Computer Music Journal Article Template

Computer Music Journal Article Template
The Chameleon Guitar—Guitar with a Replaceable
Resonator
The MIT Faculty has made this article openly available. Please share
how this access benefits you. Your story matters.
Citation
Zoran, Amit, and Joseph A. Paradiso. “The Chameleon
Guitar—Guitar with a Replaceable Resonator.” Journal of New
Music Research 40.1 (2011): 59–74.
As Published
http://dx.doi.org/10.1080/09298215.2011.554609
Publisher
Informa UK (Taylor & Francis)
Version
Author's final manuscript
Accessed
Sat Sep 16 19:35:01 EDT 2017
Citable Link
http://hdl.handle.net/1721.1/73201
Terms of Use
Creative Commons Attribution-Noncommercial-Share Alike 3.0
Detailed Terms
http://creativecommons.org/licenses/by-nc-sa/3.0/
Amit Zoran et al.
1
The Chameleon Guitar - Guitar with a Replaceable
Resonator
Amit Zoran, PhD student in the program for Media Arts and Sciences
MIT Media Lab, 75 Amherst st., Cambridge MA, 02142 USA
amitz@mit.edu
Dr. Joseph A. Paradiso, associate professor of Media Arts and Sciences
MIT Media Lab, 75 Amherst st., Cambridge MA, 02142 USA
joep@media.mit.edu
Abstract
Each acoustic instrument is one of a kind. Its unique acoustic properties are
transferred from the physical characteristics of its source materials and a
handcrafted touch. In contrast, electronic and computer-based instruments lack this
distinguishing trait. Though the technology support in music instruments offers
great flexibility, it tends to foster predictable and generic results, particularly with
common use of easily-cloned digital presets. This paper presents a new approach to
the design and fabrication of instruments that combine the advantage of acoustic
and electric instruments -hybrid instruments- that exist simultaneously in both
physical and digital environments. This approach exploits physical/acoustic
properties via a replaceable physical object complemented by a simulated shape or
other digital signal manipulation. The key concepts of this approach are presented
through an example: The Chameleon Guitar, detailed in this paper along with
evaluation from musicians and instrument makers. This work aims to demonstrate
the possibility of maintaining the qualities found in real acoustic instruments, such
New Music Research Journal
Amit Zoran et al.
2
as unique spectral and spatial behavior of wooden soundboards, with the flexibility
of digital processing.
1. Introduction
In this paper a new approach to the design of string instruments is presented,
combining digital and the physical environments though allowing the player to
seamlessly, simply, and simultaneously change both the instrument’s acoustic
resonator (a replaceable acoustic insert that function similar to acoustic guitar’s
sound board), and digital signal processing characteristics. The main goal of the
work is to merge traditional values and digital capabilities, while preserving the
resonator’s both spectral and spatial contribution to the overall timbre. This
perspective is illustrated in the implementation of The Chameleon Guitar (see Figure
1). A deeper discussion about this project, along with the state of the guitar today,
our conceptual motivation and influence on our work from other fields is presented
in (Zoran 2009). Sound examples and videos of The Chameleon Guitar are presented
in www.thechameleonguitar.com.
[Figure 1 placeholder - look at the end of the document]
In acoustic musical instruments, natural information embedded in wood can be
extremely significant to the functionality of the instrument. Traditionally, the
materials and craft qualities of acoustic instruments play an important role in
defining the instrument’s unique sound. It is difficult to find two acoustic
instruments that sound and perform exactly the same, which leads to a strong
personal connection and often a deep bond between the player and their instrument.
At the same time, electronics are playing a huge and still growing role in creating
and processing the instrument’s sounds, due to the flexibility analog and digital
processing techniques provide for sound control.
New Music Research Journal
Amit Zoran et al.
3
An acoustic guitar owes its sound quality primarily to its wooden chamber. The
timbre and volume of a guitar depends on the shape of its chamber and the
structure and properties of its material. The type of wood, its quality, the way it is
prepared and its inhomogeneous structure all create a reality where no two guitars
are the same; each guitar acts and sounds a little different. Wood can also change its
acoustic behavior over time and in different moisture conditions.
We implemented a new guitar that combines physical acoustic properties with
digital processing abilities in an innovative design - a design that benefits from the
distributed spatial-acoustic characteristics of an acoustic soundboard (unlike just
sampling the surface vibration at a single location). The concept of The Chameleon
Guitar is to separate the chamber’s shape from its material and craft quality. A
physical resonator, a replaceable piece of matter that gives the guitar a
distinguishing acoustic behavior, is situated under the guitar bridge. The Chameleon
Guitar allows the user to change the acoustic resonator without swapping the whole
instrument (and requiring just slightly re-tuning), and the array of soundboard
transducers enables a higher degree of information to be processed in the computer,
relative to the typical pair or triad of magnetic string pickups or single contact
pickup in common use on guitars today. Through this novel modular approach,
sound flexibility and a high level of resonator personality are achieved.
2. Related Work
Since the dawn of the synthesizer, significant effort was devoted to embed
synthesizer capabilities into the guitar (Paradiso 1998). Starting with envelope
followers driving active filters and other effects (Thompson 1997; Hughes 2004)
and continuing on to analog (then digital) pitch extraction that then can manipulate
an entirely synthesized sound source, musicians and engineers tried to merge the
world’s most popular instrument with state-of-the-art technologies. Examples of
this abound, coming from inventors and industry (see below), artists (ranging from
New Music Research Journal
Amit Zoran et al.
4
Derek Bailey to Fred Frith, for example) and academics (e.g., Lähdeoja 2008;
Vanegas 2007) – most high-end academic research has focused on bowed
instruments as opposed to guitars (Machover 1992; Jehan et al. 2005; Bevilacqua
2006).
One way to achieve sonic flexibility while preserving some degree of
expressivity is to first detect pitch and amplitude envelope of the acoustic signal,
then applying synthetic timbres. In this way, an array of timbral possibilities is
achieved via synthesized sound, and the sensitivity of the instrument is being
preserved through the amplitude and pitch channels. More sophisticated methods,
based on articulation detection, can be used to expressively and dynamically control
the timbre. In order to achieve this via audio analysis along, high-level signal
processing capabilities (and sometimes even artificial intelligence tools) are
required. The most complicated part of the process is to model and extract the
instrument’s dynamic transient behavior (at low latency) while preserving nuances
in its expression and perhaps some aspects of its unique sound signature.
Guitar synthesizers from the early 70’s attempted this through analog signal
processing or hardwired digital processing (e.g., the Arp Avatar or 360 Systems
products), often using a separate set of processing electronics for each string. These
devices were often unreliable, or required particular technique to play well. When
MIDI first met the guitar in the early 80’s, an easier approach evolved where the
guitar controllers sometimes did not even include strings (e.g., the Ztar controllers),
or used the strings only as sensors for fingers and to determine fret position (e.g., the
Yamaha G-10, Beetle Quantar, and devices mentioned below). The only similarity to
the guitar was the way it was held, and sometimes the way it was fingered and
perhaps plucked, but, although some interesting channels of articulation were
invented, those instruments lacked much in the way of expressivity especially when
compared to what guitars are capable of.
New Music Research Journal
Amit Zoran et al.
