501E6 vM 6`
db
THE SOUND ENGINEERING MAGAZINE
SEPTEMBER 1982
S1
501E6 vM
SNOlitcYAr
6'
95
The new advaIll:td-dt'.S1gI1 IQ-Series
I11L\E'.I'S.
RIGHT ON THE MONEY.
the \IQ:\ lk:\ hue cct hlgl:l
l(I!>ni 3tle` ltl!\e`rs that n'tll:c'
6.,la1.,n,tllSc`V ltfl ,lllcih0
:4`rtOril laI lì e. l rr c Ifh
Nrur
1
1e34 get tltEUVnaI tle'\Aifit\
-i-i.atlei F.i}c)tl c` 3c'fic.hipIlc`l
:-..c'rtc's
..
ftle\itut
`c" '-.:r?Rt't-c' l c'LlVälijO
tu cib<atik
\
sends. Plus a pair of In, stc'rprogram 9 -band graphic
equalizers. And direct interface
flexibility with per-channel dual
input sw itcl111U.
Theres also the depend ability atei sell tee onvetllence
z tn`u'e.iular L :l out lvith each
hatlllCts c tittlportents mounted
Ehi a sc.paraté tntranal printexi
c
c
it
Circle
1Q
on Reader Service
circt lit board. The \ IQ-Series.
la, ht On the n lone }: right e it ill i
the line
t
.
\ isit uur Yamah,l dealer
or write us tor mure` informltion.
Yamaha, Box 6600. Buena Park.
C\ 90622. In Canada, Yamaha
Canada-Musk: Ltd 135 Milner
Ave.. Scarb.. Ont. M1S 3R1.
Publisher
Larry Zide
Editor
John M. Woram
European Editor
John Borwick
Associate Editor
Mark B. Waldstein
Production & Layout
David Kramer
SEPTEMBER 1982 VOLUME 16. NO.
9
Advertising Coordinator
Karen Cohn
Classified Advertising
DEPARTMENTS
Carol Vitelli
Book Sales
Lydia Calogrides
Circulation Manager
Eloise Beach
Art Director
Bob Laurie
Graphics
K &S
Graphics
Typography
Spartan Phototype Co.
LETTERS
EDITORIAL
CLASSIFIED
6
32
60
Barry Blesser
DIGITAL AUDIO
12
Ken Pohlmann
THEORY AND PRACTICE
18
SOUND WITH IMAGES
Len Feldman
24
SOUND REINFORCEMENT
John Eargle
30
sales offices
Roy McDonald Associates, Inc.
Dallas, Texas 75207
First Continental Bank Bldg.
801 Marvin D. Love Freeway
Suite 303
(214) 941-4461
Denver, Colorado Area
Englewood, Colorado 80112
14 Inverness Dr. East
Bldg. 1- Penthouse
(303) 771-8181
Houston, Texas 77036
6901 Corporate Drive
Suite 210
(713) 988-5005
Los Angeles Area
Glendale, California 91204
424 West Colorado Street
Suite 201
(213) 244-8176
Portland Area
Hillsboro, Oregon 97123
510 South First
P.O. Box 696
(503) 640-2011
San Francisco Area
Emeryville, California 94608
265 Baybridge Office Plaza
5801 Christie Avenue
(415) 653 -2122
Sagamore Publishing Co.
New York
Plainview, NY 11803
1120 Old Country Rd.
(516) 433-6530
This month's cover, courtesy of Gregg
Stephens, features some of the equipment currently employed in Soundstream's Digital Editing Facility in Hollywood, California.
NEW PRODUCTS AND SERVICES
58
PEOPLE, PLACES, HAPPENINGS
62
FEATURES
THE SOUNDSTREAM DIGITAL MUSIC
COMPUTER: RECORDING, EDITING AND BEYOND
Sherman Keene
34
THE .MAN BEHIND SOUNDSTREAM
39
Richard Koziol and
Alfred W. D'Alessio
TO BUILD THE IMPOSSIBLE DREAM:
PERFORMANCE
STUDIO COMES TO THE BIG APPLE
A SOUND -INSULATED
44
A db APPLICATION NOTE:
TECHNIQUES FOR HUM AND NOISE REDUCTION
Dale Beard
47
THE FFT: BIG-TIME MATHEMATICS
COMES TO AUDIO
Robert Berkovitz
52
NEW YORK'S CENTER FOR THE MEDIA ARTS
Joseph Coencas
ADDS A SCHOOL OF AUDIO ARTS
57
db. the Sound Engineering Magasine (ISSN 0011 -7105) is published monthly by Sagamore Publishing Company. Inc. Enure contents copyright G 1982 by Sagamore PublishingCo., I12001d Country Road. Plainview, L.I., N.Y. 11803. rdephonc(516)0336530.
db is published for those individuals and firms in professional audio-recording, broadcast. audio-visual. sound reinforcement, consultants. video recording. film sound. etc. Application should be made on the subscription form in the rear of each issue. Subscriptions are $15.00 per year (528.00 per year outside U.S. Possessions; $16.00 per year Canada) in U.S. funds. Single copies are S1.95
each. Editorial. Publishing and Saks Offices: 1120 Old Country Road. Plainview. New York 1803. Controlled circulation postage
paid at Plainview, NY 11803 and an additional mailing office.
1
cri
Ta
Eladio
for
Letters
professionals
db OR NOT dB
To Tin, EDItoR:
After an exasperating hour at a local
library. doing research for the Audio Technica style book. I am turning to you
REEL TO REEL TAPE
Ampex, 3M. All grades.
On reels or hubs.
CASSETTES, C- 10 -C -90
With Agfa. TDK tape.
LEADER & SPLICING TAPE
EMPTY REELS & BOXES
All widths, sizes.
Competitive!
Shipped from Stock!
G..
Poly
ecording supplies catalog.
312/298 -5300
Corp
1233 Rand Rd.
Des Plaines. IL 60016
15
Circle
0 on Reader Servire Card
EH L-15 CO -AX
200 Watts
103 dB M/W
Double Spider
15 Inch
for your thoughts on a matter which must
be of some importance to you. since it
involves the word from which your
publication derives its name: the decibel.
I am looking for an "unimpeachable"
authority on whether a db is really a dB.
( Your own publication seems to be sitting
-
EMILAR CORPORATION
1365 N. McCan St.
Anaheim, CA 92806
cm
Circle
15
on Reader Service Card
Alpha Audio
Ampex
Audio -Technica
23
20
Bose
21
Cetec Vega
II
CTA B
Datatronics
Emilar
Haney Professional
Hewlett- Packard
22
JBL
Kimball
Klark leknik
Maxell
Orban
your name db on the masthead. but you
frequently use dB in textural matter.) The
standard reference of writers and editors.
A ,t/anual of Style. by the University of
Chicago Press. lower -cases both the
word and its abbreviations: decibel. db.
So does Webster's New Collegiate
Dictionary and the Oxford Dictionary of
.4meriran English. Most of the manufacturers and publishers in our industry.
however. still use dB.
Perhaps we should emulate the diner
who. knowing he is incorrect in doing so.
orders "turbo" for "turbo/. instead of the
approved turban. for fear those around
him will think him uneducated. Apparently. dB is wrong. but many of our
customers and colleagues would undoubtedly think us illiterate if we spelled
Otani
Peavey
Polyhne
Pro Audio Yearbook
Shure
it db.
One stylistic area in which we enthusiastically agree with Webster's is that the
proper abbreviation for microphone is
mike, rather than mho. Since the abbreviation is pronounced as a word. the rules
of pronunciation would mandate calling
it "mirk" -as in pie. hie. or tic. Another
good reason for mike is that it avoids the
god -awful term "micing. " when used as a
verb. in the participial form.
Please let me have the benefit of any
research you or your fellow editors may
have done on these terms. I'll hold upon
our style book until I hear from you and
some of the lexicographers and style
book editors to whom I'll pose the same
questions.
DON KIRKENa I,I.
Director
Marketing Communications
Audio- Technica U.S.. Inc.
db (the Mat;a=ine) got .started in 1967.
when life was a lot simpler, and the
decibel had a long history of being
abbreviated as "db." However, in current
engineering practice. units of measure
which use a person's name are spelled out
in small letters, but abbreviated with a
capital letter. with no period (unless at
25
Cover II
Beyer
squarely on the fence. since you spell
db (dB ?) replies:
For complete Information call
714- 632 -8500
800 -854 -7181
Index
16
6
19
7
9
8
10. 43
15
12
Coyer IV
28 -29
6
27
14
Studer
Swintek
33
Telex
Tentel
45
TOA
Trident
18
10
14 -35
8
Coming
Next
Month
Next month is our AES show issue.
and we'll be featuring articles on a wide
range of topics. Ham Brosious of Audiotechniques checks in with a feature outlining the pros and cons of renting pro
audio gear: Richard Factor of Eventide
brings us a progress report on the microprocessor: Murray Allen of Universal
Studios gives us some insights on what
can go wrong on both sides of the control
room window, and our own John Woram
shows how a programmable calculator
can be used to design a sophisticated
sound system. Of course, our regular departments and columnists will be on
hand. making October's db -The Sound
Engineering Magazine, a show issue to
remember.
P:
EXPERIENCE IN AUDIO
Tf IfLOGY
HP's 8903A now makes audio testing
easier than ever before.
Now a single microprocessor- controlled
instrument offers you the versatility
of an extremely wide range of audio
measurements with just a few keystrokes.
Manual or HP -IB programmable, the
HP 8903A Audio Analyzer is ideal
for testing 20 Hz to100 kHz characteristics of stereo amplifiers, pre- and line
amplifiers, tape decks, cassettes, and
other high performance audio equipment.
The 8903A measures such basics as
distortion, gain and frequency response,
power, AC or DC volts, and frequency;
as well as more complex parameters like
signal -to -noise and SINAD. It includes
a sweepable low distortion audio source
and drives an X -Y recorder directly. It's
an automatic test system all in one box,
yet is priced at only $6600*
For more information, contact your
nearby HP sales office, or write to
Hewlett- Packard Co., 1820 Embarcadero
Road, Palo Alto, CA 94303.
'Domestic US price only.
NIP 100
SYSTEMS
04105A
P
(Ink
IN
HEWLETT
PACKARD
on Reader Service Card
the end
nwe have
of a
91'
.sentence.
(not
of course).
9f)=
Thus.
9von's.
.tluhiplier prefi.yes are capita/Led only
above 106 ( = .N). This gives u.s 1.500.000
volts= 1.5 Ml'. nhile0.001 volt (not volts)
_ l nt V. Moving right along toward the big
.finale. l bel = 1 B = 10 decibels= 10 dB.
Alost audio publications made the move
loner rase. and will
probably end it that way (hopelidlr, not
soon).
As for that transducer you mentioned
ne usually cop out by completely spelling
it out. Mike technique suggests something
that Michael dues very ire //. and miring is
probably hest left to an exterminator. or
the studio rat.
Now then. can anyone tell us hou' to
pronounce EQ.'
Two independent channels
Limiter/
Compressor
P
LC -1
S695.00
A
Continuously variable slope
Stereo coupling
Illuminated VU. meters for
gain reduction or output level
High performance at low cost
Compact size (19 "x 13í4"x 8 ")
TRIDENT (USA)
Parametric
Equalizer
Q
CB9066
S650.00
SPEAKING OF MICE...
3 band eq
Variable "Q"
Separate Bypass for each band
Hi 8 Low Pass Filters with continuously variable frequency and slope
Built -in power supply
Compact size (19 "x 13/4 "x 8 ")
Order direct:
652 Glenbrook Rd., Stamford, CT 06906 (203) 357-8337
('ire/e l9 on Reader Service Card
To THE E11row
Congratulations on a fine June issue.
Just two minor items: I've checked
every condenser microphone here at
Sony and a few by other manufacturers
and I can't find one black electret. Most
of ours are gold. I'd telex Japan to ask
them. but
don't think they'd see the
humor in it.
And to answer your question in the
AES News section of "Happenings. "yes.
Sony will be there. It should be a good
show. even if the place has mice.
JOHN F. PHFIAN
Western Regional Sales Mgr.
Sony Corp. of America
1
More and more recording studios are
discovering the great sound of the
Kimball Professional Grand.
Here's why:
The Kimball 6' 7" Professional Grand derives its heritage of greatness from the world's finest piano -the
Bosendorfer. The scale and plate design are derived
from the Bosendorfer Model 200, and the plate is extra thick to assure maximum sustain and to avoid
plate noise from hammer strikes. The
Bosendorfer- derived scale and non duplexed trebles enhance tonal
clarity. and pitch perception. The
Kimball Professional Grand is (- '`IIIII
specifically designed for
clear, pure tonality. free of
spurious noise and false
t
tl
harmonics. It also offers
superior durability and
tuning stability. Its entire structure, including the soundboard,
is of precision -laminated
woods. greatly reducing
differential expansion in
changes of temperature
db replies:
Don't send that telex- there's nothing
furent' about black electrets. although
they might help our miring technique.
Sorry about the typo. N'eillire the proofreader. if we had one.
I\t
depth,\
I
t
and humidity.
For more information
/41
about the Kimball
Professional Grand,
contact Wade Bray at
k1mBa
(812) 482-1600.
`The key to sounding gres
549 Royal St. Box 460
Jasper, IN 41546
op
Circle 14 on Reader Service Card
j
Before you invest in
new studio monitors,
consider
all the
ankles.
Polar response comparison oj'a hy+ird t wo:wr coaxial studio morito and JRL} new
11.0 Ki=Radialstudio monitor from I k112
to IO
kHz.
.,
t ipiral hori_:ottal
\o une has tu tell ui he n\ important flat frequency response is in a
studio monitor: But il VOLI judge a
monitor's performance 1w its on -axis
response curve, you're only getting
part of the story.
\lost conventional monitors tend to
narrot% their dispersion as frequency
increases. So %chile their on -axis
response may he flat, their off-axis
response can roll off dramatically. literally locking you into the on -axis "sweet
spot:' Even worse, drastic changes in
the horns directivity contribute significantly to horn colorations.
%
Introducing the
J BL Bi-Radial
Studio Monitors.
\t JUL, we've been investigating
the relationship between on and off
axis frequency response for several
years. The result is a new generation
of studio monitors that provide flat
response over an exceptionally wide
range of horizontal and vertical angles.
The sweet shot and its traditional
restrictions are essentially eliminated.
The key to this intproycd performance lie, in the unique geometry of
the monitors Iii -Radial burn! I)cveloped with the aid of the latest computer design and analysis techniques.
the horn provides constant coverage
from its cross( wcr point of Hill() lz
to beyond 1ó kl Ix.l'he Iii - Radial
compound flanc configuration maintains precise control of the horn's
wide 11111° x 11)1)° coverage angle.
.lHl,
And the Iii -Radial horns performance advantages aren't limited to just
beantwidth control. The horn's rapid
flare rate. for instance, dramatically
reduces second harmonic distortion
and its shallow depth alloys for optimal acoustic alignment of the drivers.
This alignment lets the numitors fall
well below the !flatten and Laws
criteria for minimum audible tinte
delay discrepancies.
But while the Bi- Radial horn
offers outstanding performance, it's
only part of the total package.The
new monitors also incorporate JBI.s
most advanced high and low frequency transducers an cl dividing
networks. \ \' Irking together these
I/.iN:rN/rd
components provide exceptionally
smooth response, high power capacity, extended bandwidth. and
extremely low distortion.
Judge For Yourself
of couse, the only tray
judge
to really
studio monitor is to listen for
yourself. So hefore you incest in new
monitors, ask your local JB1, professional products dealer for a Bi- Radial
monitor demonstration. And consider
all the angles.
a
James B. Lansing Sound, Inc.
851)0 Balboa Boulevard
P.O. Box 2200
Northridge, California 91::20 t'.S_ \.
1
I. Patent applied tor.
JBL
Professional
Products
Division
(D
(
irrII'
I
,,n /ietlf/tT .Setrirr
(
ard
CLOSING THE CREDIBILITY GAP
ACCURATE
TAPE TENSION
MAKES BETTER
RECORDINGS
To THE EDITOR:
In my letter of May 29. I called to your
attention some misinformation published in the May. "Sound with Images"
column. This was never acknowledged.
nor were my questions answered. I would
like to see a statement by the editor or the
author in reference to the points raised. I
feel that corrections should be published
for the edification of your readers. Your
credibility shall suffer severely otherwide.
13-FN SOBIN
If something SOUNDS F /SHY it may
be your fish sca/e approach ro measuring
tension.
Ben Sobin Motion Picture
Sound Recordingand Equipment
Tentel Tape Tension
is designed to diagnose problems in your
magnetic tape equipment.
Virtually all recorder manufacturers use and recomThe
Gage
mend the TENTELOMETER
for use with their equipment.
db replies:
Go easy on us. Ben. Your letter and our
editor:c/author's replies were published
in the July issue. L'n/i,rumateh, by the
time our subscribers receive db. the ne.vt
issue is already at the primer. and the ne.vt
one after that is well on the way there. H'c'
try to hold the letters column open until
the very last moment (this one just
sneaked in under the wire), but there's
inevitably a At of a month or two,
especially if your letter requires an
author's response.
TENTELOMETER ® measures
tape tension while your
transport is in operation. so you can "see" how your transport is
The
handling your tape, high tension causing premature head and tape
wear. low tension causing loss of high frequencies. or oscillations
causing
wow and flutter. Send for the Tentel "Tape Tips Guide".
$279 - complete.
The T2- H20-ML sells for
1506 Dell Avenue
C Iv TEL
Campbell, CA 95008
(408) 379.1881
Toll Free 800 538 -6894 (ex. CA)
Circle 12 on Reader Service Card
ßrooke Siren Systems
For years our MCS200 series crossovers have been used by major touring companies
worldwide. The same high technology is available in our FDS300 series crossovers. Our
products include some very unique features, highlighted below, but perhaps our best feature
is the way we'll sound with you. Available now through professional audio distributors and
music stores nationwide.
Limiters on Each Section
24dB /Octave Slope
Subsonic /Ultrasonic Filters
Section Mute Control
±6dB Level Control
Sealed Potentiometers
Brooke Siren Systems, 262R Eastern Parkway, Farmingdale, New York 11735 (516) 249 -3660
Gerraudio Inc., 363 Adelaide Street East, Toronto, Ontario M5A 1N3 (416) 361 -1667
N
co
Q)
III
a'-
Wen
i! F
ti
(
-M-
Ir_
4
MMtOI/M011UM
MIL
t
h_`M-MOMLtvtl
-'.-,
_.
I
-6W-,
.1
f 05 340
flROUEN
:Y
ONfYNG
SYSTEM
FDS
Fréquenéy
FD5320
FDS340
o
2
3
-Way Stereo
or 4 -Way Mono
iiIdIiÑ'à' Systems
AR116 Active Direct Box
AR125 Cable Tester
Circle /3 on Reader Service Card
MCS200 Modular Series
3, 4, or 5 -Way Stereo
J
340
liant performer.
Model 82
Wireless Condenser
Hand -Held
The Model 82 condenser wireless
microphone has been added to
Cetec Vega's professional hand-held
line. The Model 82 incorporates the
popular Shure SM85 condenser element and attractive black windscreen
to provide:
Minimal handling noise, reduced
mechanical vibration, and virtually
no "boominess" (by means of controlled low- frequency rolloff).
Clean reproduction of close -up
vocals with moderate proximity
effect.
"Crispness" and presence with
high -definition midrange.
Clear, scintillating highs with crisp
upper register.
Cardioid pickup pattern for effective
rejection of off-axis sounds.
All Cetec Vega hand -held wireless
microphones (including the Model
80 with the Electro -Voice EV -671
dynamic element and the Model 81
with the Shure SM58 dynamic element) have an attractively contoured
black case with internal antenna.
Used with Cetec Vega professional
wireless receivers, the FM systems
operate on any crystal- controlled frequency between 150 to 216 MHz, at
a range up to 1000 feet or more.
Transmit -to-receive frequency
response is almost perfectly flat from
100 Hz to 12 kHz with gentle rolloffs
to 40 Hz and 15 kHz. Total harmonic
distortion is typically 1/2 percent.
System dynamic range is 90 dB when
"Dynex" (transmit compression and
receive expansion) is incorporated,
with a resulting low noise floor.
Cetec Vega hand-held wireless
microphones are newly redesigned
for 20 to 30 percent additional battery life, using a commonly available
9 -volt alkaline battery (Duracell recommended). Microphone sensitivity
is easily adjustable with an audio
Circle
16
on Reader Service Card
gain control on the bottom, with an
adjacent LED indicator to verify
optimum setup. Power and audio
on/off switches are also conveniently
located on the bottom.
Write or call for further information and location of your nearest
dealer: Cetec Vega, P.O. Box 5348,
El Monte, CA 91731. (213) 442 -0782
TWX: 910- 587-3539
In Canada:: I.C. Simmonds & Sons Ltd.
t Cetec Vega
Division of Cetec Corporation
BARRY BLESSER
Digital Audio
Digital Filter Design: Part
l'he design of digital filters has generally been left to mathematically-ori ented engineers, since a high level of technical expertise is necessary to build such
systems. This contrasts with analog filter
design, where readily available tables of
circuits and values can be used to construct filters. As time marches on, digital
filters will also become the province of
practical engineers without the mathematics. In the next series of articles we
will try to develop some of the ideas necessary to digital filter design.
I
Before beginning, I would like to note
that the mathematics actually makes the
design task much simpler rather than
much harder. Therefore, I'll try to introduce mathematics (gently!) when necessary.
LIN EAR ITY
All filters, whether digital or analog,
are based on linearity, which is a mathematical abstraction. It simply means that
if I place a time signal x(t) into a system
The Dream Equalizer:
Now mono or stereo.
When we introduced our 672A "dream equalizer" in
1979, we had an instant hit. Audio professionals loved its
versatility and clean sound. Eight parametric EQ bands
(with reciprocal curves) were combined with wide -range
tunable 12 dB /octave highpass and lowpass filters.
The result: an amazingly powerful and useful machine.
A cost -saving one too, because the outputs of both
filters are available to perform a full electronic crossover
function.
The 672A now has a stereo twin -the new 674A,
with all the power of two 672Ás in a space- saving 51/4"
rackmount package. Naturally, both equalizers are built to
full Orban professional standards. That means industrial quality construction and components, RFI suppression,
a heavy -duty roadworthy chassis, and comprehensive
backup support.
orbeA
Associates
CV
r(t) comes out, then if 1 place 2x(t)
into the system, 2y0) will come out;
doubling the input doubles the output.
Similarly, multiplying the input by any
constant will result in multiplying the
output by the same constant.
A further extension of linearity says
that if xi(t) produces ßi(0) and x_(t) produces 1'2(t), .r2(0) will produce pi (0) +y2(1).
Most audio systems approach this mathematical notion of linearity. Increasing the input signal increases the output
signal. Adding a trumpet and a violin at
and
Orban
Inc.
645 Bryant St. San Francisco, CA 94107
Telex: 17 -1480, Cable: ORBANAUDIO
Circle 21 on Reader Service Card
To the audio professional, when a
compressor or limiter is needed to
tame the potentially disastrous
consequences of uncontrolled level
or to create special effects, one
name stands out as the best: UREI.
Studio Standards for more than
a decade, the compressors and
limiters from UREI have earned their
way into thousands of recording,
mastering, and broadcast installations around the world.
Because we built our reputation
for unparalleled professional
performance and quality with our
compressors and limiters, we have
continuously advanced their
engineering and technology to offer
more reliability, features and
performance. When you need the
fastest, quietest and most flexible
gain control instruments available,
you can be totally assured that these
products will prove to you why
they've earned the title Studio
stereo coupling. The 1176LN is the
most widely used limiter in the world.
The Model 1178
A two channel
version of the
1176LN in a
co
compact
(3 -1/2 ) rack
mounting design. Featuring perfect
tracking in the selectable stereo
mode, it additionally offers
selectable VU or Peak reading meter
ballistics.
From One Pro To Another trust all
your toughest signal processing
needs to UREI.
The UREI
Compressor/Limiters
-
-
Standard:
The Model LA-4
A single channel, half -rack unit with
patented electro- optical attenuator.
Featuring smooth, natural sounding
RMS action, it offers selectable
compression ratios, a large VU
meter, adjustable output and threshold levels and stereo coupling.
The Model 1176LN
A peak limiter which features
adjustable input and output levels;
individual attack and release time
controls; selectable compression
ratios; switchable metering; and
CHANNEL
g
L al From One Pro To Another
United Recording Electronics Industries
8460 San Fernando Road. Sun Valley. California 91352 (213) 767 -1000 Telex: 65-1389 UREI SNVY
Worldwide: Gotham Export Corporation. New York: Canada: E.S. Gould Marketing, Montreal
See your professional audio products
dealer for full technical information.
Multiplier
x111)
J
Figure 1. A multiplier with an introduced
time variable.
the input produces trumpet- plus -violin at
the output.
7 -he
mathematical notion of linearity
an abstraction because it assumes absolutely no distortion. But we know from
practical experience that all systems will
have some distortion. even if it is very
small. This is the real difference between
the mathematician and the engineer. The
mathematician can imagine pure systems, the engineer knows that all real
system have defects.
is
you write it
Many readers do not realize that
they can also be writers for db.
We are always seeking meaningful articles of any length. The
subject matter can cover almost
anything of interest and value to
audio professionals.
You don't have to be an experienced writer to be published.
But you do need the ability to
express your idea fully, with adequate detail and information. Our
editors will polish the story for
you. We suggest you first submit
an outline so that we can work
with you in the development of
the article.
You also don't have to be an
artist, we'll re -do all drawings.
This means we do need sufficient
detail in your rough drawing or
schematic so that our artists will
understand what you want.
