index compares the extent to which two spectra
are different between a reference condition and a test condition under different
illuminants or observers [Nimeroff, 1965; Fairman, 1997; Chen et al., 2004]. In
particular, the metameric index using Fairman parameric decomposition [Fairman, 1997] corrects the test spectrum until exact tristimulus equality is achieved
under a reference condition. Then the metameric index is calculated using an
International Commission on Illumination (CIE) color-difference equation for a
test illuminant and observer. For the spectral curve difference metrics, we often
use the root mean square error (rms) between original and estimated spectra and
the goodness-of-fit coefficient (GFC) [Hernández-Andrés, 2001]. GFC is based
on the inequality of Schwartz having values between 0 and 1 and indicates the
Color Desktop Printer Technology
correlation between two spectral curves; a value unity corresponds to a perfect
spectral match. The metric is calculated using Equation 9.1:
(λ j )R e ( λ j )
∑ R
( λ j )
∑ R ( λ )
where Rm(λj) is the measured original spectral data at the wavelength λj and Re(λj)
is the estimated spectral data at wavelength λj. According to the developers of
the metric, GFC ≥ 0.999 and GFC ≥ 0.9999 are required for respectively good
and excellent spectral matches [Hernández-Andrés, 2001].
Because we are considering complex images and not only single color stimulus, we also have to consider psychophysical evaluation after reproduction [Day,
The multi-band image acquisition and spectral estimation process can, by
itself, produce a book involving many aspects such as estimation techniques
[Mancill, 1975; Pratt, 1976; Praefke, 1996], encoding [Keusen, 1996], design and
use of filters [Vent, 1994; Vrhel, 1995; Vora, 1997; Haneishi, 2000; Imai, 2001;
Hardeberg, 2002; Rosen, 2002] and number of samples used in characterization
[Tsumura, 1999].
Multi-band image acquisition and spectral estimation result in improvement
over traditional image acquisition systems, even if the resulting image is rendered
for a particular illuminant and all the subsequent color management is only
colorimetric-based [Saunders, 1993; Imai, 1996; Sun, 2002; Day, 2003]. Spectral
image acquisition can be a useful analytical tool for painting [Baronti, 1998;
Casini, 1999; Berns, 2002] or human skin [Sun, 2002]. However, in reproduction
terms, we can take full advantage of the spectral-based color image acquisition
if this is followed by a subsequent spectral-based color image management and
a spectra-based color separation printing.
In Chapter 8, traditional color management was discussed. Here we discuss how
a spectral version of color management might be implemented. For a spectral
input profile, Equation 8.1 and Equation 8.2 could easily be updated to a transformation such as found in Equation 9.2 and Equation 9.3, respectively.
[Cn′ ] = LUTC →C [Cn ]
(n:1 to N )
Desktop Spectral-Based Printing
 f1   a11
f  a
 .2  =  .21
 ..   ..
  
 fR   a R1
a 22
a1N   C1′ 
a 2 N   C2′ 
.   .. 
a RN  CN′ 
where N is the number of input channels, Cn are the digital counts from the source
device; ajii is the matrix coefficient to estimate spectral reflectance or radiance;
R is the number of samples per spectrum; and fr is spectral reflectance or radiance
from an object.
To update Equation 8.3 through Equation 8.5 for producing a spectral output,
one is tempted to simply use analogous structures as those seen in the following
[ fr′ ] = LUTf →f [ fr ]
(r: 1 to R )
 D1′ 
D 
 .2′  = LUTf …R′→D …M′ [ f1′ , f2′ , ..., fR′ ]
 .. 
 D M′ 
[ Dm ] = LUTD
m ′ →Dm
[ D m′ ]
(m: 1 to M )
where M is the number of output channels and Dm are the digital counts for the
destination device.
Equation 9.5 introduces a huge problem, however. One cannot apply a simple
scale factor to all the image-processing approaches that work for metameric systems
described by Figure 9.3a to create a system that works for the spectral systems of
Figure 9.3b. When R in Equation 9.5 is sufficiently large, the multi-dimensional
look-up table of that equation becomes far too large for implementation. Rosen,
Ohta, and Derhak have attacked this problem through the introduction and development of the concept of the interim connection space (ICS) for spectral color management [Rosen, 2001a; 2003a; 2003b; Derhak, 2005].
Beyond changes in processing, the differences between data capture and
throughput demands of the two systems described in Figure 9.3 motivate fresh
approaches to imaging in a spectral system.
Discussion of spectral color management strategies has begun to gain some
momentum within the community [Hung, 1999; Hill, 2000; Rosen, 2000, 2001a;
2001b; 2003a; 2003b; TAO, 2002]. The International Color Consortium (ICC)
color management based on colorimetry shown in Figure 8.3 could be updated
to spectral color management in a simple fashion, as shown in Figure 9.4.
Color Desktop Printer Technology
Although the change from this viewpoint is quite simple, the tremendous increase
in dimensionality for spectral processing will cause it to deviate dramatically
from the ICC approach in its most efficient configuration.
In 2000, Rosen et al. described the basic aspects of a spectral profile. The
device characterization found within a spectral profile needs a new profile connection space (PCS), one that may be based on reflectance, transmittance, or
radiance. A color management system supporting spectral profiles will probably
be required to know how to deal with any of these PCSs. Table 9.1 shows which
spectral PCSs would be the most appropriate different device types.
Input profiles, within a spectral color management system, could describe
any of the following transforms:
(a) Input digits to reflectance-based PCS or transmittance-based PCS
(b) Input digits to radiance-based PCS
(c) Input digits to a colorimetric PCS (not spectral, but still useful and
essential for backward compatibility).
Color processing would need to be able to make the following conversions:
(d) Reflectance or transmittance to radiance by multiplying an illuminant
(e) Radiance to reflectance or transmittance by dividing an illuminant
(f) Reflectance or transmittance to colorimetry by multiplying an illuminant, multiplying color-matching functions, and then integrating
(g) Radiance to colorimetry by multiplying color-matching functions and
then integrating.
A destination profile could describe any of the following transforms:
(h) Reflectance-based PCS or transmittance-based PCS to output digits
(i) Radiance-based PCS to output digits
(j) Colorimetric-based PCS to output digits
These basic operations could be applied in series to carry out complex tasks.
For example, consider the problem of choosing the right wallpaper for one’s
tungsten-lit living room. It is well known that the narrowband fluorescents in the
store can be misleading. A customer sophisticated in spectral reproduction techniques could take out a multispectral camera, capture a picture of the wallpaper
in the store, and take a second shot of the store’s lights. At home the two images
could be downloaded to a computer. Using the camera’s spectral profile, each
image would be converted to radiance (functionality (b), above). Using the radiance image of the store light source, the wallpaper radiance image could then be
transformed to reflectance (functionality (e), above). A subsequent picture of the
living room light would allow for calculation of what the wallpaper would
have looked like at home (functionality (d), above). Finally, to view the color
image on the home monitor, the radiance image would be converted to XYZ
Desktop Spectral-Based Printing
Transform from
Device Digit to
Transform from
Spectra to Device
FIGURE 9.4 Possible logical spectral image-processing block diagram.
Spectral PCS Basis for Device Types
Spectral PCS Basis
Reflectance or Transmittance
Reflectance or Transmittance
Alternative Spectral PCS Basis
Reflectance (for illuminant-controlled studio capture)
Radiance when viewing conditions are defined
FIGURE 9.5 (See color insert following page 176.) Spectral color management would
provide a way to impose a new light source on the image of a captured object.
Color Desktop Printer Technology
(functionality (g), above) and then the ICC monitor profile would be used to
transform to monitor RGB (functionality (j), above). Figure 9.5 illustrates a
cartoon of the user action in this example. Figure 9.6 shows the series of actions
taken by the spectral color management system.
The most significant work done to date for defining a spectral profile has
emerged from the Akasaka Natural Vision Research Center of the Telecommunications Advancement Organization (TAO) of Japan. In its 2002 research report
[TAO, 2002] a detailed proposal for many aspects of a spectral profile, was
described. Borrowing data structures and format from the ICC specification, the
Natural Vision group put together an important first step.
The Akasaka report included some of the important structures that a spectral
profile would need. Importantly missing from the report, however, was a description of a profile tag for the spectral characterization of a printing device. Multichannel input devices and multi-primary displays were included. A spectral
printer, however, brings up issues of characterization and processing that far
exceed the considerations for the previous tags.
The Natural Vision proposal included two spectral PCSs. Both were defined
between 380 and 780 nm in 1-nm increments. The first space was specified in
units of W/sr/m2/nm and the other was specified in units of reflectance factor.
Following the logic of Table 9.1, most cameras and display devices could use the
first of these PCSs. Most scanners and printers could use the second one.