5
One popular example, the SynthAxe, invented by Bill Aitken (SynthAxe website),
supported two sets of strings; one set, made from short-length strings running
across the guitar’s body, was used to detect picking, and another set ran down the
fret board to determine pitch (lower cost controllers along similar lines were
introduced by Casio and Suzuki). Zeta Music also made interesting hybrid guitars in
their Mirror series (Paradiso 1998) with a multimodal interface that featured a wired
fret board for pitch detection, a capacitive touch detector on each string for
determining the expected acoustic damping, hexphonic pickups for amplitude
detection and pitch bend, accelerometers for measuring the instrument's rigid-body
dynamics, and an instrumented whammy bar (and more).
In recent years, as signal processing capabilities have improved, there has been a
shift away from the dedicated MIDI guitar controllers (described above) and back
toward existing, standard electric guitar interfaces that identify playing features and
dynamics by running real-time DSP algorithms on the guitar’s audio stream, still
generally exploiting hexaphonic pickups that derive separate audio from each
string. The Line 6 Variax guitar, for example, maps the guitar player’s input onto a
variety of preset sounds (Line 6 website), from classic acoustic and electric tones to
sitar and banjo. It allows the player to plug into a computer and customize a chosen
timbre, while the hexphonic piezoelectric pickup, located on the bridge, transfers the
signal to a DSP unit located on the guitar. Expressive playing and sound flexibility
are enhanced with these digital guitars. Another example is Fender’s VG Stratocaster,
a hybrid electric and digital guitar (Fender website). The Gibson Robot Guitar series
also uses a DSP unit on the guitar to control the automatic string tuning mechanism
(Gibson website). Modern high-end electric guitars often come equipped with a
connector to transfer multichannel digital audio directly from the guitar to a
computer network or dedicated processing electronics.
New Music Research Journal
Amit Zoran et al.
6
Some artisans still develop in the acoustic realm – for example, Ulrich Teuffel is a
German designer who produces unique guitars. His Birdfish model is an electric
guitar that allows the player to replace wooden supports (Teuffel website). The
guitar has two metal structures connected by solid wood panels. By replacing the
wood, the damping properties of the guitar change and modify the sound.
The haptic feedback from the musical instrument, as well as the tactile qualities
of the experience, was the focus of many projects. Several projects applied a similar
concept in musical instrument design, such as in The Sound of Touch (Merrill 2007),
or with the Cicada’s Rapid Sequential Buckling Mechanism (Smyth 2002). In the
work of Cadoz et al. (2003), a vibrating device with sensing forces and
displacements at its manipulation stick was able to produce a force-feedback, and
allowed to highlight an inter-sensory phenomenon. Howard et al. (2003) describe a
physical modeling music synthesis system known that enables virtual instruments
to be controlled in real-time via a force-feedback joystick and a force-feedback
mouse.
3. Motivation and Vision
The Chameleon Guitar design preserves the unique properties of the wood used to
craft guitars, yet through its modular construction, also offers an instrument that
musicians can use to customize and modify the guitar’s intrinsic timbre and acoustic
“personality.” Traditionally, acoustic guitars cannot be modified once they are
made; it is not part of the player’s experience to “tamper with” the structure of the
instrument. Acoustic guitars are highly crafted and offer acoustic integrity, but they
offer no flexibility for sound design control. An exception is perhaps seen in the
work of certain artists who modify and extend their guitars (e.g., Derek Bailey, Fred
Frith, Elliot Sharp, Paolo Tofani), but these tend to be one-off creations that stay in a
particular timbral space.
New Music Research Journal
Amit Zoran et al.
7
Under our concept, a musician can still be involved in creating and modifying
the guitar’s acoustic timbre, as well as sharing designs with the guitar community,
using online blogs and websites for example. This combines the values of an
electrical (analog or digital) guitar with the uniqueness of a wooden acoustic guitar’s
tone. By doing so, we can achieve expressive playability through a unique tool that
also enables the player to design a desired sound under a coordinated acoustic and
signal processing approach.
The design of The Chameleon Guitar focuses on the influences of the chamber on
the guitar’s sound. The chamber’s main acoustically-relevant parameters are its
shape and material. As different chambers can be separately identified when they
are inserted into the guitar body, the digital processor will know what chamber is
installed, and a customized suite of digital processing effects can be presented that
are well suited to the current resonator – in some sense, we can consider this to be a
digital part of the particular physical chamber that is being played. Accordingly, we
have realized an interchangeable “hybrid chamber”: part of it is physical (the
guitar’s resonator) and part of it is virtual (the span of digital effects and processing
that are suited to that resonator).
All resonators are small soundboards with an arch-top guitar bridge. The strings
are tied to a conventional tailpiece. The resonator can be very easily swapped by
opening an aluminum tray in the back of the guitar. The resonators have four
piezoelectric sensors located in different places on its surface to capture different
mixtures and phases of the eigenmodes. The signal-processing unit is located at the
back of the guitar; it merges the signals into one (or, for example, a stereo pair) and
acoustically compensates the output to imitate the sound of a full acoustic guitar of
an average size, using filter banks. The signal processor is also capable of
introducing a variety of other effects, such as phase modulation, frequency
modulation, and distortion. Several different kinds of resonators were made and
New Music Research Journal
Amit Zoran et al.
8
experimented with in order to explore their difference - a change in resonator
structure leads immediately to a change in output timbre.
The Chameleon Guitar presents a three-element instrument: the body, the
resonator and the digital signal-processing (DSP) unit. The body is essentially a
skeleton of a standard guitar that holds the two other elements; it is the guitar’s
player interface. Underneath the guitar interface, there are two customizable parts:
the programmable DSP and the replaceable resonator.
By combining digital with physical in this fashion, we believe we can merge both
values. The replaceable resonator can play an important role in continuing the
traditional connection between players and their unique instruments while at the
same time allowing a coarse span of intrinsic timbral characteristics; the digital part
can be finely controlled, and augments the acoustic experience within the virtual
domain.
4. Technical Background
In his thesis work, Ra Inta explained and analyzed the physics of the acoustic
guitar (Inta 2007), and Fletcher and Rossing gave a classic physical overview on the
subject, including acoustic background (Fletcher and Rossing 1990).
The main parts of the modern flat-top acoustic guitar are the neck and the body,
which contains an air cavity. The most important part of the body is the soundboard
(the top plate), usually made from spruce or cedar wood, in which a round hole is
placed just near the end of the fret board. The neck itself is usually made of a harder
(but flexible) wood that resists the string tension. The guitar strings are tight
between the bridge and the tuners on the headstock (the end of the neck). When the
strings vibrate, they transfer vibrations via the bridge to the soundboard, which then
starts to sympathetically vibrate. The soundboard in turn drives the air, which is the
sound we hear. Although all the guitar parts vibrate, and there really isn’t a fixed
part in the guitar, the sound we hear is mostly emitted from the soundboard and the
New Music Research Journal
Amit Zoran et al.
9
chamber’s hole. The shape and dimensions of the soundboard, chamber and hole, as
well as the soundboard material, are the most significant elements to affect the
sound. The back and sides of the chamber are usually made from hard woods, in
order to prevent acoustic energy from being damped there.
