Assignment #3: Indeterminacy

Steven Backer
sbacker @ ccrma , stanford , edu

February 27, 2005

MUSIC 220B - Synthesis Techniques, Compositional Algorithms, Psychoacoustics and Spatial Processing
CCRMA, Stanford University

Sometimes, you just gotta let it go!” -Son Seals (1942-2004)

When composers create music with elements of indeterminacy, they are relinquishing control of some aspect of the composition to “Chance”. In this type of composition, the choice of the structure that generates randomness becomes very important. If one does not choose a control structure that generates enough random variations, the results can still sound very deterministic. Thus, it is the composer's job to control the selection of the elements at a higher level of abstraction that will in-turn generate what is perceived by the listener to be randomness in shorter time frames.

In Cage's lecture entitled “Indeterminacy: New Aspects of Form in Instrumental and Electronic Music”, he chooses to tell a series of short stories, all in English, one after another. Regardless of the word length of each anecdote, he articulates each story over one minute of time, speeding up or slowing down his narration to stretch or compress the length of the tale to one minute.

This lecture, along with the source material I have chosen to work with, is the inspiration for this piece. It is a series of short anecdotes about our experiences thus far in Music 220(a,b). “Granular” synthesis is used to select short sections from sound files generated by everyone in Music 220b, and these large grains are woven together, randomly, to create a larger sound source. Cage's speech is more similar to “sampling rate conversion”, not granular synthesis; however, he tells complete stories (very well), whereas I've only chosen to reiterate a word or two from each.

One option for the assignment includes creating conditionals to determine whether the random results are acceptable and should be used in the final mix. I chose not to add anything of this sort in my lisp code, as this sort of judgment about the source material does not belong in lisp code in this context. However, it was unavoidable to make these sort of decisions in the selection of the source material, just as one will essentially control the randomness of probability distribution functions by setting variables like variance.

I sincerely hope you enjoy the results, and hope everyone is comfortable with the “mangling” of their previously painstakingly created sounds – recall that letting go was one lesson taught in Music 220a. If you do not like the sound of yours I selected, then this is your cue to take my code, alter it, and resynthesize the piece. I did not incorporate any sounds from the instructors or TA's work, but they are present nonetheless. Hiroko and Matt's contributions can be heard in the creation of the overall Form of the resulting sound file. Jose contributed practical advice on dealing with granular synthesis in larger time frames. And of course, Nando's music can be heard in the lisp code itself (though I am still listening and trying to understand this piece!). You may not believe me when I say these are sounds, but as one classmate pointed out about an experience at a recent concert – we should be listening to Everything in the context of a musical environment if we desire to truly hear the müsic.

All in all, I am not sure if it is possible to integrate full indeterminacy into prerecorded “music”. It seems that we can only achieve “statistical indeterminacy” in this manner. If a composer did relinquish all control, she/he would only be left with a pair of fuzzy dice, which would make no sound when rolled. However, it should be possible to create such music of changes in a live improvisational setting, where there is realtime feedback from the audience to the performers. Consider the following (and please, let me know your opinions on the matter):

Controlled Elements

Indeterminate Elements

Number of grains selected from each source

Order of grain selection from each source

Grain duration

Grain placement in time

Parameters that seed probability distributions

Grain density

Content of grain sources

Content of grain sources

220.wav – Stereo Mix of Two “Independent” Synthesized Versions

Lisp Code:

grani.ins unlabeled.lisp

The Grain Sources: Note: I posted these here for posterity. If you hear a sound in the mix and you are curious where it came from, I recommend that instead of wading through all of these below, you ask a random classmate which sound of theirs I incorporated. Once you've spoken to everyone, you should have your answer. :)

Bill Schottstaedt's bobwhite.wav


Per Bloland.wav

Brad Burkett.wav

Juan-Pablo Caceres.wav

Jen Carlile.snd

Jack Conte.wav

Jonathan Eccles.wav

Jesse Fox.wav

Kris Havlak.snd

Dean Kitanidis.wav

Paul Larson.wav

Robert Lugo.wav


Bruno Tucunduva Ruviaro.wav

Storm Staley.wav

Mitchell Wilcox.wav

Lillian Wolfson.wav

Wilmot Yeh.wav

Additionally, the above composer's works in their original contexts can be found here and here