Corrections and Revisions

Last updated 22 Oct 2005

The Computer Music Tutorial

Corrections and revisions

Curtis Roads
Media Arts and Technology
University of California
Santa Barbara
California 93106-6065 USA
clang@mat.ucsb.edu


These comments have not been incorporated into the September 1996 revised version.


Front matter

p. xiv: 2. Interdisciplinary Spirit Section, 1st sentence,
“The knowledge base of computer music draws from composition, acoustics, signal processing, synthesis, composition, performance, computer science and electrical engineering.” – Delete second use of “composition.”

p. xx "Phillip" Greenspun should be spelled "Philip."
p. xx. Daniel "Terrugi" should be "Teruggi." (one r and two gs).

PLEASE ADD
"For the Chinese edition I would like to acknowledge corrections supplied by Keiji Hirata, Takafumi Hikichi, James McCartney?, and Graham Hadfield."

Chapter 1

p. 7. Edison is mentioned. What about Charles Cros, who the French credit with the invention of recording?

p. 18. 440, 880, 1760 harmonics. This should be 440, 880, 1320.

p. 18. If we displace the starting point by 2π...
CHANGE TO
If we displace the starting point to π radians...

p. 27. DELETE THE SENTENCE:
"This means that one cycle takes longer than the interval between samples."
Actually it is shorter (11 cycles in 10 samples), but the explanation is clearer without mentioning it.

p. 20. In the discussion on the importance of phase, cite this from X. Serra:

The output of the deterministic analysis is a set of amplitude and frequency functions with one breakpoint for every frame. From these functions a series of sinusoids can be synthesized that reproduce the deterministic part of the sound. The phase trajectories are not kept because they are unnecessary in the final synthesis, they are perceptually irrelevant in most cases, and they make it harder to perform modifications. However, we have found some situations in which the preservation of phase data has made a difference in the quality of the resynthesis. These are: badly analyzed sounds, very low instrumental tones, and some vocal sounds. In the case of badly analyzed sounds, some of the trajectories may actually be tracking non-deterministic parts of the signal, in which case the phase of the corresponding peaks is important to recover the noisy characteristics of the signal. In the case when the analyzed sound has a very low fundamental, maybe lower than 30 Hz, and the partials are phase-locked to the fundamental, the period is perceived as a pulse and the phase of the partials is required to maintain this perceptual effect. Also in the case of some vocal sounds, the higher partials have a high degree of modulation that cannot be completely recovered from the frequency and magnitude information of the partials, but that seems to be maintained when we add the phases of the peaks.

Serra, X. 1997. "Musical sound modeling with sinusoids plus noise." In C. Roads, et al. 1997. Musical Signal Processing. Lisse: Swets & Zeitlinger.

p. 45 Table 1.2 Digital audio media, add this information on DVD
DVD Random
(optical)


120-mm diameter optical disc storing a maximum of 17 Gigabytes (on the four-layer version) and with a transfer rate of 11 Mbits/sec in the initial version. A DVD disc supports a family of related optical disc formats: DVD-ROM–a data storage medium, DVD-Video–for delivery of motion picture content, DVD-Audio–for delivery of high-resolution audio-only, DVD-R (Recordable)–a variation of DVD that permits one time recording of data, DVD-RAM–a variation of DVD that is erasable, and can be rewritten (See http://www.ima.org/cp/dvdsig/faq.html)

Chapter 2

Page 72
The book "(Soustrop 1991)" is cited. CHANGE TO Stroustrop 1997

Page 77
"On page 77, there are the append programs in Prolog and Lisp.
It seems to me that those program fragments are not wrong but, rather,
are redundant. Why did you use these?
Usually, the append program in Prolog consists of two clauses: the first
and the third in your program. Likewise in the Lisp program.
I think that you should substitute them with the standard and
conventional programs (the two-clauses version)." K. Hirata

Chapter 3

p. 89. "Max Mathews completed Music II in 1958; it was written in assembly language for the IBM 7094 7090 computer, a transistorized and an improved computer along the lines of the IBM 704. The 7094 7090 ran several times faster than the older vacuum tube machines. It was thus possible to implement more ambitious synthesis algorithms.

p. 95. Delete the sentences
"This concludes our introduction to fixed-waveform table lookup synthesis. The next section shows how aspects of synthesis can be varied over time."
Substitute:
"In the recent generation of microprocessors, the speed of numerical operations is much faster than reading from and writing to memory. As a result, the economy of table lookup, which involves reading from memory, is being called into question. Certain waveforms, such as sine waves, can be generated many times faster by directly evaluating their equation (McCartney 1997)."

P. 110, figure 3.11. The computer should be attached to a mic and ADC to gather the sound to be analyzed.

p. 112, caption of figure 3.12. "Three lines of comments" -> "Two lines..."

p. 119. "Composers of musique concrète work directly with sound objects (Schaeffer 1977; Chion 1982)."
ADD: "An important example of the mature form of the musique concrète idiom is the composition De Natura Sonorum (1975) by Bernard Parmegiani. Luc Ferrari took the concept of musique concrète in the direction of what is now called ambient or environmental music. Horacio Vaggione has combined musique concrète with sophisticated digital granulation techniques (see chapter 5)."

Chapter 4

p. 149, under the heading Data Reduction in Analysis/Resynthesis
CHANGE FROM: "Data reduction is essential to efficient analysis/resynthesis."
TO: "Data reduction is important to efficient analysis/resynthesis."