In the arch-top guitar family, unlike the flat-top, the soundboard (top plate) is
curved and usually carved from a bigger wooden block, similar to the violin. The
strings are tied to the head of the guitar and the tailpiece, and they are pushing the
bridge (located in the top arch of the curved soundboard) down into the
soundboard.
When a string vibrates, it excites the bridge. The bridge then drives the
soundboard. However, not all the harmonics are actually transferred to the
soundboard; this depends on the eigenmodes (modes of vibration) of the guitar’s
body, which is a coupled vibrating element constructed from the soundboard itself
(similar to a plate behavior) and the Helmholtz resonator made by the hole.
The tension in the strings tends to rotate the bridge and deform the soundboard.
To make the instrument relatively loud, the soundboard must be thin yet span a
relatively large area, hence a structural reinforcement is required to balance the force
introduced from the string tension. This is usually in the form of wooden braces
behind the soundboard. This bracing system plays an important role in sound
production. By adding more mass to the soundboard, as well as adding stiffness to
specific locations on the plate, the braces influence the eigenmodes (natural
frequencies and spatial patterns). The art of bracing the top plate of the guitar is
highly important to the sound of the guitar - different guitar styles usually have
different bracing designs, yielding varying acoustic characteristics. The expertise of
a guitar-maker depends on his ability to control the sound through delicate bracing
craftsmanship (Siminoff 2007).
New Music Research Journal
Amit Zoran et al.
10
The low frequencies of the guitar depend on the guitar’s chamber; the Helmholtz
resonance and the soundboard size are critical to the first and second modes of
vibration (Fletcher and Rossing 1990). The soundboard’s material qualities and
braces are usually responsible for the midrange and higher eigenmodes. One of the
important jobs of the luthier is to select and prepare the wood, especially for the
soundboard.
All of the above suggests that guitar design is highly dependent upon a luthier’s
design and craftsmanship (Parker 2009; Coppiardi 2009). From the player’s
perspective, the guitar is a highly expressive instrument that can be controlled by
using different excitation methods, e.g., using a pick or fingers to pluck the strings,
plucking them at different locations, damping or bending strings with fretting
fingers, etc. Friction and mechanical properties of the finger or pick, as well as the
plucking direction, are most influential on the interaction between the string and the
guitar body. The strings or the body can be excited or damped in many other ways,
giving rise to a large range in native timbre and a multitude of playing techniques.
Together with the structural-acoustic characteristics of the wood, the player
maintains a unique relationship with the instrument that contributes to the musical
style that it fosters and sound being created.
5. Fabrication and Implementation
The Chameleon Guitar merges acoustic qualities and digital processing into a new
guitar platform. This new platform was designed in several stages that will be
presented here. The Chameleon Guitar’s signal path was the first element of the
project to be defined (see Figure 2). The string vibrates the resonator’s bridge,
similar to an arch-top guitar, and then the bridge drives the soundboard. Unlike a
normal acoustic guitar, here the soundboard is too small to drive loud acoustic
waves (especially at low frequencies; for instance, the modes influenced by the
Helmholtz resonance in an acoustic guitar). An array of sensors captures the
New Music Research Journal
Amit Zoran et al.
11
structural-acoustic behavior of the resonator (the multipoint displacement of its
surface), while a computer simulation processes the signal through a transform
derived from a programmable (virtual) shape, or any other digital sound effect: four
piezoelectric sensors, located in different places on the resonator, capture different
combination of the eigenmodes. The four sensors’ signals are buffered and amplified
by the analog circuit on the resonator PCB (printed circuit board), and then, using an
electronic connection, transferred to a digital signal-processing unit (SP unit) located
on the guitar. The signal output is then re-assembled from the four inputs, imitating
different guitar chamber sizes along the way (by changing the properties of the filter
bank presented in section 5.6), or implementing other appropriate digital audio
effects. The resultant signal is transferred to the output jack, and then sent to an
acoustic guitar amplifier. After a preliminary prototype was built to validate the
basic concepts and the signal path was defined, the digital version of The Chameleon
Guitar was designed.
[Figure 2 placeholder]
5.1. The Reference Guitar and Impulse Response Tests
A Yamaha FG330 acoustic guitar was used as a reference. The actual timbre of an
acoustic guitar depends on the acoustic properties of the surrounding environment.
The reference acoustic guitar was recorded with a single MXL USB.008 microphone
in a recording studio room, located 50cm in front of the guitar bridge, while the
instrument was placed on the floor of the studio and damped with soft foam.
A linear system’s behavior can be analyzed by its response to an impulse input.
Although a guitar is not a linear system, its behavior at low amplitudes is very
similar to one. Inta used the method of impulse response to analyze and model a
guitar’s behavior (Inta 2007). In this research, impulse response was imitated by
New Music Research Journal
Amit Zoran et al.
12
hitting the top-center of the guitar’s bridge with a plastic-coated, hand-held metal
stick. Several iterations were recorded, and the most similar three recordings were
averaged to create the system response.
5.2. Acoustic Principles and Resonator Design
The goal of the basic signal processing that we perform is to modify the
resonator’s sound so the output will sound like a full size guitar without damaging
its intrinsic audio features; the output sound will still represent the resonator’s main
acoustic properties that relate to the material and the fashion from which it is made,
while the properties that relate primarily to the vibrating system’s shape (the
chamber’s dimensions) will be modified.
An acoustic guitar’s behavior depends on its shape and material properties.
However, at lower frequencies, when the frequency is much lower than the rate of
changes in materials (changes in stiffness, density, supportive braces, etc.), these
properties can be calculated as an average. In other words, the lower eigenmodes,
depend more on shape than on wood patterns, especially those influenced by the
Helmholtz resonance, (usually around 100Hz) and the one related to the
soundboard’s lowest eigenmode (usually around 200Hz, mostly dependent on
soundboard size). However, as the frequencies get higher, their dependence on
material pattern and brace structure becomes more significant.
Based on the above, The Chameleon Guitar’s intrinsic processing should modify
the lower formants (based on the lower resonator’s eigenmodes) and keep the
higher ones as natural as possible, in order to achieve the design principle of
preserving the wood’s authenticity but modeling the output signal to sound like an
acoustic guitar (with a controllable chamber size). In general, it’s difficult to generate
reliable, digitally modeled string attacks (transients) were one to try and digitally
emulate the full guitar sound from samples of a “plain vanilla” string lacking
New Music Research Journal
Amit Zoran et al.
13
acoustic personality. In our framework, the mid-range and high frequencies
transient behaviors are preserved so the transients’ sound signatures are kept as
natural as possible.
Each wood resonator has a different acoustic behavior. The sensor locations and
DSP algorithm were defined according to the simulated and measured behavior of a
reference resonator (resonator no. 1, see Figure. 9); all other resonators used these
properties.
Several researchers have suggested the use of FEM (Finite Element Methods) for
musical instrument design and analysis (Inta 2007). The resonator’s shape was
designed in an iterative process using FEM (Comsol Multiphysics software), physical
acoustics tests, and mechanical adjustments (see Figure 3 and Figure 4). The goal
was to find the bridge location, boundary conditions, and smallest surface area that
can support vibrations that are as close as possible to the lower eigenmode of the
acoustic guitar (around 100Hz). The design process proceeded in several iterations,
started with four rigid points on the boundaries, and converged on a pseudo
optimal shape, with just three simple supportive points.