It can be prestigious to be published and it can be profitablé
too. All articles accepted for publication are purchased. You won't
retire on our scale, but it can
make a nice extra sum for that
special occasion.
Figure 2. For filter design, an audio signal
may be divided into a series of small
slivers, x(t).
Now, when you shake,
rattle and roll,
your microphone won't.
The Shock -Stopper'" from Shure.
The amazing Shock- Stopper ''
microphone mount from Shure.
Whether you're on stage or anywhere else. this microphone mount
significantly reduces mechanical
and vibration noises.
It is available in both a half-circle
version -literally a shock -mourned
swivel adapter-for quick removal.
cores in a full -circle design for
vibration-free studio or stage
instrument miking at any angle.
For more information on the A»
and A55 Series Shock -Stoppers:"
visit your Shure dealer, or call or
write Shure Brothers Inc.. 222
Harney Ave.. Evanston. IL
60204. (312) 866 -2553.
Or, it
sNURE*
THE SOUND OF THE PROFESSIONALS"..
WORLDWIDE
Circle 22 oli Reader Service Curd
)V
Audio engineers who try Maxell
won't let go.
Maxell quality woes a lot of
recording situations. N. laxell meets
your ';i " open reel and audio cassette
needs, no matter hods demanding
You can see Maxell excellence in
the cassette construction and on the
'scope or meter. The physical construction is strong enough to meet
you are. Because we're more
all professional requirements. Maxell
open reel tape and cassettes give
demanding. We've developed a
name that means unique quality all
you quality you can hear. And your
around the world. For example,
clients can hear as well.
Maxell cassettes give you a producWe'll give you all the technical
tivity boosting four-function leader
information you need to form your
with A/B side indications, directional own opinions. But if you're like just
arrows, non -abrasive head cleaner
about every audio professional that
and five -second cue to set timing
tries Maxell, you won't let go.
and level.
Remember, we warned you!
maxell
PROFESSIONAL /INDUSTRIAL DIVISION
F.;
3
IT
(D
Our success
Maxell Corporation of America, 60 Oxford Dr., Moonachie, N.J. 07074 (201) 440-8020
Circle 16 on Reader Service Card
.
is magnetic.
TIME -INVARIANCE
Like linearity, filters are also based on
the notion of time -invariance. This is a
more difficult concept. It means that if 1
place the signal x(t) into a system and
y(l) comes out, then the signal x(1- T) as
an input will produce y(t -T) as the output. In simple English, the system will do
the same thing at 9:00 o'clock as at 10:00
o'clock. If the violin appears at the ampli-
(A)
(B)
Figure 3. A typical filter output, for any
of the input slivers seen in Figure 2.
fier input I hour later, then the output
violin will also appear hour liter.
To use an analogy, a trip by automobile might be considered a time invariant process; if 1 enter the automobile I hour later, I will get to my destination I hour later. In contrast, an airplane trip may not be time invariant since
the schedule is fixed. Arriving at the airport I hour later may result in ether arriving at the same time (if I can still catch
the same plane), or arriving 3 hours later
(if I miss that plane).
Another way of looking at this issue
is by considering if time is local to the signal or is an absolute external reference.
The automobile uses local time; the airplane uses absolute time (schedule).ln
FIGURE 1, the multiplier is not time invariant since the relationship between
the input and output is dependent on the
time variable cos(wt) in the multiplier.
The output is
1
With an API totally modular console by Datatronix, you
have unlimited flexibility in size and function. Each console
is custom designed ... each component individually
selected by you. After all, the choice
is yours ... you design the console to fit
For
your needs today and tomorrow.
example, you can choose
from five different, interchangeable, plug-in EC)
modules and an array of input and
auxiliary modules -all assembled into
a mainframe designed to your requirements. As your needs change, so can the modular components.
Plus ... each console may be prewired for expansion, so that
you can easily expand your capability at minimum cost!
The API console by Datatronix offers you the quality of
sound, performance, and flexibility you demand as a professional recording engineer. So call us today... and remember
the choice is yours.
With an API
modular console
the choice is
yours
Performance you can count on
y(t) = x(t)cos(wr).
If the input is delayed by one time unit
t hen
t(1 -1) # x(t- I) cos (wt).
Thus, a multiplier is generally not a timeinvariant process. Linearity and timeinvariance are the only mathematical
concepts necessary for filter design.
CONVOLUTION
Io see how we might apply these concepts to filter design, we will begin by
considering a piece of an audio signal
x(t) as shown in FIGURE 2. 1f we wish,
we can break this signal up into tiny
slivers, each of which is very narrow. This
allows us to say that
x(t) = xi(t) + x2(i) + x3(t)...
Why have we done such an operation?
The answer is that we will find an elegant
way of looking at filters and some interesting properties of filters. To see this,
we need to follow certain arguments.
First, break the input signal x(1) into a
set of signals x,(t), x2(:), x3(t).... Mathematically we write this as x(t) = 2 x,(t),
where i = 1,2,3....
Now, consider a filter which produces
an output h(i) when a single input sliver
appears at the input. This is illustrated
in FIGURE
subsidiary of ATLANTIC RESEARCH CORPORATION
(703) 620 -5300 TWX 710 833 0365
2100 Reston Ave., Reston, VA 22091
3.
A
to
Circle 23 on Reader Service Card
The single sliver which produces the
output h(t) completely defines the filter!
Moreover, we can find the filter's response to all inputs just by knowing the
response to this one case. To do so, we
need the two principles of linearity and
time -invariance. Linearity says that if
know the response of the system to xi(t)
and to Mt), then I will know the response
to .vi(i) + x2(t). In other words, I can get
the response of the filter to the complete
signal by adding the responses to each of
the slivers.
The response h(i) to the first sliver is
scaled by the amplitude of that sliver.
The second sliver is like the first sliver
except that it has a different amplitude
and it appears I time unit later. Hence,
its output must also be scaled and it must
appear
I
TYPE VS IMPLEMENTATION
When we talk about designing a filter,
we must be caret ul to separate the design
of the impulse response from that of the
implementation. Both are important.
invariance arguments are necessary for
the solution.
In our further discussions we will refer
to a filter only in terms of its impulse response rather than its implementation.
For example, the impulse response of the
circuits in FIGURE 5 is given by
Sometimes a given impulse response is
extremely difficult to implement. The
impulse response above has a certain frequency response. We must ask, "Is that
the response we wish ?" If it is, then we
must ask about a good way to implement it.
r(1) = At
Here we have two different implementations with the same impulse response.
Hence, they are identical from a mathematical point of view. They will produce
the same outputs for the same inputs.
One may be better or worse from an engineering point of view but not from a
mathematical point of view. The impulse
response is thus a very compact notation
yll)
x(t)
for describing filters.
unit later.
Figure 5. Two circuits which give the same
impulse response.
xlhlt -I)
When somebody says that he needs a
9th-order Bessel filter, he is using the
language of type with no regard to implementation. Similarly, he could have
specified a 19th -order elliptic. As it turns
out, this one is very difficult to implement because it requires a stage with very
high Q. Mathematics does not concern
itself with a Q of 500. However, the
mathematics to compute the impulse
response of a 19th-order elliptic may also
be difficult. Many computers do not
have enough accuracy for such computations. Even 20 decimal digits may not
be enough. Nevertheless, the difficulties
2)
h(t
- 3)
OUTPUTS
Figure 4. The filter outputs for
input slivers.
a
series of
are not directly comparable.
1 his article has mixed the discussion
FIGURE 4 shows this process. On the
left we have the input signal x(t) decomposed into its slivers. Each sliver has an
amplitude corresponding to that part of
the signal from which it was taken. On
the right we have the response to that
sliver. 1 his response has an amplitude
which is determined by the input sliver
amplitude and a time function which is
Delay
SI()
-.1,
1hlay
iu
0.5
h(t) delayed by the delay of the sliver. The
x4(t) sliver begins 4 time units after 0,
hence the output response must also
begin 4 units later. This is the time -invariance argument. 1 he linearity argument says that the sum of the signal
responses on the right must be the actual
signal which would come from the full
input.
Mathematically, this summation can
be
written
YU)
as
!YU =
[xi)
h(t J)J
There are several interesting things to
note about this. First, the response to any
input signal can be determined if we
know the response to a single sliver. This
is called an "impulse response," since the
mathematical definition of a sliver is an
impulse. Secondly, by computing the
above summation, we have a systematic
way. of determining the response to a signal This activity is called convolution.
Unfortunately, solving this summation
is usually very difficult. Fortunately, it is
only necessary to appreciate that it could
be solved and that the linearity, time-
0.25
(B)
Figure 6. Two practical filter circuits
which produce the same impulse response.
The same argument can be made for a
digital filter. FIGURE 6 shows two different implementations which have the
same impulse response.
We can say that these are identical.
From an engineering point of view, however, there are subtle differences relaing to truncation noise in the multiplier.
between analog and digital indiscriminately because the issues are the same. Just
as a resistor, capacitor, and inductor are
linear -time-invariant elements, so are
scaling, delay, and addition. The kinds of
impulse responses which can be achieved
are different. The mathematics are the
same.
t
KEN POHLMANN
theory Itt Practice
Babylonian Astronomy and Digital Filters
Mathematicians went through a lot
of chalk in the eighteenth century. Following the invention of calculus by Newton and Leibnitz, there was a burst of
activity in topics of mathematical physics. There was a critical need to better understand physical phenomena with
mathematical precision, a need that has
continued to this day in every profession
-yes. even audio. One of yesterday's
hot topics of controversy continues to be
of critical importance today: the method
of representing complex functions, such
as waveforms, with simple functions.
such as sine waves.
The boundary -value problems of vibrating strings. and bars or columns of
air, led eighteenth- century scientists to
associate mathematical theories with
musical tones. It seemed natural to think
of functions arranged much like musical
tones are as mathematical multiples of each
other. the ideas seemed simple enough.
and they were. but to reach an under-
standable result required the unification
of diverse research in vibrating strings.
planetary action, and finally the conduction of heat. Even then, the conclusion
reached was met with skepticism and tday, it still sometimes seems to breed
more confusion than enlightenment. But
if math geniuses fumbled over it for fifty
years. what do you expect from us
mortals?
BATTLING SCIENTISTS
Scientific hcaoeight, d'Alembert.
Euler and Bernoulli were aggressively
concerned with the study of the vibration
of a string fixed at both ends. Both
d'Alembert and Euler obtained equations in essentially identical forms which
explained the string vibration as the
superposition of two waves travelling in
opposite directions -you pluck a string.
and it's actually two opposite waves. But
d'Alembert thought the initial forni of
the string could be given by a single ana-
Only Swintek
GIVES YOU A CHOICE
in wireless microphone systems
Choice of mics -Beyer M500,
Shure SM57, SM58, SM78, SM8
Choice of finishes -chrome,
black chrome or gold
Integral antenna or
rubber duckie
Systems for hand held mics,
lavalier mics, electric instruments or ENG/EFP/LPTV
Unsurpassed field -proven
reliability assures the high audio
quality that professionals demand. VHF
high band and narrow RF signal bandwidth
permit numerous systems to be simultaneously
used on nearby frequencies without interference
or receiver desensitizing by local Nor radio
broadcast signals. All metal construction
assures durability and RF shielding. dB -S companders
permit wide dynamic range without unnatural compression.
Swintek -the proven performers.
hW%fde/l
TELECOMMUNICATIONS DIVISION
1180 ASTER AVENUE. UNIT J SUNNYVALE. CA 94086
(408) 249 -5594 TELEX 172 -150 SWINTEK SUVL
Circle 24 on Reader Service Card
lytical expression. whereas Euler thought
of it as a continuous curve with a different analytical expression for each part.
Then Bernoulli gave the solution in the
form of a trigonometric series which.
being general. subsumed the other theories. Euler got excited, and wondered if
any function could be expressed as an
infinite series of multiple sine waves: he
decided it would be impossible- sines
are periodic and odd, they would only
work with periodic and odd functions.
Then a young and upcoming mathematician. Lagrange. in defending Euler
from d'Alembert, proposed a new look at
the problem. He hypothesized that a
string could be considered as an infinite
number of particles stretched on a
weightless line. He solved the associated
equation with a series of sines and cosines. If he had taken one step more, and
changed the order of summation and integration in his solution, then instead of
talking about Fourier functions all the
P
R
O
F
E
S
S
I
O
N
A
A
L
U
D
I
O
/
V
I
D
E
O
D
I
V
I
S
I
O
Right On The Money:"
Harvey presents the Soundcraft SCM762
24 Track with nine position autolocator.
from $19,050.
4 Track from $21,250.
16 Track
The true job of any pro sound dealer is to help you find
the equipment that is exactly right for your needs. Harvey thinks
Soundcraft may be just that kind of equipment. Designed to
eliminate costly manufacturing methods, Soundcraft equipment
nonetheless preserves every last ounce of the sound quality
and reliability you demand in your operation. Offering outstanding specs, ready user and
service access and rock -solid construction make the SCM762 -24 the right machine at the
right price from the right dealer. Harvey.
HLL FREE:
1-
25
W.
800-223 -2642
45th St., NEW YORK, N.Y. 10036 (212) 921 -5920
(
in le
2N
,n Reader- Serri(
(
iart/
N
time. we would be calling them Lagrange
functions- but that gets ahead of the
story.
Separate from their disagreements
about vibrating strings. Euler and
d'Alembert were also disagreeing about
take a deep breath- --the astronomical
problem of the expansion of the reciprocal of the distance between two planets
in a series of cosines of multiples of the
angle between the radii. Both men presented the idea of using a definite integral
to find the series coefficients, that is, the
constants that specify which simple functions represent the complex function,
and Clairaut published a paper detailing
the integrals needed. Still. no one thought
of linking these new integrals with the
work on vibrating strings, and the theory
-
of trigonometric series remained unresolved.
FOURIER TO THE RESCUE
I hen.
in 1811. Fourier presented a
paper to the Paris Academy on his
Mathematical Theory of the Conduction
of Heat -- Theorie Analttiyue de la Chaleur-still a classic in heat conduction.
He proved that certain simple functions
which he needed to explain the conduction of heat can be represented on a
bounded interval by series of sine and
cosine functions and, perhaps more importantly, he asserted that any piecewise
smooth function can be expanded into a
trigonometric series. This was nothing
new, but he was the first to assume almost
carelessly, that a series could be found for
any arbitrary function -and that was his
distinct advance. Everyone was surprised, and skeptical -even Lagrange
flatly denied that it could be done. Fourier was met mostly with scorn, and he
left it to others to later prove his contentions--he was more interested in applications and methods, not conditions
of validity. It wasn't until Dirichlet presented a proof in 1829 that Fourier's
theory became secure.
So, Fourier finally nailed down the
idea that a continuous function can be
represented by an infinite trigonometric
series. Calculator -brains out there will
clearly recognize the famous result showing the infinite series of sines and cosines:
x
JO) = A. +
(A. cos not + B. sin not)
11
SOUND
UNSEEN
THE BEYER MCE 5
Voice and appearance are
key to a news announcer's
success. Beyer benefits both
with its MCE 5 electret condenser mic. A matte -black
omni so tiny it's virtually invisible. Full 20- 20,000Hz
range projects the vital presence of the speaker, un-
colored, undistorted. Unique
floating element eliminates
clothing noise. Removable
windscreen for outdoors.
62dB S /N. 116dB overload.
700 -ohm balanced output
for professional performance. Phantom or selfpowered, also available in a
variety of connector configurations to meet every
application requirement,
including
wireless.
Visit your
ACTUAL
local Beyer
SIZE.
dealer today.
Suddenly math was a lot easier, and
complex functions could be quickly
analyzed and better understood. Ever
wonder just what's going on inside that
trumpet tone? Fourier will tell you evert.thing (almost) you want to know. That
complex periodic function can be analyzed as harmonic component frequencies -the trigonometric form is the Fourier series for the function, and the process of determining the values of the
constants is called Fourier analysis.
The values of the constants can be
informative in themselves. Symmetry
around the x-axis will give a zero value
for the constant. if the function is even.
all the sine terms will be missing, and if
odd, the cosine terms will be zero. The
familiar formulas for determining the
constants are probably also easily remembered by calculator- brains (check a
deeper memory level):
A. _
An
¡
=
B = -.
f' ///) dt
51.1(1) cos
51
fit) sin
not dt
mot clt
o
Let's look at an example, and one difficult in the respect that it requires a long
series to be accurately represented. The
faniliar saw -tooth waveform can be defined mathematically like this:
J(U
=a(I-
7'l
substituted into the
three equations to determine the coefficients. we find the results:
If that function
is
A =O.A =a
l3
=a
na
And the complete harmonic series representing a saw -tooth wave can be written:
J(t) = - (sinwt + , sin /rut +...
beyer dynamic))))
BEYER
DYNAMIC, INC.
-05 Burns Avenue. Hicksville.
NY 11801 (516) 935 -8000
5
o
(V
Circle 29 on Reader Service Card
ti
sinnest)
That might not look like much, but it of
fers a relatively easy way to mathematically represent a saw -tooth sound.
And that's a real opportunity -instead
of working with the actual signal, we can
work with the numbers that comprise
it -just like a digital system.
There's a lot of information contained
The Basé 402 Loudspeakers
The professional standard
for small -group
P.A.
systems.
ow Bose brings the
IV advanced technology of
the Articulated Array"' system
to a speaker designed spe-
cifically for high -quality reinforcement of vocals and
acoustic instruments. The 402
Professional Loudspeaker. An
affordable and truly portable
alternative to the traditional
small -group PA. speaker.
Each 402 enclosure contains four 41/2" high-sensitivity
drivers mounted on a faceted
3- dimensional baffle. This
unique Articulated Array"'
system works together with a
special Acoustic Diffractor
and built -in Directivity Control
circuitry to deliver exceptionally uniform room coverage,
without the penetrating shrillness of horns or the
muffled sound of
columns.
Tuned Reactive
Radiator slots allow
the 402 speaker to
produce surprisingly
high output levels
with low distortion. The Per
matched 402 -E Active
Equalizer assures
smooth, accurate spectral response across the
entire operating range
of the system. And the
TK -4 Transit Kit lets
you clamp a pair of
402 speakers (with
equalizer) together
into a compact unit
light enough to
carry in one hand!
The 402
Loudspeaker makes
it easy for anyone to
obtain the outstanding performance of Bose's Articulated
Array"' system. Ask for a live
demonstration at your
authorized Bose Professional
Products
sçi
Ott
.r
++
E Bose Corporation, Dept MR
7
The Mountain
Framingham. Massachusetts 01701
Please send me a copy of the Bose
Professional Products Catalog and a
complete dealer list.
_ dealer.
Name
Address
City
State
Zip
Telephone
L
Better sound through research.
J
N
if a function consists
certain number of sine and cosine
vibrations, then the analysis will contain
only that many terms. Analysis of a beat
tone will contain only two terms. The
more complex the vibration, that is, the
more abrupt the discontinuities, the more
terms needed to accurately represent the
functions. In the case of a square wave or
saw -tooth, an infinite series must be
considered to gain equivalence. Practically speaking, acoustical vibrations
are usually fairly well behaved and convergence is rapid. Only a relatively few
number of terms must be considered. In
the case of the saw- tooth, perhaps ten or
twenty terms would add up to the same
vibration, sharp teeth and all. Of course,
it should be mentioned that the Fourier
theorem can be used in reverse too. A
linear combination of simple vibrations
which have commensurable frequencies
forms a complex vibration at a greatest
common divisor frequency: we can synthesize complex tones from simple ones.
Need a string section? How many sine
wave generators on your test bench?
here. For example,
of
a
Don't laugh- Stockhausen realized
some of the world's greatest electronic
music with that trick. And more than a
few digital synthesizer manufacturers
have picked up where he left off.
That idea of a harmonic and partials is
an incredibly important one. as anyone
who has ever heard music will attest.
Consider the fact that if a low frequency
pure tone is sounded first as a low and
then a high loudness level, most perceivers will state that the second tone
has a lower pitch, despite the fact that the
frequency has not changed. However
for complex tones (say, from acoustic
musical instruments), that perceived
change in pitch is much smaller because,
as Fourier analysis reveals, even if the
fundamental lies in a pitch range subject
to that decrease, the harmonics will have
frequencies for which the pitch changes
very little, or perhaps increases. That annoying dependence of pitch on intensity
is compensated for by a dependence on
waveform.
TIME AND FREQUENCY
Finally, let's mention the Fourier
duality of time and frequency. We can
display a function in time, and we can
also look at it by frequency, that is. as a
spectrum of the time signal. The process
of Fourier transformation allows
us to
freely switch from the time domain to
the frequency domain and back. Those
transform functions are incredibly important. Once again, information is the
name of the game. It was in the eighteenth century and it still is- only more
so.
That ability to switch from one domain to the other as applied to digital
implementation -- remember, that
is
where numbers derived from our mathematically precise methods can really be
utilized -allows us to approximate con-
tinuous transform functions with DFT
(Discrete Fourier Transforms). The high
speed algorithm (great for high speed
computers) for computing the DFf is
the FFT (Fast Fourier Transform). We
can use that FFT to find out all kinds
of thngs about a waveform: How about
instant spectral analysis? Easy. Or, if we
smooth the data with a window function
before we compute the FFT, we could
digitally filter the function. And that is a
great idea. The filter is dynamic --it can
be programmed. Any time an analog
function has been converted to digital
form, these Fourier methods arc ready to
analyze and process- the chance for
some extremely complex and accurate
audio processing is at hand. Those old
phase -shifting. distorting equalizers
made of resistors. capacitors and inductors? Get rid of them. It can all be done
with beautiful. absolute analytic precision and grace.
So --the discussion of quarrelling
eighteenth century scientists led to a general theory for representing functions.
which makes it easy to mathematically
analyze and process those functions.
which has led us to digital filters to the
promise of the digital mixing console.
Oh if only for the sake of completeness, the ancient Babylonians deserve
most of the credit. Thousands of years
ago. their astronomical computers were
already using summations of sines and
cosines to predict celestial events.
THE UI
Did you ever dream of a mike that never
lies? Never adds, never subtracts?
Not even at 130 dB SPL?
Well, this is it
LC 25. The transformerless
line level condenser.
The only one of its kind.
MiLAB
Manufactured by CTAB, S -265 00 Astorp, Sweden
PROFESSIONAL MICROPHONES
BELGIUM Trans European Musk 02 569 18 23 FINLAND Studiotee 90 556 252 FRANCE MCI France
227 2595 GERMANY EMT -Frans 07825 1011 HONG KONG Slider Rewx 5 -412050 JAPAN Continental
NEW ZEALAND General Video 4-872 -574 NORWAY Pro -Technic 02- 460554 SPAIN
Singleton Production 2283800
SWITZERLAND Audio Bauer 01 64 32 30 UNITED KINGDOM AVM Audio Video
Merlreti g 0632 893092 USA we Cam International 213 821 7898 USA south Swedeks 404 881 9981 USA east
Fad 212 370 0032
-
1
Fr last 03 583 8451
-
Circle .30 on Reader Service Card
4'
-
-
In demonstrating our microphones throughout the country,
we've found a serious limitation
in most stage mixers. They are
unable to handle wide range microphones on stage. And they
just can't cut it when it comes to
making demo tapes. Which means
that the musicians need TWO
mixers and perhaps TWO sets of
microphones to get the sound
they want on stage as well as on
tape. It's a luxury not everyone
can afford!
So, to solve your problem and
ours we set out to create a
"double threat" mixer which
would be a great stage mixer, yet
still give you the sound and control you need while taping. A
mixer designed to take full advantage of every mike you own, including phantom -powered models.
Our standards (like yours) were
high. Everything had to be rugged, reliable, and very clean. With
wide basic frequency response,
plenty of headroom, and very low
distortion and noise. And the
mixer had to be very natural to
use. Finally, the price had to be
right. We invite you to examine
the new Audio-Technica ATC820
and ATC1220 stereo mixing consoles to see how well we have
-
-
Our prototypes have done a lot
of traveling. Users were im-
pressed with the features, the
flexibility, and the sound. They
liked the 3 -band EQ on every input.
And the 7 -band stereo graphic
program equalizers, plus another
graphic equalizer for the monitor
output. But most appreciated
were the variable high -pass filters for each output. They permit
you to use wide -range recording
microphones on the stage, while
exactly limiting bass response to
suit acoustics and to keep from
overloading your speakers. Yet
during recording you can go all
the way down to 20 Hz if you wish.
There's a long list of very practical features. Phantom power is
available at each of the transformer- isolated mike inputs. Two
20 dB mike input pads plus an
LED to warn of clipping on each
input. A SOLO button to check
any input with headphones without affecting the mix. "Stackable"
design when 8 or 12 inputs aren't
enough. Even an assignable talk back input. And all the logical
controls for the transformer balanced MONITOR, EFFECTS,
SOLO, PHONES, and OUTPUT
busses. In short, very flexible, and
quite complete.
With a very modest investment, you can do almost everything the single-purpose boards
can do ... and do it very well. And
get the benefit of phantom powered recording mikes on
stage as well as during recording.
The more you learn about the
ATC820 and ATC1220 the more
impressed we think you'll be. Ask
your Audio -Technica sound specialist for a hands-on tour of this
brand new breed of mixer. Or
write for literature. We may have
the ideal answer to your mixer
requirements.
accomplished our goal.
AUDIO -TECHNICA U
S
.
INC .1221 Commerce Drive, Stow, Ohio 44224
Circle 27 on Reader Service Card
(216) 686 -2600
LEN FELDMAN
Sound with Images
Pro Video Borrows Consumer Tape
For New VTR Format
ast April. RCA's Commercial Communications Division introduced a new
video recording format. They also
submitted the new system as a proposed
standard to the SMPTF. (Society of
Motion Picture and Telex ision Engineers).
