Vol. 24 Issue 2 Reviews
Eduardo Reck Miranda: Computer Sound Synthesis for the Electronic Musician

Softcover/CD-ROM, 1998
ISBN 0-240-51517-X, 208 pages
illustrated, appendices, bibliography, CD-ROM instructions, index
Focal Press (an imprint of Butterworth-Heinemann)
Linacre House, Jordon Hill, Oxford OX2 8DP, UK
telephone +44 1865-314627; fax +44 1865-314091
electronic mail bhuk.orders@repp.co.uk
Focal Press 225 Wildwood Avenue, Woburn, MA 01801-2041, USA

Reviewed by Paschall de Paor
Limerick, Ireland

This book is one of several published by Focal Press in its Music Technology Series (Francis Rumsey, ed.). It is aimed primarily at university and college courses in the field and at working professionals seeking to update or expand their knowledge (as stated in the Preface). This review is based on that stated objective.

The book is divided into eight chapters, with two appendices, references, index, and instructions for using the CD-ROM. It discusses several synthesis techniques, provides examples for implementation, and outlines how to construct one's own instruments.

The preface is, as one would expect, an overview of the book, with some tantalizing information on each chapter. However, some details are unnecessary here (for instance, stating Fourier's Theorem), and some need clearer explanation (the author states that additive synthesis is "expensive" to run–p. x and later p. 125–meaning in computational and not financial terms). Plainly obvious to the experienced reader but not to the novice, assumptions or interpretations of this kind need to be explained more fully.

In Chapter 1, "Computer Sound Synthesis Fundamentals," Mr. Miranda consigns the initial two small paragraphs to explaining the fundamental behaviour of sound and the recording chain. This is really not enough–it assumes a greater knowledge on behalf of the reader than is stated at the outset. And some of the explanations, in their simplicity, will lead to fundamental misunderstandings by the reader (such as the basic explanation of an analog synthesizer). Other topics also contain potential for misunderstanding, for example the statement that an "analog oscillator produces only a sinusoidal waveform" (p. 14). Perhaps the earliest examples did, but this is clearly no longer the case! Furthermore, some argumentative factual errors are present: "[the] first programming language specifically designed for sound synthesis was Music III" (p. 8). Music I was specifically designed for sound synthesis. Music III was the first to include unit generators (UGs) and the ability to patch these together to form a network for synthesis. A small detail, perhaps, but again, the intended reader will pick up on such things (certainly, I would wonder about Music I and II). And, of course, the book is about synthesis, so therefore one expects details on the subject itself to be accurate. One may state that Music III was the first of its kind, but it must be contextualized, and in this case it was not. It is another example of a well-intentioned, but inaccurate statement of fact. I also wonder about the order of information presented. For example, the sampling theorem and quantization noise precedes Nyquist information. The series editor (Mr. Rumsey) handles these and other related topics excellently in his own book (Sound and Recording: An Introduction, 2nd edition, 1994), and includes much needed graphical examples and explanations. Finally, other smaller problems, no less important to a student, include spelling (Herz instead of Hertz, etc.).

Chapter Two deals with the software on the accompanying CD-ROM, and in enough detail to enable the reader to effectively use it. However, one must consider the prudence of placing this so early. Clearly the author wishes the readers to be sufficiently familiar with the software to enable them to work through the provided examples. This is understandable and generally a good idea. However, it does interrupt the flow of the book. It could easily have lived in an appendix, with instructions to the reader, at the beginning, to work through it if wishing to make use of the examples. Here, the reader has a choice. In Mr. Miranda's case, they do not (skipping chapters is not a valid option). Furthermore, having just covered the fundamentals of sound synthesis, the reader is launched into an explanation of software operation and functionality, introducing terminology and techniques not yet covered. This could lead to confusion.

In Chapter 3, "Loose Modeling Techniques," Mr. Miranda for some reason devises a taxonomy of sound synthesis techniques. However, I am not sure this is the place to do it. AM, FM, waveshaping, wavetable, Walsh function synthesis, and Binary synthesis are classified as Loose Modeling. Further research by the reader will not return many references to this term, and consequently it opens the possibility for confusion. A very useful element here, though, is the inclusion of examples using the accompanying software.

In Chapter 4, "Physical Modeling Techniques," Mr. Miranda rightly points out that his inclusion of some techniques in this section might cause controversy. There exists a legitimate argument for including subtractive synthesis, but this is really not the place to do it. An introductory text such as this should be devoid of such contentious issues.