[Figure 3 placeholder]
The number of sensors and their locations were defined next. FEM simulation
was used to analyze the first 15 eigenmodes for a 2.5mm Sitka spruce plate with
boundary and support as described above. The eigenmodes were plotted and
analyzed graphically. Piezoelectric contact sensors are good for sensing surface
vibration; however, they are much more sensitive to surface bending (derivative of
the displacement) than orthogonal motion. The gradient maps of each eigenmode
were analyzed and clustered in four groups, according to their spatial patterns. The
five lower eigenmodes were given priority in defining the sensor locations, in order
New Music Research Journal
Amit Zoran et al.
14
to guarantee better coverage at low frequencies – locations were picked that had the
highest bending displacement. Our goal was to minimize the number of sensors to
be used: as such, we tried to define how few sensors were needed to capture the
highest signal to noise ratio (SNR) from all 15 eigenmodes. At the end of this
process, four sensors locations were chosen.
[Figure 4 placeholder]
5.3. Design of the Guitar Body
The resonators define constraints on the body design. The guitar’s body should
be able to embed the resonator inside it, yet still be strong and ergonomic. Several
elements are important for guitar ergonomics: weight, stability, body size, thickness
and string tension. The strings’ sustain, which needs to be as high as possible,
depends on their tension. Most of the references for this Section are based on
interviews with instrument-makers (Coppiardi 2009; Parker 2009) and useful
references for acoustic and electric guitar-making (Hiscock 1986; Kinkead 2004).
The Chameleon Guitar defines a new guitar family and could be implemented
using any guitar as its interface: e.g., classical guitar with nylon strings, acoustic
guitar with steel strings, electric guitar with nickel strings, and others. This project is
focused on evolving the guitar to a new stage; therefore we decided to base its
interface on the most popular guitar type, the electric guitar.
An electric guitar has better sustain than acoustic guitars; the solid-body
minimizes the strings’ energy loss at the bridge. The Chameleon Guitar, however, does
not have a solid body. Therefore, a long neck scale was chosen in order to maximize
the strings’ static tension for a given note, thus maximizing the guitar’s sustain. On
the other hand, a long neck scale with high tension can cause problems, such as a
resistance to bending. Making the non-vibrating parts of the strings longer can
New Music Research Journal
Amit Zoran et al.
15
minimize those problems. These factors influenced the design of the neck’s head and
the tailpiece location. A single-cut body was selected, due to the resonator’s shape
constraints on the body.
[Figure 5 placeholder]
Several 2D and 3D sketches were made in an iterative design process, before
defining the final shape (see Figure 5 and Figure 6). A 3D model of the guitar was
then built in Rhino3D. Design efforts were made to make the aesthetic look of The
Chameleon Guitar compatible with that of the current popular electric guitar family
without making it too similar to existing models, while, at the same time,
encouraging viewers to focus on the resonator. A turquoise color was chosen to
contrast with the warm resonator colors, which were expected to be brown wood
tones.
After designing the guitar outline, more detailed designs were needed: designs for
the mechanism to easily replace a resonator, selecting tuner types (Steinberger
Gearless Tuners), the tailpiece (Gotoh 510 Tailpiece) and materials for the guitar. The
mechanism for the resonator replacement, called the resonator tray, was designed in
Rhino3D. A key driver in its design was to allow a user to swap resonators in less
than ten seconds. Aluminum and delrin have low friction coefficients, which make
them good candidates for sliding elements. The tray itself was designed from
aluminum, while the rails and the lockers were made from delrin.
[Figure 6 placeholder]
The last design stage was to choose the woods for the guitar parts. Mahogany,
which is better than maple for vibrating at low frequencies (Mahogany and Maple
New Music Research Journal
Amit Zoran et al.
16
are the most popular woods for electric guitar necks), was chosen for the neck.
Poplar was chosen for the main body frame: it is light, easy to work with and shares
acoustic properties with mahogany. A carbon fiber structure is located inside the
poplar body frame to add stiffness.
5.4. Electronics
The signal starts its path with the piezoelectric sensors, amplified in the resonator
PCB analog circuit and digitally processed in the SP unit. The sensors are standard
ceramic piezoelectric disks (common for musical uses) with a resonant peak at
7000Hz (± 600Hz), 9.9mm diameter and 0.12 mm thickness. A small disk size was
preferred, in order to minimize the affected resonator surface. Voltage fluctuations
(the sensed signal) develop on the sensors when a pressure field is applied, and are
transmitted to the resonator PCB with thin coaxial wires. The ceramic disks were
hard-fixed to the resonators with standard adhesive (more detail below).
The resonator PCB circuit was designed for five channels. Here, we only used four
of them and left an extra channel for more design flexibility. In order to change the
piezoelectric voltage signals from high to low impedance, they need to be buffered
with a non-inverting op-amp. The piezoelectric sensors have a series electric
capacitance that can implement a high-pass filter with a load resistor. The lowest
standard note of the guitar is the E note (80Hz). The sensor capacitance, 10nF, as
calculated from piezoelectric equation (Fraden 2003) gives a filter cutoff (f3db) of 32
Hz when the shunt resistor value is 0.5M.
[Figure 7 placeholder]
This guarantees that the guitar’s relevant spectrum will not be suppressed. The firstorder transform function of the filter is less relevant here, and will be taken into
consideration in the DSP algorithm. The signals are biased to 2.5V, and the PCB
New Music Research Journal
Amit Zoran et al.
17
outputs are all line level signals. The PCB is powered from the SP unit with a LED
power indicator, and grounded by the resonator tray, so that all the aluminum parts
are grounded. A special socket connector is built into the resonator tray; the resonator
PCB slides into it when a resonator is inserted to the guitar. In principle, each
resonator can host a different set of electronics if desired, and it is trivial for each
resonator to be uniquely identified via an ID on its PCB, allowing for a different
suite of downstream signal processing options and effects to be enabled.
The SP unit that was chosen is the Freescale’s Symphony™ SoundBite Development
Kit, hosting a Freescale Symphony™ DSP56371 (192 MHz, 24-bit fixed point processor
- see Figure 7). The unit has 8 audio line level inputs and outputs. The unit’s
sampling rate is 48Khz, with 16-bit quantization levels. The analog to digital sampler
includes an anti-aliasing filter.
Four line level inputs are transferred from the resonator tray connector to the
inputs of the SP unit by coaxial wires. After digital processing, the output signal is
transferred from a line level to an instrument level (in other words, a low impedance
signal changed to a high impedance one compatible with standard guitar amplifiers)
and then to the guitar output jack. The SP unit can also produce multichannel audio
outputs.
5.5. Fabrication
The fabrication of The Chameleon Guitar body was done in two stages. Digital
fabrication, based on computer modeling of the guitar, and hand-made modification
(sanding, gluing, adjustment and varnishing). The guitar’s neck and body frame
were milled separately using a milling machine. The carbon fiber structure was
glued inside the body with epoxy. The fretboard was made by hand; adjusting its
dimensions, then inserting, trimming and sanding the fret wires. Then the neck was
adjusted and glued to the body with epoxy. The guitar was sanded, varnished and
New Music Research Journal
Amit Zoran et al.
18
polished, then the tailpiece, output jack, and tuners were assembled. The resonator
tray, which was made in a milling process, was the last element to be assembled.