One aspect of this new system is its use of
half-inch videotape. housed in standard
250 meter (T -120. 2 4 6 hour) VHS
format videocassettes. RCA calls the new
system Chroma 'Trak. and they are
already using it litr ENG (Electronic
News Gathering) and other video recording applications where light weight and
portability are important factors.
The group of products related to the
new Chroma Trak system are part of
RCA's new "Hawkeyc" product line. As
illustrated in FIGURE
a pair of H R -2
Hawkeye videotape recorders are controlled by a Model HF -I edit controller
in the studio. A combined camera video
recorder. Model HCR -I (not shown). for
in- the -field use has also been developed.
as have separate cameras and portable
video recorders using the new format.
I
I.
WHY NOT VHS OR BETA?
v
N
It you aren't intimately involved with
videotape recording. you may be wondering why professional video production houses didn't simply modify either of
the two popular consumer videotape
formats to suit their needs. After all.
VHS and Beta VCRs also use half-inch
wide tape and many of the new portables
now available for these popular systems
are amazingly sophisticated and, at the
same time, light enough to meet the needs
of video professionals who must often
carry their equipment for long periods of
time and into inhospitable environments.
While both VHS and Beta VCRs are
amazingly intricate and technologically
advanced video systems, the truth is that
the picture and sound quality that they
deliver is simply not good enough for
broadcast use. Remember, in broadcast
applications the videotape that's actually
sent out over the air may well be a few
generations removed from the original
tape shot by the news crew, which only
serves to further degrade the quality of
both picture and sound. If you doubt it.
just look at the picture produced by a
2 -inch wide videotape system (or even the
'/a -inch U -Matic system used by most
small professional video production
houses) to appreciate what good video
color pictures can look like -even when
limited by the U.S. NTSC standards.
Figure 1. Products developed by RCA for
the Chroma Trak video recording system
include the two recorders shown flanking
an edit controller.
Better still, audition a laser -type video
luminance (brightness) registration is
disc. using a good TV monitor and you'll
improved by a factor of more than 3 -to -I.
see
pictures having
a
video frequency
response up to 4.0 M Hi or more. as
against response that rolls off quickly
above 2.0 MN/ in both the Beta and the
VHS home videotaping formats.
CHROMA
TRAK'S ADVANTAGES
Even compared with /,-inch U- Matie.
the color quality of the new Chroma Trak
system is dramatically better. Freedom
from color noise, streaky colors and even
improved color purity is very apparent.
Chroma Trak has twice the informat ionpacking density of U- Matie. Translated
to visual terms, this greater packing
density means greatly improved picture
quality. Chrominance (color) resolution
is more than three times as great as in a
3A -inch video system. Chrominance
signal -to -noise ratio is improved by
approximately IO dB. Chrominance-toAudio
2
Audio
I
track
Audio
3
track
Luminance small -image detail is maintained. affording a sharpness of picture
not previously obtainable except in widetape studio machines. Perhaps best ()fall.
the Chroma Trak signals are designed to
be recorded on standard. familiar VHS
cassettes which are readily available and
reasonably priced, compared with other
exclusively-professional tape packages.
HOW CHROMA TRAK WORKS
Basically a dual -rrai k helical -scan
system. Chroma Trak records the
luminance video signal on a track 7 mils
wide, while the color signal is recorded on
an accompanying, parallel track 2.5 mils
wide. These side-by -side tracks are laid
down by a pair of heads mounted very
close to each other (See FIGURES 2 and
3). Each track is 3.7 inches long and.
taken together. they contain information
rack
Control track
Reference cdac
Figure 2. Track layout for Chroma Trak
format.
= III =111= III =III=
II-
=111 =I
111111=
1_111 =111= III
=III=111
111= 111
111
._
.
=III=
=III
I =111= III = III =III111 = III = III=''
11=111 =1'
11
=III=
SCONIEX
FOR CLEAN SOUND
After WLEE installed
SONEX, people called about
the "clean, new sound!" Was it
new equipment?
No. The answer is simple
and inexpensive:
SONEX kills
all the stray
reflections, so the mike
processes a clean signal of the
announcer's voice. The
announcer hears himself in a
warm environment, and there's
COLOR
OPTIONS
AVAILABLE
less clutter to waste watts. The
result is a noiseless, clear
signal and greater market
penetration.
Get the facts today. SONEX
is available in five colors, in
three thicknesses, and in
standard four -ft. square
panels. Use it in broadcasting,
recording, videotaping or
performing areas for really
pure, clean sound.
Alpha
Audio
2049 West Broad Street
Richmond. Virginia 23220 (804) 358 -3852
Acoustic Products for the Audio Industry
Sonex is a product of Illbruck /USA
Circle 32
cm
Reader .Serrire' Card
N
cn
for one complete field of video. The tape
is wrapped 180 degrees around a drum
which is rotating at 29.97 revolutions per
second. The drum is 2.44 inches in
diameter. There are four longitudinal
tracks in addition to the angled video
track -pairs. Tracks I and 2 are used for
audio (stereo is taken into account), track
3 may be used for time code for post production editing and audio synchronization. and track 4 is the control track.
Now, here comes the important
additional difference between consumer type VHS video and the Chroma Trak
system. Longitudinal tape speed of the
new system is 8 inches per second or
more than six times as fast as a VHS tape
system operating at its fastest (SP'2
hour) speed. A T -120 videocassette will
therefore provide about 20 minutes of
record play time. Since the main purpose of the system is for use with cameras
(as in remote news-gathering operations),
this should not prove too much of a
problem. And in any event, if more
recording time is needed in the field, the
cameraman can simply pop in another
VHS cassette. All of the remotely -shot
videotape will be electronically edited
before broadcast anyway.
Figure 3. The drum containing two pairs of
luminance and chroma heads rotates in the
same direction as the linearly- moving tape.
MAKEUP OF SIGNAL ON THE TAPE
Effective head -to -tape "writing speed"
works out to be about 225 inches per
second and, since the Chroma Trak
system boasts a recording capability of
27.000 Hz per inch. that means a
bandwidth capability of slightly in excess
of 6 M Hz! W hen tied in with a "Hawkeye"
camera recorder system, pure luminance
(Y as well as separate chroma I and Q)
signals are derived directly from the
camera encoding matrix system. When a
camera is not used, a comb -filter type of
and Q
decoder derives separate Y,
signals. The three separate signals are
converted to frequency modulation prior
to being recorded onto the tape. The Y
signal causes FM deviation from 4.3
M Hz to 5.9 M Hz (peak white signal) and
occupies a band which extends out to
about IO MHz. Because this signal is
purely monochrome, strong color sub carrier systems which have caused moire pattern interference in other video
recording schemes are totally absent.
The
and Q color signals are also
converted to FM. The signal deviation
or modulation is from 5.0 to 6.0 MHz.
PICTURE QUALITY COMPARISON
lla"Lee
Chrominance Resolution
Chrominance Noise -SNR
Chrominance Registration
3rd Generation
Figure
5.
of 0.79 ips.
The chart shown in FIGURE 4 shows
how video recording densities have
improved since the first video recorders
were produced more than two decades
ago. Recording density is shown in cycles
per inch of tape -to -head writing speed.
and the new RCA Chroma Trak system is
compared with the popular U -Matic 34inch tape system and with earlier systems
such as Type C. Quadruplex and Quad.
The table shown in FIGURE 5 presents a
speed
30.000
1
20.000
r,.ouu
10,000
7000
o
Quad
C- format
,
inch
Ilawkeye
1
1
co
N
'.
0.3
inch
MHr
48 db
38 dB
90 nsec
300 nsec
Picture Quality Comparison.
while Q color signal deviation is from
0.75 to 1.25 MHz. After suitable
preemphasis these signals are combined
by simple addition, and fed to the color
track of the tape. The I FM signal is at a
level high enough to serve as the record
bias for the lower-level Q signal. which is
recorded linearly.
Several of the problems that have
plagued previous videotape recording
systems are eliminated with this system.
We have already mentioned the elimination of the moire -pattern thanks to the
fact that the luminance and chrominance
channels are separately laid down on the
tape. In addition. since there is no sub carrier on the color track to be time modulated by jitter. there is a complete
absence of visible noise streaks often seen
on other video recording systems.
Furthermore, since differential phase
and gain are a result of intermodulation
between chrominance and luminance
signals, the Chroma Trak system, which
separates luminance and color signals.
does not suffer from these additional
signal faults. Although RCA's description of the Chroma Trak system says
nothing about audio fidelity of the sound
tracks of the new system, it doesn't take
much calculation to realize that tape
moving at 8 ips is going to yield a lot
better audio fidelity and signal -to -noise
ratio than it does moving at the standard
VHS tape speed of 1.31 ips or the Beta
10.000
1.0 MHz
Figure 4. Comparison of Recording
Densities.
comparison of picture quality between
the RCA Hawkeye system and the
popular U -Matic system with regard to
color resolution. video signal -to-noise
ratio. and color-to- luminance registration. As the table shows. the major
improvement is in the quality of color
offered. In more technical terms. chrominance noise has been reduced by more
than 3 to I. This translates to a IO dB
improvement in measured noise. and an
even more noticeable improvement in the
subjective effect. Chrominance -to -luminance registration has been improved by
more than 3 to I. Finally. in the
luminance signal itself. small -image
detail is preserved. which tends to
eliminate the "cartoon" look found on
many current 4 -inch VCRs.
EASE OF EDITING
Earlier tape systems. which placed the
color subcarrier on the actual composite
videotape recording. forced the editor to
worry about color framing when editing.
The freedom to choose points at which to
edit was therefore limited to one of four
fields. Chroma Trak, which completely
eliminates the subcarrier from the tape
itself. allows editors to choose editing
points with the same freedom that existed
years ago. before the advent of color.
Also, the chance of making a bad
color-field edit, with its attendant'"jump
left" or "jump right" is totally eliminated.
According to RCA. Chroma Trak has
been enthusiastically received by a large
number of producers. directors, editors
and technical personnel who have
reviewed its specifications and observed
its performance.
AUDIO PARAMETERS
In writing the specifications for their
proposal, RCA spelled out the specs for
the audio signal tracks. Recorder
reference level for 0dB. usinga I kHz test
signal. is specified as 100 nWB /m tape
flux per unit track width and a standard
volume level indicator (per ANSI /IEEE
Std. 152-1953-R1976) is also called for.
The Audio I track (in FIGURE 4) is
assigned for mono audio. For stereo
audio, audio track I is designated as left
channel while audio track 2 is right
channel. The time code track may also be
designated as a third audio track.
With the video industry seeking to
standardize a l/-inch (8 mm) format for
consumer use, one wonders whether
the half -inch videotape formats now so
popular in consumer circles may not.
some day, become "strictly-pro" tape
formats
as
progress.
video technology continues to
OUT NOW!!
The latest editions of two important and well respected reference books
on sound and television broadcasting and engineering.
AN ABSOLUTE MUST FOR THE BUSY PROFESSIONAL.
AUDIO
(PRO
YEARBOOK 198'1/83
138mm . 216mm
660 Pages (Appro.l
ISBN 0.900524 98.7
Over 600 Black &
White Photographs
ISSN 0260.8537
Casebound
The PRO -AUDIO YEARBOOK is an annual guide to products and
services for engineers and technicians operating in the world of
professional recording and sound broadcasting around the world.
Published in hardback and containing over 650 pages, it contains
sections covering every conceivable pro -audio requirement. In
addition to the many product sections, there are sections covering
Engineering and Consultancy Services, Jargon and Journals,
Computer Services and Training, and an important section
providing full details of Mains Power Supplies in almost 200
countries.
The 1982,3 edition of the PRO -AUDIO YEARBOOK has been
completely revised and updated, providing even better coverage
of the ever expanding pro -audio market.
VIDEO
YEARBOOK 1982/83
138mm. 216mm
660 Pages (Appro.l
Over 600 Black &
ISBN 07137 1144
ISSN 0140.2277
2
White Photographs
Casebound
To anyone engaged in the business of cummunicating via
television, the INTERNATIONAL VIDEO YEARBOOK should need
no introduction.
This lavish publication has, over the years, become an institution
to the buyer of television equipment or services around the world.
The first part contains over 70 separate sections covering every
conceivable type of video equipment ranging from cameras,
through Monitors, Effects, Editing to Video Recorders. In addition
there are new sections covering Airborne Video, Satellite Stations,
Antenna and Masts, Video Hard Copy and Portable Audio Mixers,
and of course the renowned International Television Standards
section.
The second part of the book contains the indexes, cross referenced
to the product sections, and giving full addresses, phone and telex
numbers and principal contact for over 2,500 companies in the
video and related industries around the world.
:
raEVlsoffislEssu1D
Please fill in the coupon and return with remittance to:
Sagamore Publishing Co, 1120 Old Country Road, Plainview, New York 11803, USA.
(In NY State, add applicable sales tax)
r-
PLEASE SEND
U
copies Pro- Audio Yearbook 1982/83
copies Video Yearbook 1982/83
ii
$54.00 each
$54.00 each
(All Prices include Surface Delivery. Please add $7.00 per book For Air Mail)
L]
Name
Company Name
Address
Phone Number
Signed
_.,....
Please charge my Access American Express Diners Club Mastercard Visa card number (please delete as appropriate). Card holders address must be used
Card No
11l
Please send by Air Parcel (see above)
I
enclose
a
cheque for
Please return the completed order form enclosing your full remittance (including delivery) or giving your credit card number, to:
Sagamore Publishing Co, 1120 Old Country Road, Plainview, New York 11803, USA.
(In NY State, add applicable sales tax)
s
RICAN
EXPRESS
MARK
IV
MONITOR
Features and specifications subject to change without notice.
It is the purpose of any musical
performance, live or recorded, to
successfully communicate with the
listener. To attain that goal is often
even for the most
a challenge
experienced musicians, sound
personnel, and stage crew. At Peavey
we realize the criteria to be met
before this goal can be obtained.
-
talkback system will help alleviate
the problems musicians sometimes
have in establishing the proper onstage mix, especially if a previous
sound check was not possible.
First. the musician must be
satisfied with the blend and balance
of the on -stage monitor mix. In
most concert type situations, the
musicians may demand anywhere
from two to six separate monitor
mixes. Our new Mark IV" Monitor
Mixer can supply this need with up
to eight individual monitor mixes.
Each channel of the Mark IV"
Monitor Mixer features LED status
indication of -10 dBV and +10 dBV,
an input gain control, 4 -band
equalization, built -in mic splitter,
phase reversal switch, PFL and mute
switches, and 8 color -coded rotary
level controls which correspond to
color -coded slider level controls in
the output section.
To make the most out of the
Mark IV" Monitor Mixer's
capabilities, we have equipped the
mixer with two separate built -in
communication systems. By utilizing
our optional headset or "gooseneck
microphone," the monitor mix
engineer can communicate with the
musicians through any of the 8
separate monitor mixers. This
sum.
Each channel of the Mark IV'"
mixing console features an input
gain control, two pre- monitor sends,
4 -band equalization, effects / reverb
send control, pan control,
"push/push" channel assignment
MARK IV'" MONITOR MIXER
The Mark IV" Monitor Mixer is
available in 16 x 8 or 24 x 8
configurations and features
transformer balanced inputs and
outputs, 8 unbalanced outputs,
PFL /Solo headphone system, 10segment LED ladder displays for
each of the 8 outputs, auxiliary
inputs and low -cut controls for each
mix and a unique PFL/Solo patch.
The PFL /Solo patch is a highly
desirable feature that enables the
monitor engineer to patch any of the
mixes back into the switched inputs
so that externally equalized or
processed signals can be monitored.
This is a feature which is not usually
found on custom -made monitor
mixing systems costing $15,000 or
more.
power supply, variable low -cut
controls on each sub (20 Hz to 500
Hz), and in -line patching facilities
between the sub outputs and the
switches, pre and post EQ,
send/reverb patching and PFL (pre fade listen) switch.
The Mark 1V" Professional
Mixing Console has two
complimentary communication
systems for use with our Mark IV"
Monitor Mixers, headsets, gooseneck
microphone and Talk /Comm "slave"
A second communication
can also be established by the
link
monitor mix engineer between the
stage crew and lighting personnel by
utilizing the optional Talk /Comm
"slave" units. The Mark IV"
Monitor Mixer's front panel utilizes
an LED indicator to alert the
engineer as a call function and also
shows when intercom is active.
MARK IV- MIXING CONSOLE
Next, the house (main) system
must be able to deliver crystal clear,
noisefree sound reproduction to the
associated equalizers, power amps
and horn /loudspeaker enclosures.
For the main PA, our new Mark RV'
Professional Mixing Consoles offer
the sound engineer the necessary
performance, flexibility and
functions to do almost any sound
job.
The Mark IV" Professional
Mixing Consoles are available in 16
or 24 channel versions (16/24 x 4
x 1) and feature transformer
balanced inputs and outputs, PFL
headphone system, 10- segment LED
ladder display for all outputs,
channel and sub output LED
indication ( -10 dBV and +10 dBV),
internal reverb and effects / reverb
return to the monitors. The console
also utilizes a 24 volt phantom
('irele 33 on Reader Servire Card
units. The Mark IV" Series intercom
system allows communication
between the "house" and monitor
mix engineers as well as stage,
lighting and other associated concert
personnel.
Both the Mark IV" Monitor
Mixer and the Mark lV'"
Professional Mixing Console feature
gooseneck lamp connectors (BNC)
with dimmer controls for use with
our optional gooseneck lamps. This
option allows superb visibility of the
mixers in poor lighting situations.
The Mark IV" Series Monitor
Mixers and Professional Mixing
Consoles are the successful result of
our extensive research and
development efforts as well as
constant "monitoring" of the needs
of professional sound reinforcement
companies and soundmen. This
outstanding series of mixers
represents, we believe, truly
exceptional and professional
products that will outperform
competitive products retailing for
many times the price.
For complete information on
the Mark IV" Series write to:
Peavey Electronics Corp., P.O. Box
2898, Meridian, MS 39301.
PEAVEY ELECTRONICS CORP.
711 A Street
Meridian, MS 39301
c 82
JOHN EARGLE
Sound Reinforcement
Combining HF and LF Elements
The first six columns in this series have
dealt with the building blocks of sound
reinforcement loudspeaker systems: low frequency components. high -frequency
components. and dividing networks. In
this month's column. we will see how
these components are specified to meet
certain performance criteria and how
they are matched.
In general, it is best to stick with a
single manufacturer in arraying loudspeaker components, since it is likely that
a given manufacturer's items will work
together with minimal compromise. Although good interchangeability exists
between like items made by several manufacturers. only experienced designers
should attempt to "mix and match" between manufacturers.
As examples of how systems go together, we will consider two designs: a
high -level music monitoring system and a
speech reinforcement system for a large
auditorium. In working out these examples. we will be making two kinds of
sound pressure level calculations:
1. Fixed distance, variable -power input. The equation for this is: Level (dB) =
IO log (P; P0), where P is the input
power and P is the reference power.
2. Variable distance, fixed -power input.The equation for this is: Level (dB) =
20 log (D D), where D is the distance
from the loudspeaker and D is the reference distance. This last equation is an
example of the inverse square relationship. where sound pressure level is observed to fall off 6 dB per doubling of
distance from the sound source in a free
field, one essentially free of reflections.
MUSIC MONITORING SYSTEM
°
The specification is given here:
I. A two- channel stereo system must
be able to deliver sustained levels of
105 dB at a distance of 3 meters in a relatively -dead acoustical environment.
2. The system bandwidth must extend
from 30 Hz to 18 kHz.
3. The system must exhibit uniform
horizontal dispersion of 90 degrees above
I kHz.
Generally, the low- frequency portion
of a system sets its outpost limits, so let
us begin with those requirements. Typical low -frequency transducers made for
monitoring systems have a I -watt.
I -meter sensitivity of about 93 dB and a
thermal power input rating of 150 -200
watts. Such a transducer would be ca-
pable of generating the following levels:
Input Power
Distance
Level
(watts)
(meters)
(dB)
I
10
93
I
I
150
150
3
we need.
103
I
100
Employing two loudspeakers in stereo
would add 3 dB to the overall level capability, bringing it up to 98.5 dB. However. we are still about 6.5 dB shy of what
113
115
I
Let us employ a pair of low- frequency
units in each stereo channel and see what
95.5
I
108
watt
tIF
104
l'dB
attenuation
100
I
wait
UI.
LE
Attenuation requind for
power response correction
f,'Mlris()V'r"N
IAI DRIVE REQUIREMENTS
Capacitor for power
response correction
300 watt.
L_
Loss p_d
J
LE
Low -pass
network
T
IBI SYSTEM LAYOUT. HIGHLEVEL DIVIDING NETWORK
High pans and
power response
correct ion
50 watts
o
Low
pass
.00 watt,
(Cl BIAMPLIFIED CONFIGURATION
Figure 1. A High -level music monitoring
system.
the resulting capability is. A pair of low Irequcncy units can obviously handle
ice the input power of a single unit. and
t
this ill result in a 3 dB increase in level
capability. In addition to this. we have
the phenomenon of muuurl coupling. in
which
a
pair of low -frequency trans-
ducers behave as a single larger transducer. increasing the output capability
by another 3 dB. Thus. using a pair of
low -frequency tranv'ucers in each channel and driving each channel with 300
watts will produce a combined level of
14.5 dB at a distance of 3 meters. This is
sufficiently close to our requirement of
115 dB. and the low- frequency design is
1
satisfactory.
Ported low- frequency enclosures
should he used for maximum performance in the 30 -35 Hz range. Enclosure
volume should be chosen so that 30 -Hz
tuning of the enclosures gives smooth
response. Here, a knowledge of the
hicle -Small parameters of the driver
he essential.
We have designed this system around
the continuous input power capability of
the drivers. In practice, there should be
more power available, since drivers of
the caliber used here would be capable
of handling peaks at least 3 dB greater
than the continuous power rating.
I he high -frequency demands are relatively simple. For a crossover frequency.
of NO II/. there are a number of radial
horns. constant coverage horns. and
horn -lens combinations which have excellent 90- degree horizontal coverage.
I
would
Iypicall . these des ices, coupled with
their appropriate high -frequency drivers.
exhibit sensitivities of about 108 dB. Iwatt at I- meter. Accordingly. the high frequency part of the system will he
padded. or attenuated. significantly to
match the low -frequency part of the
system, and this is shown in FIGuRi I. A
mid -band attenuation of 12 dB reduces
the effective high -frequency sensitivity
from 108 dB to 96 dB so that it matches
the sensitivity of the low- frequency pair of
drivers. Note that, at mid -band frequencies. one watt at the high -frequency input
results in only
watt reaching the high -
/
frequency driver; therefore. system
power input of 300 watts is safely reduced
to about 38 watts, well within the rating
of a typical high- frequency driver.
If a constant coverage high -frequency
horn
used, then there should be some
degree of power response correction
above 3 kHz. This is shown in FI(;CRE I B
in the form of a bypass capacitor around
the loss pad. This of course results in
more power being available for the high frequency driver, and care must be taken
that excessive high -frequency program
does not enter this driver. Here, the
is
choice of it smaller 4y, cm diaphragm
driver or a larger 10 cm diaphragm driver
becomes important, because of the
greater power handling capability of the
latter.
If the system were to be stressed in the
high -frequency range, then it might be
appropriate to go to a three-way design,
crossing over to compression tweeters
above. say. 8 kHz.
Most manufacturers provide networks
with the requisite loss and frequency division for matching the components of
the system described here. Some provide,
in their networks, the required high frequency boost for correcting power response as well. F1GURL IC shows the
biamplificd form of the system, which
is preferred in terns of overall lower dis-
tllv
i
I
ws
watt
LP
Iellt,»IIVLR
SYSTEM LAYOUT
Figure
2. A
SPEECH REINFORCEMENT
SYSTEM
The requirement here is for a single channel system capable of producing
peak levels of 100 dB at a distance of 30
meters. For this application, a large low frequency ported double -horn system
would be a logical choice, since high output capability below 100 Hz is not required. A typical enclosure loaded with
medium -sensitivity, high -linearity low frequency drivers will have an overall
sensitivity of perhaps 105 dB: -watt at
I- meter. and will be able to handle 400
watts of input power. Level calculations
are given below:
1
Input Power
(watts)
Distance
(meters)
1
Speech reinforcement system.
Level
(dB)
105
I
IO
I
115
100
I
125
I
131
400
400
400
10
III
30
101
As we note, the low -frequency ported
double horn easily fits the requirements.
Again, we are discounting any increase in
level due to reflected sound in the room.
Turning our attention to the high -frequency part of the system. let us assume
that a 90 -by -40 degree nominal coverage
pattern will be appropriate. A typical
constant coverage horn for this purpose
coupled to its recommended high -frequency driver will have a sensitivity of
about 113 dB, -watt at I meter. FiGRI. 2A shows the basic drive relationships for the system. Note that the high frequency section is padded 8 dB relative
to the low- frequency, and this means
that the power reaching the high -frequency driver will be just about one-sixth
that presented at the input. Therefore, a
500 -watt amplifier driving this system
would never present more than 80 watts
to the high -frequency driver. This is
within the range of a typically 'rugged ízed" IO cm diaphragm high -frequency
driver. and the system should be' coasting" most of the time.
FIGURE 2B shows how the system
would he laid out. No flat power response
would he called for here, since it is customary to roll off speech -only systems
above about 2 kHz at the rate of about
3 dB, octave. While biamplification
would improve system performance, it
would not be necessary in such a system
as this.