Three profile types important to spectral color management were defined in
the Akasaka report. The N-component spectrum-based input profile and the
N-component spectral matrix-based input profile both contained tags supporting
transformation described by Equation 9.7 and Equation 9.8. The M-primary
display profile supports the transformation described by Equation 9.9 and Equation 9.10.
linearCn = mTRC[Cn – bn] (n: 1 to N)
 f1   n
f   0
 .2  = 
 ..   ..
   .
 fR  
0 
  a11
0   a 21
...   .
1   R1
k n 
a 22
a1N   linear C1 
a 2 N   linear C2 
. 
a RN   linear CN 
where bn is the bias digital count; Cn is the digital count of the input device (0–1);
aij are the matrix coefficients to estimate spectral reflectance or radiance; kn is
the coefficient of sensitivity-level correction; R is 401; and fR is spectral reflectance or radiance from an object.
Desktop Spectral-Based Printing
Source Device
with Respect
to Radiance
Wallpaper Captured
by Multi-Channel
Source Device
Store Light
by Multi-Channel
Source Device
Wallpaper in Store
Input Image
Living Room Light
by Multi-Channel
Source Device
Store Light
Input Image
Living Room Light
Input Image
Apply Source Profile Transform to Radiance
Wallpaper in Store
Radiance Image
Store Light
Radiance Image
Divide by
Light Spectra
Living Room Light
Radiance Image
Multiply by
Light Spectra
Wallpaper in
Living Room
Radiance Image
with Respect to
Colorimetry (ICC)
Multiply by Color
Matching Functions
and Integrate
Wallpaper in
Living Room
XYZ Image
Apply Destination
Transform to
Monitor Digits
Wallpaper in
Living Room
RGB Monitor
FIGURE 9.6 Block diagram illustrating the logical color management system actions that
would accompany the wallpaper example described.
Color Desktop Printer Technology
linearCn = mTRC[Cn] (n: 1 to N)
 r1   p11
r  p
 .2  =  .21
 ..   ..
  
 rR   p R1
p 22
pR 2
p1N   linear C1   p1 
p 2 N   linear C2   p 2 
+ .
  .. 
. 
  
p RN   linear CN   p R 
( 9.10 )
where Cn is the digital count of the input device (0–1); pRN is the spectral radiance
of the nth primary measured at maximum digital count; pn,r is the bias spectrum;
R is 401; and rR is the radiance of an object on the display.
In the spirit of the spectral color management systems previously described
by Rosen [Rosen, 2000, 2001a], Yamaguchi and coworkers in the TAO report
illustrated the use of the spectral profiles in color management systems that
support both spectral and colorimetric profiles [Yamaguchi, 2002].
The Cellular–Yule–Nielsen–Spectral–Neugebauer (CYNSN) model [Wyble,
2000] was used to investigate a spectral model for printers. The Neugebauer
model [Neugebauer, 1937] is an additive model for multi-ink printing in which
a macroscopic colored area is a weighted sum of the individual microscopic
colors. The weights are determined from the halftoning algorithm. Because of
light scattering within the paper, the relationship between the macroscopic and
microscopic colors becomes complicated. Yule and Nielsen [Yule, 1951] found
that exponentiating reflectance, in similar fashion to converting reflectance to
optical density, greatly improved prediction accuracy. Viggiano [Viggiano, 1985]
further improved performance by considering the optical mixing over a narrow
range of wavelengths. The resulting Yule–Nielsen–Spectral–Neugebauer (YNSN)
model is shown below:
Rλ = 
Fi R λ ,i 
where Rλ,i is the macroscopic spectral reflectance of the ith color type at 100%
area coverage, n is the Yule–Nielsen exponent, and Fi are the fractional area
coverages of each microscopic color type. The maximum value of i depends on
the number of inks and the halftoning algorithm. For three-color printing that
uses rotated screens or frequency modulation, the maximum number is eight (e.g.,
cyan, magenta, yellow, red, green, blue, black, and paper white). That is, three
inks printed randomly result in eight unique colors; for six-color printing, the
result is 64 colors. These colors are known as the Neugebauer primaries. The
fractional areas are determined as a product of random variables, shown in
Desktop Spectral-Based Printing
Equation 9.12. These probabilities when used for printing are attributed to
Demichel [Demichel, 1924].
Fi =
∏  (1 − a )
if ink j is in Neugebauer Primary i 
if ink j is not in Neugebauer Primary i
where aj is the effective area coverage of ink j. (The term effective is used because
this area coverage is determined statistically, not optically using reflection microscopy [Wyble, 2000].) Area coverage is a function of the digital signal, dj, controlling the amount of ink delivered to the substrate, defined in Equation 9.13.
aj = f(dj)
The n value and the effective area coverage relationships are determined
statistically, typically using one-color ramps.
From a geometric viewpoint, the YNSN model performs multi-dimensional
linear interpolation across Rλ1/n space, the interpolation weights calculated using
Equation 9.12 [Rolleston, 1993; Balasubramanian, 1999]. Heuberger [1992] recognized that interpolation performance is always improved by reducing the interpolation area. This is easily achieved by creating subspaces called cells. Rolleston
and Balasubrananian evaluated the cellular method when using the YNSN model
for printer characterization, creating the Cellular–Yule–Nielsen–Spectral–Neugebauer or CYNSN model. Improvement was shown to be significant. In particular,
the cellular approach greatly reduced the need for highly accurate analytical
models beyond what was typically achieved using the YNSN model.
The value of the cellular approach is shown graphically in Figure 9.7. Here,
the outer square and solid circles represent the YNSN model, and the point O1
is calculated by interpolation from four outer corner points, P1, P2, P3, and P4,
which are represented by the solid circles, the Neugebauer primaries. The whole
figure, including solid and dashed circles, represents the CYNSN model in two
dimensions. If each ink is printed at four levels, there are more known values
and can be used to create cells (subspaces). The corners of each cell are the
cellular Neugebauer primaries, or simply cellular primaries. If the cellular model
is used to predict the point O1, we can use the nearest four cellular primaries
(P11, P22, P33, and P44). The accuracy improvement of the cellular model is
significant because interpolation is performed in a much smaller subspace. Of
course, the cost is that more colors need to be printed and measured. Agar and
Allebach [Agar, 1998] showed the relationship between prediction error and
number of cellular primaries. The accuracy of the cellular model can be improved
significantly as more primaries are considered, though, as noted by Balasubramanian, there is a diminishing return.
Color Desktop Printer Technology
FIGURE 9.7 Graphical interpretation of the YNSN and CYNSN models.
Iino and Berns [1998a, 1998b] used the YNSN model to characterize an inkjet
printer and a proofing system for offset printing with good success. Balasubramanian [Balasubramanian, 1996, 1999] evaluated the spectral Neugebauer (SN),
YNSN, and CYNSN models when characterizing a four-color electrophotographic printer. He used a typical black-printer strategy so that the number of
samples required to create a four-color model was not excessive. As expected,
accuracy improved significantly by adding the Yule–Nielsen n value. The addition
of the cellular subspaces resulted in modest incremental improvement. Balasubramanian also investigated using weighted linear regression to optimize the spectral
properties of the Neugebauer primaries. Rather than using macroscopic measurements, he optimized the spectral properties of those primaries not containing
black ink resulting in the best average colorimetric performance. Samples near
the primary in colorant space were weighted more heavily than samples far away.
The improvement to the YNSN model was similar to adding the cellular approach.
Most recently, Imai, Wyble, and Tzeng [Imai et al., 2003b] used the YNSN model
to characterize a CMYK inkjet printer as part of an end-to-end spectral color
reproduction system with reasonable success.
Tzeng [1999a] and Tzeng and Berns [2000] used the YNSN model for sixcolor proofing using cyan, magenta, yellow, black, orange, and green inks. One
goal was to use the proofer to simulate different ink sets. Accordingly, a spectral
model was desired so that proofs could be produced that were minimally
metameric. That is, a printing system defined spectrally either computationally
or by direct measurement could be proofed via spectral color reproduction. This
required numerically inverting the YNSN model. In order to ensure convergence
and reduce processing time, Tzeng subdivided the six inks into 10 four-ink
models. For any color, only four inks would be printed. Model accuracy depended
Desktop Spectral-Based Printing
on the particular four inks. Taplin and Berns [Taplin, 2001] extended Tzeng’s
research and considered all six inks simultaneously. They used an inkjet printer
with a small dot size, an error diffusion halftoning algorithm, and heavyweight
art paper. This combination enabled all of the 64 Neugebauer primaries to be
printed without ink blotting. The YNSN model was used again with reasonable
performance accuracy. More sophisticated optimization algorithms were used for
model inversion as well as a continuous-tone model to provide reasonable starting
values. The printer was used for a spectral color reproduction system for artwork
[Art-SI, 2004].