5.6. Signal Processing
5.6.1. General
The signal-processing algorithms that we used in our work were developed and
tested using Matlab and implemented on the above-mentioned SP unit in C code,
using the Freescale’s Symphony™ Studio Development Tools and based on Freescale’s
Eight-channel-C-template C code software (48KHz 16-bit, one sampling cycle latency).
The development tools included DSP memory and device mapping, as well as
analog-to-digital convertor and digital-to-analog convertor drivers.
The main goal of the DSP algorithm that we designed is to implement a virtual
chamber based on the physical resonator, i.e., to manipulate at least one resonator’s
signals (resonator no. 1) and re-construct them to minimize the difference D between
The Chameleon Guitar’s output impulse responsei signal (captured by a microphone,
20 cm in front of an Acoustic AG15 15W 1x8 Guitar Combo Amplifier) and the reference
guitar impulse response (see section 5.1):
.
This is an equalization problem, finding the correct frequencies and amplitudes
of band pass filters. The reference guitar’s signal is sr, each sensor’s signal is si (i is
signal’s index, from 1 to 4), the band coefficient per signal channel is cij, and the IIR
filters are represented by hi. The minimization of D was achieved here through an
experimental, brute-force iterative process as described below. The D value can be
minimized by a proper filter bank design (hj values) and the choice of correct
i
Impulse Response here is as described in section 5.1.
New Music Research Journal
Amit Zoran et al.
19
coefficients (cij). First, each of the raw sensor signals (after sampling) is processed
through a filter bank with its bands tuned according to the reference guitar’s
formants: for each band, the filter cut-offs were tuned (by eye, based on a Matlab
graph) to fit the reference acoustic guitar formants, and cij was tuned to fit the
formant amplitude. For the minimization of D, for each band just the ideal si was
chosen - the best cij candidates were selected, and for the rest were tuned to zero.
The amplitude and decay rates of each band were scaled in order to best achieve the
required reference level. However, when more than one sensor signal (si) produced
a good candidate for a specific band, the one with the higher SNR was chosen. After
tuning the signal-processing algorithm to minimize D, adjusting it to a sound like a
smaller or bigger guitar chamber was relatively easy.
As the acoustic waves in the guitar approach its resonance modes, the decay
rates at the corresponding frequencies are slower. An infinite impulse response (IIR)
filter can imitate such a behavior coherently; the distance of the filter’s poles from its
region of convergence (ROC) tunes the resonance behavior of the IIR. The IIR can
add a slower decay rate to the transferred band, i.e., by tuning the filter bank’s IIR
coefficients, we can fit artificial reverberation to selected bands. Our filter bank was
implemented by a Second Order Section Direct Form II filter (see Figure 8). The filter
bank implementation is simple, and is based on summing all of the bands in time
domain (while ignoring phases).
[Figure 8 placeholder]
The impulse response of resonator no. 1 was used for tuning the filter banks. The
IIR coefficients were optimized in Matlab’s FDAtool using a brute-force manual
process. This Matlab system required 14 bands and mainly processed resonance
modes below 1KHz. It was implemented on the guitar with fewer bands (starting at
New Music Research Journal
Amit Zoran et al.
20
seven and leading down to four). In practice, the resonators projected an acoustic
sound that could not be ignored, which was mixed with the processed sound.
Morphing between the guitar’s acoustic sound (attenuating directly from the
resonator) and the SP unit output (after amplification) tends to give interesting
overall results: the sound in the recording studio has a stereo quality, depending on
the positioning of the guitar and the amplifier in the room. Overall, we can say that
the guitar digital processing contributes mainly to the lower modes, and the sound
reflected directly from the resonator contribute more to the middle and high
frequency range (see Figure 9).
[Figure 9 placeholder]
5.6.2. Alternative Algorithms
The algorithm described above is a suggested implementation of a virtual
chamber. However, the use of a DSP enables implementation of a variety of sound
processing techniques and synthesized algorithms. Lazzarini, Timoney and Lysaght
for example, describe adaptive frequency modulation synthesis based on an acoustic
oscillator instead of an electrical oscillator (Lazzarini, Timoney and Lysaght 2008).
Julius O. Smith III in Physical Audio Signal Processing describes many different sound
effects that can be implemented digitally, such as virtual distortion (Smith 2008).
The Chameleon Guitar resonator has four authentic channels. All of them represent
the same acoustic event, but each has a different timbre, as the pickup responsible is
at a different location. Each of these signals is a different superposition of the
resonator’s eigenmodes (with phase difference that can be ignored for low
frequencies). The distance between the signals can be represented as Dnm:
.
New Music Research Journal
Amit Zoran et al.
21
Dnm is highly dependent on the medium; it is a unique property of the resonator,
which contains rich acoustic information that can be used to synthesize or control
the sound; for example, using the output of one sensor to manipulate another
sensor’s signal.
A suggestion for such processing is described in this relation:
S = s4 * sin( b1 * IIR( s2, Hd14 ) + b2 * s3 ) + b3 * s2;
This is a combination of phase modulation with a distortion effect, depending on
the bi coefficient values. The amplitude of s4, which is low-frequency oriented,
controls the clipping of the non-linear part of the formula (inside the sinusoid). At
the same time, s2, which is mid frequency oriented, provides a natural signal. In case
of a resonator that allows easy, non-coupled manipulation of s4, such as in resonator
6, which is divided to two plates that are connected with a 4x6x50 mm wooden bar
in the backside (see Figure 9.6 - the small plate, which is flexible, and easily
manipulated contains s4), an interesting overall effect is given. As mentioned earlier,
the signal-processing suite accompanying each resonator can easily be customized to
the sonic palette that the resonator suggests.
5.7. Resonators
The resonators’ designs were a long process of trial and error. All resonators
have four piezoelectric sensors located in the same place. The first four resonators
are more conservative; all of them include wooden soundboards supported by
braces and a glued bridge, varying only in their structure and materials (see Figure
10 for a detailed description). The last four resonators test different ideas –
embedding springs, an ABS plastic chamber, screws or complex boundaries and
connections. Different players have tested all of these resonators.
3D models of the soundboard and the bridge were built in Rhino 3D. Wooden
blocks were prepared, sometimes by gluing two pieces to make a joint block, where
the wood-cuts and grain direction were selected in a traditional way. Then, the
New Music Research Journal
Amit Zoran et al.
22
resonator’s shape was milled using Shopbot CNC machine and cut with a Universal
Laser Cut machine. The bridges were made in a similar way, and glued with epoxy
to the resonators after location adjustment.
All the resonators were hand-finished, first sanded or trimmed with a scraperii,
then varnished using different techniques for protection and aesthetics. The resonator
PCBs were glued to the resonator with epoxy. The sensors were glued with special
ethyl cyanoacrylate adhesive, and were protected with a thin balsa ring. Coaxial wires
connected the sensors with the resonator PCB, sometimes guided by small plastic
elements. All the resonators have plastic or wood bindings at their edges to protect
them from damage.
[Figure 10 placeholder]
6. Evaluation
Fifteen guitar players took part in the evaluation of the project. The players
varied in their usual playing time, from 0.1 hour to 17 hours per week. The average
weekly playing time was 4.3 hours. The players were asked about their favorite
music style and the type of guitars they normally use. For our study, each of the
players used the guitar for an hour in an acoustic recording studio.