1
(AI DRIVE REQUIREMENTS ICROSSOVER TYPICALLY 800H1)
IBI
Manufacturer's recommendations regarding component placement
and polarity (phasing) should be carefully followed for optimum performance
in the crossover region.
Finally, we have ignored the effects
of room reflections on output level. Even
in a relatively -dead environment there
would be some reinforcement in overall
level, perhaps as much as 2 or 3 dB.
tortion.
Editorial
Where there's smoke, there's an editorial
\II' I IytIs t HIS l'AGF is easy to write. Sometimes
it's not. Sometimes. it seems there isn't a fresh
thought within miles of this typewriter. ( It rutddn't
be the typist.)
Or so it seemed one Friday afternoon. when the
publisher strolled by and hinted gently as only a
publisher can do - that time was slipping away. He
expressed his concern subtly: "Where the hell is your
S
&# editorial!! ?" or words to that effect.
As the reverberation died away. the phone rang
(maybe it was ringing all the time). offering a chance to
escape from all of this. and perhaps to return with an
editorial as well. The caller represented an insurance firm
investigating a damage claim for a truckload of audio
hardware that was just barbecued. some 1000 miles from
here.
In less time than it takes to say "double indemnity. "we
were off to the airport and, not that much later. wading
through the earthly remains of a Ford F-700 truck.
searching for traces ofa vanished sound system. Still later
the same day (actually. very early the next). we were back
home again. with an instant suntan. some very dirty
clothes. a few rolls of depressing pictures. and enough
fresh thoughts to fill up at least some of this page.
As we all know. fires are one of those little
inconveniences that happen only to other people. who
should have known enough to be prepared. The trouble
is. many of those other people think that they too are
immune from disaster. which is known to happen only to
still other people. Every now and then. reality intrudes, as
it did to the once -proud owner of this mess. who was now
11
is
glumly picking through the rubble. looking for
salvageable traces of his empire.
Fortunately. he had a first -class policy. and an
insurance representative was on the scene almost as the
smoke cleared. Contrary to what the ads imply, the agent
was not shovelling money into the insured's pockets. but
was methodically gathering what evidence he could to
arrive at an equitable settlement. Fortunately. a detailed
CNI
M
inventory of almost 20 pages was available (just like
yours. right ?). The insurance company had insisted on
this list some time ago. when the policy was first issued.
Now, all that remained was to find sufficient trace of
the major items on the list to verify that they were indeed
on the truck before (and more important. during)the fire.
and. that they were worth what was being claimed.
Some 90 percent of the items were no problem to find.
Even in their new charcoal enclosures. the speaker
systems were recognizable (not pretty, though). Multi pair microphone cables. now the texture of fettucine.
were also easy enough to verify.
In the middle of the heap. a Calzone equipment case
seemed to have survived both the fire and the fire
department. With some anxiety. the case was opened and
there sat the owner's favorite Tektronix scope. a trifle
damp but otherwise none the worse for the experience.
Time out for a minor celebration. in the midst of
mourning.
Now comes the hard part: a major claim for lots of
dollars -worth of custom- designed hardware. But how
much is a custom -built black box worth? The price of the
components may be negligible but the price of the idea
behind them may be beyond calculation. The price of the
finished product, when (and if) it reaches the market will
be somewhere in between.
If it's that commercially- available speaker system. a
few will know the actual value of the components. Still
fewer will know just how much development money went
into it. But everyone knows the market price tag.
Now let's look at your handful of charred components.
A parts catalog will tell anyone who cares that we're
looking at say. 5127.50 worth of scorched resistors. ICs
and such. But what catalog will tell us their worth. once
you've molded them into some wondrous new machine?
How much will you claim for this mess? How much will
the insurance company want to give you? The two figures
may not he even close.
Perhaps it's not a black box full of hardware after all.
Maybe it's just a white box. full of 24 -track master tapes.
How much is that worth? It's at about this point that the
chummy relationship between you and your agent starts
showing some signs of wear.
On the off- chance that he really wants to help you.
consider his problem. if you can forget about your own
for the moment. He's got to convince some bean -counter
back at the home office that this junk is worth the
millions you claim, and not the petty cash that everyone
else says it's worth.
Do you have a good photo. taken before the disaster?
Is there a schematic available? Have you already sold
one, or something similar, to a satisfied customer? Has an
independent outside source recently visited your lab.
studio, or whatever? Can such a source verify the
approximate worth of your gear? Do you have receipts
for goods purchased?-for goods sold?
If the answer to most of these questions is "No." you
may be in for some rough going. Oh. not you. to whom
such disasters never come: we meant the others. Maybe
you'd better warn them. before it's too late. JMW
Like father,
er^
like son.
roughly $10,000, you
can own the ultimate
analog masterin deckThe Studer A8ORC halfinch two-track recorder
Beyond
budget? AB
Well, for about 1/5 the price MI
you can own aRevoxPRPg
compact professional
recorder It's made by thdlMil
same company, it draws on
the same wealth of
engineering expertise, and it
ref lects the same philosophy
of design and construction-a
philosophy established by Willi
Studer over three decades ago.
_.
The PR99's bloodlines are
evident in every detail...in the
4
precision-machined headblock, the
rugged die-cast chassis, The servocontrolled capstan motor, and the Studermade heads Professional design features
include a flat faceplate for easy head
access; edit switch to defeat tape lifters and
fast wind latching; tape dump button;
balanced XLR inputs and outputs switchable for
calibrated or uncalibrated mode; and two-way
self-sync with auto input switching. The PR99 may
ba ordered with 346'7}6or7!6'15ips tape speeds.
Von-speed, full remote control, and monitor panel
available as options
The PR99 now comes in console, rack mount
and transportation case versions Check it out. Call
or write today for the location of your nearest
dealer
The PR99 From the worlds most
respected name in recording
For
R
STU[ÆncÆvOx AMERICA, |mC ^1425 Elm Hill Pike
`
Nashville, Tm3721O^(u5]254-5ó5
(x.'//úv,t, ,\,. .<^..'
SHERMAN KEENE
The Soundstream Digital
Music Computer.
Digital recording is said to have reached the beginning of its
own era-and if the Soundstream Digital Editing Facility is any
indication
has...
-it
FN6AGEDindigital recording equipment
manufacturing has gone a slightly different way in
developing the mechanism of digital encoding. the
computer codes used, and the features of the machines
they make. For example, some companies have developed
digital editing by the technique of actually cutting the digital
master tape with a razor blade. Digital multitrack machines
EACH COMPANY
Sherman Keene is the author
for the Recording Engineer.
of Practical
Techniques
exist with punch in/ out capabilities (allowing overdubbing)
and machine -to- machine digital mixing has been developed
even including digital equalization which does not require
dropping back to analog. Soundstream. on the other hand, has
more or less bypassed the competition by not competing at all
in those areas of digital sound development which are holdovers from analog multitrack recording techniques. Rather,
Soundstream has aimed squarely at simplicity, super -high
audio quality. and reliability in their recording machines. They
have. however. developed a complex editing system. which is
way ahead of its time and its competition.
Recording, Editing
and Beyond
TALKING TO HAL
Soundstream has developed a "blue box" which houses their
exceptional digital circuitry. The digital information is then
recorded using a stock- off- the -shelf. Honeywell instrumentation recorder. Using Soundstream electronics and Honeywell
transports. the company offers two -track, four -track and eight track digital recordings. The operation of their machines is
straightforward: when you record, you record on all tracksno overdubbing. You can't record one track at a time and layer
your production: the machine simply records an event a live
one or a mixdown from (hopefully) a digitally -recorded
multitrack project. Then. you return to the control room of the
Soundstream computer facility and begin a dialog with "Hal"
(fial, the computer you know).
Hal allows you to digitally transfer your music into his
random- access, hard -disk, computer- storage devices from
which you can accomplish any number of simple or vastly
complex post -production tasks, simply by asking Hal (via
computer terminal) that it be done.
The music storage capacity of two Soundstream computer
disk -paks is 34 minutes of stereo music but, by simply switching
one disk -pak while Hal is busy with the other one, an unlimited
length of continuous music is possible. You can hear the
polished. finished version of your post -production changes.
improvements and corrections almost at once. If, upon hearing
your alteration, you like the effect, you incorporate the new edit
function into the music. Once entered into Hal's "edit table." the
computer will he able to play the corrected version from now on.
Since the music is in digital form, sampled in I /50,000íh
second (20 µsec.) segments, you can be very precise about
where you want a computer- orchestrated change to begin and
end.
The most remarkable thing about computers (especially
hard -disk, random -access computers) is the speed with which
Crafted in Japan.
Proven in The States.
TOA Electronics, Inc.
_Z
r1' v .-rj
rrrle
i?
on Redder .Sen
480 Carlton Court
S. San Francisco. CA 94080
(415) 588.2538
Telex: 331 -332
In Canada.
TOA Electronics, Inc.
10712 -181 Street
Edmonton. Alberta T5S 1K8
(403) 489 -5511
Toronto: (416) 624 -2317
the computer can access the information (in this case. the
musical section) the operator wants from literally trillions of
information bits. Consider: a three minute stereo recording, on
Soundstream's hard -disk would process 2 (tracks) : 3 (minutes)
. 60 (seconds per minute) 50.000 (samples per second) or 18
million digital samples. The Soundstream system uses a 16-bit
digital word. One bit is a positive or negative waveform
excursion "flag" and the remaining 15 bits indicate what value
(amplitude) the waveform has above or below the zero-crossing
line. Each digital word. therefore. can have a positive or
negative value of the number 2 raised to the 15th power. which
represents a range of music waveform values from +32.767 to
32.768 or a total range of 65.536 possible music waveform
values stored for each track of music during each sampling. This
would he 18 million (digital words) 16 (hits per word) = 288
million digital hits of music waveform information carefully
stored away for each 3 minutes of stereo music. To he sure that
the computer does not lose any of these tiny information
particles. health check -up is done each day. in which the
computer performs a complex edit function and then goes hack
and checks itself. to see that it did not lose even so much as a
single hit for the test piece of music.
l'hc process of music production with the Soundstream
system is: (
record the original master tape from a live
performance or the playback mix session of (hopefully) a
multitrack digital project. (2) transfer this reel -to -reel recording
to the disk system of the editor computer. (3) perform edits
hem ell takes. within takes and apply polishing touches (special
effects. tightening ups. elegant cross-fades. etc.) to create the
finished product and (4) have the computer -using the
inst ructions created M the editing session transtertheoriginal
music (modified by your editing changes) to a fresh reel -to -reel
digital tape creating an edited digital master. This edited master
tape can he copied. shipped, or carried to any disk mastering
facility. dubbing facility or wherever it is needed. It is important
to keep in mind that all copies of this edited master are of equal
quality: there is no difference between the original and any
generation following it. provided the digital tape recorder and
player machinery is working properly.
The Soundstream disk storage area is divided into three
partitions a main region where the original music is stored, a
scratch region where experiments are performed and the work
region where edits, cross fades and the edit table are stored.
Soundstream editing sessions never alter the original music
itself. Instead. only tables of numbers are created by the
computer in response to simple commands from the editor. The
computer later uses this Edit Table to remind itself which takes
were chosen, what edits were performed. what effects were
decided upon, and where they all go in the music. Amazingly.
the computer can think fast enough to do all these
computations and perform all the required edits, while playing
hack the still physically unaltered music it has in its main region.
I
.
ISSO
31
-62
PH 17P
SP
(I
.1eB;e
a
-7639
7639
75025 re
IS t56
1- JUL -82
Ìi r .
SP
11;
7639
750250
75124(1023
75075;0
MOVIE THEMES (BSOfMILLIRMS)
75075 :8
(23.1000.
ilLi
75126 :1023
Figure 1. Display of stereo image. Duration: 75 records
or 1.5 seconds. Room ambience leading to initial
attack of program material.
When the computer is playing your music with the edits you
have created. this is what goes on behind the scenes. The
computer looks up which song is to be played first and. if there
are no polishing adjustments to the intro area. begins by playing
directly from the original music disk area (you can be taping this
playback or just listening). It follows the original musicstream in
the main region while watching the work region for upcoming
instructions on edits. fades. etc. When it comes to a place where
an edit or other instruction is needed. it jumps to another disk
location to pick up the proper music according to the
instructions it left for itself in the edit table. No gaps in the music
occur because of: (I) the extreme speed at which the computer
"thinks." (2) the computer's ability to do seseral things at once
(like spooling music data at the same time it
is looking up the
next edit task elsewhere on the disk) and (3) the use of the Buffer.
THE BUFFER
The Buffer is an electronic device which lets the computer
feed in data in chunks (the way computers like to supply
information) "hile allowing the data (the music) to come out
connected and smooth (the way humans like to hear it). Long
before (long in the frame of reference of the computer. that is)
the edited section is even played hack. the computer has looked
up the instructions necessary to jump out of the "play- editedmusic" mode and has gone hack to supplying original.
unaltered music to the buffer or whatever its instruction table
has told it to do. A short time later this original music section.
smoothly joined to the edited section. comes out of the buffer
and is heard. Thus. the computer's parts and pieces are
connected into ultra -smooth music to he output to the listener.
What can you do with the editor? \fell. here is an up-to-theminute list of standard features: (I) Nutt edit. (21 cross -fade edit.
(3) cross fade. (4) clone a sound. (5) readjust the Iesels of either
or both sides of the edit. (6) edit or reposition one track while
not editing any other. (7) visually in cstigate a music section to
set an edit point by using the graphic display and the Bit Pad. (8)
interpolate hem een two pieces of moderately or wildly different
music data so that a "rough" splice which might never have
worked using analog methods. can be made to work. (9) replace
distracting ambience with more suitable ambience (or dead
silence. if it is preferred) around dialog or other intermittent
program.
DEFINITIONS
By tiny of explanation of the forgoing list:
Butt edit. This is what the engineer does when he cuts an
analog tape and edits the new head piece to the old tail piece. A
big difference with Soundstream digital editing is that you can
try moving the edit point forwards and backwards on both the
head and tail piece in microscopic steps without actually
damaging the tape as you would when trying this with actual
tape. In addition. since the Soundstream system keeps a
separate edit table and does not make edits by actually altering
the original music. any edit can he re -done at any time without
having to re-do the remainder of the musical program. as is
necessary with machine -to- machine digital editing. Some tape to -tape editing systems require you to re-do a //edits from where
you are until the song's end. should you decide to alter or add an
edit within the song!
Cross -fade edit. A quick cross -fade at the edit point. between
similar musical sections. to help the edit work smoothly. This is
what you would get if you could manage a very long diamond shaped razor-blade cut. he tail of. say. the first piece would
come to a narrow point in the center of the tape. while the head
of the next piece would forma long V- shaped notch. With such
a remarkable cut. both stereo tracks would "slide" from the
previous old piece to the edited -on new piece. This is. of course.
impossible to do manually but very easy with the computer. A
similar result might be achieved using analog methods at some
loss of quality and patience by dubbing the two sections to be
cross -faded to two SMPTE -controlled auxiliary machines. By
finding the proper SMPTE offset. synchronizing the two
machines and moving four faders smartly. one might be able to
duplicate in 30 or so minutes a procedure which takes all of
I
1[53Y:s79
MOVIE r.rtn 5
1550.nILL1W.51
Figure 2 Display of left channel lead -out and lead -in
locations Duration: one Ill record or 1 50th of a second
Allows waveform comparison between takes at splice
point Assists in determining a smear width value to
provide an inaudible splice.
%everal seconds to do by Sounds' ream's computer. This
technique makes for beautiful- sounding edits because. in most
cases. cross -fades are much more "invisible" to the ear.
Cross -fade. A transition between different musical programs:
an alternative to the lade -to- silence and re -intro scheme of
song -to -song transition. You can request a cross -fade from Hal.
and hear it almost immediately'. To he fair. long cross -fades (20
seconds and more) take a while to process because of the
astronomical number of hits which the computer is intermixing
for you while it is "away" doing its tireless bookkeeping. On the
other hand. the computer's cross -fades are beautiful: crisp.
clear. and the music comes hack hard as nails. Since the music is
blended digitally. you still have first -generation music quality
(it your project has been completely digital. that is). Entire
albums without a single "dull moment" are now possible by
cleverly superimposing background sounds over introspective
moments. adding crystal clear ambience and cross- fading
between cuts or interludes.
Clone a sound. Something only the most energetic engineers
have ever tried - usually because they had accidentally
damaged a small piece of the master tape which they desperately
needed to replace. They would do so by copying (cloning) a
piece from another similar part of the master and "fudging" it
into the position of the crumpled original piece. To clone at
Soundstream. you indicate the time frame (a note. a bar. a
phrase. etc.) you would like to clone. specify the number of
times you want the computer to clone it and instruct Hal where
you want the clone placed in the music. Several adjustments
may he performed during cloning like level match. duration
trimming and entry and exit point manipulation. Remember.
these location decisions can he assigned or moved around
within the music to an accuracy of 50.000th of a second.
Usually though. a wider edit -point window is more practical
something like
100th of a second or so.
Readjust the levels. Something which you just can't do while
manually editing analog tapes (unless you want to drop a
generation while experimenting with altered levels leading into
or out of the edit. that is). The levels of the music at each side of
an analog tape splice are set at mix time: if a fader gets
accidentally humped between mixes. then the mixes are not
level- compatible and that's that. lsen with digitally selectable
edit points and cross -fade techniques. a difficult edit may still
sound jarring until you try a level offset at the splice point. I his
level correction can be smoothed hack to normal levels oser a
suitable time period. recouping any level lost or gained.
Note -by -note level adjustments can be made if desired.
Suppose you had a piano solo recorded with a particularly
rough passage which really taxed the performer. One of the
notes in the passage was hit too softly. You can "invade" the
existing piano recording note-by-note and re -set the level of any
I
I
note or notes to any lescl you like (hv setting a new level
parameter at the computer keyboard ). Along these same lines.
by' cloning (to lengthen) or"nihhling "1 to shorten). hits of a note
you can adjust the t.ENGI 11 of any note of the piano solo.
taking up the musical "slack "of gained or lost time elsewhere by
simple arithmetic.
Edit or reposition one track. This would be like trying to edit
or time -displace track 3 of an 8-track tape while not affecting
tracks I. 2 and 4 through K. Try 'h6 one w ith analog tape and a
raror blade! !liars right. the computer doesn't mind it you want
to tidy up the entrances of the background singers( on say. track
3) who came in late at each chorus You can delete some silence
prior to their entrance and add an identical quantity of silence
elsewhere to square the time up again. If that doesn't sound too
good. try deleting the opening silence and then lengthening the
singer's first note by cloning it to just the right length so that
their first phrase stretches out tow here it used wend originally.
If your sense of humor gets the hest of you. you can "lay in"
another word instead of the one they sang substitute a noise or
musical note really. anything you can imagine. the computer
will do. And if the computer's stock repertoire of features
doesn't has': a function the client wants. an emergency
programming session in Salt lake City may he feasible to
invent the software needed to "fix it in the edit.-1 his is the real
power Soundstream has oser its competition: their editing
computer
almost infinitely expandable in its abilities.
requiring Limy a rood suggestion and a cleser computer
progranunei to implement it.
Figure 3. Display showing a click within program material.
Scale is two 12) records or 25th of a second.
1
Visually investigate (in addition to listening. of course) a
musical section to set an edit point. Have you everdonea sound
check with dead quiet ambience only to heara mysterious flurry
of clicks and pops once the actual recording phase of t he session
begins.' The Soundstream computer is a champ at cleaning up
ticks. clicks and pops. You can use a graphic s isual display.
point out the little troublemakers to the computer. and let Hal
get rid of them. You point out edit locations by using a Bit Pad
in conjunction with the sideo display. -I he Bit Pad is a small
magnetically -sensitise drawing hoard on which you "draw"
with an electronic pencil. This method allows editing by eye as
well as by ear. Editing begins by requesting that the computer
"paint" a given time period of the nnisical waveform on a video
display. Instead of typing in a flurry of instructions indicating
the parameters of the edit. simply mosea small indicator (about
the sire of a cassette) around on the Bit Pad while watching the
video display. As the computer senses the signal from the
indicator. it displays a bright dot on the screen along with the
music display.
Now move this dot to the beginning nl the musical waselorm
edit location you're uueiested in (by sliding the bit -pad stylus
around) and press the hit -pad indicator button. Now. musing
the hit -pad indicator dot to the exit point on the waseform. you
press the button once again and the computer knows exactly
Co)
t
iiii to nntmr \II the edit -point ntornution the
computer needs is calculated autonn itica
he graphics which
are displayed during this unction :ire particularly beautiful and
curious: a look at music Iron a heretofore Una\ailable point of
what \ou
I
I
view.
Since the music which was buried by the "pop" is invisible to
the computer. the waveform values on either side may not line
up. To just connect the two ends of the broken waveform with a
straight sertical line would cause a "mini- squarewa\e" that
would he nearly as audible as the pop which was removed. In
order to connect the gap caused by the removal of a noise. one
simpl asks the computer to study the waveforms on each side
of the gap. and interpolate.
Interpolate between two pieces of moderately or wildly
different music data. If you try to splice two pieces of tape that
really don't go with each other using the old analog method, you
might try to make the edit work by shaving little pieces of tape
off one or both ends of the splice area. After a little
experimenting. your tape ends become so shortened that you
can't es en put the original music back together again. should
you decide that the splice isn't going to work aRcrall. Phis kind
of analog hit- and -miss "fix-it" editing is particularly annoying.
With the computer on your side. howeser, a splendid and
complex mathematical computation is put to work to adjust the
audio wasefornts to match each other in spite of themselves. A
software package was recently completed which automatically
looks at the w :reforms at either side of the splice and makes up
a "hest guess" as to what the removed waveform would have
looked like. Thus. many pops can he removed. one after the
other. without disturbing (shortening) t he temporal quality of
the music. This precision methodology seldom fails to work.
and even if it shouldn't. there is still the battery of other
functions which you can try.
All these methods can he tried on a particularly difficult
splice in the time it takes to do even one "exploratory" splice
with a raior blade.
A major record label. while in the process ofarchising their
analog tapes. found that hundreds of splices had crept apart.
Once the music was entered into the Soundstream computer.
these wayward splices were found and closed up. Both the
archise original copy I with the gaps) and the archise repaired
digital masters Iwith the gaps removed) are now stored away
safe from the degradation which analog tape recordings suffer
over long periods of time.
Ambience Replacement. This process is particularly
annoying because you must edit around every note or word.
replacing the original ambience with a more suitable or
acceptable background sound. Consider a narration recording
where two actors have an argument. The background ambience
is totally quiet. Later. the producer re- records one half of the
argument (one actor's part) and has these responses edited into
the original recording. It is then that he notices that the edited
master sounds artificial because when one actor speaks. there is
silence. while when the other speaks. there is (let's say) a water running noise in the background. In order to proceed. the
producer will have to cut the ambient tape pieces out of the
"repaired "' actor's recordings and replace these pieces with
"dead" air to match the background of the original recording.
What's even worse is if the entire original recording has (it is
discovered later) unacceptable background ambience! Then all
the silences between each word (or at least between sentences)
must be replaced. At Soundstream, this process is quite simple,
using the hit pad. The unwanted ambience can he either
removed, creating dead silence between speeches. or the
original ambience can be replaced with a suitable background
sound borrowed from another recording.
BLACK BOXES AND DIGITAL MIXERS
Digital music tronc other manufacturers' digital tape
machines can also he loaded into the Soundstream computer
for editing. phis is possible because the computer has a "black
box." containing a user -adjustable sampling frequency
converter which can be set to whateser value is needed to
properly read the foreign digital information. hew digital
tapes can now he edited w ith all the power of the Soundstream
software brought to hear on them. Once full edited. the
finished music may he transferred hack to the host recording
machine for further use in the outside world.
I
-
SS.
15136 31- .1uL -02
Sn II].0,200.
r
-SS0
76962,500
151
6
I-.lul-62MOVIE
76962:600
TMEMES (BSOIMILLIBrISI
S
74962
16962,69$
Sc C21
e- 00.
'4962f0610
Figure 4 Expanded view of #3 w'th a Qate of 200 samples
or 250th of a second Allows for either precise location of
edit points or interpolation points for removal
1
Figure
5.
The bit pad and storage scope
display
Soundstream is now working on new hardware to enable
their editor and tape machines to act in a "slave" role for
SM.PTF synchninirer work. Adding this feature to their
equipment will open up a whole new field of work to them
including advanced video sound and digital sound for film
work. Also. they are hard at work to perfect a remarkable.
inexpensive digital playback des ice for the home which uses a
thin plastic card as the music storage medium. The music is
recovered from`the card by a laser. (More on this Audiohile
card system in a later issue of db Ed.)
I would like to close with an incredible Soundstream story
illustrating their ability to respond to special circumstances.
For the old Caruso masters which Soundstream worked on.
they invvented.a really amaiing computer program to restore
naturalness to the sound of the music which had been originally
recorded onto disk through a megaphone. Megaphones
(mounted backwards large end toward the music) were once
used to concentrate the sound so that it would properly vibrate a diaphragm which then cut the music onto a wax disk.
This made everything sound horn-like. Soundstream taught the
computer the actual physics of sound passing through a horn
and then instructed the computer to work backwards to "dehorn" the sound. It worked.
My special thanks to Jim Wohington (chief pilot of the
Soundstream editing computer) and Gregg Stephens (chief
technical engineer) in Los \ngeles and to Bob Ingehretsen (a
company Vice President) and Dane Brewer (electronic engineer
extraordinaire) in Salt Lake ('it} tor their time and cooperation.
The Man Behind
Soundstream
potential for ahsoluteh superb audio
fidelity offered by the use of digital methods can
he eomple telr lost in the A -D and / or D -A conThe
verters employed unless sonie basic. hut not-so-obvious. pet:!Orinancespecifications are met.
wORI)S \Rt- ROM an abstract of an ALS paper
presented at the Society's 41st convention. in 1971.