Chapter 5

p. 165. Change equation (1) from

wave(x,y) = (x-y)  (x-1) (x +1)(y-1)  (y+1)
to
wave(x,y) = sin(π/2  x)  sin(π/2y)

p. 165. Use this as the new figure 5.6



p. 173, Figure 5.12 Fourier grid. Make the vertical resolution more detailed, making it more realistic. There are only 12 frequency bands. Or say "Detail of..."

p. 175. "We say "quasi-synchronous" because the grains follow each other at more-or-less equal intervals. "
CHANGE TO: "We say "quasi-synchronous" because the grains follow at a slightly varying time interval."

p. 197. SYTER was used as an engine for time-varying subtractive synthesis by composers such as Jean-Claude? Risset in his composition Voilements Sud Voilements realized in 1985 (Wergo recording 2013-50).

p. 184. AFTER THE PARAGRAPH: Case (3) resembles case (2) except that the playback rate can be varied by a parameter that controls the speed at which synthesis advances through the samples. The playback can vary from normal speed to a slowed-down rate in which a single sample is repeated over and over again. Hence this method can be thought of as an interpolation between case (1) and case (2). "
ADD THIS NEW PARAGRAPH:
"The music of Horacio Vaggione, in compositions such as Schall (1995), Nodal (1997), and Agon (1998), are outstanding examples of time granulation of sampled sounds. Vaggione has developed a technique of shifting between different time scales by changing the size of the grains at pivotal points in these pieces. Granulation techniques also appear in several compositions in my POINT LINE CLOUD (Roads 2004)."

p. 184. "Granular synthesis constitutes a diverse body of techniques that share only the concept of sonic grains."
CHANGE TO: "Granular synthesis constitutes a body of micro-sonic techniques that share only the concept of sonic grains."

p. 184. "Finally, in combination with time granulation and convolution (Roads 1993a),"
CHANGE TO: "Finally, in combination with time granulation and convolution (Roads 1993a, forthcoming),"

p. 201-2. Check to make sure that figures 5.35 and 5.36 are numbered correctly on the original figures.

Chapter 6

p. 223, figure 6.7
Replace the existing figure with this new figure:



p. 261. "FM panameters" -> FM parameters.

Chapter 7

p. 270. "impendence" should be "impedance."

On Page 282
I think that the title of subsection "Waveguide Model of Plucked
Strings" should be replaced by "Waveguide Model of Struck Strings",

p. 293. "The computational resources for basic KS synthesis are modest (no multiplications necessary)." Delete parenthetical remark because this technique is not explained and the text uses multiplications.

p. 309. The caption of figure 7.22 should be: "A VOSIM pulsetrain. The parameters are explained in the text."

Chapter 8

p. 342. Add after final paragraph on GENDY.

Stochastic compositions come into being "out of nothing," by probabilities only. As Peter Hoffmann (1997) has pointed out, the initial conditions and time-varying parameter settings by which these blind probabilities are tamed are decisive in determining the character of the resulting sound. Hoffmann's reimplementation of Xenakis's program lets users adjust the synthesis parameters in real time. With only slight adjustments, however, the results may vary enormously. This is both the difficulty, but also the opportunity of stochastic synthesis.

Then add to the references
Hoffmann, P. 1997. "The New GENDYN Program."

p. 343

Figure 8.15
The time and amplitude barriers (P, N. T) defining a mirror constrain the next vertex generated from the vertex marked by an asterisk. If the next vertex generated stochastically (the initial projection I) falls outside the barriers indicated by the box, the barrier P overrides this choice, reflecting the vertex back into the box (reflection R).

Chapter 9

p. 363. Figure 9.6. The label for the "Auxiliary return section" appears to apply to the output section. Move the label to the side, showing that it applies to only one row.

Chapter 10

p. 424 "We can think of the coefficients a, b, ... i as elements in an array hi"
"i" is a poor choice of index. Change hi to hj and change xj to xk, etc.

Chapter 11

p. 463, Doppler shift, first described by the astronomer C. Doppler (1842).
CHANGE "astronomer" TO "physicist."

p. 475. Figure 11.17. The horizontal axis is Time, not Frequency.

p. 478. "loopDelay is the delay D expressed in seconds = D/R "
Add: ", where R is the sampling rate."

p, 487. "It is well-known that clouds in the atmosphere contribute a reverberation effect."
CHANGE TO: "It has been speculated that clouds in the atmosphere contribute a reverberation effect."

In the chapter on additive synthesis I mention John Chowning's Stria as an example. In the program notes with the Wergo version it says it is an example of FM synthesis. Thus another example of algorithmic control of additive synthesis is needed.

Chapter 12

p. 552. "When "normalized" to the range of -π and π it is called the wrapped phase representation. "
CHANGE TO: "When the phase spectrum is "normalized" to the range of -π and π it is called the wrapped phase representation."

Chapter 13

p. 575. "This can be demonstrated by subtracting the resynthesized version from the original signal to yield a residual." Actually they subtract the TPV spectrum from the original spectrum. See Serra 1996.

Chapter 14

p. 649. figure 14.11. "...one can varying" -> one can vary.

Chapter 15

p. 685. G. Englert -> name index.

Chapter 17

p. 809, "There are minor syntactic differences between Music 1000 and the Music N language previously described. Whereas Music N gives the signal to the left of the <- sign..." The text should refer to Music 0, not Music N.

P. 788 and 792. Check to make sure that figures 17.1 and 17.2 are numbered correctly on the original figures.

Chapter 19

p. 860. State 4 should be State 5 in figure 19.3.

Chapter 20

p. 921 figure caption 20.4 missing. Yamaha DX7II FD synthesizer.

p. 936. Chapter 20. "At a sampling rate of 50 KHz...within 1/50,000th of a second or 20 msec." Should be microseconds or µsec.