Today, there are specialized areas where capturing and reproducing the spectra
of an original scene or document are considered important goals. These include
reproduction and archiving of artwork, proofing, medical imaging, remote sensing, and catalog sales. Many other potential opportunities exist for the use of
spectral information in a color reproduction workflow. Given the groundswell of
interest and investigation into spectral imaging capabilities, it is likely that developments will accelerate quickly. Some consumer camera systems are already
increasing the number of input channels, and many desktop printers have six to
eight inks. There is every reason to believe that spectral reproduction will move
to the desktop as the technology continues to mature.
Section 9.5 is reprinted from [Chen, 2004], with permission from IS&T: The
Society for Imaging Science and Technology, sole copyright owners of The
Journal of Imaging Science and Technology.
A.U. Agar and J.P. Allebach, An interative cellular YNSN method for color printer characterization, Proc. IS&T/SID Sixth Color Imaging Conference, pp. 197–200
J.S. Arney, A probability description of the Yule–Nielsen effect I, J. Imag. Sci. Technol.
41, pp. 633–636 (1997a).
J.S. Arney, A probability description of the Yule–Nielsen effect II: The impact of halftone
geometry, J. Imag. Sci. Technol. 41, pp. 637–642 (1997b).
J.S. Arney, C.D. Arney, and P.G. Engeldrum, Modeling the Yule–Nielsen effect, J. Imag.
Sci. Technol. 40, pp. 233–238 (1996).
Art-SI,, Art Spectral Imaging (2004).
M. Bacci, Fibre optics applications to works of art, Sensors and Actuators B 29, pp.
190–196 (1995).
Color Desktop Printer Technology
M. Bacci, S. Baronti, A. Casini, F. Lotti, M. Picollo, and O. Casazza, Non-destructive
spectroscopic investigations on paintings using optical fibers, Proc. Materials Res.
Soc. Symp. 267, pp. 265–283 (1992).
R. Balasubramanian, The use of spectral regression in modeling halftone color printers,
Proc. of IS&T/OSA Optics and Imaging in the Information Age, pp. 372–375
R. Balasubramanian, Optimization of the spectral Neugebauer model for printer characterization, J. Electronic Imaging 8, pp. 156–166 (1999).
S. Baronti, A. Casini, F. Lotti, and S. Porcinai, Multispectral imaging system for the
mapping of pigments in works of art by use of principal-component analysis,
Appl. Optics 37, pp. 1299–1309 (1998).
R.S. Berns, Spectral modeling of a dye diffusion thermal transfer printer, J. Electronic
Imaging 2, pp. 359–370 (1993).
R.S. Berns, Challenges for colour science in multimedia imaging systems, in L. MacDonald and M.R. Luo, Eds., Colour Imaging: Vision and Technology, John Wiley
& Sons, England, pp. 99–127 (1999).
R.S. Berns, Billmeyer and Saltzman’s Principles of Color Technology, 3rd ed., John Wiley
& Sons (2000).
R.S. Berns and M. Shyu, Colorimetric characterization of a desktop drum scanner using
a spectral model, J. Electronic Imaging 4, pp. 360–372 (1995).
R.S. Berns, J. Krueger, and M. Swicklik, Multiple pigment selection for inpainting using
visible reflectance spectrophotometry, Studies in Conservation 47, pp. 46–61
H. Boll, A color to colorant transformation for a seven ink process, Proc. IS&T Third
Technical Sympos. Prepress Proofing Printing, pp. 31–36 (1993).
P.D. Burns, Analysis of Image Noise in Multi-Spectral Color Acquisition, Ph.D. Thesis,
R. I. T., Rochester, NY (1997).
A. Casini, F. Lotti, M. Picollo, L. Stefani, and E. Buzzegoli, Image spectroscopy mapping
technique for non-invasive analysis of paintings, Studies in Conservation 44, pp.
39–48 (1999).
Y. Chen, R.S. Berns, and L.A. Taplin, Six color printer characterization using an optimized
cellular Yule–Nielsen spectral Neugebauer model, J. Imaging Technol., 48, pp.
519–528 (2004).
J.L. Dannemiller, Spectral reflectance of natural objects: how many basis functions are
necessary?, J. Opt. Soc. Am. A9, pp. 507–515 (1992).
E.A. Day, The Effects of Multi-Channel Spectrum Imaging on Perceived Spatial Image
Quality and Color Reproduction Accuracy, M.S. Thesis, R. I.T., Rochester, NY
M.E. Demichel, Procédé 26 pp. 17–21, 26–27 (1924).
M.W. Derhak and M.R. Rosen, Spectral colorimetry using LabPQR — an interim connection space, J. Imaging Sci. Technol., in press.
P. Emmel and R.D. Hersch, A unified model for color prediction of halftoned prints, J.
Imaging Sci. Technol. 44, pp. 351–359 (2000).
H.S. Fairman, Metameric correction using parametric decomposition, Color Res. Appl.
12, pp. 261–265 (1997).
H. Haneishi, T. Suzuki, N. Shimoyama, and Y. Miyake, Color digital halftoning taking
colorimetric color reproduction into account, J. Electronic Imaging 5, pp. 97–106
Desktop Spectral-Based Printing
H. Haneishi, T. Hasegawa, A. Hosoi, Y. Yokohama, N. Tsumura, and Y. Miyake, System
design for accurately estimating the spectral reflectance of art paintings, Appl.
Opt. 39, pp. 6621–6632 (2000).
W. Hanson, Color photography: from dream, to reality, to commonplace, in E. Ostroff,
Ed., Pioneers of Photography, Their Achievements in Science and Technology,
SPSE, Springfield (1986).
J.Y. Hardeberg, F. Schmitt, and H. Brettel, Multispectral color image capture using a liquid
crystal tunable filter, Optical Engineering 41, pp. 2533–2548 (2002).
M. Hauta-Kasari, K. Miyazawa, S. Toyooka, and J. Parkkinen, Spectral vision system for
measuring color images, J. Opt. Soc. Am. A16, pp. 2352–2362 (1999).
R. Herbert, Hexachrome color selection and separation — model for print media, Proc.
IS&T 3rd Technical Sympos. Prepress Proofing and Printing, pp. 28–30 (1993).
J. Hernández-Andrés, J. Romero, J. L. Nieves, and R. L. Lee Jr., Color and spectral analysis
of daylight in southern Europe, J. Opt. Soc. Am. A18, pp. 1325–1335 (2001).
K.J. Heuberger, Z.M. Jing, and S. Persiev, Color transformations and lookup tables, Proc.
TAGA/ISCC, pp. 863–881, (1992).
B. Hill, Color capture, color management and the problem of metamerism: does multispectral imaging offer the solution? Proceedings of SPIE, 3963, pp. 2–14 (2000).
B. Hill, (R)evolution of color imaging systems, Proceedings First Europ. Conf. Color
Graphics, Imaging Vision, pp. 473–479 (2002).
P.C. Hung, Color reproduction using spectral characterization, Proc. Intl. Sympos. Multispectral Imaging Color Reproduction Digital Arch., pp. 98–105 (1999).
P.C. Hung, T. Mitsuhashi, and T. Saitoh, Inkjet printing system for textile using Hi-fi
colors, Proc. PICS Conference, pp. 46–50 (2001).
R.W.G. Hunt, The Reproduction of Colour, 5th ed., Fountain Press, Kingston-uponThames, U.K. (1995).
K. Iino and R.S. Berns, Building color management modules using linear optimization I.
Desktop color system, J. Imaging Sci. Tech. 42, pp. 79–94 (1998a).
K. Iino and R.S. Berns, Building color management modules using linear optimization II.
Prepress system for offset printing, J. Imaging Sci. Tech. 42, pp. 99–144 (1998b).
F.H. Imai, N. Tsumura, H. Haneishi, and Y. Miyake, Principal component analysis of skin
color and its application to colorimetric color reproduction on CRT display and
hardcopy, J. Imaging Sci. Tech. 40, pp. 422–429 (1996).
F.H. Imai, Multi-Spectral Image Acquisition and Spectral Reconstruction using a Trichromatic Digital Camera System Associated with Absorption Filters, MCSL Technical
Report (1998).
F.H. Imai, R.S. Berns, and D. Tzeng, A comparative analysis of spectral reflectance
estimation in various spaces using a trichromatic camera system, J. Imaging Sci.
Technol. 44, pp. 280–287 (2000a).
F.H. Imai, M.R. Rosen, and R.S. Berns, Comparison of spectrally narrow-band capture
versus wideband with a priori sample analysis for spectral reflectance estimation,
Proc. Eighth Color Imaging Conf., pp. 234–241 (2000b).
