Saturday, April 29, 2000
So I’ve got the new CSound book from Amazon.com now. I’ve had it for several weeks, actually, but I’ve really been diging into this last week.
My big discovery in it has been the technique of “granular synthesis.” I’m not exactly sure how it works yet, but apparently any sound can be described as a mutlitude of little sonic “grains.” A grain is just a wave with an envelope so short that, when strung together, individual envelopes are imperceptible — like individual frames on a film strip.
CSound has a couple of different opcodes to help create granularly synthesized sounds. I haven’t experimented with these much yet, but once I get a better grasp of what’s going on I most certainly will. They’ll be posted on my upcoming CSound page, for those of you who are watching from home.
An important thing about CSound for me is its ability to create sounds that are impossible to make with normal synth gear. Granular synthesis seems to be a good way to do that — although the examples from my book that I’ve heard so far really all sound either like a regular noise generator or a sample-and-hold filter. But I think there’s potential for granular synthesis to do much more.
One thing I think will be freaky will be applying granular synthesis to a long sample, maybe of someone speaking. I don’t understand the process exactly, but I gather that one can use granular synthesis to randomly rearrange chunks of an incoming sample. With normal little granules, you’d probably get more noise or sample-and-hold sounds, but with big grains, you might be able to achieve a strange effect where the individual phonemes of speech remain intelligible, but being rearranged are totally unintelligible together. I don’t know yet. Requires experimentation.
That’s all for now in this very first real post in Josh Knowles’ brand new Blogger log.