Chapter 22

p. 1007. Under "Bandwidth limitations" 320 msec and 960 msec should be µsec.

p. 1016. The International MIDI Association... Their address is: The MIDI Manufacturers Association may be contacted at this address:
MIDI Manufacturers Association
P. O. Box 3173
La Habra, CA 90632-3173 USA
www.midi.org

Appendix

p. 1075. "...finally published a theory..." ..."Fourier's original lecture was only published for the first time some 165 years later."
Change "finally published a theory" to "presented a revised theory"

p. 1100. Figure A.17 caption. Description of spectrum says 22 to 442 Hz, but image shows 4000 Hz. Redo this figure to show the variations with window size.

References

p. 1146. Bird, J. 1982. Percy Grainger. This is listed under the letter G. It should be under the letter B.

Levitt, D., 1981. "A melody description system for jazz improvisation." M.S. Thesis. Cambridge, Massachusetts: Artificial Intelligence Laboratory.
-> Add "M.I.T." before Artificial.

p. 1164. The reference to Marino et al. 1992. in Perspectives of New Music says "to appear." THE VOLUME AND ISSUE NUMBER ARE 31(1).

David, E., M. Mathews, and H. McDonald?. 1958. "Description and results of experiments with speech using digital computer simulation." Proceedings of the 1958 National Electronics Conference. New York: Institute of Radio Engineers. pp. 766-775.
SHOULD BE
Proceedings of the National Electronics Conference. Volume IV. Chicago: National Electronics Conference. pp. 766-775.

Mathews, M., J. Miller, and E. David, Jr. 1961. "Pitch synchronous analysis of voiced sounds." Journal of the Audio Engineering Society of America 33: 179-186.
SHOULD BE
Mathews, M., J. Miller, and E. David, Jr. 1961. "Pitch synchronous analysis of voiced sounds." Journal of the Acoustical Society of America 33: 179-186.

Manning, P. 1987. Electronic and Computer Music. Oxford: Oxford University Press.
UPDATE TO
1993, Second edition.

McAdams?, S. 1981. "Spectral fusion and the creation of auditory images." In M. Clynes, ed. Music, Mind, and Brain: The Neuropsychology of Music. New York: Plenum.
SHOULD BE 1982.

ADD:
McCartney?, J. 1997. "Synthesis without table lookups." Letters section. Computer Music Journal 21(3): 5-6.

ADD:
Roads, C., S. Pope, A. Piccialli, and G. DePoli?, eds. 1997. Musical Signal Processing. Lisse: Swets and Zeitlinger.

ADD:
Roads, C. 2002. Microsound. Cambridge, Massachusetts: MIT Press.

Roads, C. 2005. POINT LINE CLOUD. CD + DVD. San Francisco: Asphodel Limited.

Soustrop, B. 1991
SHOULD BE
Stroustrup, B. 1997. The C++ Programming Language. Third edition.
Reading, Massachusetts: Addison-Wesley?.

Name Index

ADD:
Strawn, J. p. 131, 150
Parmegiani, B., p. 119
Vaggione, H. p. 119, 184
Ferrari, L., p. 119

Subject index

ADD:
De Natura Sonorum (B. Parmegiani), p.
Schall (H. Vaggione), p.
Nodal (H. Vaggione), p.
Agon (H. Vaggione), p.

---

6 February 2000

Dear Dr. Osaka,

Thank you for your message.

I have corrected Figure 6.7 and I am sending it to
Tokyo Denki Daigaku Shuppankyoku tomorrow.

---

p. 256 bottom
x=cos(q) -> x= cos(theta)?

---

Yes.

---

p. 261 in the equation
csc -> sin

---

I cannot verify this. Moorer 1976 is a JAES paper
...

---

Dear Dr. Roads,

I am Naotoshi Osaka, who has almost finished traslation of your
"Computer Music Tutorial" into japanese, together with Keiji Hirata
and other members.

In December 1998, we have talked on some mistakes of a figure.
But somehow I did not correspond to your kind proposal.
As you know, we are now in the final stage of the edition.
Here again, I want to remind you a kind proposal.
I attatch the communication log at last.

Could we get a corrected version of Fig. 6.7?
I first thought about correcting the figure ourselves, but
the quantization of the sine wave and some apperance might come
differently. I would appreciate if you could send a corrected version
of the figure. Could we get a printed one?

The postal address to be sent is:
(Publisher's address)
Mr. Tohru Tokutomi
Tokyo Denki Daigaku Shuppankyoku
Kanda Nishikicho 2-2, Chiyoda-ku, Tokyo, 101 Japan

As for those author's mistakes, we have not systematically
reported to you.
My understanding is: Hirata informed you of all mistakes that he
found. Other members did not stacked those and have forgotten to report
them to you. Sorry about that.
I attach only some other minor mistakes in chapter 6.

Fig. 6.26, 6.27, 6.28 and Fig. 6.29
W should be italic w. All the function w in the body is itallic w.

p. 261 in the equation
csc -> sin

p. 256 bottom
x=cos(q) -> x= cos(theta)?


Thank you in advance.

Naotoshi Osaka
Leader
Media Representation Research Group
Media Information laboratory,
NTT Communication Science Laboratories.