F.H. Imai, S. Quan, M.R. Rosen, and R.S. Berns, Digital camera filter design for colorimetric and spectral accuracy, Proc. Third Intl. Conf, Multispectral Color Sci.,
University of Joensuu, Finland, pp. 13–16 (2001).
F.H. Imai, M.R. Rosen, and R.S. Berns, Comparative study of metrics for spectral match
quality, Proc. IS&T’s First European Conf. Color Graphics, Imaging Vision, pp.
492–496 (2002a).
Color Desktop Printer Technology
F.H. Imai, L.A. Taplin, and E.A. Day, Comparison of the Accuracy of Various Transformations from Multi-Band Images to Reflectance Spectra, MCSL Technical Report
F.H. Imai, L.A. Taplin, and E.A. Day, Comparative Study of Spectral Reflectance Estimation Based on Broadband Imaging Systems, MCSL Technical Report (2003a).
F.H. Imai, D.R. Wyble, and D. Tzeng, A feasibility study of spectral color reproduction,
J Imag Sci Tech 47, p. 543 (2003b).
T. Jaaskelainen, J. Parkkinen, and S. Toyooka, Vector-subspace model for color representation, J. Opt. Soc. Am. A7, pp. 725–730 (1990).
G.M. Johnson, Computer Synthesis of Spectroradiometric Images for Color Imaging
Systems Analysis, M.S. Thesis, R.I.T., Rochester, NY, (1998).
T. Keusen, Multispectral color system with an encoding format compatible with the
conventional tristimulus model, J. Imaging Sci. Tech. 40, pp. 510–515 (1996).
T. Kohler and R.S. Berns, Reducing metamerism and increasing gamut using five or more
colored inks, Proc. IS&T Third Technical Sympos. Prepress, Proofing and Printing,
pp, 24–28 (1993).
F. König and W. Praefcke, A multispectral scanner, in L. MacDonald and M.R. Luo, Eds.,
Colour Imaging: Vision and Technology, John Wiley & Sons, Chichester, pp.
129–144 (1998).
P. Kubelka, New contribution to the optics of intensely light-scattering materials. Part I,
J. Opt. Soc. Am. 38, pp. 448–457 (1948).
H. Kueppers, Process for manufacturing systematic color tables or color charts for sevencolor printing, and tables or charts produced by this process, U.S. patent number
4878977 (1989).
L.T. Maloney and B.A. Wandell, Color constancy: a method for recovering surface spectral
reflectance, J. Opt. Soc. Am. A3, pp. 29–33 (1986).
C.E. Mancill, Digital Color Image Restoration, Ph.D. Thesis, University of Southern
California, Los Angeles (1975).
D.H. Marimont and B.A. Wandell, Linear models of surface and illuminant spectra, J.
Opt. Soc. Am. A 9, pp. 1905–1913 (1992).
K. Miyazawa, M. Hauta-Kasari, and S. Toyooka, Rewritable broadband color filters for
spectral image analysis, Optical Review 8, pp. 112–119 (2001).
Y. Murakami, T. Obi, M. Yamaguchi, N. Ohyama, and Y. Komiya, Spectral reflectance
estimation from multi-band image using color chart, Opt. Commun. 188, pp. 47–57
H.E.J. Neugebauer, Die theoretischen grundlagen des mehrfarbendrucks, Zeitscrift fur
wissenschaftliche Photographie [Reprinted in Proc. SPIE: Neugebauer Memorial
Seminar on Color Reproduction 1184, pp. 194–202 (1989)], (1937).
I. Nimeroff and J.A. Yurow, Degree of metamerism, J. Opt. Soc. Am. 55, pp. 185–190
N. Ohta, Structure of the color solid obtainable by three subtractive color dyes, Die Farbe
20, pp. 115–134 (1971).
N. Ohta, The color gamut obtainable by the combination of subtractive color dyes IV.
Influence of some practical constraints, Photograph. Sci. Eng. 28, pp. 228–231
V. Ostromoukhov, Chromaticity gamut enhancement by heptatone multi-color printing,
Proc. of SPIE 1909, pp. 139–151 (1993).
J. Parkkinen, J. Hallikainen, and T. Jaaskelainen, Characteristic spectra of Munsell colors,
J. Opt. Soc. Am. A 4, pp. 318–322 (1989).
Desktop Spectral-Based Printing
W. Praefcke, Transform coding of reflectance spectra using smooth basis vectors, J.
Imaging Sci. Tech. 40, pp. 543–548 (1996).
W.K. Pratt and C.E. Mancill, Spectral estimation techniques for the spectral calibration
of a color image scanner, Appl. Opt. 15, pp. 73–75 (1976).
R. Rolleston and R. Balasubramanian. Accuracy of various types of Neugebauer model,
Proc. Color Imaging Conference, pp. 32–37 (1993).
M.R. Rosen, Navigating the Roadblocks to Spectral Color Reproduction: Data-Efficient
Multi-Channel Imaging and Spectral Color Management, Ph.D. Dissertation, RIT,
M.R. Rosen and N. Ohta, Spectral color processing using an interim connection space,
Proc. 11th CIC, 187–192, 2003b.
M.R. Rosen and X. Jiang, Lippmann 2000: A spectral image database under construction,
Proc. International Sympos. Multispectral Imaging Color Reprod. Digital Arch.,
Chiba University, Chiba, Japan, pp. 117–122 (1999).
M.R. Rosen, E.F. Hattenberger, and N. Ohta, Spectral redundancy in a six-ink jet printer,
J. Imaging Sci. Technol., 48, 192–202, 2004.
M.R. Rosen, M.D. Fairchild, G.M. Johnson, and D.R. Wyble, Color management within
a spectral image visualization tool, Proc. Eighth Color Imaging Conf. pp. 75–80
M.R. Rosen, F.H. Imai, X. Jiang, and N. Ohta, Spectral reproduction from scene to
hardcopy II: image processing, Proc. of SPIE 4300, pp. 33–41 (2001a).
M.R. Rosen, L.A. Taplin, F.H. Imai, R.S. Berns, and N. Ohta, Answering Hunt’s web
shopping challenge: spectral color management for a virtual swatch, Proc. Ninth
Color Imaging Conf., pp. 267–273 (2001b).
M.R. Rosen, F.H. Imai, M.D. Fairchild, and N. Ohta, Data-efficient methods applied to
spectral image capture, J. Soc. Photogr. Sci. Technol. Japan 65, pp. 353–362
T. Sato, Y. Nakano, T. Iga, S. Nakauchi, and S. Usui, Color reproduction based on low
dimensional spectral reflectance using the principal component analysis, Proc.
IS&T/SID Fourth Color Imaging Conf., pp. 185–188 (1996).
D. Saunders and J. Cupitt, Image processing at the National Gallery: the VASARI project,
National Gallery Tech. Bull. 14, pp. 72–86 (1993).
L. Sipley, A Half Century of Color, Macmillan, New York (1951).
L. Sipley, Photography’s Great Inventors, American Museum of Photography, Philadelphia
R.W. Slawson, Z. Ninkov, and E.P. Horch, Hyperspectral imaging: wide-area spectrophotometry using a liquid-crystal tunable filter, Publ. Astronomical Society Pacific
111, pp. 621–626 (1999).
E.J. Stollnitz, V. Ostromoukhov, and D.H. Salesin, Reproducing color images using custom
inks, Computer Graphics Proc., Annu. Conf. Ser., pp. 267–274 (1998).
Q. Sun and M.D. Fairchild, Statistical characterization of face spectral reflectances and
its application to human portraiture spectral estimation, J. Imaging Sci. Technol.
46, pp. 498–506 (2002).
L.A. Taplin and R.S. Berns, Spectral color reproduction based on a six-color inkjet output
system, Proc. Ninth Color Imaging Conf., pp. 209–213 (2001).
Telecommunications Advancement Organization of Japan, R&D Report on Image Presentation and Transmission System for Next Generation, pp. 22–79, in Japanese
Color Desktop Printer Technology
S. Tominaga, Spectral imaging by a multichannel camera, J. Electronic Imaging 8, pp.
332–341 (1999).
N. Tsumura, H. Sato, T. Hasegawa, H. Haneishi, and Y. Miyake, Limitations of color
samples for spectral estimation from sensor responses in fine art painting, Optical
Rev. 6, pp. 67–61 (1999).
D. Tzeng, Spectral-Based Color Separation Algorithm Development for Multiple-Ink
Color Reproduction, Ph.D. Thesis, R. I.T., Rochester, NY (1999).
D. Tzeng and R.S. Berns, Spectral-based ink selection for multiple-ink printing I. Colorant
estimation of original objects, Proc. IS&T/SID Sixth Color Imaging Conf., IS&T,
Springfield, VA, pp. 106–111 (1998a).