The paper was read by an associate professor of
Computer Science at the University of Utah.
In the intervening years. the professor continued his pursuit
of superb audio fidelity, and eventually as professors
sometimes do started up a little company of his own. By now.
some 30 conventions later. the company and its founder have
become quite well -known in the recording world.
Soundstream. Inc. has digitally recorded about 200 albums.
and I)r. Thomas G. Stockham will become AES president next
month, at the Society's 72nd convention. Apparently. what
began as a part -time hobby has become a full -time pursuit.
Perhaps Stockham's most enduring (and endearing) trait is
his use of the language. While others may prefer to spin webs of
verbiage every time the mouth is open (which is often), the
Stockham technique is one of instant access, no doubt a result
of his computer background. Often, a Stockham response
begins even before the inquiry has ended- especially at public
gatherings when a question from the floor shows signs of
turning into a doctoral dissertation. At a recent seminar on
digital technology. Stockham delighted his audience by
delivering more information in five minutes than some of his
learned colleagues were able to expound in fifty.
TIII
I
computer. we could simulate the effect that these acoustics
would have on the original recording. without actually going
through a real loudspeaker. In other words. we could 'degrade'
the audio with the characteristics of the room only. Then. we
could compare these recordings with recordings made in the
room using a loudspeaker.
"Eventually, we could separately determine the effects of the
room and of the loudspeaker. and get a feeling for where the
problems were. To do all this. we made digital recordings of
phonograph records and tapes. and then processed them with
the room characteristics.
"The project got underway with 'home -brew' equipment
early I I- and I2 -bit A/D and D; A converters. our own digital
filters and amplifiers. and a computer memory. We connected
the electronics to the converters, the converters to the
computer. and we were in business.
"By 1962. we were digitizing audio. Our first "high-quality"
recordings had a 25 kHz sampling frequency. a IO kHz
bandwidth, and were only minutes long. We weren't interested
in ultra -high frequency response. although we did have the
capability to sample at 37.5 kHz and did some experiments at
this frequency. However. we kept the sampling frequency lower
for the acoustic experiments. in order to keep the computing
time down.
"We simulated the effect of putting a sound through a room
over and over again. After five or six passes through the room
characteristics (via computer). all sounds were alike. It sounded
like pink noise. with a cacophony of ringing."
And so began Stockham's early digital work. in a search for
new ways to process information. The computer was his ideal
laboratory tool for testing new ideas. for moving forward in
research. and best of all. for "playing" with audio recording.
Soon enough. he realized that the way to make a recorder that
would transcend the traditional limitations of the medium was
to go digital. But at the time. there was just no practical way to
get all those bits onto a piece of tape. It was possible. but the
cost was astronomical. As Stockham puts it, "It took 15 more
years until the astronomical absurdity had been reduced to
merely being outrageously expensive."
"The early Caruso restorations of old 78s (released by RCA)
was another test of our theories. We blended signals by
convolution. Imagine a blurred photograph in which the 'signal'
you want has been altered by a linear system whose
characteristics are unknown. To get rid of the effect, you must
do a Fourier analysis, and the Caruso project was a
demonstration that this kind of thing will work. "(Forotherdemonstrations, see Robert Berkovitz's feature in this issue -Ed.)
-
"I was
never really satisfied with recording
quality... The tape recorder was
larly frustrating instrument."
a
particu-
After watching the doctor in action. it occurred to your
reporter that it was time to go off to Salt Lake City to learn a
little more about the Soundstream operation. and of course,
about Thomas G. Stockham.
THE EARLY YEARS
Stockham's interest in audio began in high school. where he
made the usual recordings of the school band. But even then he
was vaguely dissatisfied with what he heard. "1 was never really
satisfied with the recording quality, and the tape recorder was a
particularly frustrating instrument to me. It just didn't seem
possible to capture all of the 'live' feed. Back in the early '50s. I
knew this would all change. Tape recorders would improve
vastly. But it didn't happen."
FM stereo was a disappointment, as Stockham perceived a
hack -sliding of broadcast audio quality. The rapid growth of
television didn't help either. Thinking back. "In 1946, the big
event in broadcasting was two consecutive Friday evening
performances of 'La Traviata." Thinking not -so -far back.
"What were the big events of the last ten years? They were
certainly televised, with audio playing an almost -zero (maybe 5IO percent) role."
Although recording technology didn't actually back -slide, it
didn't move ahead fast enough either. Stockham looked (and
listened) in vain for vast improvements. Although good stereo
recordings were now being made, he was personally interested
in other things- removing wow- and -flutter aberrations,
cleaning up headroom. getting rid of hiss, low- frequencies
anamolies and print- through.
In short. Stockham was searching for the recorder -the
digital recorder -that did not yet exist. And, as his career in
teaching and computer science developed, he found less and less
time to pursue his frustrating hobby. But things have a way of
changing.
Ed.)
In 1976, Stockham arranged the necessary financing, and
started Soundstream. His first digital tape recorder was
demonstrated to the Audio Engineering Society in late 1976,
THE BOSE EXPERIMENT
"In the early '60s. was doing some research at MIT. and
discovered some things that made me think it would be possible
to improve tape recording. I was working with Dr. Amar Bose.
1
°
A few years earlier, he had started his speaker research as an EE
project. At first, it was sort of a Saturday, Sunday hobby thing.
He began investigating some basic acoustic issues which also
interested me a lot (they had some big computing implications).
The idea was: What does a room do to a sound?
"At the time, the contention was that much of the
dissatisfaction with loudspeaker performance was because you
were 're- launching' the waveform in a second room. We wanted
to find out just how significant that second room was. Bose
hypothesized that the effect was considerable. And I
hypothesized that we could simulate the notion of a perfect
loudspeaker, even if we couldn't build one. Instead, we could
'probe' a room, and measure its acoustics. Then, with a digital
At
MIT, I discovered some things that
made me think it would be possible to
improve tape recording."
and Richard Warnock presented a paper on "Longitudinal
Digital Recording of Audio" (AES preprint 1169). A second generation system was demonstrated to the AES the next year.
and it attracted wide interest, but no customers. Coming from
the academic/ research world, Stock ham wasn't prepared for
the economics of the recording industry. And the recording
industry wasn't prepared to gamble hard cash on Stockham's
machinery. At the end of six months, the recorder was
withdrawn from the market, and Soundstream was reoriented
from recording sales to recording service.
With the Soundsteam business now exclusively serviceoriented, later-generation hardware has been tailored towards
service, and away from sales. This means going heavy on
systems reliability, and light on the bells- and -whistles (see
FIGURE I). "Our machine is designed and manufactured
exclusively by us. our concept, our designs, our engineering, our
maintenance. Our studies have shown that if we were to now
offer a machine for sale, we would be forced to make a different
set of engineering choices. For example, we'd have to work
towards lower -cost, mass -producible circuits."
THE SOUNDSTREAM RECORDING SERVICE
For Stock ham, "This worked out beautifully. We're
personally responsible for more than one-quarter of the digital
recordings that are now being made world -wide.
"Our present service is in three segments: recording. editing,
and disc mastering. For recording and editing, we always
provide an engineer. We don't insist any more, but none of our
clients have done without our engineer. who is not the session
engineer. though many of them have that capability. Generally.
our clients prefer to keep that responsibility for themselves, so
our man simply assures that everything is functioning
properly."
As for getting ready for a digital session. Stock ham advises,
"Just pull out your (analog) recorder and substitute ours.
Beyond that, be aware that the medium has a much -higher
dynamic range, and much -lower distortion. Don't throw away
these advantages by imposing the traditional operational
functions of analog -gain riding, for example. And be sure
your microphones have dynamic range and recording
characteristics that are up- to- digital."
At a more subtle level, the Soundstream engineer may make
some suggestions about microphone placement. Current
practice is based somewhat on the limitations of analog tape,
and these limitations are largely removed in digital recording.
Stock ham suggests miking for a direct -to-disc, or live -type of
session, and capturing the sound of the performance. Later on,
worry about the limitations of the disc. This could be a good
hedge against future improvements in disc quality, which will
allow the engineer to deliver greater dynamic range to the
RECORDING FORMAT
Soundstream offers a maximum of eight channels. Most of
the digital work so far is classical, or at least is recorded in the
classical manner. Of course, the top-40 market needs more
channel capacity, but Soundstream is not yet in a position to
provide it.
Looking ahead toward greater channel capacity, Stockham
feels that it's not possible to put more channels on a piece of tape
and still maintain the Soundstream level of data reliability. His
machines have a three -level error detection- and -correction
system built in. On the front panel, a series of LEDs (one per
channel) indicate error concealment. If one single bit is wrong.
one of these lights comes on. and stays on. "Maybe one light
comes on in a week, generally due to a tape problem."
A second bank of lights indicate every incident of errors that
are corrected, and a third set warns of the possibility of an error
occurring.
Il
customer.
At first, Soundstream engineers quickly found that clients
did not solicit advice on how to record their sessions. However,
the "I can do it myself" attitude is gradually improving, as
engineers and producers become more comfortable around this
new breed of equipment.
"We're personally responsible for more than
one -quarter of the digital recordings that
are now being made."
CRITICIZING DIGITAL AUDIO
Anyone who reads the Sunday supplement on Arts and
Leisure has surely seen lots of dumb descriptions of what's
wrong with digital recording. The favorite "explanation" by
authors in search of authority is to say they miss all those little
bits of beautiful music that get lost in- between samples. Such
critics might be well-advised to stop trying to explain things
they don't understand, and be content with "1 don't like
period. " However, even if we ignore all the nonsense that gets
into print, there is still a recurring "something -is- wrong" feeling
out there.
Stockham doesn't lose much sleep over this. "In 1925, the
critics didn't like electric recording. Later on, they didn't like the
Ip. Still later, they didn't like stereo." Apparently, if you believe
everything you read, the art of recording has been steadily
declining, ever since it was invented. Actually, part of the
criticism is simply a resistance to something that's different. It's
a new medium, and therefore not as familiar as the old.
"Maybe, the recording is better, and the customer doesn't like
that." For any art form, getting closer to life is a risky business.
Early photographers were not warmly received by viewers who
had grown up seeing the world in a sketch or painting. Today,
it-
Figure 1. The Soundstream digital tape recorder's
electronics emphasize data reliability first, with a minimum
of bells -and -whistles on the front panel.
If not, Stockham reminds us: "Like any other
art, the imperfections of the recording medium are used. and
become an almost -subliminal factor. When these limits are
changed, people object."
we know better.
INSTANT- ACCESS EDITING SYSTEM
lo this observer at least, ten minutes on
a Soundstream
editing session is more than enough to make a digital disciple
out of any non -believer. What can it do? At risk of answering a
question with a question, what would you like done?
The system is nothing more (or rather, nothing less) than a
computer. Like any other computer. its boundary limits are
pretty much a function of the programmer's imagination. (For
more specific details, see Sherman Keene's feature article in this
each, this will probably not happen until the prices come down.
In practice. there's little or no need for additional drives. since
each disk pack can easily be replaced while the other is being
accessed (FIGURE 3).
In the editing room, the operator sits at a computer terminal
with the usual keyboard and CRT (FIGURE 4). An adjacent
scope displays waveforms and a digital tablet moves a cursor
around the waveform for micro -surgery on those really tough
editing sessions.
issue.)
Simply stated. thecompletesystem consists of a transfer room.
from which the eight (or less) channels of the master recording
are loaded into the computer memory. With apologies in
advance to Dr. Stockham for putting it this way, it's just like
loading Pac -Man into your Atari (only more fun).
Figure
4.
An editor's eye -view of Instant- Access
digital editing.
As tapes are loaded into memory. they are named (as in,
SAVE TAKE 3. or whatever). During editing, any program
segment may be played back without waiting, simply by typing
in the name of the segment and depressing the PLAY button. It
Figure 2. Dr.Stockham, with his PDP -1160 in the
background, and an earlier prototype system up- front.
The computer (FIGURE 2). which lives in
nearby room. is a
PDP -1 160 from -where else? Digital Equipment Corp. Its
memory consists of two Century Data 300 Megabyte hard -disk
drives. capable of holding 34 minutes of on -line stereo program.
01' course, more drives could be added. but at 20 kilobucks
a
certainly is instant- access editing. and the next segment can be
on -line just as soon as you can type its name. If you're really in
a
hum ten segments can
.
Pedts:
EM131;
53,
be assigned to single- stroke soft keys.
10:23
52
13- 2-0N -82
F';des
-
1101E3
f911)
Ahi
11
A5D
IRI
(MI
IMI
(MI
ASE
A5F
A35D
A58
AISE
10124
AS8
A15F
ASI
IRI
(MI
(R)
(111
II
(9)
(1
(811I
(R)
(I
(R)
15804;0
16804;0
18304;0
21224;0
21724;0
23852;0
24352;0
27696;0
30696;0
31696;0
32776;0
33076;0
(MI
(MI
IMI
(MI
(M)
IMI
(MI
1000;0
1500;0
2920;0
500;0
2128;0
500;0
3344;0
3000;0
1000;0
1080;0
300;0
1632;0
Ne.t available location (M1: 34708;0
3940
3041
3150
3150
3289
3280
3400
3400
3400
3675
3675
3720
139
140
139
139
139
140
139
140
140
139
140
139
3116
3299
3175
3388
3305
3571
3553
0
3730
3690
3803
'27792;0 (9:29.180161,
eA >>
:
USE:
'13900;1
?00;0
LO:
(Ml 84;428
522 :2
LI:
(MI
'1904;0,
1935;675
6MY: 0;1000
(8 Trks: Multi -track)
1560':0'
Figure 5. The CRT screen, with a facefull of information
that only a digital editor could love.
As the operator decides on suitable edit points. the short
segues between outgoing and incoming takes are stored in a
N
Figure 3. Senior editor Denis Mecham replaces the disk
pack on one of the Century Data drive systems.
separate memory location. Aside from that. there is actually no
edited music -not yet. anyway. So far. the only program is the
one you've written into the computer. As usual. it's just a series
of sequential instructions which the computer will dutifully
obey, when and if you type RUN. Then. the computer plays
your edited program. Except it isn't really edited at all. In fact, it
doesn't even exist. What you are hearing is the computer
instantly accessing the various program segments you've
requested. If you like what you hear. you can save it on tape. If
not, you can do some more editing. Best of all, having second
thoughts about an edit in the middle of the program does not
require re -doing everything that follows. Just change the
necessary instruction, and its done.
Like Tom Stock ham. the Soundstream Instant -Access
Editing System doesn't waste time yours or its.
`No noise, nor silence,
ut one espial musicr
ohn Donne
The new Klark-Teknik highperformance DN30130 graphic
equaliser offers much more than
just a quiet ability to balance
channels right across the audio
spectrum. Thoughtful ergonomics
are backed by a new circuit design
breakthrough using ultra-stable
microelectronic filter networks to
set performance standards
comparable with Klark- Teknik s
'golden oldie the DN27A. The
DN30/30 is the equaliser to boost a
studios reputation, meet
broadcasting specs in less
rackspace, cut costs and
equipment failures on the road
because
...
-
It fits two matched high
specification graphic channels into
a single unit, each providing V.3
octave equalisation over a full 30
ISO centre frequencies.
It gives fine fingertip low frequency control covering the
subwoofer range down to 25Hz
with touch -sensed centre detents,
selectable cut boost level range and
fail -safe design giving extra
certainty during live events.
-
Its advanced design, tough
construction, stringent testing and
long burn-in exceed even KlarkTeknik's previously high standards
for reliability and consistent
performance on the road.
For technical information ask for.
Our DN60/RT60 Data Sheet.
Our DN30 /30 Data Sheet.
Our Application Notes on
equalisation.
KRIM TE11111k
sound science
Klark- Teknik Electronics Inc.
262a Eastern Parkway, Farmingdale, NY 11735. USA.
Telephone: (5161 249.3660
Omnimedia Corporation Limited
9653 Côte de liesse /Dorval, Quebec H9P 1A3, Canada.
Telephone: 15 141 636 9971
Manufactured by Klark. teknik Research limited. England.
[013001/C 5(1121
11.
SOME MORE
EQUAL THAN OTHERS
The microprocessor -based DN60
Spectrum Analyser and RT60
Reverberation Analysers provide
LED matrix display of system,
performer and studio response at
30 frequencies identical with those
of the DN30 /30.
, 1, , 1 ;
'
1
41%4
1
4441,, 1
w
I
\#a
e
d
Circle 43 on Reader Se vice Card
oni
RICHARD KOZIOL and ALFRED D'ALESSIO
,
4"
\!%*'
,)
To Build
the Impossible Dream:
A Sound -Insulated
Performance Studio Comes
to the Big Apple
An inside look at one of New York's newest concert halls.
couLDN-t tt.nve had a more hostile environment
in which to create a live performance studio for
WNCN. Yet by opening night. April 21. 1982.
WNCN unveiled a finished product fit for the likes
of Aaron Copland. Beverly Sills. Ruth Laredo and the host of
other classical artists who took part in the four -hour concert
which was broadcast live from the studio.
The studio had been promised to WNCN in 1976. when GAF
Corporation purchased the station. It was to be fashioned from
raw space located to the rear of WNCN's broadcasting facility
at 1180 Avenue of the Americas in Manhattan. The space had
been so- designated at the time of purchase. But until we got the
go-ahead to build the performance studio. the space. occupying
approximately 3.000 square feet on the building's fifth and sixth
floors. had been primarily used for storage.
WF
Richard Ko_iol is Chief Engineer for WNCN.
Alfred D'Alessio heads A. W. Dit /essio Associates
of New York Cin, a systems and acoustic consultingJìrm
v Or the television and sound recording media.
As we got started on our plans for the new studio, we were
immediately faced with our first challenge. The studio would
have to be built in two stages. Phase I of the project would entail
creating a studio environment which would be conducive to live
broadcast performances by small groups. Since the Phase I
project was somewhat experimental, the budget would be much
smaller than for Phase II -in which we would build a control
room, as well as complete the performance space by floating the
entire studio on a suspended floor, acoustically isolated from
the rest of the building.
In undertaking the project. we had to draw plans for both
phases. even though Phase ll might be years away. We were
charged with ensuring that plans for each phase would
coordinate with the other, so that in building Phase I1, the
Phase I work would not be destroyed.
MAGNECORD
MC -11
Compare Quality at this Price
THE BIG PROBLEM: NOISE
Before even doing our preliminary measurements. we knew
that our biggest problem in constructing the studio would he
insulating it against noise. which came from two major sources.
l he hack of the building, which is where the studio would he
located. faces an alley, and part of that alley faces a street. In
addition to the street noise, which echoes in the alleyway. we
also found noise dumped into the alley by the air conditioning
systems of other tenants in the building.
A more threatening source of noise was located in the studio
itself. a "wet" column which runs the length of the building. In
this column are located steam pipes. drain pipes. and water
pipes hooked into the building's lavatory flushometers. in
addition to smaller. but still quite noisy. pipes. We had to find a
way to insulate the noise from that column, or wed never he
able to broadcast from the studio.
Later on, we discovered there would he a third noise source.
I
t
irermi
INIZEIIIAMecoa
O
syóa6
0,0a.
Meets or exceeds NAB standards, with
IEC equalization on request.
DC servo, flutter -filter drive runs true
regardless of line voltage fluctuation.
Cool operation; no ventilation required.
Full remote capability.
Long life heads and phase locked tape
guides.
Mono or stereo play models field
convertible to record.
Automation cue tones (stop, secondary,
tertiary) with LED's and external
switching contacts.
Cue track access for FSK logging.
Universal mic /line imput.
Immune to RFI and EMI.
Rugged design in the Magnecord tradition
-made
in USA.
'Suggested Pro Net Price
Quality Products for the Audio Professional
TELEX
MAGNECORD
)MMUNICATIONS.
Composer Aaron Copland with pianist Leo Smit.
INC
Circle 37 on Reader Service ('ard
About midway through construction of the studio. with all of
our final measurements ta'ken and planned for acoustically. the
building management overhauled the air conditioning system
throughout the entire building. Now it was running much better
than befre. but it also made more noise! We found we had to
add more insulation.
ORIGINAL PLANS WENT AWRY
We weren't tar into our planning before we realized that we
had another obstacle to contend with. The original designer of
the space had apparently planned fora studio on the fifth floor.
with the control room located on the sixth floor. overlooking
the studio. But the only access between the two floors was
through the building's elevaror system! There were the obvious
logistical problems with this arrangement. since without adding
an expensive stairway. it would simply take too long to respond
to situations which might arise.
In addition. a mechanical equipment room for the station is
located on the sixth floor. This noise -producer is only surpassed
by still another mechanical equipment room located above it.
which scrses the entire building.
It soon became clear that we would have to put the control
room on the fifth floor. next to the studio. behind a windowed
wall. While it caused a reduction in the studio's performance
area. lol_istically it turned out to he
a
much better setup.
SOUGHT LIVE RECORDING ENVIRONMENT
In planning the studio. we wanted to create a fairly -live
recording environment as opposed to a controlled. multi -track
recording studio. Performers. typically playing without a
conductor in small ensembles or chamber groups. would have
to he able to hear each other's instruments. The studiowould be
more likea mini concert hall. typically includinga live audience.
In recording performances. we would treat the studio
environment just like the remote concert sites from which we do
many broadcasts every year.
The reverberation time in a 30.000 cubic-foot multi -track
studio (the size of the WNCN room) would run 250 -300
milliseconds. but we figured that the ideal reverb time for
WNCN's studio intended for a coincident mike pair should
full second.
Before adding any acoustical treatment, we tested the space
to find out exactly what we had. We found that the reverb time.
which was fairly constant over the audio spectrum. was within
20 milliseconds of where we wanted it.
However, we did have a slap echo problem because of the
studio's characteristic parallel surfaces. Our problem was to get
rid of the slap echo without disturbing the reverb time and
without spending enormous sums of money to build splays and
diffusers which would take up precious space in the studio.
he one
ADDED ACOUSTICAL MATERIAL
The solution was fairly simple: hang acoustical material at
performance level, with the absorbers located at intervals across
from reflective surfaces. This solved the studio's slap -echo
problem without destroying the reverberation time. (We had no
slap echo from floor to ceiling, thanks to an enormous stroke of
luck. W hen the building had been constructed. fiberglass panels
were placed on the ceiling.) And that's exactly what we'd have
done to solve the problem. if there had been one!
BUILT ACOUSTICAL ENCLOSURE
To tackle the problem of the wet column, we insulated each
pipe. using a high -mass damping compound for the biggest
noise producers. After we had treated each pipe. we designed an
acoustic enclosure for the column. And (knock on wood!). it
has been quiet ever since.
In order to insulate the studio from the alley noise, we were
able to use high -mass dry -wall construction. building several
feet in from the back of the building. and successfully eliminate
the problem.
co
A WRENCH IN THE WORKS
We thought we had covered the major sources of noise. and
then we inadvertently found another.one. In addition to the wet
column, the studio contained two structural columns. which
had not concerned us as they produced no noise. But as
construction proceeded. and the building's air conditioner was
repaired. a nearby water return pipe from the cooling system
caused the two structural columns to vibrate. We were forced to
treat those columns as well.
MODIFIED AIR CONDITIONING SYSTEM
Air conditioning was a concern to us tor another reason. In
its unfinished state. the studio area had not been included in
WNCN's air conditioning system: now it would have to he
added. But when we took estimates from vendors who could
design a system for the studio. we found the costs were
prohibitive. relative to the project's budget. We'd have to find
another way.
It's no secret that a chief engineer at a radio station must. of
necessity. wear many hats. And this one was no exception.
Through experience. it was apparent that one position of
WNCN's dual air conditioning system was not working to full
capacity. That reserve capacity could he diverted and used for
the studio.
A duct and louvre system was designed to tap into the
existing air duct system. Because the studio's ducts are long.
and turn several times. they tended to isolate noise generated by
the air conditioning units. and this was not a concern.
Since the studio has a 22 -foot ceiling. it acts as a reserve for
this cool air. and the positioning of the ducts is such that as new
air comes in. it naturally circulates throughout the room. The
acid test of the system was during the opening Gala Concert.
when we found the temperature remained constant over the
four -hour duration -television lights notwithstanding.
Because we designed our own system. we were able to have a
separate contractor come in_just to do the ductwork. The cost to
WNCN was substantially less than what would have been
incurred had we had an air conditioning contractor design a
new air conditioning system.
THE FINISHED PRODUCT
In designing the finished size of the room. we again had luck
on our side. We found that the resonant modes were well
distributed across the audio spectrum. not hunched up at
certain frequencies.
The studio. when finished, measured 1.150 square feet.
Because of the center column in the room. we designated one
quadrant of the room as the performing area, allowing for
audience members in two additional quadrants. The
performance area is located next to one of two accesses- -which
artists use as a private entrance. The fourth quadrant.
obstructed by the columns. is used as an access area for the
audience. and is located at the site of the second door.
The control room area --which was set up as a remote
recording area during Phase I of the project -is located directly
across the room from the performance area. affording a clear
view of the entire room through a large double -glazed window.
The doors themselves are steel-clad and sound- retardant,
with expansion seals in the jambs.
-
WNCN ACTED AS CONTRACTOR
Despite the usual and unusual problems encountered in
building the studio. the finished product was a resounding
success. Total construction time was about nine months, due in
part to the fact that the chief engineer -- with many duties
acted as project manager with the station while serving as
general contractor for the project. Fitzgerald Construction did
much of the major construction work and Jorge Cao was
interior designer for the project.