Tel: +81-462-40-3655 Fax: +81-462-40-4725 e-mail: osaka@brl.ntt.co.jp
http://www.brl.ntt.co.jp/people/osaka/index.html

---

Date: Thu, 10 Dec 1998 14:45:06 -0500
From: Curtis Roads <100407.574@compuserve.com>
Subject: A question on "Computer Music Tutorial"
Sender: Curtis Roads <100407.574@compuserve.com>
To: Naotoshi Osaka <osaka>
Message-Id: <199812101446_MC2-6331-E3AD@compuserve.com>
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

To: 100407.574@compuserve.com
Subject: A question on "Computer Music Tutorial"
Date: Thu, 10 Dec 1998 19:12:27 +0900
From: Naotoshi Osaka <osaka@idea.brl.ntt.co.jp>

Dear Dr. Roads,

My name is Naotoshi osaka, who is translating your "Computer Music
Tutorial"
together with Keiji Hirata and other members.

I want you to check Figure 6.7 on page 223.

This explains the amplitude modulation.
I do not think the time scale in (c) does not correspond to those of (a)
and (b). That is, (c) should be written stretched double in horizontal
direction. It seems to me that you misunderstood (b) is bipolar.

Thank you.

---

Naotoshi Osaka
Leader
Sound Representation and Computer Music Research Group
NTT Basic Research Labs. Information Science Research Lab.
Tel: +81-462-40-3655 Fax: +81-462-40-4725 e-mail: osaka@idea.brl.ntt.co.jp
http://www.brl.ntt.co.jp/people/osaka/index.html

---

Dear Naotoshi Osaka,

Thank you for your message. Yes, this is my error.
I can supply a new figure, if you wish. This could be in
PICT format or printed on paper.

Or you could substitute a figure that you make.

Please let me know if you would like me to supply a new figure.

Thank you very much for your careful work.

Best regards,

Curtis Roads

---

Date: Fri, 11 Sep 1998 00:17:09 +0900
From: Keiji Hirata <hirata@nefertiti.brl.ntt.co.jp>

Dear Prof. Curtis Roads,

I'm forwarding another message concerning an error correction of CMT
that was pointed out by a co-translator, Mr. Takafumi Hikichi.

Regards,
Keiji Hirata
====

I found that the following description in Chapter 7 of The Computer
Music Tutorial would include a typo and a mathematical error.

On Page 309:

I think that "seven pulses of 200 $\mu$sec duration and a delay equal
to 900 $\mu$sec, the total period is 3 ms so the fundamental frequency
is 333.33 Hz. The formant centers at 5000 Hz." should be corrected,
for 200 x 7 + 900 = 2300 = 2.3 ms.

I think that replacing the duration of the delay to 1.6 msec is the
easiest to correct, for there is no effect on the fundamental frequency
and the formant frequency.

However, Figure 7.22 shows six (not seven) pluses and a delay, so I'm
not sure which do you think the best way to correct them.

I'd appreciate it very much if you could confirm this and inform us of
the correct description.


Date: Mon, 31 Aug 1998 12:56:44 +0900
From: Keiji Hirata <hirata@nefertiti.brl.ntt.co.jp>

Dear Prof. Curtis Roads,

I'm forwarding the message concerning an error correction of CMT
that was pointed out by a co-translator, Dr. Masataka Goto.

Regards,
Keiji Hirata

=
I found that the following description in Chapter 23 of The Computer
Music Tutorial would include a typo and a mathematical error.

On Page 1068:
I think that "the difference tone (|F2-F1|) and the cubic difference
tone (2F1-F2 or 2 x F2-F1)" should be "the difference tone
(|F2-F1|) and the cubic difference tone (2 x F1-F2 or 2 x F2-F1)".

I'd appreciate it very much if you could confirm this and inform us of
the correct description.


Dear Prof. Curtis Roads,

I think I've found a small mistake in CMT.
In page 1022, in the 9th line from the bottom, we see "~ 60 Ohm".
Probably the correction will be "~ 600 Ohm". Right?

Regards,
Keiji Hirata


Dear Prof. Curtis Roads,

Sorry for bothering your work often. This email contains
some questions and suggestions for error corrections to CMT.

Regards,
Keiji Hirata
=
Chapter 18

Page 822, Figure 18.1:
It is also possible create a personal style and play any song in
this style.

Didn't you miss "to" just before "create"?

Page 824, Figure 18.4:
The Roman numerals over the eight columns of the refer to the eight
parts of the waltz, ...

".. of the refer to.." -> ".. of the table refer to .." ?
^^^^ -> i?

Page 863, 8th line from the bottom:
Another common automaton is the five-cell Life automaton, with a
neighborhood r of two, but in two dimensions (figure 19.5b).

I have the same suggestion as above, and I think "one" is correct,
instead of "two".

Page 873:
Although we see the index item "Cumulative selection", should it be
"Cumulative distribution"?

Page 873, around 4th line from the bottom:
if random(seed) is less than ptable(i) then
play(i);

Should we insert a line "break;" between "play(i);" and "end"?
That is, I think that the correct program fragment be
if random(seed) is less than ptable(i) then
begin
play(i);
break;
end
end

Page 877, Table 19.1:
In the equations for various P(X), while the second P(X) equation doesn't
contain "dx", the other equations contain "dx". Which notation is
correct?

Page 885, 2nd line:
This method uses N random number generators labeled from 0 to N.

"0 to N" -> "0 to N-1"?

Page 885, 12th line from the bottom:
Bit 3 (generator 4) has the lowest frequency.

"generator 4" -> "generator 1"?

Page 885, 9th line from the bottom:
Notice from table 19.1 that ...

"19.1" -> "19.2"?

Page 886:
Should the index item "Levy Flight" be "L\'{e}vy Flight"?

Page 888, 16th line:
For 1 < l < 3, a fixed limit value of 1-1\lambda attracts all initial
values of x.

"1-1\lambda" -> "1-1/\lambda"?
That is, should you insert "/" between "1" and "\lambda"?

