Pure Data: Electronic Music and Sound Design 1

FOREWORD

FOREWORD
by Miller Puckette

There was a time when programming a digital computer (or an analog one, which also existed) involved using switches, knobs, and patch cords to program the thing and to feed it its input data. As late as 1980 or so, the Digital Equipment Corporation PDP11 had rows of switches on the front. You could put numbers directly into the machine’s registers or memory.

This was the machine on which Music 11, Barry Vercoe’s music compiler that later morphed into CSOUND, once ran. To use Music 11 was to function at two levels of software abstraction from the raw PDP11 hardware: the music compiler itself, and the operating system that ran underneath it. These two layers of abstraction saved the user the trouble of learning the PDP11 instruction set or its memory addressing scheme, or even the binary number system that all its digital circuitry was based on. But on the other hand, the user had to buy into various assumptions that the authors of both of these layers made about what their users might want to do.

Some of these assumptions were overt: for instance, to store data persistently one organized it into “files”, which weren’t real files at all — those were made of paper — but were a convenient abstraction built on an easily grasped metaphor. At the level of Music 11, an assumption was made that the composer’s input could divide neatly into a “score”, a sequence of “notes” that presumably was of the composer’s primary and individual concern, and an “orchestra” consisting of virtual instruments that composers and researchers frequently swapped around among themselves.

Music 11 itself played the role of the performer, and did so exactly as a computer program could be expected to: it caused the instruments in the orchestra to play the notes in the score, one by one, exactly as prescribed in the score. Vercoe himself saw this as a compromise that had to be made because computers of that time weren’t able to generate sound samples interactively, as we would need to do to allow a human performer into the process. The available hardware thus constrained the design of the software, which in turn constrained the composer.

To be sure, the composer had other options. Other electronic instruments of the time offered different trade-offs: “tape studios” offered a much more tactile experience and far more sound storage capacity; analog synthesizers could be played in real time; and hybrid computer/analog systems such as GROOVE combined computers and analog gear to achieve some of the specific advantages of each of the two technologies. To choose to work with Music 11 was to decide that its particular advantages were important and its particular constraints were acceptable.

With this historical example in mind, readers of this book could think about its own particular choice of technology in a similar way. The technology consists of three layers: first, the Virtual Sound library, a toolset for synthesizing, analyzing, and processing digital sounds in real time along with many example applications and audio materials. This library was designed specifically for users of this book. Second, underneath Virtual Sound, the Pure Data environment (abbreviated as “Pd”) provides the necessary fundamental audio building blocks, and the real- time interactive programming capabilities, upon which Virtual Sound is built. Pd itself is used in a wide variety of other situations (forinstance, inside video games or audio “plug-ins”, in concert with video and graphics extensions, and/or with customizations appropriate to particular production studios) and often appears with various specialized libraries; so the reader of this book is seeing Pd from one specific angle. Third, underneath Pd is the operating system and hardware, which is presumed to be a laptop or desktop computer with a keyboard, mouse, and display.

I can think of three good reasons to use a computer (instead of, say, a collection of music-specific hardware devices such as synthesizers, hardware controllers, drum machines, and/or turntables). The first is cost. Computers are cheap, and since they can also be used to do many other tasks, there’s a good chance a reader of this book already has had to get one for some reason. Second, computers are far more flexible and open-ended in their capabilities than any special-purpose audio hardware could be. Third, and of particular importance: computers are transparent in the sense that you can know what they’re doing. Nothing that matters is secret about them. This is often not the case with hardware instruments.

As to Pure Data, its design also aims for the highest possible level of generality and transparency. Unlike much computer software, Pd is open-source, so that there are no secrets about its inner workings. And while many of the design choices underlying Pd will seem constraining to users, they are much less so than those of most music software, which is typically designed with particular workflows and/or musical styles in mind.

And finally, among the many libraries that can be run atop Pd, Virtual Sound is particular in that it is primarily designed for pedagogy. Although a student using this book can’t escape having to learn many of the particularities of the Pd environment itself, the authors have worked hard to make the process as straightforward as possible for the beginner, who can then focus on the audio and musical concepts treated in this book, avoiding much of the trouble that besets the student trying to learn computer music technique using Pd alone.

All this is not to claim that this particular stack of technologies is the answer to everyone’s needs at all times, but only that it has been crafted with the goals of this book in mind: primarily, as I see it, to open a path along which the reader (whether a student or an independent reader) can get a theoretical and practical understanding of the fundamental techniques for making music with a computer, to make this process as direct and straightforward as possible, and to require a minimum of specialized knowledge in advance.

Returning to the historical context into which I’ve thrown this introduction, one last point is worth thinking about. What did composers in the decades from 1957 to about 1983 see in the computer as a music-making tool? I think the computer workflow of that day was one that was familiar to Western-trained composers of that time, with their attention to planning, organization, and above all detailed manipulation of numbers and sets, and their ability to work for months or years on a project without getting to hear the music itself until afterward (if even then). Most humans would be discouraged from working like that. Today, on the other hand, the sound pours from your loudspeakers while you’re making it up. In the world of electronic music, the special role of “composer” disappears, or perhaps more precisely, the roles of composer, performer, and listener have converged. Music making is a much freer and wider-ranging activity than it was before the advent of electronics, mediated as they are almost always today by computers. There are many possibilities out there to explore, and studying both the theory and the practice of audio programming is the best first step of the way.

(Miller Puckette is the original author of Max and Pure Data)