After introducing the guitar’s concept and its technical aspects, the participants
were asked to play the guitar and use all of its eight resonators, for about seven
minutes per resonator. The participants tried three different digital processing
options: big chamber digital processing, small chamber digital processing (section
5.6.1) and a digital effect (section 5.6.2). Each participant played in his or her
preferred musical style. The participants were asked to replace at least one resonator
ii
A sharp steel plate, used for trimming wood surfaces.
New Music Research Journal
Amit Zoran et al.
23
by themselves and to examine the tangible qualities of the resonators (such as
knocking and scratching the resonators’ surface). The evaluation results are
summarized here – more detail is presented in (Zoran 2009).
6.1. Concept and Contribution
The first question the participants were asked was if they would like to have The
Chameleon Guitar. Answers were rated on a scale from one to seven (one represents
“no” and seven represents “definitely”). The average response was 6.33, with
standard deviation of 1.18, which is a very positive answer. Then the participants
needed to select the most important properties of The Chameleon Guitar, the role that
best fits the guitar, and the musical style that best fits the guitar.
The participants were asked about the contribution of The Chameleon Guitar to the
guitar and the field of music in general. The dominant answers were the guitar’s
new acoustic sound possibilities and new expressive ways to play the guitar
(together with new sound effects). Several players mentioned the new craft and
hobby options: players would be able to reconsider form and materials, and could
experiment with many different sounds by themselves. Few answers referred to the
emotional connection and unique narratives linked to each of the resonators. Some
said this could lead to new sound possibilities and to a new collecting culture (most
serious players have several guitars, and develop a particular bond with each – here,
a player would collect unique resonators, each representing different acoustic/sonic
properties and aesthetic values). Going further, some participants predicted that
new iconic original sounds could appear, associating the new sounds with
resonators instead of with the guitar itself.
The participants were asked if it is easy to replace a resonator, on a scale from
one to seven (one represents “no” and seven represents “definitely”). The average
answer was 4.6, with a standard deviation of 1.4; this means it is not easy enough,
although the replacement itself does not take more than fifteen seconds. Some of
New Music Research Journal
Amit Zoran et al.
24
them complained that the SP unit (the DSP on the back of the guitar) needs to be
covered and protected from the player’s body, an obvious modification were this
instrument to be produced.
6.2. Digital Abilities
The participants were asked if the digital abilities are interesting. While all of
them answered positively, their explanations differed one from another. The ability
to manipulate digital effects with a tangible acoustic interface, such as in resonator
no. 6 (see the explanation in section 5.6), seemed to be very compelling to the
majority of players. Some of them wrote that this allows them to expand the playing
experience. Several participants believed the expressive control of digital effects
from the resonator can serve as an interesting alternative to guitar pedals.
For a few of the participants, the digital capabilities suggested a big potential to
create new sounds. Some of them, however, indicated that this potential still needed
to be developed further. One participant said the current algorithm could be
implemented with analog processing, and that the degree of digital control was not
developed enough. There are some disadvantages in using a computer for sound
processing (e.g., there is a prejudice that computers are not aging nicely as wood,
and the guitar community still rejects the integration of digital technology into the
guitar body) and one participant was not sure about the argument of implementing
the computer over an analog circuit.
6.3. Resonators
The participants were asked how many resonators they felt they might use
(unlimited by our collection of 8 resonator). The average answer was 4.2, with
standard deviation of 2.6. Then they were asking to choose their preferred three
resonators, and to explain their selections. More analysis of the reaction to
conventional vs. novel resonators is presented below.
New Music Research Journal
Amit Zoran et al.
25
6.4. Instrument-makers Evaluation
Instrument-makers also qualitatively evaluated The Chameleon Guitar. The guitar
was presented to individual guitar-makers and to violin-makers. The guitar-makers
gave an inside perspective, while the violin-makers provided a comparatively
unbiased point of view. The project was discussed freely, regarding its concept,
technical issues, and optional contributions. The majority of instrument-makers said
they would be willing to develop resonators if the market would demand it. Some
suggested to keep on improving the sound quality, and to have at least one
resonator that was acoustically optimized.
7. Discussion
The Chameleon Guitar was designed and built to implement a new concept of an
acoustic-digital hybrid instrument, based on discussions with instrument-makers,
experience from the preliminary prototype and the capabilities of digital simulation.
While fifteen players formally evaluated it, at least fifteen more players tried it
within a total period of ten weeks. During that time, two groups of resonators were
made and fifteen different resonators were tested, and the resonator tray was opened
at least three hundred times. The project was discussed with designers and
engineers, demoed multiple times, and received significant press coverage.
The concept seemed to be compelling to the majority of participants. Players
from different musical styles, as well as instrument-makers, understood the idea and
its potential well. It seems that the new combination of digital and hybrid features
interest the majority of those in the study, while allowing maximum flexibility in
both domains.
The player participants responded very positively to the question “would you
like to have such a guitar”, and were consistent on the uses and role they believe this
guitar would be appropriate for. The players that liked the guitar less are generally
New Music Research Journal
Amit Zoran et al.
26
players who are not used to this guitar interface (such as players who prefer nylon
strings). It seemed that hollow body guitar and electric guitar players enjoyed The
Chameleon Guitar more.
After analyzing the survey forms, no correlation was found between the
participant’s favorite musical styles and the role they saw for The Chameleon Guitar.
While the most important property of the guitar, according to the players, was its
new sound potential together with the option to replace resonators, players tended
to associate the guitar to rock, jazz, contemporary music and folk music styles,
which are more experimental musical fields (unlike classical music or blues).
The participants referred to the new sonic possibilities as one of the most
important contributions of The Chameleon Guitar, but did not correlate this property
with the guitar itself. They correlated sound qualities with resonators. None of the
answers to the questions about improving the guitar dealt with sound, and although
all the players and instrument-makers believed its digital abilities were important,
most felt that it contributed as a sound effect rather than a sound source.
In general, this means that the guitar’s digital processing abilities do not yet have
a significant identity. More work thus needs to be done to enlarge the digital
capabilities. The question of how to control the software, by tangible interfaces on
the resonators themselves or by electronic controllers on the guitar, still needs to be
tackled. When considering the high degree of complexity that digital processing
adds to the sound, we also need to discuss the off-line sound design interface: how
an external computer interface takes part in simulating, modifying and controlling
the preferred sound, and how the unique properties of each resonator can be
leveraged and maximized with digital sound design.
All the subjects correlated sound qualities with the resonators and recognized
that the resonators’ replacement, together with the sound possibilities, are the most
important properties of the guitar. However, when asking the players how many
New Music Research Journal
Amit Zoran et al.
27
resonators they would like to have, the average answer was 4.2. A lot of guitar
players own more than one guitar, and 4.2 does not seem to maximize the
innovation potential. On the other hand, instrument-makers suggested
experimenting with as many resonators as possible. When discussing this conflict
with players, the most common answer was that they choose resonators from the
collection that they have seen, and they believed that after trying more resonators
for a longer period, they may want to have more. It was difficult for the average
player to imagine a new type of resonator. However, instrument-makers could
easily discuss new resonator designs.