At its opening. hundreds of musicians. music -lovers and
members of the press attended the open house and Gala
Concert which featured more than two dozen classical artists in
a four -hour program. WNCN General Manager Matt Biberfeld
proclaimed the studio "a new concert hall for New York City."
In his column in .Veo York magazine. critic Peter Davis referred
to it as "a stunning new live -perfrmance studio."
takt
Ap lkation Notes
DALE BEARD
Techniques for Hum
and Noise Reduction
INTRODUCTION
Ht purpose of this application note is to provide the
reader with a basic understanding of commonly encountered causes for, and cures of, hum and noise in
professional sound systems. Included will be a discussion of signal sources and their peculiar characteristics, noise
sources, mechanisms of noise energy transmittal, ground loops.
power distribution, and good cabling practices.
Sometimes the pursuit of hum -and -noise reduction appears
to he a mixture of witch- hunting and mystical rites to the
electron gods. But keep the faith and try to develop a firm
conviction that for every observable phenomenon there is a
r
(A)
(B)
rational scientific explanation.
Figure
SIGNAL SOURCES
Before describing the peculiarities of specific sound sources, a
discussion of the two primary categories of sources and their
associated interconnections is in order.
F161121: I shows both balanced low- impedance and
unbalanced high impedance systems. Their power-transfer
capability is roughly eyuisalent; the former being low
voltage high current. while the latter is high voltage low
current. Another obvious similarity is the use of an outer shield
conductor. So much for similarities.
THE UNBALANCED LINE
If the outer shield isa perfect conductor and also achieves 100
percent electrical shielding, then the laws of physics decree that
an external electrostatic field will not produce any net charge
(or voltage) on the inner conductor with respect to the outer
conductor. since it is completely surrounded by an equipotential surface. However, nature is never very generous, and has
failed to provide us with a perfect conductor. and a 100- percent
electrical shield is difficult to attain while retaining flexibility
and low cost. The shield helps considerably, but there is always
1.
Balanced (A) and unba lanced (B) lines.
Aligned
Stay Aligned
Get
with
precision
magnetic test tapes
STL
These dependable tapes are used by
broadcasters, recording studios, equipment manufacturers, governments and
educators throughout the world. Widest
variety...Alignment, Sweep, Pink Noise,
Level Set, Azimuth and Flutter/Speed.
Available on reels, in cartridges and in
Also, the Standard Tape
cassettes
Manual & the Magnetic Tape Reproducer
.
Calibrator.
Phone for fast delivery or free catalog.
T0
STANDARD TAPE LABORATORY, INC.
Dale Beard l.4president ofLinear and Digital SI'3/eln5,
V
Circle 38 on Reader Service Card
some net charge that leaks through. and therefore room for
improvement exists.
THE BALANCED LINE
Enter the twisted pair. Since we are concerned with imperfect
shields. let's examine the limiting case- -that of no shield at all
(refer to FIGURE 2 for the following discussion). Assume that we
wish to transmit an instantaneous signal voltage of 2 volts. For
the unbalanced system. this 2 -volt signal is applied to the center
conductor. In the balanced scheme. the signal is divided into
equal-and -opposite I -volt signals. These voltages are typically
generated by a center -tapped transformer with the tap
grounded. thereby creating a sort of electronic teeter -totter. At
the receiving end, it is the difference in voltage on the pair of
wires which is interpreted as the signal. rather than the absolute
value of the voltages.
So what does all this do for us? Again referring to FIGURE 2.
let us further assume that an external electrostatic field is
present and that it affects all three wires by inducing an
additional 2 volts. It is not important here to understand the
field theory which makes this happen or even to know what
kind of field it is (call it a watermelon field if you wish). What is
important is that from experience we know that these fields do
exist and intuition tells us that they do induce unwanted
voltages on signal lines. since their audible replicas eventually
reach our ears. For the unbalanced line this induced voltage is
added to the original 2 -volt signal for a total of 4 volts at the
receiving end of the line. obviously a gross distortion. In the
case of balanced transmission, our original signal levels of + I
volt and I volt have been altered to +3 volts and + I volt by the
effects of the external field. However, the difference is still 2
volts and so our original signal has survived unscathed. The
balanced method therefore derives its tremendous advantage
from the differential operation of the receiving circuitry. This
property is known as common -mode rejection, since voltages
common to both wires are rejected as much as possible. When
the twisted pair is shielded there is an additional improvement
which yields a "belt and suspenders" solution to hum and noise
reduction.
Another advantage of balanced low- impedance is that very
long cable runs can be made with no appreciable loss of highfrequency content. The center conductor(s) are in close physical
proximity to the shield and therefore create a stray capacitor to
ground, albeit a very small one. However, the value of this
capacitance is directly proportional to cable length (in fact cable
manufacturers specify it in picofarads per foot) and can
therefore become significant for long cable lengths. FIGURE 3
shows the resulting equivalent circuit which, due to the source
resistance, is a classic low -pass filter whose cut -off frequency is
given by.
= I;
(RSOI'RCE:
CïrRAY)
One can deduce from this equation that, for a given cable-length
(or CsrRAV ), a lower source impedance will produce a higher cutoff frequency. Conversely, for a given cut-off frequency, a lower
source impedance allows a larger C'srRAS and therefore longer
cable lengths.
Having developed a firm grasp on the operating characteristics of signal sources and transmission methods in general, we
will now discuss the peculiarities of specific signal sources.
ELECTRIC GUITARS AND AMPLIFIERS
The first step in solving hum and noise problems with
electrics is to logically isolate the source of the problem. There
are basically three possibilities: the guitar. the interconnecting
cord. and the amplifier. The cord should be checked first since it
is the easiest to isolate and repair.
Next determine whether the guitar or amplifier is at fault by
simple swapping tactics. If the guitar is at fault it may be due to
one of several causes. Pickups may be the source. particularly in
older guitars. Most pickups are of the variable- reluctance type
which operate on the principle that the strings present a variable
magnetic path length (or reluctance) to the pickup. which is
excited by a permanent magnetic field. thereby generating a
voltage on the integral coil. By being inherently responsive to
magnetic -field variations, these devices are obviously affected
to some degree by external fields. Modern variable- reluctance
pickups incorporate hum -bucking coils which tend to minimize
external field sensitivity. Unfortunately, the cure for the pickup
problem is sometimes the purchase and installation of a modern
pickup. Local music dealers can generally be of assistance in
locating a source for the pickup as well as installing it.
Inadequate shielding of the compartment housing the tone
and pickup controls can also be a culprit. Aluminum foil tape
can be used to shield the compartment and should be grounded
to the ouput jack. A foil shield under the pickup(s) is also frequently employed, so check its connections.
21rRSODRCrCSTRAY
Figure 3. The effect of cable capacitance on frequency
response.
As with guitars. age is often a problem with amplifiers. most
notably the vacuum tube designs. One inherent problem with
tube equipment in general is the AC voltage on the filament
which can couple to the signal electrodes as hum. There is no
cure for this problem short of installing a DC supply for the
filaments. and this is not a simple fix. Another problem that
occurs in older amps is excessive ripple on the power supply
lines, which also manifests itself as hum. This problem is often
due to dehydration of the electrolyte in the filter capacitors and
can usually be identified by a visual examination of the
capacitors (they're always big and usually in metal cans) where
+3v
i
VSICNAL = 3V
+IV
2V
4V
IRL
---r
VsicNAr = 4v
1
co
Figure 2. The balanced line offers superior common -mode
rejection.
-
IV = 2V
the leads exit the case. Any seepage or corrosion -like solid
deposits indicate a problem which should be corrected by
replacement.
Power
source
Figure
4. A
To
amplifier
circuitry.
commonly -used power input circuit.
A power supply input circuit which was virtually universal
during the two -prong power plug era is shown in FteuRE 4. This
circuit has been the cause of more than its share of problems
despite its relatively innocent appearance. The original intent of
the designers apparently was to by -pass noise on either line to
circuit ground. A noble cause, but there are no guarantees that
circuit ground is earth ground (in fact it is Very likely not), which
is where the noise should be by- passed to begin with.
Furthermore. when the ground is floating. ('t and ('2comprise a
capacitive voltage divider placing circuit ground at a 50 -60 volt
AC potential. This has caused the lingers of more than a few
guitar players to tingle when some part of their anatomy is
simultaneously in intimate physical contact with a grounded
mic stand. The capacitors are generally small enough in value
that the current they supply is not lethal. but someone who
wishes to perform in a tub full of water may have a serious
problem. Finally. with respect to noise performance. the circuit
is also highly undesirable since the capacitors guarantee that
any noise on either line will he dumped into circuit ground.
which is the last place you want it to go. So the recommended
fix is to remove the damned capacitors, throw them on the
ground. and stomp on them to display your disgust.
MICROPHONES
Dynamic microphones are typically balanced low -impedance
sources. and when handled with a reasonable amount of care
are highly reliable devices which generally do not exhibit hum
or noise problems. Ira problem does occur. a cable or connector
should immediately he suspected and can he readily isolated
simply by swapping the suspect component with a known good
one.
Condenser microphones are sometimes utilized due to their
extremely -wide frequency response. While these are inherently
high -impedance devices, they typically employ built -in
preamplifiers and or matching transformers such that their
electrical output characteristics are identical to the dynamic
type. with all the attendant advantages.
you to repair a cable when it does fail (which is inevitable in
road use) -unlike the molded types which should simply be
discarded.
Since the acoustic pickup has a very -low- voltage. highimpedance output. a preamp is generally employed to amplify
and buffer the signal. Often a direct box will be placed after the
preamp to make the source appear electrically equivalent to a
balanced low- impedance source such as a microphone. In fact,
the so-called "direct box" is really nothing more than a
transformer having primary and secondary impedances in the
vicinity of 10,000 and 500 ohms respectively. By utilizing a
direct box in this application one can enjoy the benefits of
balanced low impedance transmission along with the electrical
isolation of the guitar and preamp from the mixing console.
Two potential problems arise at this point however. First, most
preamps use garden -variety operational amplifiers such as the
LM358 (National) and the MCI458 (Motorola) in the signal processing stages. While these devices perform admirably in
many circuits, they are less than ideal for audio preamps due to
significant amounts of crossover distortion and relatively poor
input noise specifications. This can often he cured by replacing
the op -amps with better devices such as National's I.F442 (a
BIFET device with ultra -low current drain) or Signclics'
NE5512. The second problem is that some of the newer preamps have a built -in direct output which consists ofa solid -state
circuit instead of a transformer. The reason manufacturers do
this is to reduce cost. despite the claim that they did it to reduce
distortion (which it probably does). This feature comes at the
expense of transformer isolation which can result in ground
loops and or increased levels of hum. More about direct boxes
as well as ground loops later.
KEYBOARDS
With a few exceptions. keyboards are largely electronic in
nature. and typically do not present any new problems. Some of
the previously mentioned problems do occur however. For
example, a home-brew conversion of an acoustic piano to an
electric environment will often he performed by installing
pickups and preamp(s) similar. or even identical. to those used
in acoustic guitars.
Clavinets are also prone to noise problems. particularly older
ones. These problems are similar to the electric guitar
difficulties. since the principle of operation is the same. The
problem is compounded due to the huilt-in preamp employed.
Often much of the problem is due to inadequate shielding and
lack of grounding of electrically unused metal parts. such as the
front panel and switch brackets. As with electric guitars,
shielding the preamp control compartment with grounded toil
tape and grounding all metal surfaces will go far in reducing
noise.
Synthesizers and other purely electronic instruments are
generally clean signal sources. Noise that does occur is due to
equipment design shortcomings and as such is beyond the scope
of this discussion. However. hum problems may occur with any
of the keyboards (as well as just about any instrument) due to
ground loops. Ground loop phenomena will he discussed later
in a separate section.
DIRECT BOXES
ACOUSTIC GUITARS AND PREAMPS
Acoustic guitar pickups are generally less troublesome than
electric pickups since they usually transduce the guitar body
movement into an electrical signal rather than including the
strings in a magnetic circuit. Cable problems are. as always,
fairly common particularly since the connector is usually a
miniature phone type. It is wise to avoid the molded cables and
make your own. using high -quality wire and metal -shell
connectors. This is well worth the extra effort since it will enable
As was previously mentioned. the direct box is a device which
converts from unbalanced high-impedance to balanced low impedance transmission. A schematic for a full- feature direct
box is shown in FIGURE 5. with the main ingredient being the
transformer. The purpose of the high-impedance in and out
jacks is to allow the signal to be run to an on -stage amplifier for
monitor purposes as well as to the mixer. The pick -up; amp
switch simply changes the gain much like pad switches on a
mixer so that a high -level signal from an amplifier, as well as a
CD
PICKUP
htZin
200K
300pF
(
Iti
Z
o
IItTIN
out
I'I.AT
XLR connector
Ground
lilt
switch
Lighting and other natural atmospheric electrical
disturbances.
Switching of inductive loads (motors, transformers. etc.).
Motor brush noise (cash registers and small appliances).
Lighting equipment, especially dimmers.
The second type of noise source is characterized by primary
transmission through air such as:
Radio transmissions of various frequencies. e.g. AM.
FM. CB. and TV.
Microwave sources (ovens, radar, etc.).
Computers and other digital devices.
The final category of noise is that which is generated
internally by electronic equipment used in the sound system
itself. This type of noise is generated both by semiconductors
(diodes, transistors, and integrated circuits) and by passive
devices (resistors, capacitors, etc.). The primary causes are the
following:
Thermal noise.
Shot noise.
Figure
5.
A full -feature direct box.
low -Iesel signal from a pickup or preamp, may be accommodated with equal case. A filter switch is provided so that the
signal may be run essentially flat. Otherwise. the treble may be
rolled off. This feature is useful for electric guitars. especially at
distortion settings. This is due to the fact that guitar amps
usually employ speakers in the I2-inch diameter range. which
have relatively poor high -frequency response. while the
electronic signal from the amp is generally full- range. This
causes the signal to the mixer and ultimately the main speakers
to sound much more "raunchy" than what the guitarist hears on
stage. The remaining feature is the ground -lift switch which
allows complete isolation of the input and output circuits. The
best position for this switch is simply that which yields the
lowest hum and or noise. This can be determined experimentally.
For those so inclined, construction of a direct box is a
relatively straightforward matter which can save considerable
cost if many channels are required. The most expensive item is
the transformer. such as a Triad T-I X. which will run about
$10 -15. Total cost should be under $30. Another approach the
author has found useful is to actually build the direct box into
electric guitar amplifiers. Room can usually be found on the
rear panel for the mic connector and on the inside for mounting
the transformer. This technique is not recommended for the
novice and care should be taken to use proper wiring practices.
Mount the matching transformer as far away as possible from
the power and output transformers. It is also highly desirable to
pick off the signal prior to the volume control so that changes by
the performer will not alter the volume in the mains. Finally.
some experimentation with the capacitor value is suggested so
that the tone of the mains is roughly equivalent to that of the
amplifier (the values shown are typical but at least represent a
starting point). The advantages of building the direct box into
the amplifier are as follows:
Elimination of the phone connectors and the associated
patch cord.
Reduction of on -stage clutter due to more boxes and
cords underfoot.
Less possibility of switch settings being inadvertently
altered.
Ability to "tune" each box to match the amp for truer
frequency response characteristics.
The advantages of balanced low- impedance transmission
with the ease of simply inserting a mic cable.
NOISE SOURCES
íßn
Noise sources can he classified into three different types. The
first type represents sources whose noise energy is transmitted
via the power line and include the following:
I
F noise (also referred to as
flicker noise).
The first two types of noise are relatively common to
everyday experience. manifesting themselves in TV reception or
stereo operation. For example, we hate all been watching our
favorite episode of Star Trek for the 17th time when someone in
the kitchen fires up the electric mixer, producing picture tearing
and an obnoxious whine in the audio. Ditto for hair dryers and
vacuum cleaners. A heavy inductive load such as an air
conditioner motor will often produce a momentary shrink or a
bright flash in a TV picture. Light dimmers will sometimes
produce a buzz in a stereo system.
With this brief and informal introduction to noise sources, let
us now deal with the question of how the unwanted signals are
transferred to the sound system and, more importantly. what
the devil to do about it.
NOISE ENERGY TRANSMITTAL
As we mentioned previously. noise sources may be
categorized by the manner in which their energy is transferred
to the sound system. The primary mechanisms are. of course.
air -borne and line- borne. However. life is never simple. and in
practical situations noise usually enters the system by a
combination of the primary mechanisms. For example, noise
from a light dimmer may travel through the power line and into
an electric guitar amplifier's power supply. thereby disturbing
sensitive signal stages. However. an alternate path also exists
since the building's power lines act like a large grid -like antenna
which radiates a portion of the noise energy. This airborne
energy enters the system through the guitar pickup or the
sensitive front -end stages of the amplifier. Of course, there's
other airborne noise, such as a truck driver on the interstate
with a 1000 -watt linear amplifier strapped to his $17 CB (over modulated of course) thereby splattering electronic replicas of
unintelligible babble for miles in all directions. Talk about air
pollution! Again. the power lines act as an antenna. only this
time they are receiving rather than transmitting. The noise
travels down the line and into the amplifier as before. Even
more complicated interactions can arise. particularly in the case
of radio -frequency interference (RF1). Let us re-examine the
CB case. where we had air -borne, then line -borne interference.
Once the noise enters the metal case surrounding the amplifier
(which acts as a shield) via the line cord, it is possible (in fact not
uncommon) for the noise to be re- radiated and again be
amplified by sensitive circuits.
By now many readers are undoubtedly developing a feeling of
despair, since noise clearly obeys Murphy's Law. Those of you
who are still awake. however. may have deduced that, with the
exception of cases where air -borne noise energy enters a pickup
Utility
or amplifier directly. there is a common denominator in all this
mess. I hat common denominator is the power lines.
transformer
To
utility
mains
CLEAN UP YOUR ACT
Since me lase identified a common culprit in noise problems.
it would he logical to address this unhappy condition. he first
step is to apply filtering to the power lines. This should he done
tt ith a high quality RH filter of the type used in computers and
instrumentation. Do not use the cheap TV varieties from your
local radio shop for two reasons: I t hey can't handle the current
and 2) they seldom, if ever. work.
I.ine oltage transients due to lightning have been previously
II
R3
l
mentioned as noise sources. What was left unsaid however. is
that these transients can reach potentials which are extremely
damaging to electronic equipment. During thunderstorms. the
120 -volt outlet frequently exceeds 1000 V and levels higher than
50(X) volts will occasionally occur. Fortunately the duration of
these transients is typically very short (on the order of 10 microseconds) but still long enough to destroy our rather unforgiving
semiconductor det ices. Equipment can be protected against
transients of this nature by devices known as logically
enough transient suppressors. One particularly effective
transient suppressor is the metal -oxide varistor (MOV) which
behaves like a voltage -dependent resistor (hence the term
varistor). Below a certain voltage, known as the threshold. the
des ice conducts only a Ictt milliamperes. but above the
threshold conducts very heat
In this manner the MOV
tends to clamp the line at reasonable oltage levels. Circuit
placement of the des ice is between line and neutral.
It is also good practice to physically separate lighting power
wiring and audio cabling as much as is practical. This is due to
the fact that light controllers employ phase control circuits and
typically generate a considerable amount of power line noise.
usually manifested as the infamous sound system burr.
il.
GROUND LOOPS
he term ground loop is probably the most often quoted yet
least understood of all the terms in the professional audio
sernacufar. Let us first describe the mechanism of the ground
loop and then explore its significance tsith respect to sound
I
systems.
Ideally. a ground is a ground is a ground. Unfortunately. in
the real world this is not the case. the ground loop
phenomenon arises because conductors and connectors ha%e
finite resistance and will therefore exhibit a voltage drop when
current flows through them. We therefore have a situation in
which the ground terminal at the stage power outlet is different
from that at the mixer. giving rise to hum and possibly noise
problems. The astute reader w ill at this point probably say "wait
a minute. grounds are not supposed to he current- earrying
conductors since they are only there for safety." A valid
objection. However. one must consider what happens back at
the Ruse box. Referring to Fi t't i 6. we see that the neutral and
ground connections both go to the center tap of the transformer
and to earth ground through the grounding stake (sometimes a
water main is used). Also shown are some parasitic wiring
resistances which cause differences in ground potentials. For
example. let us assume that R = 0.1 ohm (normally considered
by electronic types to he negligible) and that Branch 2 has a load
current of 10 amperes. According to Ohms Law (E = IR) a
oltage of volt will therefore appear across R I. This -volt
drop will not noticeably affect the brightness ofa light bulb, but
could he catastrophic to the sound professional! Worse yet, if
the stage is on Branch and the mixer on Branch 2, then the
ground(s) between the stage and the mixer will be in series with
R and actually share some of its current. Now you can see why
it is referred to as a loop.
i
I
I
I
i
Ground stake
Figure
6.
The ground -loop phenomenon.
Now what to do about it. You obviously don't have time to
rewire every building in which you are going to set up for a
three -day gig. One must therefore somehow break the loop.
There are two basic solutions. One is to float everything at one
end of the system or the other. This is probably the most commonly employed approach. However. there is one important
aspect to he considered and that is safety. l.et us examine the
purpose of the ground conductor in the first place. Assume you
are standing knee deep in water in a basement and you are going
to drill a hole in something. Lct us further assume that your drill
has developed an internal short such that the metal case is
placed at or near line potential. Now if you "float" your drill by
using a three- prong -to- two -prong adapter (cheater) the case of
the drill will he at line potential and therefore your body will
present a path to ground. and you'll soon wind up under the
ground. Hoa ever. if the outlet is properly wired and no cheater
is used then the same faulty drill will present essentially a short circuit from line through the case hack to ground. thereby
drawing large amounts of current and blowing the fuse or
throwing a breaker for the circuit. safely removing the lethal
condition.
At this point. one could argue that even if the mixer end were
floated. the signal conductor's grounds still constitute a safety
ground. However. these wires normally carry only milliamps
and are therefore relatively smaller wires. Therefore if a fault
develops (which can momentarily exceed 100 amps). the small
ground wires could fuse before the fuse does! To counteract this
problem you should run a separate safety ground of very heavy gauge wire (about #12) and make sure it is tied to ground at an
outlet at one end and to all the chassis at the other.
The other approach to breaking the loop is to use direct boxes
on every signal source which is line -powered. This is very
effective and allows equipment at both ends to he grounded
normally. to keep costs down you can easily construct your
own.
SUMMARY
We have covered a lot of territory so let's step back and review
the basics:
I. Clean up the line. The significance of this should be
obvious if you've managed to read this far.
2. The importance of good cables cannot be overemphasized.
For best results. make your own with high -quality wire
and metal -shell connectors being sure to use good soldering practice (good solder too -60 40 resin -core only).
3. Isolate problems by simple swapping techniques. Swap
one component at a time so that you're confident you've
isolated the problem. If you become confused. swap
things hack until you are sure of your observations.
4. Use direct boxes whenever possible and eliminate ground
loops at all costs.
5. Correct faulty or inadequate shielding in instruments
employing magnetic pick -ups.
6. Upgrade equipment when necessary, such as ancient
pickups and tube amps.
One last word of wisdom -Don't succumb to superstition. I
will repeat a statement from the introduction. For every
observable phenomenon there
explanation!
is a
rational scientific
ROBERT BERKOVITZ
The FFT: Big -Time
Mathematics Comes
to Audio
Part Two: Making audio measurements with the FFT.
THF FAST FOU'RIER TRANSFORM (FFT) is a
particularly
efficient way to calculate the spectrum of a digital signal
by the discrete Fourier transform method. Part I of
this article dealt with the underlying principles in a
simplified way: now we take up practical applications.
A discrete Fourier transform is executed by repeatedly
multiplying the numerical values ofa digitized signal by values
of sines and cosines. The FFT achieves its efficiency -typically
hundreds of times faster than a direct Fourier transform -by
setting an important special condition. If the computed
spectrum consists of a set of multiples (harmonics) of a single
frequency. many of the multiplication operations required will
he identical. In an FE T. each such operation is done only once.
and the result is saved and moved about in the computer's
memory as needed. The FFT is important in many fields of
research and engineering because it provides answers that are
not readily available in any other way. Although it has
limitations, like all measurement methods, the FFT has come to
dominate the world of digital signal processing because it allows
users to look into a rich new world of signal characteristics.
especially where transient signals are concerned.
In this second and final part of the article, we will look at
some practical measurements made with an FFT system, and
see how its special properties influence the way in which results
appear. Although use of the FFT as an analysis tool is still
largely restricted to laboratory computer systems, a new
system' allows any owner of an Apple I I computer to install an
FFT analysis system that can carry out most digital signal
processing functions of interest in audio, medical and other
fields. All of the measurements and plots shown here have been
made with an Apple II with this system installed. At the end of
the article, there is an FFT program in BASIC and readers who
are interested will find it easy to learn more about the FFT
simply by running the program with different data inputs. The
program is usable with very little change on any computer
running BASIC. For Apple II owners, a simple graphic display
sub -routine is included.
a
1E.
Figure 1. Producing an impulse. (A) a single low frequency,
(B) the summation of four frequencies, and (C) the
response after the summation of 16 frequencies.
Robert Berkurir_ is Director of Research. Teledyne
Acoustic Research.
than 20 microseconds. Looking at such an impulse, it
sometimes seems intuitively difficult to believe that it has the
THE TIME DOMAIN AND THE FREQUENCY DOMAIN
Audio measurement is ordinarily carried out by comparing
the output of the tested device to the characteristics of the input
signal. For example, analog measurements of frequency
response are made by driving the tested unit with a signal that
changes frequency uniformly while maintaining a uniform
level. By tracing the level of the output on chart paper
synchronized to the frequency of the input signal, a plot is
obtained showing the variation in output with frequency.