Page 902, 12th line:
In figure 19.21b the network is overconstrained.

"19.21b" -> "19.24b"?

Page 906, 15th line,
... "Enough of this, let's try something different" and relay feedback
parameters that change the generative logic.

Is ".. relay feedback .." correct?

Chapter 22

Page 1025, Figure 22.3:
... four outputs (in this case three different recorders).

Should it be replaced by "three outputs"?

Page 1033, 16th line from the bottom:
For example, a typical MIDI cable transmitting 31.25 Kbits/second
is limited to a distance of 6.6 meters.

Is "6.6 meters" correct? Should it be 15 meters? Or, does it happen to
be an empirical rule?

Page 1034, 8th line from the bottom:
Controller Change -> Control Change

Page 1035, 11th line:
ideo $B"*(B video

Page 1046:
Although the index item "Internet" refers to page 1046,
the word "Internet" does not exist on this page.



Dear Prof. Curtis Roads,

> Not included in the figures I sent were the following,
> which are from rare books or journals:
>
> 1. 11.13
> 2. 11.14
> 3. 12.2a & 12.2b
> 4. 15.2
> 5. 15.4
> 6. 16.6
> 7. 16.15
> 8. 16.16
> 9. 18.5
> 10. 18.8
> 11. 18.11
> 12. 19.16
> 13. 20.10
>
> Please check if you have figure 14.15, a photograph of Berlioz.

So, I would like you to send me photocopies of the following figures:

1. 11.13
2. 11.14
4. 15.2
5. 15.4
8. 16.16
9. 18.5
10. 18.8
11. 18.11
12. 19.16
13. 20.10

Figure 5.3
Figure 10.6
Figure 13.3b

Please tell me the way we remit the cost ($78) convenient for you.
I apologize for bothering your work, and thanks a lot for your effort
in advance.

Best Regards,
Keiji Hirata

---

Date: Fri, 16 Jan 1998 16:41:39 +0900
From: Keiji Hirata <hirata@nefertiti.brl.ntt.co.jp>

Dear Dr. Curtis Roads,

There're several matters related to the Japanese translation project of
The Computer Music Tutorial. It'd be happy to receive your reply.

The errors that I've found out so far are
on page 27:
This means that one cycle takes longer than the interval
between samples ^^^^^^
-> This means that one cycle takes shorter than the interval
between samples

on page 566:
sample-by sample -> sample-by-sample

And there're the following questions:
on page 31:
* You write "synthesis algorithms generate nothing other than sine
waves above 11KHz (44.1 KHz sampling rate) or 12 KHz (48 KHz
sampling rate". Why is the ratio is approximately 1/4?
Is this value 1/4 empirically valid?
* In the same paragraph, there are no sentences that explicitly
mention oversampling. Nevertheless, does your discussion assume
that synthesis algorithms employ oversampling?

And I have a couple of things to be confirmed:

Your email on Feb. 12nd (more than a year ago!) said:
> The photographic figures are in two places. Half are in France
> at the French publisher Emap-Alpha?. The other half are here with me.

Well, how and when can we (the Japanese publisher) get the photographic
figures?

Your email on Nov. 18th said:
> I would like to send you some corrections.

I think the corrections doesn't arrive at me yet...

Thanks a lot in advance for your cooperation to our translation.

Best Regards,
Keiji Hirata (NTT)

---

Date: Wed, 05 Aug 1998 17:23:51 +0900
From: Keiji Hirata <hirata@nefertiti.brl.ntt.co.jp>

Dear Prof. Curtis Roads,

Today I've received your postal mail containing the corrections and
revisions for CMT. We are soon going to reflect it to our translation.
Thanks a lot.
Our translation team still keeps some other errors that we've found via
the translation; before long, I'll send its list to you.

Best Regards,
Keiji Hirata
Date: Thu, 13 Aug 1998 16:20:08 +0900
From: Keiji Hirata <hirata@nefertiti.brl.ntt.co.jp>

Dear Prof. Curtis Roads,

> This means that these figures were NOT INCLUDED, so
> I realize that you do not have them.
>
> ...
>
> I am not surprised that there are few
> discrepancies. I am sure that we can resolve them via email.

The guy of the Japanese publisher summarized to make the list
regarding the photographies and drawing that you had sent to me.
This enclosed below is a report from him, and I would just let you
inform of the list. Soon he will make another list of figures that
he needs to make plates.

Best Wishes,
Keiji Hirata
====

>_
>Chapter 1
>
>1.2 Crop
>1.6 Crop
>1.8 Layout as indicated
>1.13 Add labels, draw box
>1.14 Add labels, draw box
>1.15 Layout as indicated, add labels, draw box
>1.17 Draw box
>1.20 Add labels

All in Tokyo.

>_
>Chapter 2
>
>2.1 Crop the bottom as indicated
>

All in Tokyo.

>_
>Chapter 3
>
>3.3 Add labels, draw box
>3.4 Change font to conform to rest of book, if
>necessary
>3.9 Change font to conform to rest of book, if
>necessary
>3.10 Change font to conform to rest of book, if
>necessary
>3.11 Possibly halftone

The figures in Section 3 are enumerated to 3.12 (none from 3.13);
3.1 to 3.12 in Tokyo.

>3.13 Possibly halftone
>3.14 Draw box around
>
>_
>Chapter 4
>
>4.1 Halftone.
>4.2 Halftone
>4.3 Possibly halftone
>4.7 Try to preserve the line detail in the reduction.
>4.9 Add labels, draw box
>4.10 Add labels, draw box.
>4.11 Draw line around.
>4.12 Halftone
>4.14 Add labels, draw line.
>4.15 Add labels, draw line.
>4.17 Add labels, draw line.
>4.18a + b Possibly halftone
>4.20 Possibly halftone

Figure 4.18 doesn't have the ramification a + b.
All in Tokyo.

>
>_
>Chapter 5
>
>5.3 Halftone
>5.6 Add labels, draw box
>5.7 Draw box
>5.8 Draw box
>5.9 Draw box

Figures 5.1 to 5.9 and 5.17 are not in Tokyo.
5.3 is a halftone photograph.

>5.12 Possibly halftone
>
>_
>Chapter 6
>
>6.4 Draw box; reduce.
>6.11 Add labels, draw box.
>6.13 Add labels, draw box.
>6.15 Halftone.
>6.19 Add the indicated labels at a larger size. Draw
>box.
>6.20 Add the indicated labels at a larger size. Draw
>box.
>6.23 Add the indicated labels at a larger size. Draw
>box.
>6.24 Add labels, draw box.
>

All in Tokyo.

>_
>Chapter 7
>
>7.2 a + b + c Layout together
>7.3a Possibly halftone
>7.3b Draw box
>7.5 Possibly halftone
>7.7 Please reduce
>7.9 Please halftone
>7.11 Do not reduce, if possible.
>7.18 a + b Add labels, draw box.
>7.21 Add labels, draw box
>7.22 Add labels, draw box
>7.23 Add labels, draw box
>7.24a Add label "(a)"
>7.24b Add labels, draw box
>7.27 Add labels, draw box.
>

Figures 7.1, 7.2a + b + c, 7.4, 7.7, 7.12, 7.15 and 7.24b are not in Tokyo.

>_
>Chapter 8
>
>8.2 Reduce 8.2a to about the same size as 8.2b.
> Draw box around both.
>8.5 Draw box around. Keep fairly large.
>8.6 Reduce, draw box around
>

Figures 8.2ab, 8.5, 8.6, 8.13 and 8.16 are not in Tokyo.

>_
>Chapter 9
>
>9.1 Possibly halftone
>9.3 Possibly halftone
>9.4 Possibly halftone
>9.8 Color slide.
>9.11a Halftone
>9.11b Halftone
>9.12 Possibly halftone
>9.13 Color transparency
>

All in Tokyo.

>_
>Chapter 10
>
>10.6 Halftone original in Gravesaner Bl=E4tter 14 , p.
>137; draw box
>10.7 Possibly halftone
>10.10 Add labels, draw box
>10.12 Add labels, draw box
>10.14 Add labels, draw box
>10.16 Add labels, draw box
>10.17 Add labels, draw box
>10.20 Add labels, draw box
>10.22 Add labels, draw box
>10.27 Possibly halftone
>10.36 Possibly halftone
>10.38 Possibly halftone
>10.40 Possibly halftone
>

Figures 10.38, 10.39 and 10.40 are not in Tokyo.
10.6 is just smeared.

>_
>Chapter 11
>
>11.1 Halftone
>11.2 Halftone
>11.3 Possibly halftone
>11.4 Halftone
>11.8 Possibly halftone
>11.13 Halftone photograph from p. 52 of Gravesaner
>Bl=E4tter
> 18, spherical loudspeaker.
>11.14 Halftone photograph from p. 73, figure 37 of
> Schwingende Elektronen by Morawska-B=FCngeler.
>11.18 Possibly halftone
>11.25 Possibly halftone
>

11.13 is not in Tokyo (as you said).

>_
>Chapter 12
>
>12.1 a Reduce
>12.2 a + b Use p. 72, figs. 55 & 56 in D. Miller The Science
>of Musical
> Sounds
>12.7 a + b Add labels, layout together, and draw box around
>12.9 Draw box.
>12.11 Add labels, draw box
>12.16 Draw box
>

12.2ab is not in Tokyo (as you said).
Figures 12.14, 12.16 and 12.17 are not in Tokyo.

>_
>Chapter 13
>
>13.1 a + b + c Add labels, layout together, draw box around.
>13.3a Halftone this color slide.
>13.3b Halftone

13.3b not in Tokyo, but 13.3a in Tokyo.

>13.5 Halftone this B&W photograph
>13.7 Possibly halftone
>13.8 Possibly halftone
>13.9 Add labels, draw box around
>13.10 Possibly halftone
>13.12 a + b Possibly halftone
>13.15a + b + c Add labels, latout together, draw box around
> possibly halftone
>13.23 Draw box around
>13.25 Reduce, draw box around
>13.26 Draw box around
>13.27 Reduce, add labels, draw box
>13.29 Add labels, draw box
>13.30 Add large labels, draw box
>13.31 Possibly halftone
>13.32 Halftone
>13.33 Possibly halftone
>13.34 Draw box
>13.25 Halftone this color print
>
>_
>Chapter 14
>
>14.1 Halftone.
>14.3 Halftone leftmost photograph. Make higher contrast.
>14.4 Halftone.
>14.5 Photocollage. See accompanying folder for
>instructions.
>14.8 Crop and halftone.
>14.9 Possibly halftone.
>14.10 Halftone and crop.
>14.11 Possibly halftone.
>14.12 Halftone.
>14.13 Halftone and crop.
>14.15 (*** Berlioz. TO BE SUPPLIED.***}
>14.16 Halftone.
>14.17 Halftone. Please reduce to small size.
>
>_
>Chapter 15
>
>15.1 Possibly halftone
>15.2 Halftone from A. Buchner Mechanical Music
>Instruments,
> plate 15
>15.3 Possibly halftone this engraving.
>15.4 Halftone from Plate VIII, before p. 109, in A.
>Moles
> Les musiques exp=E9rimentales
>15.10 Crop and halftone
>15.12 Halftone
>15.13 a + b Halftone and layout (a) above (b)
>15.17 Gray out part of the circle, as indicated.