D. Tzeng and R.S. Berns, Spectral reflectance prediction of ink overprints by KubelkaMunk turbid media theory, Proc. TAGA/ISCC Sympos., pp. 682–697 (1999b).
D. Tzeng and R.S. Berns, Spectral-based ink selection for multiple-ink printing II. Optimal
ink selection, Proc. IS&T/SID Seventh Color Imaging Conf., pp. 182–187 (1999c).
D. Tzeng and R.S. Berns, Spectral-based six-color separation minimizing metamerism,
Proc. IS&T SID Eighth Color Imaging Conf., pp. 342–247 (2000).
D.S. Vent, Multichannel Analysis of Object-Color Spectra, M.S. Thesis, R.I.T., Rochester,
NY (1994).
J.A.S. Viggiano, The color of halftone tints, Proc. TAGA 37, pp. 647 (1985).
P.L. Vora and H.J. Trussell, Mathematical methods for the design of color scanning filters,
IEEE Trans. Image Processing 6, pp. 312–320 (1997).
M.J. Vrhel and H.J. Trussell, Color correction using principal components, Color Res.
Appl. 17, pp. 328–338 (1992).
M.J. Vrhel and H.J. Trussell, Optimal color filters in the presence of noise, IEEE Trans.
Image Processing 4, pp. 814–823 (1995).
M.J. Vrhel, R. Gershon, and L.S. Iwan, Measurement and analysis of object reflectance
spectra, Color Res. Appl. 19, pp. 4–9 (1994).
D.R. Wyble and R.S. Berns, A critical review of spectral models applied to binary color
printing, Color Res. Appl. 25, pp. 5–19 (2000).
M. Yamaguchi, T. Taraji, K. Ohsawa, T. Uchiyama, H. Motomura, Y. Murakami, and N.
Ohyama, Color image reproduction based on the multispectral and multiprimary
imaging: experimental evaluation, Proc. SPIE, 4663, pp. 15–26 (2002).
J.A.C.Yule and W.J. Nielsen, The penetration of light into paper and its effect on halftone
reproductions, Proc. TAGA 3, pp. 65 (1951).
S. Zuffi, R. Schettini, and G. Mauri, Using genetic algorithms for spectral-based printer
characterization, Proc. SPIE 5008, pp. 268–275 (2003).
Additive primaries, 42
Addressability of printers, 35–38
Adhesion phenomena, between toner image,
photoreceptor, 170
American Standard Code for Information
Interchange (ASCII), 87
Angles for halftone printing, 51
Apple, 88
Apple dot matrix printer, 88
Apple ImageWriter, 88
Apple StyleWriter, 89
Area sequential method, 197–198
ASCII. See American Standard Code for
Information Interchange
Ballistric wire computer printers, 87
Banding level, acceptability of, 181
Bible, printing of, 6. See also Gutenberg, Johann
Blanket-to-blanket press, 15
Blanket-to-blanket-type perfecting press, 17
Canon printers, 89–90
Character printers, 86–87
Characteristics of letters, 7
Chroma, 41
Clay, development of type characters from, 4
Clephane, James, 85
Coarse screen engraving, 7
Coated inkjet papers, 142
Color control, 54–64
Color densitometry, 44
Color density, methods of reproduction of, 206
Color halftone printing, 49–51
Color laser printers, 92–94
Color management, 235–277
color reproduction input/output device
combinations, 238
desktop spectral-based printing, 249–268
digital devices, stimulus/response pairings
for, 238
International Color Consortium color
management, 240–245
International Color Consortium history,
spectral color management, 254–260
Color mixing, 41–45
Color reproduction input/output device
combinations, 238
Colorimetry coordinates, 44–45
Computer printing, 90–105
Continuous inkjet printing, 95, 99–101
Continuous tone image, with halftone image,
Controller, 206–207
characteristics for application, 207
color gamut, 207
product range, 207
resolution, 206
running cost, 207–208
tone reproduction, 206–207
user aspects, 207–208
Conventional etching, 7–8
Conventional printing presses, ink transfer to
substrate, 9
color gamuts in, 182
type comparisons, 176
Copper etching, 8
Copperplate engraving, 25–26
Crosfield, lasergravure process, 21
Cross-line gravure screen, 23
CRT/FOT exposure system, 214
Dahlgren type of direct-feed dampening system,
Damages to planographic plate, 19
Dampening systems, 17–18
Dark Age, termination of, 7
Densitometer, 39–40
advantages over visual matching, 40
with light sources, 39
monochrome, 40
output displays, 40
readings, 48
RGB, 44
Densitometer readings, 48
Densitometry, 44
Descriptive hue circle, 43
Desktop computing, printing, 35
Desktop printers, 85–110
American Standard Code for Information
Interchange, 87
Apple, 88
Apple dot matrix printer, 88
Apple ImageWriter, 88
Apple StyleWriter, 89
ballistric wire computer printers, 87
business, 85–110
Canon, 89–90
Canon/Centronics, joint development
program of, 86
Canon LBP-CX, 88
Centronics 730, 88
Centronics 779, 88
Centronics Data Computer Corporation, 86
character printers, 86–87
Clephane, James, 85
color laser printers, 92–94
computer printing, 90–105
continuous inkjet printing, 95
dot matrix printers, 86–87
mechanisms used, 86
drop-on-demand printing, 95–96
Epson, 88–89
Epson MX-80, 88
Hertz, Dr. C. Hellmuth, 95
Hewlett Packard, 90
laserjet laser printer, 88
historical overview, 85–90
Howard, Robert, 86
inkjet printing, 95–105
continuous inkjet printing, 99–101
inkjet papers, 103–105
technological developments, 102
LaserWriter, 89
Macintosh, 89
market, 85–110
MX-100, 88
Scitex VersaMark system, 97
Silentype printer, 88
stored-energy printer, 87
thermal wax transfer, 94–95
Color Desktop Printer Technology
Trendcom, 88
Wang, Dr. An, 86
Wang Laboratories, 86
Desktop spectral-based printing, 249–268
metameric systems, 249–250
spectral color management, 254–260
spectral image acquisition system, 251–254
spectral model for printers, 260–263
spectral PCS basis for device types, 257
traditional vs. spectral-based systems,
Developer for laser printers, classification of,
Device-dependent models, 34
Diazonium salt compounds, spectroscopic
photosensitivity, 229
Digital devices, stimulus/response pairings for,
Digital halftoning, 68. See also Dithering
classification, 73–75
techniques, 73
Digital halftoning mask, 67–70
Direct-feed dampening system, Dahlgren type
of, 17
Direct lithography, 12
Direct thermal printer, 205
Dither, 74
Dither mask, 74
Dithering, 74
Dot gain, 54–56, 61–64
mechanisms of, 55–56
Dot matrix printers, 35, 86–87
mechanisms used, 86
Dot shapes, examples, 37
Downing, William, 46
Driography. See Waterless lithography
Drop-on-demand printing, 95–96
Drop size, 117
Drop volume, 117
Drum press, 16
Drypoint engraving, 25
Dye-donor-layer-absorbing printing media type,
Dye sublimation printers, 199
structure, 199
sublimation due, sheets, 199
Dye thermal-transfer printers, 195–210
area sequential method, 197–198
configuration, 198
controller, 206–207
characteristics for application, 207
color gamut, 207
product range, 207
resolution, 206
running cost, 207–208
tone reproduction, 206–207
user aspects, 207–208
direct thermal printer, 205
driving mechanism, 197–198
dye sublimation printer, 199
structure, 199
sublimation due, sheets, 199
electrosensitive transfer printer, 204
improved printer engines, 203
improvement in processing speed, 203
improvement of durability of thermaltransfer printer, 203
laser thermal-transfer printer, 204–205
light-sensitive microcapsule printer, 204
line sequential methods, 197
other printers based on thermal transfer,
stability, 208
recording density, 208
thermal issues, 208
wear, 208
thermal head, 200–202
concentrated thermal-transfer head, 202
requirements, 200
structures, features, 200–201
temperature control, 201
thermal rheography, 203–204
wax melt printers, 200
sheets, 200
structure, 200
Dyes, 135
Electric typewriter, 35
Electromechanical engraving, 21–22
with diamond styli, 23
Electrophotography. See Laser printer
architecture comparisons, 162
development methods, 168
Electrosensitive transfer printers, 204
Emulsion aggregation toner
particle size distribution, 190
process flow of, 189
Engraving, 7
coarse screen engraving, 7
copperplate, 26
copperplate engraving, 25
drypoint engraving, 25
electromechanical, 21–22
with diamond styli, 23
halftone photoengraving, 7
Helioklischograph, 21
line engraving, 26
line photoengraving, 7
Messotint engraving, 25
photoengraving, 7
steel-die, 26
Epson printers, 89
Exposure systems, 213
Fabric, in screen printing, 27
Ferrite carrier, 185