By analyzing the popularity of different resonators, we learned that the second
group of resonators (no. 5 – no. 8, the more experimental resonators), were
preferred. Here, the most popular resonators were no. 6 and no. 8, and the main
argument for their popularity relies in their novelty, unexpected behavior (although
sometimes this behavior was actually produced by the digital effects), higher
expressivity, and experimental options. Loose elements and embedded chambers
have a lot of potential – although experimental guitar artists have explored such
ideas, they are still relatively unknown to the guitar-playing mainstream. The
Chameleon Guitar provides a platform well suited to exploration of such concepts. It
can be interesting, for example, to combine these with other mechanical elements
(such as wheels or wires) and to redesign the use of the spring (resonator no. 7, see
Figure. 9).
The first group of resonators (no. 1 – no. 4, the conservative group) received less
popularity from the players, although almost all participants chose at least one
resonator from that group in his or her selection. Here, the main reasons for
choosing a resonator were sound and aesthetic qualities, referring to acoustic guitar
standards. However, the preferences varied for each participant, where the selection
of good sound or “the most beautiful” resonator depended on personal preference.
New Music Research Journal
Amit Zoran et al.
28
Each one of the resonators got the title “sounds best” or “the most beautiful” from
different players. For the instrument-makers, the conservative group was more
interesting than the experimental one, perhaps because of their bias towards quality
in standard guitar design. In general, sonic qualities, interface issues, narrative
properties, and aesthetic qualities were related to the resonators, more than to the
guitar.
The tuning of the guitar, which varied when resonators were replaced up to halfa-tone, can be stabilized by adding fine-tuning mechanism to the resonator’s bridge,
allowing the user to set all the resonators to the exact location (and string action)
under the strings’ load. Another option is to use automated mechanical tuning,
which automatically re-tunes the guitar after replacing a resonator.
Although a resonant chamber with acoustic pickups can aggravate feedback
problems, in practice it was not a significant problem with The Chameleon Guitar.
While feedback accrued when amplifying the resonators’ raw signals directly, the
low-pass nature of our digital processing prevented it from becoming a practical
problem. In that context, the guitar was recorded and evaluated in a sound studio,
and it still needs to be examined in concert conditions, which may introduce
different challenges.
8. Conclusions and Future Work
The Chameleon Guitar is the product of a year and a half of development, inspired
from both the digital and the physical musical instrument landscapes. This new
approach to designing guitars was tested successfully, and proved itself over time.
The guitar and its resonators functioned well, were evaluated by fifteen players, and
tried by many more. Several mechanical changes need to be made to the current
guitar model, such as a new design for the resonator tray and new tuners. Other than
that, the guitar is stable, ergonomic, and offers an open-ended selection of timbres.
New Music Research Journal
Amit Zoran et al.
29
The main goal of this work was to merge traditional values with digital
capabilities. Based on our evaluation results, we can say that it was fairly successful.
The main contribution of The Chameleon Guitar lays in its innovative solution to use
replicable, acoustic resonators with electronic processing, while enjoying a higher
degree of acoustic information captured from these resonators by several sensors
(relative to the single surface sensor that is commonly in use in acoustic guitars).
While any digital algorithm to create sound can be easily reproduced and copied,
each wooden piece is unique and has a spatial-acoustic behavior. Here, we combine
this acoustic uniqueness with the digital environment, leveraging the unique
acoustic signature of each resonator with a palette of appropriate digital sonic
transformations. In addition to this success, the process of experimentation and risk
taking in the design of resonators resulted in innovative expressive abilities. It
seemed that the community of instrument-makers felt more attached to the
traditional approach, while the guitar players were more excited by the
experimental approach. We believe that both of these endeavors need to continue
being developed together by introducing traditional values into experimental
solutions. As a quality criterion, it is important to have at least one resonator that
sounds like a good acoustic guitar. While the sound produced by several resonators
(together with the digital processing), is already being recognized as an acoustic
guitar sound (by all listeners) more work, acoustically and digitally, needs to be
done in order to compete with the sound of high quality acoustic guitars. This goal
may be achieved by improving the resonator design, sensor positioning, and digital
processing algorithm.
The instrument’s digital capabilities have a huge promise that interests users,
especially because of their potential to enlarge and expand upon the resonator’s
unique physical properties. However, richer digital processing options still need to
be investigated. A visual feedback and control scheme is also needed, dependent on
New Music Research Journal
Amit Zoran et al.
30
the guitar and resonator design – each resonator lends itself to different expressive
affordances.
More generally, this approach could be easily implemented in other string
instruments, such as the violin family, and with a bit more effort could even be
developed into a piano solution.
The external computer interface for modifying the digital content of the
instrument is a different topic that requires more research. One can envision, for
example, a simple high-level API, that would enable each resonator to bring up a
particular set of options and adjustments on an attached PC, allowing the player to
appropriately modify the guitar’s sound based on meaningful parameters (as
opposed to adjusting filter coefficients or directly writing code, although that’s
always an option for those so inclined). The potential is huge: in this system, we can
connect a sound-making object to virtual environments in a very fluid fashion. This
connection can demonstrate how physical objects can share in the same media
revolution as digital objects, and opens up new possibilities for future forms of
interactive entertainment. Such a connection can lead the way in combining craft,
tradition, and acoustics with the digital environment, opening up a new future for
hybrid design of objects.
9. References
Bacon, T., Hunter, D. (2004). “Totally Guitar; The Definite Guide”. Thunder Bay Press.
Chapters 1-3 & Chapter 5.
Bevilacqua, F., Rasamimanana, N., Fléty, E., Lemouton, S., Baschet, F. (2006). “The
augmented violin project: research, composition and performance report”. 6th
International Conference on New Interfaces for Musical Expression (NIME06).
Paris, French.
Cadoz, C., Luciani, A., Florens, J. L., Castagné, N. (2003). “Artistic Creation and
Computer Interactive Multisensory Simulation Force Feedback Gesture
New Music Research Journal
Amit Zoran et al.
31
Transducers”. 3rd International Conference on New Interfaces for Musical
Expression (NIME03). Montreal, Canada.
Carfoot, G. (2006). “Acoustic, Electric and Virtual Noise: The Cultural Identity of the
Guitar”. Leonardo Music Journal. Vol 16. P. 35-39.
Coppiardi, M. (2008). Videos, personal and e-mail interviews.
Green, D. W., Winandy, J. E., Kretschmann, D. E. (1999). “Mechanical Properties of
Wood”. Wood Handbook—Wood as an Engineering Material. United States
Department of Agriculture, Forest Service. Forest Products Laboratory.
www.fpl.fs.fed.us/products/publications.
Cuzzucoli, G. and Lombardo, V. (1999). “A Physical Model of the Classical Guitar,
Including the Player's Touch”. Comput. Music J. 23, 2 (Jun. 1999), 52-69.
Fastl, H; Zwicker, E. “Psychoacoustics, Facts and models”. Springer. 2007. Chapter 7.
P. 175-202.
Fender website. (2009). www.fender.com. Last accessed July 15th, 2009.
Fletcher, N. H., Rossing, T. D. (1990). “The Physics of Musical Instruments”. Second
edition. Springer. Chapter 1, P. 3-153.
Fraden, J. (2003). “Handbook of Modern Sensors: Physics, Designs, and
Applications”. Third edition. Springer.
Gibson website. (2009). www.gibson.com. Last accessed July 15th, 2009.