In making such a conventional measurement, we send
frequency- domain information to the tested unit, and we get
frequency- domain information directly from the output. When
the FFT is used to obtain data corresponding to the frequency
response, the input signal is in the time domain, and so is the
output. The function of the FFT is to convert the data from the
time domain to the frequency domain.
What are we talking about? In analog frequency response
measurements, the test signal does not change with time, for all
practical purposes. The frequency must change slowly enough
so that the plotted result will be the same as if individual sine
waves that were the same forever were sent through the tested
device. With the FFT, as we will see, the important property
of the test signal is not its stability in time, but precisely the
way it changes with time. Later, we can review some of the
interesting peculiarities of the FFT's mathematics. First, let's
look more closely at the kind of test signal usually used with the
FFT. We'll take a loudspeaker as the device to be tested.
THE TEST IMPULSE
A widely -used test signal for FFT measurement of audio
equipment is an impulse of extremely small duration, often
synthesized by the same computer that carries out the FFT, and
lasting less than one sampling interval. For full audio -range
testing, where the sampling rate would be on the order of 50
kHz, for example, the impulse would have a duration of less
.,
7
nl1II1
rid 3
r
iJ
_._
f+
-Ñ
-12
-13
- 24
L
L:Hz
27 24
1
I-
Hz
Figure 2. A single impulse (A) produces an absolutely
flat magnitude response (B) and a flat group delay response
(C) as well.
23. 24
low frequency content needed to make a useful test signal.
However, the impulse contains all frequencies from d.c. to half
the sampling rate in equal measure. The falloff, in any case, is an
inconsequential reduction in high frequency content; the reason
will become clear in a moment.
One way to show that an impulse is indeed the sum of every
frequency that can be represented at a particular sampling rate
is to let a computer generate and sum waves of these
frequencies. Naturally, nobody would want to wait around
forever to prove the point, but it takes only a few minutes and a
simple program to see that the proposition is valid.
To get an impulse, the waves of identical amplitude and
gradually increasing frequency need to be summed with their
highest levels coincident. that is, "in phase" at one central point
where t = O. FIGURE IA shows the first wave, which is not d.c..
but can be imagined to be some very low frequency. Letting the
computer run for a few seconds. and stopping the program after
three more waves have been added. we get the picture in FIGURE
I B. It is clear that the added waves have done some cancellation
at the sides of the plot, but they can do nothing but add at the
center. where every wave is going to have a value of 1.0.
Allowing sixteen waves to be summed produces FIGURE IC.
after which there can be little question of how things are going
to go. If the program runs until several hundred waves have
been added. the central peak becomes a rather thin spike. and
the wavelets at the sides flatten out to give a quite credible
impulse. quad eral demonstrandtun.Another way to evaluate the frequency characteristics of the
impulse is (of course) to transform it from the time domain to
the frequency domain using the FFT. If we carry out an FFT
operation on the impulse shown in FIGURE 2A. we obtain the
magnitude plot shown in FIGURE 2B, which is absolutely flat
over the entire range.
To go a little further. we can look at the group delay, which
the computer gives us a few moments later if we ask it to do so.
shown in FIGURE 2C. Group delay is defined as the rate of
change of phase shift as a function of frequency. Phase shift is
more directly computed by the FFT. but group delay
corresponds more closely to the intuitive idea of time delay.
The group delay is 0.1 millisecond at all of the frequencies
shown. If we look closely at FIGURE 2A. we can see that the
impulse is in fact 0. I millisecond from the start of the time scale
at the left side of thedisplay, accountingforthedelayshown. The
important point is that there is no delay of any frequency
relative to all the rest, as predicted. No surprise there.
In the first part of the article. in reviewing the procedure for
calculating a Fourier transform, we multiplied the test
waveform by successive values of a sine and cosine wave of each
frequency. sampled at the same rate as our test waveform. If we
now consider what result this would give with the impulse. the
answer is (almost) obvious. Every cosine wave will start with a
value of I, regardless of its frequency, so the first product will be
I
I = I. After that, we will get nothing but zeroes as products,
because the impulse lasts for only one sample period. For sine
waves, which start with a value of zero. we will get only zero asa
result, in the first sample position and everywhere else. That
means that every frequency that can be represented at this
sampling rate will have the same magnitude: I.
A LOUDSPEAKER UNDER TEST
Let's look at an impulse that has passed through a
loudspeaker (FIGURE 3A). The impulse, generated on the IQS
circuit board, has actually passed through a power amplifier,
loudspeaker, microphone and analog-to- digital converter. The
loudspeaker used was a small, metal-cased extension speaker
with only the woofer working. The impulse was sent to the
loudspeaker sixteen times at precise intervals. Each time, the
computer waited for the sound to travel across the room to the
microphone before starting to take in and digitize about 23
milliseconds of data (2048 samples). Because the timing can be
controlled to within
a
millionth of a second without difficulty,
w
the successive impulse responses can be added and averaged
synchronously by the computer. This procedure reduces the
amount of interface from traffic noise and other environmental
disturbances, and is one reason that the FFT method can be
used in relatively noisy environments. Averaging sixteen
impulses gives an effective reduction of 12 dB in ambient noise;
averaging 128 impulses. as many as the IQS program allows.
gives an effective improvement of 21 dB in the signal -to -noise
ratio.
Astute readers will have noticed that I have not referred to
FIGURE IA as "the impulse response of the loudspeaker." Had
the measuring microphone been placed elsewhere, a different
plot would have been obtained. Indeed, there are as many
impulse responses as there are possible microphone positions.
There is no easy way out, because a real loudspeaker and the
27
¿_S
m
i
l
l
i
iec:onds
-
C
-a
i11
E.
ail.
Hz
.
.
23_24
+1 8
+1 2
+F
I-Hz
+12
+A;
signal it receives are topologically mismatched. The one dimensional. time -varying signal coming through the wires is
transformed to an N-dimensional output. with very large N.
Engineers who ignore this fact sleep more soundly. but the
loudspeakers they design leave something to be desired.
Pushing "I " on the keyboard of the Apple II gets us a 128 point FFT. putting FIGURE 3B on the display screen. 128 points
are only half of the plot of the impulse. so we are using as data
only the first 2.75 milliseconds. We obtain a 64-line spectrum,
with a resolution of about 360 Hz. This means that the first line
after d.c. is 360 Hz. the next one is 720 Hz. then 1080 Hz. and so
on. In terms of the logarithmic scale we are all familiar with in
audio measurement. the resolution at low frequencies seems
poor. However. by using a lower sampling rate -the standard
IQS system will sample at a rate as low as I kHz and an optional
version at 200 Hz -as much as ten seconds of low frequency
data can be taken in with a resolution of 0.1 Hz! Pushing "3"
provides a more detailed display (FIGURE 3C). by doing a 512 point FFT. while pressing "4." generating a 1024 -point FFT.
shows us FIGURE 3D. In the last case. only half the frequency
range is shown at a time. The last two FFTs used II and 22
seconds of data. and provide resolution of 90 Hz and 45 Hz
respectively.
How do we arrive at these figures? We start with the sampling
rate. which is46.4875 kHz. Half this figure is the upper limit of
measurement. 23.24 kHz, but in fact. only the range to 20 kHz is
accurate because of the need to cut response above 23.24 kHz as
sharply as possible to prevent data -sampling aliasing errors. To
continue. dividing half the sampling rate by half the number of
samples used in the FFT gives the frequency resolution. It is
often simpler to divide the sampling rate by the length of the
FFT for the same result.
FIGURE 3E shows the group delay measurement for the lower
half of the frequency range. based on a 1024 -point FFT, with
the vertical scale calibrated at the right -hand side of the plot in
milliseconds. Clearly. there is considerable delay at low
frequencies and a small, probably imperceptible amount
between 7 kHz and 8 kHz. The extreme variation above 10 kHz
is almost certainly due to reflections on the cabinet from the
metal grille. the mounting hardware. and the raised edge
around the front of the cabinet. These reflections scatter the
phase by interference and produce group delay results that look
like random noise.
PSEUDO -ANECHOIC MEASUREMENTS WITH THE FFT
The impulse response plot in FIGURE 3A was made by
carefully placing the loudspeaker on a stand one meter high at
the center of a small laboratory cleared of any large furniture.
The reflections produced by the surrounding walls arrived so
late that they are not in the picture. FIGURE 4A shows a more
typical situation. produced when the loudspeaker is in a
position in which it is likely to be used by a listener at home. that
is. against a wall with its driver units close enough to the floor to
produce prominent reflections. At about 2.5 milliseconds we
have a nearly perfect broadband reflection of the original
impulse. Taking as a rule of thumb one foot and I inches per
millisecond (my thumb is dimensionally precise). the path
length of the reflection is 2 feet IO inches longer than the direct
path from the loudspeaker to the microphone. In addition to
this reflection, a number of others can be seen.
A 256 -point FFT produces the spectral result in FIGURE 4B,
ragged and noisy. because of the effects of the reflections. In
fact. this is the total spectral result of everything seen to happen
in the data shown in FIGURE 4A for the first 5.5 milliseconds
after acquisition began. because the plot shows exactly the 256
points used for the FFT. By truncating the data, as in FIGURE
4C. we simulate anechoic conditions to some degree. To do so
exactly, we would have to depend on the termination of the
impulse response before the point of truncation, and locate the
silent interval between the end of the impulse and the first
reflection. This is very difficult to do when the loudspeaker
produces ample low frequency output and is near the walls of
the room, as low frequency components of the spectrum bounce
around for quite a while. However, as FIGURE 4D shows, the
/
-b
c,
!-Hz
+
.!+
kHz
Figure 3. An impulse passed through a loudspeaker (A)
produces the FFT response seen in the succeeding illustrations (8 -E).
11.62
`,f: ".FA
LY
m
l
i
l
i
1r-
seconds
s
5
+24
+1 8
+6
ttl
r
rs
M
i
l
l
23
Hz
IS
C.
SPECTRUM- VERSUS -TIME PLOTS
One nice feature of the IQS system is its ability to produce
plots of the spectrum with changing time, with the starting point
and the interval between spectra adjustable. This kind of plot,
suggested by Shorter of the BBC long before FFT days, has
been extensively used since Fincham and Berman of KEF
showed elegantly how well a computer could produce displays
of this kind. By attaching the Apple II to the new, low -cost
Hewlett- Packard 7470A plotter, FIGURES 5A and 513 were
produced from the same data shown in FIGURE 3A.
The method used is to do an FFT using the 512 samples of
data starting at sample 0, then the 512 samples beginning with
sample I. and so on. To produce the "mountain range" effect.
however. it is necessary to plot the curves backward -a minor
detail. Comparing the two plots. we see an interesting
difference. FIGURE 5A shows some high ridges at high
frequencies that suggest the occurrence of events which are, in
fact, unreal. When we start the FFT with a sample well into the
impulse response. we may begin in the middle ofa section of the
waveform with substantial amplitude. The FFT sees a steep
vertical rise as the beginning of an impulse and reads this as
having very large high -frequency components as a result. To be
quite correct, the FFT has the property of circularity: it sees the
test waveform not as a single event encompassed in the 256
points (or whatever) being tested. but as a wave that goes on
indefinitely with its beginning tied to its end. Starting the FFT
in the middle of the impulse creates a waveform that has this
eternally repeated steep rise in every interval of the lowest
frequency represented.
S.F
rids
+12
1_\
y
40
~'
+6
P
v
li
II
ll,
kHz
l
.
.
,
.
T
tJ
23.24
Figure 4. Impulse response of a speaker in a typical listening
room (A). and a 256-point FFT (B). Anechoic conditions may
be simulated by truncating the data (C). producing the
FFT seen in (D).
first 2.4 milliseconds of the impulse response are not quite as
bleak as might appear from the FFT of the un- truncated data. It
is also clear that movement of the loudspeaker would probably
improve stereo imaging substantially by eliminating strong
cancellation effects occurring throughout the frequency range.
Reducing the length of the data sequence used, whether by
truncation or by use of a smaller EFT. produces a smoothing
effect that reduces resolution. In the truncation example, just
shown. for example. the number of data points was reduced
from 256 to 112. Fven though the number of lines in the
spectrum remains 128. there is a spreading of the data in the
results that corresponds to changing from narrow -hand to
slightly wider -hand filters in a conventional spectrum analyser.
This is completely separate from the remosat of the reflections.
which introduces another kind of smoothing by eliminating
interference. However. the same kind of smoothing. due to
shortening of the data sequence. takes place when a
loudspeaker has an impulse response that is of very brief
duration. The frequency response curve will be very smooth.
showing few perturbations and none which are very small. In
such circumstances. taking larger and larger FFTs cannot add
to the information presented. but does give a more presentable
plot.
Figure 5. Computer- generated plots generated from the
data taken from Figure a
3U
The use of windows of different length and shape changes the
appearance of the plots rather drastically, and suggests the need
to find an optimum method of windowing. Whether or not such
an idea is mathematically meaningful is unclear, but it is clear
that the ear and brain --for which loudspeakers are designed
work in a quite different way. W hen we listen to a sound.
windowng of a kind takes place, but the length of the window
and to some extent its shape are different for every frequency.
The way to study the performance of audio systems would seem
to be to mimic this process, rather than to use a fixed window.
We have been doing this at Acoustic Research for the past few
years now... but that will have to be the subject of another
-
article.
FOOTNOTES:
The IQS 401 FFT Analysis System is available from computer stores
that deal with the technical community or from IQS. Inc.. 5719
Corso di Napoli. Long Beach. California 90803. The system consists of a circuit board that plugs into the Apple II. with built in
inpulse generator. analog -to- digital and digital -to- analog conI
.
version, sampling rate and anti -aliasing filters under software
control. complete software (including a speedy machine code FFT
stored on a chip) and a detailed manual.
2. Here is a chance to pick up on something very interesting. The
increased precision of location of the impulse in time as more frequencies are added has its exact counterpart in theories about
the fundamental properties of the universe. that is. quantum
mechanics. The example we have just been talking about makes it
easier to deal with the idea that waves and particles are somehow
the same thing if one is dealing with sufficiently minute quantities.
In quantum theory every particle has associated with it a frequency
proportional to its energy. Energy momentum equals frequency
times Planck's constant, remember? If the position of a particle in
space is well defined by some experimental observation, its energy
(frequency) will be spread over a large range of possible values.
This corresponds to our impulse being exactly defined in time.
by adding many frequencies together. On the other hand, if the
energy (frequency) is established exactly, the position of the particle
will be unknown. The same is true of the impulse: by assigning a
definite frequency to it, we get a sine wave going to infinity past
and future with no change... our impulse is nowhere. Setting either
quantity to any precise value necessarily involves losing precision
in the other quantity.
A BASIC FFT Program
INES 40 to 48 generate and store a decaying pulse
by sampling the value of E ' sin (t). The results
go into the real array N, while the imaginary
array P is zeroed at the same time (line 46).
Lines 50 to 140 contain the part of the FFT program
often called "the shuffle," in which the data is rearranged from its original sequence to one that simplifies
and therefore speeds up movement of the products calculated later. Sometimes, a shuffle is done after the actual
transform, instead of before, as in this example. Lines 150
to 270 are the transform, with stepping of sine and
cosine values in line 200, complex multiplication in lines
240 -242, and accumulation of products in lines 244-249.
Lines 280 to 310 calculate and print out the results which
have been stored in array S as the absolute values of the
real and imaginary array elements in N and P. Users with
computers other than the Apple II should skip to line 390,
where a simple reversal of the data order (404-414) prepares foran inverse FFT. Flag V is set to I in line 420, to
signal that it is an inverse FFT, and the program then
jumps back to its start.
Lines 320 to 384 provide Apple 11 users with graphic
output on the monitor screen, alternately showing the
spectrum and the source signal itself as the program does
forward and inverse FFTs, one after another.
The easiest way to experiment with this program is
to substitute a new function in lines 40-48, possibly rewriting the program to allow manual entry of actual data.
The automatic alternation of forward and inverse FFTs
can be defeated by dropping lines 420 and 430 and
moving the data reversal (lines 390 to 416) near the start
and specifying the direction of the transform from the
keyboard.
In practical FFT programs, the sine and cosine values
are normally stored in a table and looked up by the program, rather than being called from the interpreter as
here , to save time. This program has been adapted by the
author from one written in FORTRAN and given in
Stearns' "Digital Signal Analysis," published by Hayden,
and available in most computer stores. It runs quite rapidly when compiled. However, even use of a BASIC compiler does not approach the speed of a program written
directly in assembly language.
10
TEXT
20
30
':INPUT
PRINT "VALUE OF
DIM N(N), P(N), S(N)
42
50
60
70
80
90
100
110
120
122
124
130
132
134
140
150
160
170
18a
190
200
210
215
220
23a
240
242
244
246
248
249
250
255
260
270
280
290
295
300
305
310
315
320
330
340
350
360
362
364
370
375
380
382
384
390
400
402
404
406
408
410
412
414
416
420
430
440
0
7.
40
44
46
48
:V =
FOR
I
T
.375
=
=
N
TO N
1
(I
00
1)
-
NCI) = EXP ( - T)
P(I) = 0
NEXT
MR = O:NN = N FOR M = I TO NN
r.
SIN (T)
I
I
L
=
N
L
=
L
2
/
(MR +
=
MR
IF MR <
TR
=
N(M
N(MR
TI
>
L
::
=
M THEN
N(M
+
1)
=
+
+
I)
P(M
=
=
(
(NT (MR
140
L))
/
+
L
1)
N(MR
+
I)
+
I)
TR
=
+
P(M + I)
P(MR + 1)
NEXT M
L
NN THEN 80
L)
-
IF
MR
1)
P(MR
=
=
TI
1
IF
IS
L
>
=
N THEN
=
2
^
L
EL
=
L
280
FOR M
=
TO L
1
(1 - M)
COS (A)
NI = SIN (A)
FOR
= M TO N STEP IS
A
HR
3.14159265
=
/
EL
=
I
J
=
+
I
TR
=
TI
=
N(J)
P(J)
N(I)
PCI)
NEXT
NEXT M
L
L
NR " N(J)
MR ^ P(J)
N(1)
PCI)
N(I)
P(I)
=
=
=
=
+
-
TR
-
TI
+
TR
WI
WI
X
P(J)
N(J)
TI
I
IS
=
GOTO
160
T
=
IF
TO N/ 2+
SQR (N(K) ^ N(K) + P(K) ^ P(K))
K=
FOR
V=
I
1
T= T/
THEN
1
N
S(K) = T
PRINT 5(K)
NEXT K
INPUT "PRESS ANY KEY TO DISPLAY GRAPHICS.
Q = 0:R = 279 / (N / 2 + 1)
FOR I=
TO N/ 2+ 1
IF SCI) > Q THEN Q = 5(I)
NEXT
FOR I=
TO N/ 2+ 1
5(I) = 159 - 159 K (S(I) / Q)
NEXT
MGR
1
I
1
I
HCOLOR
FOR
=
I
V =
1
IF
FOR
X
=
3
I=
HPLOT
NEXT
1
-
I
=
2
TO N
N
-
I
+
=
I
1,
2
/
1
S(I) TO
THEN 430
N(X)
N(X) = N(I)
N(I) = Y
Y = P(X)
P(X) = P(I)
P(I) = Y
NEXT
V = 1: GOTO 50
V = 0: GOTO 50
END
Y
N/ 2+
TO
100 R
2
oe
I
R
-
1,
159
"
;K$
JOSEPH COENCAS
The New York Center for Media Arts
Adds a School of Audio Arts
Ill: HIG APPI.r has long been a major draw for young
people seeking careers in the recording. television.
photography and advertising industries. However,
few of these people have much more than desire going
for them. and finding work can he difficult. frustrating, and
finally. a disappointing experience. Even the four -year college
eduction makes no guarantee of employment.
Enter the Center for the Media Arts, a two -year -old
consortium of schools for video (formerly RCA Institutes).
photography (The Germain School). and advertising art and
design (The Pels School). The Center recently purchased a S4
million building on Manhattan's West 26th Street. to combine
these schools. and the new School of Audio Arts, under one
roof. A SI million renovation of the ten- story. 100.000 sq. ft.
building has been underway since last spring. with the Center
planning to he operational in time for the fall '82 semester.
The School of Audio Arts was designed. and will he directed.
by Harry Hirsch, founder designer of two of New York's
leading recording studios. MediaSound and Soundmixers.
Hirsch is first vice- president and chairman of the Education
Committee of the New York chapter of NARAS (National
Academy of Recording Arts and Sciences). as well as adjunct
School of M usic Business Technology.
professor. NY
Stressing hands -on learning. the school will provide students
with individual work stations for music recording. sweetening.
editing and mixing. and for equipment maintenance.
According to Hirsch. Scott Cannel! (the Center's VP of
program development) wanted to develop a broad -based
curriculum that would help prepare students for audio work in
film, video, radio and multi -media, as well as for all facets of the
recording studio industry. "He was not particularly interested
in just buildinga recording studio and producing mixers. So, we
developed a 700 -hour program highlighted by 430 hours of
hands -on workshops emphasizing the three major areas of
audio training: craft and creativity, business and management.
and repair and maintenance.
"Each work station will be equipped with a Ramsa I2 -in. 4 -2 -1
out multi -track console, fed by a central 16 -track recorder that
will play back music programs in various instrumentations
(including piano and voice, rhythm section. big bands and small
groups with chorus). The program will be fed into each work
station, where the student will do a mixdown assignment on a
reverb -equipped console, monitoring by headphones. The mix
will be done on a stereo cassette, which will be graded. The
students will spend 100 hours completing this phase of the
program.
"One course that's sure to be popular is 'Mixing for Video
and Film, "" Hirsch adds. "Our mixing lab will have a large
MGA four -foot video projector and synchronized tape
recorder. Using SM PTE time code, our students will gain some
first -hand experience syncing picture with sound."
in the background, Harry Hirsch (complete with an
official Otari tee- shirt) tries to imagine what it will all look
like when it's finished.
The school's edit lab will be equipped with 15 Otani 5050B
two -track machines. Prepared edit assignments will incorporate work and master reels with headphone monitoring. A state of-the -art recording studio. large enough to hold 40 musicians.
will boast an MCI 32 -in 24 -out hoard. 24 tracks of Dolby noise
reduction, an Otani MTR -90 24 -track recorder and a variety of
microphones.
A small announcer's booth and two isolation areas will be
cued into the Center's School of Television Arts, so that visual
information may be sent through video switchers providing the
capability of producing programming for cable television. The
school is presently negotiating with the music departments of
several area colleges to provide musicians for recording
sessions. The new building will even house a 2.000 sq. ft.
video/ sound stage, hooked into a 24 -track control room
complete with loudspeaker monitoring, intercom and video
circuits. Conceivably, a symphony orchestra could he recorded.
then edited and mixed upstairs in the Audio Arts labs.
Students will be required to complete 100 hours of
Electronics Lab, including sessions in AC DC circuitry, and
semi -conductor electronics. This is to provide the foundation
that Hirsch feels is necessary in order to go on into troubleshooting, repairing and maintaining the complex audio
equipment found in the contemporary recording studio.
Hirsch has promised (are you reading this. Harry?) to deliver
a full construction story as soon as the work is over. (Well.
maybe the day after.)
m
New Products
STEREO SLAVE
4 -BAND
PARAMETRIC EQUALIZER
GRAPHIC EQUALIZER
Integrated Sound Systems has recently
introduced the TI)M -8200 Stereo Slave.
When coupled with the TDM -8000
Audio Time Compressor. stereo sound
tracks can he compressed without altering the original pitch and tone. The
TDM -8000 8200 produces a stable.
time -synchronized stereo image by
making intelligent logic splicing decisions between channels- Vocal and
instrumental sounds that are common
to both channels will remain stable with
respect to stereo image. and processed
stereo sound tracks can be played in
the monaural mode, without cancellations or other adverse effects. Radio real time applications include not only stereo
FM. but stereo AM as well, since the
audio processing is completely compatible with any of the stereo AM systems.
The TD M-8.000 8200 is used with Type C
broadcast video recorders. 'A-in. variable speed video cassette decks. variable speed turntables, and audio tape machines. The TDM -8000 8200 compresses
stereo music up to 1.5 times, and maintains high frequency response and
dynamic range, while allowing very little
distortion. At any point from I to 1.5
times compression of original material.
the frequency response is 20 Hz to 15 kHz.
the dynamic range is 81 dB. and theTH D.
IM. and noise is never greater than
0.3 percent.
t1/i: Integrated Sound Systems. lnr.
Price: Tl).11 -8000: 54,995.00:
T1).11-8200: S2.800.00
Circle 40 on Reader Service Card
tT141144a
111eISt:j
1-
The PE-40 rack -mountable parametric
equalizer is a high resolution alternative
to more common graphic and quasi parametric type EQ. The PE-40 has four
identical channels which may be used to
process four discrete programs. or the
channels can be cascaded when more
extensive frequency correction is necessary. Each of the PE -40's four channels
has four overlapping bands. Center frequencies may be swept from 40-800 Hz.
500 to IO kHz, and 800 -16 kHz. A concentric knob adjusts the "Q" (sharpness)
of each band from 1.1 to 5 so that a
broader or narrower band of frequencies
is affected. A separate knob adjusts each
band's gain for up to 15 dB of boost or
cut. In PA work, the PE -40 can be tuned
to the exact center of each feedback node.
In addition to the parametric EQ. each
channel has three push button -selectable
filters, two high pass and one low pass.
The 60 Hz. 18 dB octave filter cuts out
rumble, motor noise and other sub sonics while the 160 Hz, 6 dB octave
filter reduces wind noise and further
increases gain before feedback to avoid
howling in PA work; both high pass
filters can he combined for an even
steeper low frequency roll -off. The
15 kHz, 12 dB octave I.P filter can be
used to reduce hiss, cut leakage from
adjacent instruments. etc.
T- IS(. -1.t/
Circle 4/ on Reader Service Card
The Model 160 Graphic Equalizer
provides a ±12 dB of correction capability at the center frequencies of the
IO octaves which encompass the musical
spectrum. For matching the output
between equalized and bypass modes.
±8 dB of overall level adjustment is
provided. Each of the slide controls is
center tapped to ground in the flat posi-
tion. so that all frequency selective
networks are balanced out of the signal
path. The 160 also provides an optional
microphone and a test disk of band limited pink noise. The microphone
plugs into the equalizer. which provides
a Iront panel LED readout of level as
each band is played. enabling correction
of each. independently for each channel.
at the listening position. In addition to
the inclusion of a tape monitor function.
a Record switch enables the equalizer to
he inserted into either the record or the
playback path of the tape machine connected to the monitor. The 10-I.ED
front panel display provides standard
increments from 20 to +3 dB. and when
the microphone is not connected. they
show the overall output level. Two
sensitivity ranges are provided.
.11/r:
/)arid llaller Company
Circle 42 on Reader Service Card
25,000 copies in print
Fifth big printing of t e
definitive manual of recording tec n
John Woram has filled a gaping hole in the audio literature. This is a very fine book ... I recommend it highly.
High Fidelity. And the Journal of the Audio Engineering
Society said, 'A very useful guide for anyone seriously
concerned with the magnetic recording of sound.
-
So widely read ... so much in demand .. that we've had
to go into a fifth printing of this all- encompassing guide to
every important aspect of recording technology. An indispensable guide with something in it for everybody
to learn, it is the audio industry's first complete
handbook on the subject. It is a clear, practical,
and often witty approach to understanding what
8 clearly- defined sections
18 information -packed chapters
The Basics
vec.ou,
Sound
I.
I
nip
Il. Transducers: Microphones
and Loudspeakers
Microphone Design
Microphone Technique
Loudspeakers
III. Signal Processing Devices
Echo and Reverberation
Equalizers
Compressors. Limiters and
Expanders
Flanging and Phasing
IV. Magnetic Recording
Tape and Tape Recorder
Fundamentals
Magnetic Recording Tape
Tape P,-nrder
makes a recording studio work. In covering all
aspects, Woram. editor of db Magazine, has provided an excellent basics section, as well as more
in -depth explanations of common situations and
problems encountered by the professional engineer.
Noise and Noise Reduction
Noise and Noise Reductio"
Principles
Studio Noise Reduction Syste--
V.
VI. Recording Consoles
The Modern Recording
Studio Console
VII. Recording Techniques
The Recording Session
The Mixdown Session
VIII. Appendices
Table of Logarithms
Power, Voltage. Ratios and
Decibels
Frequency. Period and
Wavelength of Sound
Conversion Factors
NAB Standard
Bibliography
Glossary
r
Its a "must.' for every working professional ... for
for every audio enthusiast.
every student
ELAR PUBLISHING COMPANY, INC.
1120 Old Country Road, Plainview, N.Y. 11803
copies of THE RECORDING
STUDIO HANDBOOK. $37.50. On'15 -day approval.
Yes! Please send
Name
Address
City /State /Zip
Total payment enclosed $
(In N.Y.S. add appropriate sales tax)
Please charge my
Use the coupon to order your
copies today at S37.50 each.
And there's a 15 -day
money -back guarantee.
Account
it
Master Charge
BankAmericard /Visa
Exp date
Signature
(charges not valid unless signed)
Outside U.S.A. add $2.00 for postage.
Checks must be in U.S. funds drawn on a U.S. bank.
c
Classified
Closing date is the fifteenth of the second month preceding the date of issue.
Send copies to: Classified Ad Dept.
db THE SOUND ENGINEERING MAGAZINE
1120 Old Country Road. Plainview, New York 11803
Minimum order accepted: $25.00
Rates: $1.00 a word
Boxed Ads: $40.00 per column inch
db Box Number: $8.50 for wording "Dept. XX,' etc.
Plus $1.50 to coser postage
Frequency Discounts:
6
times, 15 %;
12
times, 30%
ALL CLASSIFIED ADS MUST BE PREPAID.
FOR SALE
THE LIBRARY...Sound effects recorded
in STEREO using Dolby throughout. Over
350 effects on ten discs. S100.00. Write
The Library, P.O. Box 18145, Denver,
Colo. 80218.
FREE
32pg Catalog & 50 AudloiVideo Avplic.
.otx.
e..
e.meilH
e dsc
VIDEO.
e,,.,
FOR SALE: YAMAHA PM 1000 -24 mixing
console. 32 main frame. Carlo ATA case.
gooseneck lights. excellent condition.
Must sell -best offer. Mark Lowrance
1033
N
-
USED RECORDING equipment for sale.
Dan (415) 441 -8934.
if
-....
TO ELIMINATE FEEDBACK
HOWL WITHOUT GAIN REDUCTION...
WANT
ACTUALLY WITH INCREASED GAIN?
FIND OUT ABOUT OUR FEEDBACK
STABILIZERS MOD. 741XR & 742XR.
(D
BODE SOUND CO.
1344 ABINGTON PL.
NO. TONAWANDA, N. Y. 14120
716- 692 -1670
ROCKET STUDIOS -All inventory, Teac
90-16, Pioneer 4 -2 track. Allen -Heath
modified 16 x 8 x 16, Neumann, Ecoplate,
AKG, much more. complete studio $50.000.
(303) 567-2965, (303) 526 -1881.
SCULLY 284 -8 TRACK tape machine
w, sync master. Best offer. Birch Recording Studio, Secretary, Maryland. (301)
943 -8141. Reason for selling. gone 16 track.
REELS AND BOXES 5" and 7" large and
small hubs. heavy duty white boxes.
W -M Sales. 1118 Dula Circle, Duncanville,
Texas 75116 (214) 296-2773.
(901) 885 -4504.
O
TV Audio & Rod Prod consoles
(213) 934 -3566
Ar
Sycamore
LOS ANGELES CA. 90038
a A.a.. 0321 Ampio
OOPAMP LABS INC
sues
FOR SALE -WHITE 3900 narrow band
filter system for control of feedback and
ring modes. 95% complete. S3.000.00.
Call for details. Ron (216) 699-9975.
xe......
-ANNOUNCING
"LOUDSPEAKER ARRAYS
A GRAPHIC METHOD OF
DESIGNING"
-
Now the design of loudspeaker arrays is
as easy as the design of the rest of the
sound system. Our method is easy to
use, is fast and accurate!
We will design your array or do any of the
steps you don't want to do in the array
design process.
Either way, you end up with a description
of the array, and drawings from as many
views as you want.
Call
or write:
(612) 871 -6446
north star sound UMBULUS
1406 First Avenue South
Minneapolis, MN 55403
Thomas McCarthy. Cruel Engineer
ORBAN and LEXICON. All products in
stock. FOR IMMEDIATE DELIVERY.
UAR Professional Systems, 8535 Fairhaven, San Antonio, TX 78229. 512 -6908888.
FOR SALE AKG C -24 and other tube type
condenser mics. (415) 441 -8934 or 527-
6167.
PROFESSIONAL AUDIO EQUIPMENT
Shop for pro audio from N.Y s leader. no
matter where you live' Use the Harvey
Pro Hot Line (800) 223 -2642 (except N Y..
Ak. & Hi.). Expert advice. in -depth parts
video systems available.
Broadest selection such as Otani, Quad
Eight. Soundcraft. Tascam and more.
Write or call for price or product info:
dept..
Harvey Professional Products Division
25 W. 45th Street
New York, NY 10036
(212) 921 -5920
l
ACOUSTIC TEST INSTRUMENTS
S. OIE19
USE TO ADMT COMMIS.
WEARER PEACEMBIT.
Comp.* Tan Sa
(1e6,aad Sound
m.
$449
lrwl Mar
$199
imagism] Oct... P,n1
Nr. Grew.
9oCt0l on Amua,[ Tonto.*
no pall WAIL ENGINEERING.
P o
1339
590
o*91
05
9o. 596. Manmmn. W 09!3$
12011 6474377
MCI JH -416 CONSOLE 18 x 16 x 24, all
factory updates. headroom kit. selectable
mid -range dip. (EC) comparable to 428)
full patch facilities, matching walnut
producer's desk, 511,000 or offer. Also,
extra JH -416 input modulesand Mellotron.
THE MOST FLEXIBLE MICROPHONE
& SNAKE CABLE
Available...1 to 24 pairs. No jacket
shrinkage while soldering. Call /Write
for FREE CATALOG. BI- TRONICS Inc.,
Dept. D, Box 125, Tuckahoe, NY 10707.
(212) 585 -1645.
UH-OH!
FOR SALE: U.R.E.I. Model 100 -A Soni-
pulse acoustical audio system analyzer
with AKG C- 451 -E calibrated microphone.
Excellent condition. $600 or best offer.
Call Tom at (215) 374 -2784.
(612) 774 -4857.
REMANUFACTURED ORIGINAL equipment capstan motors for all Ampex and
Scully direct drive recorders, priced at
$250., available for immediate delivery
from VIF International, P.O. Box 1555,
Mtn View, CA 94042, phone (408) 739 -9740.
BLANK AUDIO AND VIDEO CASSETTES
direct from manufacturer, below wholesale, any length cassettes: 4 different
qualities to choose from. Bulk and reel
mastertape -from 1 -inch to 2-inch. Cassette duplication also available. Brochure.
Andol Audio Products, Inc., Dept. db,
42 -12 14th Ave., Brooklyn, NY 11219. Toll
free 1- 800 -221 -6578, ext. 1, NY residents
AMPEX, OTARI, SCULLY -In stock, all
major professional lines; top dollar trade ins; write or call for prices. Professional
Audio Video Corporation, 384 Grand
Street, Paterson, New Jersey 07505. (201)
523 -3333.
CUSTOM LENGTH CASSETTE BLANKS
made to order featuring AGFA, Scotch,
and Magnetic Media tape, any length from
C -2 through C -122. For pricellst write
M & K Recordings, Box 195d, Mt. Morris,
MI 48458 or call (313) 687 -7610.
SERVICES
!xK
(212) 435-7322.
MASTER THE RECORDING SCIENCE.
Sherman Keene's "Practical Techniques
for the Recording Engineer" is a book
about the real world of studio recording.
Acclaimed by magazines. reviewers, college teachers, studio owners and engineers. 381 pages, 28 chapters (4 on computer mixing). To order send $29.75 plus
6°ï9 (Calif. only) and $2.75 shipping to.
Sherman Keene Pub., 1626 N. Wilcox No.
677 -D, Hollywood, CA 90028.
MAGNETIC HEAD relappìng -24 hour
service. Replacement heads for professional recorders. IEM, 350 N. Eric Drive,
Palatine, IL 60067. (312) 358 -4622.
Have you mis-
placed your db
again? Our
high quality,
royal blue vinyl
binders keep
ACOUSTIC CONSULTATION- Specializing in studios, control rooms, discos.
Qualified personnel, reasonable rates.
Acoustilog, Bruel & Kjaer, HP, Tektronix,
[vie, equipment calibrated on premises.
Reverberation timer and RTA rentals.
Acoustilog, 19 Mercer Street, New York,
copies of
db neat and
handy for
ready
12
reference.
Just $9.95,
NY 10013 (212) 925 -1365.
SURPLUS JBL TWEETERS. NEW. Ideal
for disco tweeter arrays or improving
sound reinforcement and home loudspeaker systems. Specify choice of bullet
or slot lens type. $80 plus $2 shipping
each. Add sales tax if in NY. Rosner
Custom Sound, 11 -38 31 Ave.,-L.I. City,
NY 11106. (212) 726 -5600.
AUDIOTRONICS CONSOLE-#501 son
of 36 grand, 24 track, 26 inputs. custom or
regular E.Q. (your choice). $15.000.00.
With factory installation, $16,000.00.
Neumann Cutting System with SX 68 hd,
P.D.M. Limiter, Variable Pitch, Variable
Depth, Control Panel, Tape Recorder with
Preview HDS., and speakers. $45.000.00.
Westlake speakers $2,000.00. 16 Channel
dbx $5,000.00. Call Paul (312) 225 -2110
or (312) 467 -9250.
available in
North America
only. (Payable
in U.S. currency
drawn on U.S.
INSTRUCTION
SHERMAN KEENE CORRESPONDENCE
course. Author of acclaimed textbook
"Practical Techniques for the Recording
Engineer" invites you to study recording
at home. Course includes reading and
homework assignments in two textbooks
with personal dialog via cassette. Eight
lessons per level, 3 levels. $250 per level.
For info write: Correspondence Course,
1626 N. Wilcox No. 677D, Hollywood, CA
90028.
banks.) ORDER
YOURS NOW!!
Sagamore Publishing Co., Inc.
1120 Old Country Road
Plainview. NY 11803
YES! Please send
db binders
(u. $9.95 each, plus applicable sales tax.
Total amount enclosed S
Name
WANTED
Company
Address
WANTED: TRANSCRIPTION discs, any
size, speed. Radio shows, music. P.O.
BOX 724 -db, Redmond, WA 98052.
City
StateiZlp
_
People.. Plaies..
Digital Sound Recording is proud to
announce the appointment of Sandy.
Taylor as vice president director in
charge of Technical Marketing Services
including film, video, music and digital
recording. The announcement was made
president Van Webster who
reports that. "...due to heavy bookings
of recent months in video and music
projects, the addition of a technical
marketing director is essential to providing our clients with the best services
possible." Taylor was formerly administrative vice president general manager of
Anchor Leasing Corporation.
by DSR
Robert Pabst, president of ElectroVoice. Inc.. has announced the following
staff changes. Paul McGuire has been
promoted to National sales manager:
Greg Hockman has joined Electro -Voice
as Marketing Manager, Music Products:
and Jesse Walsh comes on board as
Central Region sales manager. McGuire's
tenure with Electro -Voice dates back to
1972. In his position as National sales
manager, he will oversee all ElectroVoice sales activities in the Pro Sound.
Music Industry, Broadcast and Recording. Commercial Sound. and Consumer
Hi -Fi markets. McGuire's most recent
position with Electro -Voice was as
director of Marketing Music Products.
Jerome C. Smith has been named
('erwin -Vegás director of Digital De-
N
co
velopment. The appointment was announced by Larry Phillips. President of
Cerwin -Vega International. Smith will be
responsible for the development and
marketing of digital products. primarily
loudspeaker systems for both residential
hi -fi use and recording studio monitors.
Smith's immediate duties are to develop
and assist in marketing plans for the
recently introduced Digital Series of
residential loudspeakers and also to
consult in the development and marketing of a new line of "digital ready"
recording studio monitors. Previous to
joining Cerwin -Vega. Smith was an
independent consultant specializing in
recording studio acoustics and playback
systems. He was also a founder and
owner of Express Sound Company, a
firm that designed and installed sound
reinforcement and playback systems for
a variety of recording facilities. Smith
was a training manager at Teac and
founded The Sound Factory -a retail
store specializing in musical instrument
and sound reinforcement products.
Dan Tynus, vice- president and general
manager of Sound Studios, announced
recently that he and a group of investors
have purchased the registered trademark
name of "Quantitape" and its tape
duplicating equipment from its parent
company in New York. Diversa-Graphics, Inc.. and have formed a new corporation called Quantitape Duplicating, Inc.
Quantitape is a state -of- the-art facility
capable of duplicating all industry
formats. including critical stereo music
and pulsed AV presentations. A complete
line of private label blank cassettes will be
introduced for a wide range of user
applications. In addition to tape services.
another dimension to the company is a
new division called Quantidisc. It will do
record mastering and produce pressings.
Floppy disc reproduction plans are in the
offing.
Acting on recommendations made by
the Board of Directors, the Executive
Committee of the National Radio Broadcasters Association has voted to appoint
an executive vice president to manage the
affairs of the association and to facilitate
a planned expansion of NR BA's activities. A recent survey and study conducted
among NR BA Board members by Board
Chairman Bill Clark developed a consensus for taking prompt action to
accelerate the association's recent. rapid
growth. The Board suggested an expansion of member services and an intensification of the already highly successful
member recruitment campaign. The
Board also directed a maximum effort to
achieve true and full deregulation
through legislation.
Frank Santucci, well -known in both
the audio and video industries. has
founded Advanced Marketing. Advanced
Marketing is an independent manufacturer's representative that is dedicated to
serving the audio video marketplace by
offering top -quality equipment for
broadcast, production and post- production. Mr. Santucci brings more than 20
years experience to AM. having been
senior product manager for Ampex,
marketing manager for Orban and most
recently National sales manager for
Harris Video Systems (('VS). Initial
product lines to be offered by Advanced
Marketing will be Hedco (routing
switchers. distribution equipment).
Asaca /Shibasoku (color monitors. signal generators. test equipment) and
United Media (computer -assisted editors.
SMPTE time code equipment).
Altec Lansing president William Fowler
recently announced the hiring of Mr.
William Chambers as new vice president
of Marketing and Strategic Planning for
the Anaheim -based manufacturer of commercial and home sound system products.
With extensive experience in marketing
planning and analysis. Chambers comes to
Alter after 19 years with Black and Decker
and subsidiary company McCulloch. At
Alter. Chambers will be involved in analyzing the Company's current business
activities and their relation to optimizing
future market opportunities.
WNVC, a new non-commercial educa-
tional TV station serving Northern
Virginia. will begin broadcasting this fall
with RCA transmitting systems valued at
approximately SI million. The new
equipment. on order from RCA Commercial Communications Systems Division. includes a TTU-601). 60- kilowatt
UHF transmitter. and a TFU -33.1 pylon
antenna. Also included in the equipment
order are four RCA studio and electronic
newsgathering cameras which will he
used in WNVC's new studio facilities.
WNVC will he operated by the Central
Virginia Educational TV Corp.
BGW Systems. Inc. has appointed
Theatre Projects Limited as exclusive
distributor for all BGW Systems professional and commercial power amplifiers
in the United Kingdom. Theannouncement
was made by BGW Systems sales manager. Irwin Laskey. As exclusive distributor. Theatre Projects Limited will
offer the complete BGW Systems line.
including the PROL IN E" power amplifiers.
Lynette Robinson has been promoted
to Executive Secretary of the Society of
Motion Picture and Television Engineers
(SM PTE). the top SM PTE staff position.
it was announced by SM PTE President
Charles E. Anderson, Ampex Corp. As
Executive Secretary. Mrs. Robinson will
be in charge of SMPTE Headquarters
with responsibility for supervising the
SMPTE staff and acting as liaison between SMPTE officers and Headquarters.
She will also be involved in coordinating
SMPTE conference activities. including
finances, registration. and exhibits.
Mrs. Robinson has been on the SMPTE
staff for eight years. Prior to her promotion to Executive Secretary. she was
manager of Conference Programming.
Scheduling. and Sections.
... 81 Happenings
Computing on the Road
HP -75 with the TV. using the HP -82163
Video Interface. The interface will also
work with the HP-41 series of calculators.
this project conies to fruition. it will
solidify our place in the industry as the
certifying agency for digital music. Who
is better qualified to tell the consumer the
difference between analog and digital
music than SPARS studios? We have a
chance not only to authenticate the
recording of tomorrow's music. but also
to have the SPARS logo become a
recognised symbol of excellence to assure
the consumer that the product has been
properly produced.
Apparently. not everyone thought this
was a good idea. Digital Sound Recording's Van Webster had this to say to
SPARS president Chris Stone:
If you've been putting off going on the
road because you just can't bear to be
separated from your computer, you have
a friend at Hewlett- Packard. Over the
past few months, the company has introduced a variety of almost- pocket -sized
computers, some of which may be interfaced with a color TV set or video
monitor.
The latest entry is the battery- operated
HP -75 Portable Computer, which will
handle 169 instructions, including 147
BASIC commands. You can work out your
programs on the road, and store them on
magnetic cards, using the built-in card
reader, writer. When you get back to the
hotel at night, you can read up to 16 lines
of 32 characters each, by interfacing the
labels the opportunity to transfer their
"hot" library albums from analog to
digital for a processing charge only. If
The HP -75 features a miniature
keyboard for entering
BASIC
owearr
instructions.
I am deeply concerned by your comments suggesting that SPARS and a
"leading digital equipment manufacturer"
are planning to enter into a "massive
project" to offer record labels the
opportunity to transfer their "hot"
library albums from analog to digital for
SPARS News
processing charge only. The implications of such a project are staggering to a
a
digital audio marketplace which
.1 he following news item appeared in a
recent SPARS newsletter:
We are pursuing a massive project with
The video interface (lower left) lets
you see your programming efforts on any
TV set or video monitor.
leading digital equipment manufacturer who is considering providing us
with the equipment necessary to offer
a
is
already suffering from an excess of
production capacity. It is outrageous to
me to think that SPARS would offer
such a service with free equipment when
studios, including your own, have
purchased digital recorders and must
support them while the club" offers the
same service for a nominal charge. Two
..More Happenings
of the featured speakers at your road
shows have promoted the virtues of
digital archiving as a business enterprise
and now SPARS wants to cut off that
business from the very studios it was
formed to represent.
In addition. the unnamed manufacturer is short -circuiting his own sales
by going into competition with his
customers. This practice is not without
precedent. even in the digital audio field,
but should hardly be encouraged. let
alone supported by an organization such
as
SPARS.
On Tour with DISKMIX
DISKMIX is easily interfaced with
MCI JH -50. Sound Workshop ARMS,
and Valley People 65K systems, and is
Sound Workshop's new floppy disk based DISK MIX automation storage,
editing system recently came home from
a demo tour of the major recording
centers, including New York (Atlantic
Recording Studios). Nashville (Sound shop Recording Studios). San Francisco
(Harbor Sound) and Los Angeles (Pasha
Recording Studios).
engineer's normal mixing moves. The
system uses the multi -track recorder's
SMPTE time code track to lock all
automation data stored on the disk to the
master tape.
Sound Workshop plans to supply
continuous software updates. which will
be free for the first year after system
purchase. A special video production
software package is now being designed.
If SPARS
wants to promote digital
archiving by having its members who
have purchased and paid for digital
recorders offer to the labels a low-cost
sample run, then the marketplace will be
well served. If manufacturers, whose
principal interest lies in retail sales, want
to provide digital recorders to all and
rebates to those studios who long ago
made a financial commitment to digital
audio, so much the better. But for a
society that labels itself professional to go
into competition with its own members
and the community that it is pledged to
serve. is greedy and unethical.
I remain sincerely yours.
VAN W EBSTER
President
Digital Sound Recording
c.c. Music Connection Magazine
Pro Sound News
Billboard
Recording Engineer/ Producer
db
Journal of The Audio Engineering
Society
Sony Corp.
Studer
J VC
3M
Neve
co
E
R.
a>
ÿ
co
Webster's remarks are quite well -put,
and in a written reply. SPARS' Chris
Stone noted that "...the proposal has
been over- ridden and cancelled by our
Board of Directors, after discussions
with our Advisory Associate members."
According to Bart DiGrazia, SPARS
Administrative Director, the Board's
action was taken several weeks before
Webster's letter arrived, suggesting that
he was not the only one who objected to
the proposal. And, on a cheerier note...
a
"chaser" system which follows the
Seated at the keyboard of the mighty
DISKMIX system, Sound .Workshop
president Michael Tapes begins a demo
mix for an SRO audience at New York's
Atlantic Studios.
The Needle Knows
Even when your ears
can't tell the difference, your
VU meters can.
Which is why we test every
reel of our 2" Grand Master
456 Studio Mastering Tape
end -to-end and edge-to-edge.
make certain you get a
rock -solid readout with virtually
To
no tape -induced level variations from one reel of 456 to
another or within a single reel.
No other brand of tape
undergoes such rigorous testing. And as a result no other
brand offers you the reliable
consistency of
Ampex Tape.
consistency that lets you
forget the tape and concentrate on the job.
A
Ampex Corporation. Magnetic Tape Division
401 Broadway. Redwood City. CA 94063
(415) 367 -4463
AMPEX
Ampex Corporalon
One of The SKIridi Comparons
4 out of 5 Professionals Master
on Ampex Tape:
.iwo.tw.1
1981 -1982
Billboard Maga
Brand Usage Survey
Circle 36 on Reader Service
FROM OUR HANDS
TO YOUR HANDS
The Otari MTR-90 Series II
8, 16, 24 Channel. Master Recorders
At Otani, the focus of our work is on innovation and problem
solving. These values are carefully reinforced by our dedication to
quality; they are inherent in every tape recorder we engineer.
The new, second generation MTR -90 Series II multi-`
channel recorders are the embodiment of this philosophy. We
have refined the features and extended the performance and
capabilities of the MTR -90 by working closely with industry
leaders who demand the extra measure of technology and commitment. With
recording and film /video post-production facilities depending on the MTR -90, we've
stayed close to the needs of today's media production houses. The new Series II
machines are the logical result; a microprocessor- controlled recorder specifically
designed to easily interface with any SMPTE -based video editing
system, machine controller or synchronizer.
Once again, we've advanced the industry's most advanced
recorders. And, kept the same dedication to the craftsmanship ,,
we've always had.
-AMPFrom our hands to yours, the new MTR -90 Series II recorders are engineered
like no other tape machines in the world; with the quality you can hear and feel.
agfOLLTecrindogy you Can Touch.
Oran Corporation, 2 Davis Drive, Belmont,
TeL
592 -8311 Telex 910-376 -4890
CA
94002
(
rrr4
11
un Reader Sen'h r
i
aid
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

advertisement