Film-based printers, 211–234
emerging technology, 221–232
heat-responsive microcapsule, 229–230
high-quality imaging technology,
image forming process, 222–224
material composition, 222–224
pictrography, 221–226
pictrography printer, 224–225
principle of recording, 222
recording principle, 226–228
thermo-autochrome method, 226–232
thermo-autochrome paper, 228–229
thermo-autochrome printer, 230–232
historical overview, 211–215
photo printers, 216–221
area CCD image reading technology,
exposure technology, 220–221
film scanning technology, 216–220
line CCD image reading technology,
marking technology, 220–221
Fixing lamps, emission spectra, 231
Flat-bed cylinder printing press, 8
Flexography, 9–10
Floyd–Steinberg, error diffusion, 70–73
Floyd–Steinberg diffusion kernel, 73
Fourier, Jean Baptist Joseph, 76
Fourier analysis, noise power, 76
Frontier 330 (area CCD), film scanner unit, 218
Frontier 350, laser exposure unit, 221
Frontier 350 (line CCD), film scanner unit, 217
Fusing system, 172
FVP600, 214
components, spectral characteristics, 215
Graininess, 179
paper surface smoothness, relationship
between, 191
Granularity, printing device, 81–82
Granularity constants, objective measures,
Gravure cylinders, 21
Gravure plate, 20
structure of, 24
Gravure plate-making process, 22
Gravure printing, 19–25
application, 25
elements of, 20
historical overview, 20–21
intaglio printing, difference between, 25
plate making, 21–24
chemical method, 21
electromechanical engraving, 21–22
pad transfer printing, 23–24
variable area–variable depth plate,
printing process, 24–25
Gutenberg, Johann, 4, 10
Bible, printing of, 6
Halftone algorithm, 68
Halftone color calculations, 51
Halftone graininess, paper surface smoothness,
relationship between, 191
Halftone images, 47–48, 64–75
digital halftoning mask, 67–70
Floyd–Steinberg, error diffusion, 70–73
noise distribution technique, 70
pre-press process photography, 64–67
Robert’s method, 70
Halftone patterns, 71
Halftone photoengraving, 7
Halftone printing, 47–49
angles for, 51
Halftone process photography, 66
Halftone relief printing, 8
Halftoning, digital
classification, 73–75
techniques, 73
Harris, A.F., invention of offset lithography, 12
Heat-responsive microcapsule, 229–230
Helioklischograph engraving, 21
Hertz, Dr. C. Hellmuth, 95
Hewlett Packard printers, 90
Color Desktop Printer Technology
Hollow character, 169
Howard, Robert, 86
Hue, 41
Hue circle, descriptive, 43
Human visual system. See Visual system
ICC. See International Color Consortium
Image density, measurement instrumentation,
Image forming process, 222–224
Image quality, 31–84, 173–180
additive primaries, 42
chroma, 41
color control, 54–64
color densitometry, 44
color mixing, 41–45
colorimetry coordinates, 44–45
cyan density, 44
defined, 32
device-dependent models, 34
digital halftoning
classification, 73–75
techniques, 73
dither, 74
dither mask, 74
dot gain, 61–64
dot-gain mechanisms, 55–56
dot-gain phenomenon, 54–55
electric typewriter, 35
frequency, size, conversion table, 75
granularity, printing device, 81–82
granularity constants, objective measures,
halftone images, 64–75
digital halftoning mask, 67–70
Floyd–Steinberg, error diffusion, 70–73
noise distribution technique, 70
pre-press process photography, 64–67
Robert’s method, 70
hue, 41
human visual system, 76–79
color gamut, 42
image quality circle, 32
image quality models, 35
color, 41–45
primary, 43
letter-quality image, 35
mechanistic models, 61–64
metrics, 32–34
modeling, 34–35
Munsell system, for naming colors, 41
Murray–Davies equation, 54
n-factor, 56–57
Neugebauer equations, 54
noise power
Fourier analysis, 76
spectrum, 75–82
pictorial image printing, 45–54
color halftone printing, 49–51
densitometer readings, 48
halftone color calculations, 51
halftone printing, 47–49
Moire Effect, 49–51
Murray–Davies equation, 49
Neugebauer colors, 53
primary colors, 42–43
usage of term, 42
RGB densitometer, 44
system models, 34
text printing, 35–41
addressability of printer, 35–38
densitometer, 39–40
advantages over visual matching, 40
desktop computing, printing, 35
dot-matrix printers, 35
electric typewriter, 35
image density, measurement
instrumentation, 39
impact typewriter, 35
ink density, 38–41
inkjet printer, 35
laser printer, 35
letter-quality, usage of term, 35
letter-quality image, 35
line jags, 38
monochrome densitometer, 40
National Bureau of Standards
Resolution Test Chart, 36
offset-litho quality, 35
offset-quality text, letter-quality text,
distinguished, 38
printer resolving power, 35–38
test patterns, 36
test procedures, 36
United States Air Force Resolution Test
Chart, 36
tone, 54–64
trichromatic theory of color vision, 41–42
value, 41
visual models, 34–35
Yule–Nielsen color calibration, 60–61
Yule–Nielsen correction, 56–60
Yule–Nielsen equation, 56
Yule–Nielsen Murray–Davies equation, 56
Image quality circle, 32–33
Image quality models, 35
Impact typewriter, 35
Impression cylinder web press, 16–17
Improvement in processing speed, 203
In-line open press, 15
Ink density, 38–41
Ink metering system, in flexo printing, 10
Ink spread, geometric model, 61
Inking systems, 18
Inkjet papers, 103–105
Inkjet printer, 35
configuration menu, 149
Inkjet printing, 95–105, 111–156
continuous inkjet printing, 99–101
dark ink, 151–154
image processing, 153–154
halftoning, 148–151
high-fidelity color, 145–154
color separation, 147–148
historical overview, 112–117
inks, 117
major technologies, 113–114
Piezo inkjet, 115–117
print media, 117
thermal inkjet, 115–117
years of color, 114–115
ink storage, delivery, 128–131
inkjet media, 138–143
mechanical properties, 139
inkjet papers, 103–105
inks, 133–138
dyes, 135
liquid inks, 133–135
pigments, 135–137
solid inks, 137–138
light ink color, 150–151
media, 138–143
imaging properties, 139–143
coated inkjet papers, 142
overhead transparency films, 143
photo papers, 142
plain papers, 142
specialty media, 143
mechanical properties, 139
print modes, 143–145
printhead service, maintenance, 131–133
technological developments, 102
technologies, 118–128
piezoelectric inkjet, 125–128
thermal inkjet, 118–125
Inks, 133–138
advances in, 117
color, 41–45
dyes, 135
liquid inks, 133–135
pigments, 135–137
primary, 43
solid inks, 137–138
Intaglio printing, 25–26
application, 26
gravure printing, difference between, 25
line engraving, 26
copperplate engraving, 26
messotint, 26
steel-die engraving, 26
Intermediate belt transfer, 171
International Color Consortium, 240–245
International Color Consortium history,
ISO printing technology, 5
Italian Renaissance, 20
Japanese dying process, yuzen-zome, 27
Keyless tower printing system, 16, 18
Kliche, Karl, 20–21
inventor of modern gravure printing, 20–21
Laser exposure, optical system for, 222
Laser printer, 35, 92–94, 157–194
banding, colors to colors miss-registration,
carrier, 185–187
ingredients, 181–187
charging, 164
cleaning, 172–173
color, 92–94
color fidelity/stability, 174–175
color gamut, 179–180
color uniformity, 175
consumables, 189–192
defects, 178
developer, 165–169
types of, 180–181
development, 165–169
electrophotographic development methods,
fusing, 171–172
Color Desktop Printer Technology
graininess, 178
image quality, 173–180
laser ROS exposure, 164–165
marking process, 160–162
marking technology, 159–180
media, 189–192
photoreceptor, 162–164
process control, 173
reproduction of fonts, 177
technology elements, 162–173
thermal-transfer, 204–205
tone, 175–177
toner, 180–189
ingredients, 181–187
toner manufacturing process, 187–189
transfer, 169–171
Laser ROS optics, 165
Laser thermal-transfer printers, 204–205
Lasergravure process, by Crosfield, 21
Letter-quality, usage of term, 35
Letter-quality image, 35
offset-quality image, distguished, 38
Letterpress printing, 4
Leuco dye, color formation reaction, 229
Light-sensitive microcapsule printers, 204
Line engraving, 26
copperplate engraving, 26
messotint, 26
steel-die engraving, 26
Line jags, 38
produced by discrete dots, 38
Line photoengraving, 7
Line sequential methods, 197