Hiscock, Melvyn. (1986). “Make Your Own Electric Guitar”. NBS Publications.
Howard, D.M., Rimell, S. (2003). “Gesture-Tactile Controlled Physical modeling
Music Synthesis”. Proceedings of the Stockholm Music Acoustics Conference,
SMAC-03, Vol. 2, August 6-9.
Hughes, T. (2004). “Analog Man’s Guide To Vintage Effects”. For Musicians Only
Publishing.
Inta, R. (2007). “The Acoustics of the Steel String Guitar”. PhD Thesis. School of
Physics, The University of New South Walls, Sydney, Australia. Chapters 2 and 4.
New Music Research Journal
Amit Zoran et al.
32
Ishii, H., Ullmer, B. (1997). “Tangible Bits: Towards Seamless Interfaces Between
People, Bits and Atoms”. SIGCHI Conference on Human Factors in Computing
Systems. Atlanta, Georgia, USA. March 22 - 27.
Jahnel, F. (1981). “Manual of Guitar Technology”. First English edition. The Bold
Strummer Ltd Publication. Chapter A, P.11-52.
Jehan, T., Yound, D., Bell, J., Lunn, C. V. (2005). “Hyperviolin”. MIT Media Lab.
Kauraniemi, J., Laakso, T. I., Ovaska, S. J. (1998). “Delta Operator Realizations of
Direct-Form IIR Filters”. IEEE Transactions on Circuit and Systems II: Analog and
Digital Signal Processing. Vol 45, No. 1. January. P. 41-52.
Kinkead, J. (2004). “Build Your Own Acoustic Guitar”. Hal Leonard publication.
Materials Chapter, P. 22-29.
Kvifte, T. and Jensenius, A. R. (2006). “Towards a coherent terminology and model
of instrument description and design”. In Proceedings of the 2006 Conference on
New interfaces For Musical Expression (Paris, France, June 04 - 08). New Interfaces
For Musical Expression. IRCAM — Centre Pompidou, Paris, France, 220-225.
Lähdeoja, O. (2008). “An Approach to Instrument Augmentation Guitar”. 8th
International Conference on New Interfaces for Musical Expression (NIME08).
Genova, Italy.
Lazzarini, V., Timoney, J., Lysaght, T. (2008). “Asymmetric-Spectra Methods For
Adaptive FM Synthesis”. 11th International Conference on Digital Audio Effects
2008. DAFx-08. 2008. September 1st-4th.
Leissa, A. W. (1969). “Vibration of Plates”. NASA. Ohio State University, Columbia,
Ohio. Chapter 1, P. 1-6.
Line 6 website. (2009). www.line6.com. Last accessed July 15th, 2009.
Machover, T. (1992). “Hyperinstruments: A Progress Report 1987-1991”. MIT Media
Lab.
New Music Research Journal
Amit Zoran et al.
33
Merrill D., Raffle H. (2007). “The Sound of Touch”. CHI 2007. San Jose, California,
USA. April 28 – May 3.
New England Luthiers. (2009). Steve Spodaryk, John Svizzero, Jeb Hooker, Alan
Carruth, DJ Parsons, Donald Lambe, Thomas Knatt and Daniel Wangerin. Video
interview. May 11th.
Paradiso, J. (1997). “Electronic Music Interfaces: New Ways to Play,” IEEE Spectrum
Magazine, Vol. 34, No. 12, pp. 18-30 (Dec., 1997).
http://web.media.mit.edu/~joep/SpectrumWeb/SpectrumX.html. Last accessed July
15th, 2009.
Parker, K. (2009). Public talk in the MFA, personal and e-mail interview. February 5th.
Recording Basics, MXL website. (2009). http://www.mxlmics.com/support/recbasics.html.
Last accessed June 10th, 2009.
Siminoff, R. (2007). “The Art of Tap Tuning”. Hal Leonard.
Smith, J. O. III. (2007). “Introduction to Digital Filters”. W3K Publishing.
Smith, J. O. III. (2008). “Physical Audio Signal Processing”. Beta vertion. W3K
Publishing.
Smyth, T., Smith, J. O. III. (2002). “Creating Sustained Tones with the Cicada’s Rapid
Sequential Buckling Mechanism”. 2nd International Conference on New Interfaces
for Musical Expression (NIME02). Dublin, Ireland.
SynthAxe website. (2009). http://www.hollis.co.uk/john/synthaxe.html. Last accessed
August 13th, 2009.
Teuffel website. (2009). www.teuffel.com. Last accessed July 15th, 2009.
Thompson, A. (1997). “Stompbox – A History of Guitar Fuzzes, Flangers, Phasers,
Echoes & Wahs”. Miller Freeman Books.
Vanegas, R. (2007). “The MIDI Pick - Trigger Serial Data, Samples, and MIDI from a
Guitar Pick “. Conference on New Interfaces for Musical Expression (NIME07),
New York, NY, USA.
New Music Research Journal
Amit Zoran et al.
34
Zoran, A. (2009). “Chameleon Guitar - A Physical Heart in a Digital Instrument”.
Master Thesis. MIT Media Lab.
New Music Research Journal
Amit Zoran et al.
35
Figure 1. The Chameleon Guitar with four different resonators.
New Music Research Journal
Amit Zoran et al.
36
Figure 2. The Chameleon Guitar signal path.
Figure 3. In order to find the lowest eigenmodes for a given surface area, different shapes were
simulated (here we present just 7 of them). The shapes were modified slightly between
simulations to search for a pseudo- optimal one. As part of the iterative process, we also looked
for good support locations (blue dots). We stopped this process after 20 shape simulations,
selecting the shape at lower right with its 3 support points. From this, we define the location
of the resonator PCB to be on the top support point. The bridge location was similarly defined
next (see text).
New Music Research Journal
Amit Zoran et al.
37
Figure 4. FEM simulation of eigenmodes for the final resonator shape. This visual data was
used to define the 4 sensor locations (at right). For each eigenmode, we detect the best sensing
location (regarding maximal surface bending); then we cluster the eigenmodes into four
groups according to a specific sensor location (shown on each eigenmode as an orange dot).
New Music Research Journal
Amit Zoran et al.
38
Figure 5. Dual and single cuts in the body of electric guitars.
Figure 6. The Resonator tray operation.
New Music Research Journal
Amit Zoran et al.
39
Figure 7. The Chameleon Guitar’s backside showing the
resonator tray and the SP unit.
New Music Research Journal
Amit Zoran et al.
40
Figure 8. The digital signal processing chain.
New Music Research Journal
Amit Zoran et al.
41
Figure 9. Logarithmic, smoothed spectrograms of the impulse response for each
sensor, showing also the final SP output as well as the reference. These plots
(excepting that of the reference) are for the Sitka spruce resonator (Figure 10.1).
New Music Research Journal
Amit Zoran et al.
42
Figure 10. Top: A group of eight resonators. The first four are more traditional, made
from wood only. The last five are more experimental, including loose screws, springs,
free plates or plastic chamber with rice or water. Resonator no. 1 was used as the
reference for algorithm development. Bottom: Resonator backside (number 4). This
resonator is arched, with all the sensors on its back.
New Music Research Journal
Amit Zoran et al.
New Music Research Journal
43
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

advertising