Figures 15.2 and 15.4 are not in Tokyo (as you said).

>_
>Chapter 16
>
>16.1 Draw box.
>16.2 Draw box.
>16.3a Draw box
>16.3b Draw box
>16.4 Halftone.
>16.5 Halftone
>16.6 Original is in ComputerMusic? Journal 5(3) p. 58,
>fig. 2.
>16.10 Draw box.
>16.11 Draw box.
>16.14 Reduce and draw box.
>16.15 Original is in Computer Music Journal 7(3), p. 35,
>Fig. 3.
>16.16 Three color slides from GRM. {*** TO BE SUPPLIED.
>***}
>16.17 Do not reduce too much; must be readable.
>16.24b Possibly halftone.
>16.31 Increase the contrast; make this more black and
>white.
>16.32a Draw box
>16.35 Halftone.
>16.38 Halftone.
>16.40 Possibly halftone.
>16.42 Possibly halftone.
>16.43 Color slide; halftone.
>16.44 Add labels, draw box.
>16.46 a + b Add labels, draw box, layout as shown.
>16.47 a + b Add labels, draw box, halftone. Layout as shown.
>16.48 Add labels, draw box.
>16.49 Possibly halftone.
>16.50 Halftone.
>

Figures 16.6, 16.15, 16.16 are not in Tokyo as you said.
16.43 is not in Tokyo.

>_
>Chapter 17
>
>Tables Possible change fonts and layout
>17.3b Possibly change font; draw box around.
>17.4 Possibly change font; draw box around.
>17.6 Possibly change font; draw box around.
>17.7 Possibly halftone.
>17.8 Possibly change font; draw box around.
>17.10 Reduce and draw box around.
>
>_
>Chapter 18
>
>18.2 Please reduce to eliminate jaggies.
>18.3 Draw box around
>18.4 Halftone
>18.5 Get original photograph from p. 29 of Cybernetic
>Serendipity
> Halftone
>18.6 Halftone
>18.7 Halftone
>18.8 Get original photograph from p. 113 of Computers
>and
> Computation. Halftone.
>18.10 Try to keep original size. This cannot be reduced
>much more.
>18.11 Get original photograph from frontispiece of Les
>Musiques
> Artificielles by N. LaChartre?
>18.13 Halftone
>18.15 Halftone
>18.18b Draw box around.
>

Figures 18.5, 18.8, 18.11 are not in Tokyo as you said.

>_
>Chapter 19
>
>Table 19.2 Try to make the dots bigger while retaining the
>column
> alignment.
>
>19.1 Draw box.
>19.6 Add labels.
>19.7 Draw box.
>19.14 Draw box.
>19.16 Photostat image from B. Mandelbrot. Fractals, p.
>39.
> "Triadic Koch island." {*** TO BE SUPPLIED. ***}
>19.17 Draw box around.
>19.18 Add labels, draw box.
>19.20 Draw box.
>19.25 Possibly halftone.
>

19.16 is not in Tokyo as you said.

>_
>Chapter 20
>
>20.1 Halftone this color print.
>20.2 Halftone this B&W print.
>20.3 Halftone this B&W print.
>20.4 Halftone this B&W print.
>20.5 Halftone this B&W print.
>20.6 Possibly halftone.
>20.10 Halftone page 214 of The Electronic Music
>Instrument
> Manual by A. Douglas
>20.17 a + b Possibly halftone
>20.19 Possibly halftone
>20.21 Halftone this B&W print.
>20.22 Two B&W prints are provided.
> Choose the best for publication and halftone it.
>20.23 Possibly halftone.
>

20.10 is not in Tokyo as you said.

>_
>Chapter 21
>
>21.10 a + b Layout together and draw box
>
>_
>Chapter 22
>
>All line art.
>
>_
>Chapter 23
>
>23.2 Redo labels
>23.5 Redraw and label
>23.6 Redraw and label

23.6 is not in Tokyo.

>
>
>_
>Appendix A
>
>A.4 Add labels and draw box.
>A.8a + b Possibly halftone.
>A.18 Add labels, draw box.
>




Date: Tue, 04 Aug 1998 02:26:14 +0900
From: Keiji Hirata <hirata@nefertiti.brl.ntt.co.jp>

Dear Prof. Curtis Roads,

> Here are my instructions to the publisher on the figures.

Thanks a lot for your support. The guy of the publisher is at first
investigating the correspondence of your list and the photos and graphs
that you've sent to me.
Unfortunately, the result at the moment is quite different from your list.
I'm bringing part of the result of his investigation as follows:

> Not included in the figures I sent were the following,
> which are from rare books or journals:
>
> 1. 11.13 O
> 2. 11.14$B!!(B O
> 3. 12.2a & 12.2b X
> 4. 15.2 O
> 5. 15.4 O
> 6. 16.6$B!!(B O
> 7. 16.15$B!!(B O
> 8. 16.16$B!!(B O
> 9. 18.5 O
> 10. 18.8 O
> 11. 18.11 O
> 12. 19.16 O
> 13. 20.10 X

Here "O" means that we have it; that is, it was actually included in the
package as opposed to your email. "X" means that we don't have as you
suppose.
But, since 12.2a and 12.2b are rather simple drawings, probably we don't
need the original. On the other hand, the photocopy of 20.10 is certainly
necessary.

Does it happen that you and we look at the different versions of CMT booK?
Or, did you forward the package from France to Japan without confirming
the contents? (If I'm wrong, I'm sorry.)