Liquid inks, 133–135
Lithographic plate, 11
Lithography, 11
Luther, Martin, catalyst for Reformation by, 4
Macintosh. See Apple
Management of color. See Color management
Market for printers, 105–107
conditions of, 107
future developments in, 107
sizing, 105–106
Messotint engraving, 25–26
Metameric color reproduction system, imaging
flow diagram, 252
Moire effect, 49–51
screen rotation, 50
Monochrome densitometer, 40
Movable type, 6
Multi-path color laser printer, 161
Multi-path intermediate belt transfer color laser
printer, 163
Multicolor web-fed gravure presses, 24
Munsell color wheel, 41
Munsell system, for naming colors, 41
Murray–Davies equation, 49, 54
National Bureau of Standards Resolution Test
Chart, 36
Neugebauer colors, 53
Neugebauer equations, 54
Noise distribution technique, 70
Noise power, 76, 78, 81
Fourier analysis, 76
spectrum, 75–82
Offset-litho quality, 35
Offset lithography, 12
elements of, 14
Offset lithography press
elements of, 15
types of, 14
Offset-quality text, letter-quality text,
distguished, 38
One-pass printer engine, 203
Optical dot gain, 56
Overhead transparency films, inkjet, 143
Pad transfer printing, 23–24
Paper surface smoothness, halftone graininess,
relationship between, 191
Particle-tolerant ink supply channels, 122
Photo papers, inkjet, 142
Photo printers, 216–221
area CCD image reading technology,
exposure technology, 220–221
film scanning technology, 216–220
line CCD image reading technology,
marking technology, 220–221
Photoengraving, 7
halftone, 7
line, 7
Photographic process halftoning, 38
Photography. See also Electrophotography;
Rotary photogravure printing
pre-press process, 64–67
Photopolymer films, 8
Physical dot gain, 61–62
Pi-Sheng, development of type characters from
hardened clay, 4
Pictorial image printing, 45–54
color halftone printing, 49–51
densitometer readings, 48
halftone color calculations, 51
halftone printing, 47–49
Moire Effect, 49–51
Murray–Davies equation, 49
Neugebauer Colors, 53
Pictrography 3000, internal diagram, 225
Pictrography printers, 221–226
process diagram, 223
Piezo inkjet, advances in, 115–117
Piezoelectric inkjet, 125–128
Pigments, 135–137
Pixel shifting, 220
Planographic plate, damages to, 19
Planographic printing, 11–19
blanket-to-blanket-type perfecting press, 17
classification of, 12
damages to planographic plate, 19
dampening systems, 17–18
direct-feed dampening system, Dahlgren
type of, 17
direct lithography, 12
features of, 11–12
Harris, A.F., invention of offset lithography,
historical overview of, 11
impression cylinder web press, 17
inking systems, 18
keyless tower printing system, 18
lithographic plate, 11
lithography, 11
offset lithography, 12
elements of, 14
offset lithography press, elements of, 15
offset lithography presses, types of, 14
planographic process, 12
plate making, 12–14
laser plate making, 13–14
presensitized plates, 12–13
waterless plates, 13
printing presses, 14–17
Senefelder, Aloys, invention of lithography,
sheet-fed press, 14, 16
water–ink balance, lack of, 19
waterless lithography, 13
web-fed offset press, 16
web letterpress, disadvantage of, 15
web offset press
advantage of, 15
blanket-to-blanket press, 15
disadvantage of, 15
drum press, 16
impression cylinder web press, 16
in-line open press, 15
keyless tower printing system, 16
types of, 15
web press, 14
Planographic process, 12
Plate making, 7–8, 12–14
gravure printing, 21–24
chemical method, 21
electromechanical engraving, 21–22
pad transfer printing, 23–24
variable area–variable depth plate,
laser plate making, 13–14
presensitized plates, 12–13
Platen printing press, 8
Plates, 8
Platforms, 109–234
dye thermal-transfer printer, 195–210
film-based printers, 211–234
inkjet, 111–156
laser printer, 157–194
Powderless etching, 8
Pre-press process photography, 64–67
Presensitized plates, 12–13
Presses, 8–9, 14–17, 64–67
blanket-to-blanket, 15, 17
conventional, ink transfer to substrate, 9
drum, 16
flat-bed cylinder, 8
flexography, 10
with four units, 9
impression cylinder web, 16–17
in-line open, 15
ink transfer to substrate, 9
keyless tower printing system, 16
letterpress, 4, 15
multicolor web-fed gravure, 24
offset lithography
elements of, 15
types of, 14
platen printing, 8
power-operated, types of, 29
rotary, 8
sheet-fed, 14, 16
Color Desktop Printer Technology
types of, 8
web, 14
web-fed offset, 16
web offset, 15–16
advantage of, 15
disadvantage of, 15
types of, 15
web press, 14
Primary colors, 42–43
usage of term, 42
Print density, visual measurement of, 39
Print media, advances in, 117
Print modes, 143–145
Printer market, 105–107
conditions of, 107
future developments, 107
sizing, 105–106
addressability, 35–38
color laser, 92–94
desktop, 85–110
business, 85–110
market, 85–110
direct thermal, 205
dye sublimation, 199
dye thermal-transfer, 195–210
electrosensitive transfer, 204
engines, 203
film-based, 211–234
laser, 157–194
laser thermal-transfer, 204–205
light-sensitive microcapsule, 204
photo, 216–221
pictrography, 224–225
printer engines, 203
resolving power, 35–38
spectral model for, 260–263
thermal-transfer, 203–205
thermo-autochrome, 230–232
wax melt, 200
Printhead service, maintenance, 131–133
computer, 90–105
dye sublimation printer, 199
gravure, 19–25
intaglio printing, distinguished, 25
inkjet, 111–156
intaglio, 25–26
gravure printing, distinguished, 25
letterpress, 4
with light inks, 148–151
photo printers, 216–221
of pictorial images, 45–54
planographic, 11–19
printer engines, 203
recess, 19–26
relief, 4–11
spectral model, printers, 260–263
of text, 35–41
thermal transfer printers, 203–205
through-printing, 26–30
wax melt printers, 200
Printing presses, 8–9, 14–17, 64–67
blanket-to-blanket, 15, 17
conventional, ink transfer to substrate, 9
drum, 16
flat-bed cylinder, 8
flexography, 10
with four units, 9
impression cylinder web, 16–17
in-line open, 15
ink transfer to substrate, 9
keyless tower printing system, 16
letterpress, 4, 15
multicolor web-fed gravure, 24
offset lithography
elements of, 15
types of, 14
platen, 8
power-operated, types of, 29
rotary, 8
sheet-fed, 14, 16
types of, 8
web, 14
web-fed offset, 16
web offset, 15–16
advantage of, 15
disadvantage of, 15
types of, 15
web press, 14
Pulverized toner, 187
Recess printing, 19–26
classification of, 19
conventional gravure plate, 20
conventional gravure plate-making process,
copperplate engraving, 25
cross-line gravure screen, 23
drypoint engraving, 25
electromechanical engraving, 21
with diamond styli, 23
gravure, 19–25
application, 25
historical overview, 20–21
plate making, 21–24
chemical method, 21
electromechanical engraving, 21–22
pad transfer printing, 23–24
variable area–variable depth plate,
printing process, 24–25
gravure cylinders, 21
gravure plate, structure of, 24
gravure printing
elements of, 20
intaglio printing, difference between, 25
intaglio, 25–26
application, 26
line engraving, 26
copperplate engraving, 26
messotint, 26
steel-die engraving, 26
Italian Renaissance, 20
Kliche, Karl, 20–21
lasergravure process, by Crosfield, 21
messotint engraving, 25
multicolor web-fed gravure presses, 24
rotary photogravure printing, 19
rotogravure, origination of term, 20
Recording principle, 226–228
Reformation, development of printing as
catalyst for, 4
Relief printing, 4–11
application, 10–11
characteristics of letters, 7
coarse screen engraving, 7
conventional etching, 7–8
conventional printing presses, ink transfer to
substrate, 9
copper etching, 8
Dark Age, ending of, 7
features of, 7
flat-bed cylinder printing press, 8
flexography, 9–10
Gutenberg, Johann, 4, 10
Bible, printing of, 6
halftone photoengraving, 7
halftone relief printing, 8
ink metering system, in flexo printing, 10
ISO printing technology, 5
letterpress printing, 4
line photoengraving, 7
Luther, Martin, catalyst for Reformation by,
movable type, 6
photoengraving, 7
photopolymer films, 8
development of clay type characters, 4
development of type characters by, 4
plate making, 7–8
platen press, 8
platen printing press, 8
plates, 8
powderless etching, 8