One of the main tenants of computer music and sound synthesis is Fourier analysis/synthesis. After skipping through it in Chapter 5 ("Time-Modeling Techniques"), Mr. Miranda comes back to it in Chapter 6 ("Spectral Modeling Techniques"), which is the right place to do it. However, additive synthesis is generally one of the first synthesis methods used to initiate novices into the field of sound synthesis; it involves the use of Fourier analysis and is easily understood. Mr. Miranda leaves additive synthesis quite late in the book, for no apparent reason, thus depriving the reader of a gentle introduction as opposed to a jolting one.

In Chapter 7, "Instrument and Sound Design," Mr. Miranda seems to disparage a "highly irrational approach to sound design" (p. 148) by suggesting one should base new instruments on well-documented acoustic theories. This is acceptable only if used in the initial stages of learning. After all, one of the key enticements and powerful applications of sound synthesis (and stated as such elsewhere by the author) is the ability to use any source of information for synthesis parameters. However, the author states closely thereafter that these new instruments should serve as a "point of departure," so it is a small detail, but one which could shape how a reader perceives instrument design. The inclusion of a detailed working example (based here on formant synthesis using a subtractive technique) is excellent, but might have been even better if some further experiments were also included (for example, encouraging the reader to try adding another formant here, a different number there, and so on).

In the final chapter, "Towards the Cutting Edge", the author's research area (artificial intelligence and sound synthesis) is clearly demonstrated. Also covered is parallel computing and its ramifications for synthesis. Both of these are very useful additions, and add considerably to the value of this book.

The choice of software on the CD-ROM is ultimately the author's choice, and he strikes a good balance between the two platforms his target audience will probably use (Macintosh and PC). There is much else available, but this set should give the reader a good indication of the possibilities.

An area too often overlooked is the music, and there is no discography included. Mr. Miranda rightly states that "the art of sound synthesis is as important for the electronic musician as the art of orchestration is for composers of symphonic music" (though why he chose "musician" instead of "composer" for the electronic field remains to be discovered). However, actually hearing examples of certain synthesis techniques is extremely valuable, not least because it demonstrates application and musical context, which can help toward choosing an appropriate technique and confirming theory with practice. Even a small list would have sufficed. However, there are several musical references in the book which should allow the committed reader to more fully explore synthesis applications and musical consequences.

This type of book should provide pointers to other information. Moreover, one expects any tome which purports to "introduce computer sound synthesis techniques and synthesis programming to students, researchers, musicians and enthusiasts" to include reference to some, if not all, of the standard texts in the field. Here, this is not the case. One of the most significant (and widely-used) books in the field is missing–Computer Music, by Charles Dodge & Thomas Jerse, is certainly still extremely relevant (in its 2nd edition)–yet Mr. Miranda chose not to include it. Another notable absentee is Curtis Roads' Computer Music Tutorial, also highly relevant to the target reader.

One of the biggest disappointments with this book is the absence of a conclusion or summary chapter, uniting the various strands. With such a vast array of techniques and information presented, there is clearly a need to sum up all that has gone before. The reader suddenly finds him or herself at the end, without any sort of summation. This further detracts from the usefulness of the book for its stated purpose.

The absence of Csound is notable. Such an introductory text could certainly have included the most widely used, understood, referenced, and freely available synthesis engine out there. Practically all the synthesis techniques discussed in the book are available in Csound, and it runs on a wide range of platforms. Mr. Miranda deems it unimportant in absentia, and this, of course, is his prerogative. To be fair, there are two references to Csound in the sections on parallel programming and on the Composer's Desktop Project (CDP). The basis of Csound, the so-called Music N languages, is discussed via PCMusic (though the student will find little reference to this in the general music technology literature). A positive point, however, is the discussion of CDP. This very important project is too often overlooked, and Mr. Miranda gives it credible coverage.

There are so many bumps in the information and its presentation that I find the book to be uneven. Certainly, it is a useful addition to the repertoire, especially the artificial intelligence section, but its inconsistencies detract from its ability to serve as a good introduction; it really needs to be used in conjunction with another text. There is no question but that Mr. Miranda knows his sound synthesis, and this is obvious to an experienced reader. However, he does not know how to communicate this clearly enough, which is unfortunate, as he has a lot to offer and his book has so much potential.