Anyway, there are many discrepancies between your list and the
existing photos and drawings.

> Please check if you have figure 14.15, a photograph
> of Berlioz.

Yes, it arrived at us.

And, the photocopy that we'd need is 5.3.

To summarize, we would need photocopies of 5.3 and 20.10, up to now.

> Please let me know if the publisher agrees to pay this expense.

Yes, the publisher agrees to pay it ($6 for each);
could you tell me how to remit the money for making photocopies to you?
I'm writing this email on behalf of (the guy of) the publisher who
subscribes to Japanese cmt-ml (see the cc: field of this email).
Thanks in advance for your cooperation.

BTW, the guy is afraid that the suspicion of lost rare figures falls on him.
But, in fact, he just checked them out and kindly reports the existing
state of things.

Best Regards,
Keiji Hirata (NTT)

---

Changes to the French edition

Dear Jean,

Here are some additional corrections. I quickly read through several chapters
of the translation. It is important to try to mention some historically
important French composers, even if this is not a book on
composition. Please also list them in the index.

---

Chapitre 3

After:

Les compositeurs de musique concrète travaillent directement avec des objets sonores (Schaeffer 1977 ; Chion 1982).

Insert:

An important example of the mature form of the musique concrete idiom is the composition De Natura Sonorum (1975) by Bernard Parmegiani.

(De Natura Sonorum is italic.)

These compositions nécessitent de nouvelles formes de notation graphique, hors des limites des partitions classiques pour orchestre (Bayle 1993).

Add new paragraph:

This research has led to new musical forms, such as the ambient
environmental music of Luc Ferrari. His series of pieces
entitled Presque rien stands as an outstanding example of this genre.

(Presque rien is italic.)

---

Chapitre 7

Le "quantum" (Gabor 1946, 1947), le "signal élémentaire gaussien" (Helstrom 1966; Bastiaans 1980), le "segment de courte durée" (Schroeder et Atal 1962), la "fonction de pondération de courte durée" (Flanagan 1972), la "fenêtre" (Arfib 1991; Harris 1978 ; Nuttall 1981), la "fenêtre coulissante" (Bastiaans 1985), "l'impulsion par fonction de fenêtre" (Bass et Goeddel 1981), "l'ondelette" (Kronland-Martinet et Grossmann 1991), la "fonction d'onde formantique" ou "FOF" (Rodet 1980), "l'impulsion VOSIM" (Kaegi et Tempelaars 1978), le "paquet d'ondes" (Crawford 1968), "l'explosion sonore" (Blauert 1983; Pierce 1990), "l'impulsion sonore" (Whitfield 1978), et même le "top sonore" (Buser et Imbert 1992) peuvent tous être considérés comme des représentations granulaires de signaux musicaux.

---

I would not recommend translating all these terms.
"tone burst" and "tone pip" should probably be left in English.

Under "Évaluation de la synthèse granulaire"

La synthèse granulaire constitue un corps de techniques diverses n’ayant que le concept de grains sonores en commun.

Change to:
un corps de techniques micro-soniques n'ayant...

---

After:

Le premier cas est le plus flexible, car l'on peut extraire les grains du fichier dans n'importe quel ordre. Par exemple, on peut extraire un seul grand grain d'une caisse claire et cloner une séquence périodique de centaines de grains pour créer un roulement de caisse claire (illust. 7.11a). Le générateur de grains peut également échantillonner aléatoirement les grains d'un fichier plus long, pour ensuite les réordonner (illust. 7.11b). Une extension de cette technique consiste à échantillonner aléatoirement plusieurs fichiers son et à entrelacer leurs grains pour créer des textures multicolores (illust. 7.11c). Ces tissus sonores entrelacés varient en grande partie selon la hauteur et le timbre des grains individuels utilisés.
Le second cas s’applique à une granulation en temps réel d'un son continu avec l'ordinateur agissant en tant que ligne de retard ou fenêtre, qui peut être configurée pour fournir les différents grains. Voir la description et la configuration de lignes de retard au chapitre 19. Dans ce cas, les effets spectraux secondaires distordent et enrichissent le son de façon contrôlable.
Le troisième cas ressemble au précédent à l'exception du taux de lecture qui peut être varié par un paramètre contrôlant la vitesse à laquelle la synthèse s'avance à travers les échantillons. La lecture peut varier d'une vitesse normale à un taux abaissé dans lequel un seul échantillon est répété en continu. Cette méthode peut ainsi être considérée comme une interpolation entre le premier et le deuxième cas.


Please add new paragraph:

The recent music of Horacio Vaggione, in compositions such as Schall (1995) and Nodal (1997), are outstanding examples of time granulation of sampled sounds. Vaggione has developed a technique of shifting between different
time scales by changing the size of the grains at pivotal points in these pieces.

(Schall and Nodal are italic.)

---

Chapitre 8

La presque totalité du logiciel SYTER a depuis été portée sur une carte de traitement du signal pour ordinateur personnel (INA/GRM 1993).

Actually this now works without a card, I believe. So please
say that it is now a purely software-based system for personal
computers.

At the end of this section, please add:

Of course, time-varying subtractive synthesis has a long history in the
analog domain, for example, Christian Clozier's Par Pangloss Gymnopede (1984).

(Par Pangloss Gymnopede is italic.)

---

Please note this correction to the chapter on reverberation:

The Computer Music Tutorial

p. 475. In the figure 11.17, the horizontal axis is Time, not Frequency!

Thank you Jean ! ! !

Chinese Edition:
Subject Index:
Principle components analysis is not found on page 1 or 50.