printing press, 8–9
printing presses
with four units, 9
types of, 8
printing processes, 5
Renaissance, 7
rotary printing press, 8
screen halftone plates, 7
web presses, for flexography, 10
wraparound plate, 7
Renaissance, 7, 72
Resolving power of printer, 35–38
RGB densitometer, 44
Robert’s method, 70
Robert’s method of error addition, 72
Rotary photogravure printing, 19
Rotary printing press, 8
Rotary screen printing, 29
Rotogravure, origination of term, 20
Screen halftone plates, 7
Screen printing
by hand, 28
ink transfer in, 29
Senefelder, Aloys, invention of lithography, 11
Sheet-fed press, 14, 16
Silk screen, 27
Simon, Samual, 27
Soft-tuck cartoon, 244
Solid inks, 137–138
Spatial frequency, 74
Specialty media, inkjet, 143
Spectral-based printing, desktop, 249–268
metameric systems, 249–250
spectral color management, 254–260
spectral image acquisition system, 251–254
spectral model for printers, 260–263
spectral PCS basis for device types,
traditional vs. spectral-based systems,
Spectral model for printers, 260–263
Stability characteristics, light sources, 219
Steel-die engraving, 26
Color Desktop Printer Technology
Steel engraving, 25
Surface-absorbing printing media type, 205
Surface temperature of thermal head, 201
Suspension polymerization toner, 188
Tandem color laser printer, 162
Tandem intermediate belt transfer color laser
printer, 163
Text image quality, 35–41
addressability of printer, 35–38
defined, 32
densitometer, 39–40
advantages over visual matching, 40
desktop computing, printing, 35
device-dependent models, 34
dot-matrix printers, 35
electric typewriter, 35
image density, measurement
instrumentation, 39
image quality circle, 32
image quality metrics, 32–34
image quality modeling, 34–35
image quality models, 35
impact typewriter, 35
ink density, 38–41
inkjet printer, 35
laser printer, 35
letter-quality, usage of term, 35
letter-quality image, 35
line jags, 38
monochrome densitometer, 40
National Bureau of Standards Resolution
Test Chart, 36
offset-litho quality, 35
offset-quality text, letter-quality text,
distinguished, 38
printer resolving power, 35–38
system models, 34
test patterns, 36
test procedures, 36
United States Air Force Resolution Test
Chart, 36
visual models, 34–35
Thermal head, 200–202
concentrated thermal-transfer head, 202
requirements, 200
structures, features, 200–201
temperature control, 201
Thermal inkjet, 115–125
Thermal inkjet configuration, 118
Thermal inkjet drop ejection process, 119
Thermal inkjet printhead, 123
electrical interconnect, 124
Thermal rheography, 203–204
Thermal-transfer head, 202
Thermal-transfer printers, 203–205
configuration of, 198
Thermal wax transfer, 94–95
Thermo-autochrome method, 226–232
Thermo-autochrome paper, 226–229
Thermo-autochrome printers, 228, 230–232
Thermo-autochrome system, digital color home
printer, 231
Thick-film head, 201
Thin-film head, 201
Thinkjet print cartridge, 113
Through-printing, 26–30
classification, 27–28
features of, 27
plate, 28
power-operated presses, 29
printing process, 28–30
rotary screen printing, 29
screen fabric, 27
screen printing
by hand, 28
ink transfer in, 29
silk screen, 27
Simon, Samual, 27
Yuzen-zome, traditional Japanese dying
process, 27
Tone, 54–64
Toner, 180–189
ingredients, 181–187
laser printer, 180–189
manufacturing process, 187–189
viscoelastic behavior, 185
Trichromatic theory of color vision, 41–42
Typewriter, electric, 35
United States Air Force Resolution Test Chart,
Value, 41
Variable area–variable depth plate, 22–23
Variable drop volumes, drive waveforms, 127
Viscoelastic behavior, toner, 185
color, trichromatic theory, 41–42
color gamut, 42
Visual system, 76–79
human, 42, 76–79
total color gamut, 42
Wang, Dr. An, 86
Wang Laboratories, 86
Water–ink balance, lack of, 19
Waterless lithography, 13
Waterless plates, 13
Wax melt printers, 200
melt wax, 200
sheets, 200
structure, 200
Wax melt thermal transfer, 196
Web letterpress, 15
Web offset press
advantage of, 15
blanket-to-blanket press, 15
disadvantage of, 15
drum press, 16
impression cylinder web press, 16
in-line open press, 15
keyless tower printing system, 16
types of, 15
Web presses, 14
for flexography, 10
Wraparound plate, 7
Yule–Nielsen color calibration, 60–61
Yule–Nielsen correction, 56–60
Yule–Nielsen equation, 56
Yule–Nielsen Murray–Davies equation, 56
Yuzen-zome, traditional Japanese dying
process, 27
Xerox M750
Epson 740
HP DJ970
’01 ‘98-’02 ‘00
Epson 980
Epson 900/
Lexmark Z65
HPDJ cp 1160
Canon 8200
Epson 760/860
Lexmark Z51
HP DJ2000
HP DJ722
Canon 7000
Lexmark 5700
HP Photosmart
Dot Diameter
(microns, average of horizontal and vertical)
FIGURE 4.2 Evolution of drop volume and drop size.
Inkjet Technology Performance Trends
DesignJet 10ps
N1000 ?
C F900
L J110
L Z65
E PM900C
FIGURE 4.3 Thermal inkjet configuration.
E C80
E C40
E C60
L Z53
L Z23
L Z52
L Z51
Million Drops per Second (KCYM)
Top Plate
Ink Inlet
Ink Inlet
FIGURE 4.4 Thermal inkjet drop ejection process.
Bubble Collapse
and Drop Breakoff
Bubble Nucleation
A superheated vapor
explosion occurs by
heating at 100°C/µ sec
Bubble Growth
Bubble expands
forming a drop
Bubble collapses
drawing in fresh ink
FIGURE 4.5 Top view of particle-tolerant ink supply channels.
Orifice meniscus
settles and refill
orifice plate
ink barrier
CVD SiO2 &
field oxide
NMOS transistor
Legend: (conductor)
Al...............................aluminum (conductor)
CVD...........................chemical vapor deposition
field oxide..................silicon dioxide (insulator)
ink barrier..................photoimageable thick film
orifice plate................nickel or polyimide
SiC............................ silicon carbide (insulator & anticavitation barrier)
SiO2..........................silicon dioxide (insulator & thermal barrier)
SiN............................ silicon nitrite (insulator & anticavitation barrier)
Ta.............................. tantalum (anticavitation barrier)
TaAl...........................tantalum aluminum (resistor)
FIGURE 4.7 An HP thermal inkjet printhead and electrical interconnect.
Polyimide Tape
Silicon Substrate (see-through tape)
Gold-Plated Electrical
Interconnect Pads
FIGURE 4.8 Epson MLP piezo inkjet drop generator.
Metal plates
Zirconia ink
FIGURE 4.10 Drive waveforms for variable drop volumes.
FIGURE 4.13 Image processing pipeline.
FIGURE 4.16 Gamut of CMY and CMYGB printers.
Development Unit
Laser ROS
Transfer Unit
Paper Stack
FIGURE 5.4 A schematic diagram of multi-path intermediate belt transfer color laser
Transfer belt
1st Transfer Roller
Developing Unit
Laser ROS
Paper Stack
FIGURE 5.5 A schematic diagram of tandem intermediate belt transfer color laser printer.
FIGURE 7.7 Film scanner unit of a Frontier 330 (area CCD).
FIGURE 7.10 Automatic scratch and dust restoration function.
Without pixel shifting
(1448 × 2172 pixels)
With pixel shifting
done once, scans twice
(2048 × 3072 pixels)
FIGURE 7.11 Pixel-shifting method.
With pixel shifting done
three times, scans four times
(2096 × 4344 pixels)
Photosensitive Material
F θ lens
B-SHG Laser
R Semiconductor
G-SHG Laser
FIGURE 7.13 Optical system for laser exposure.
Finished copy
Light Source
(1) Exposure
Piatro film
(3) Thermal
(2) Moistened with water
(4) Peel-off
Used film
(to be discarded)
FIGURE 7.14 Conceptual diagram of pictrography.
FIGURE 7.18 Basic structure of TA paper.
FIGURE 7.25 Basic configuration of a high-speed, three-head tandem digital color
FIGURE 9.2 Garrett Johnson’s “Metameric Cows” [Johnson, 1998]. This demonstration
simulates how the same reflectance properties under a single light source can appear
different to different observers. The front half and the back half of the cow are different
reflectances. To the two-degree observer (left), the front and the back have the same color.
To the ten-degree observer (right), the front and the back have different colors.
FIGURE 9.5 Spectral color management would provide a way to impose a new light
source on the image of a captured object.
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF