.CHAPTER 3
.CHAPTER_TITLE "Convolution"
.HEADING 1 Convolution
\"  what convolution is, and which impulse responses produce interesting results, how to avoid feedback on them, unorthodox uses of them.
\" Not explaining what convolution is, but rather creative uses of it, e.g. LFO modulating between two IRs, using non-IR samples or synthesizing those IRs -- which wave structures are interesting. Could also be interesting using a wavetable patch as an IR -- oh wait, you can't do that with convolution since it has to be a static wave structure. Best not use the convolution ugen itself and SC and use Alik Rustamoff's methods.

.PP
Convolution applies the acoustic properties of one sound—typically a space or a system—to another. This is achieved using an impulse response (IR), which captures how a system reacts to an impulse signal. Convolution can replicate the reverb, coloration, and spatial qualities of the original environment or hardware.

.LIST
An impulse response is typically recorded using methods like:
.ITEM
Playing a sine sweep (a tone that gradually moves through all audible frequencies) into the space or device, then recording its response and mathematically deconvolving it to extract the IR. This method is the most accurate and has a high signal-to-noise ratio.
.ITEM
Using short, broadband sounds, usually a balloon pop, to approximate an impulse directly.
.LISTEND

.PP
The early part of an IR tends to define the stereo image and spatial location cues, while the later part conveys the reverberant tail, which contributes to a sense of space and depth.
In traditional convolution, the IR is static. To speed up real-time convolution, many implementations pre-calculate and store filter values derived from the IR.
.PP
However, in more experimental uses—like modulating between multiple IRs, or using unconventional sources such as a wavetable synth patch as an IR—traditional convolution tools fall short. These techniques require non-linear or time-varying convolution methods, which demand different mathematics and tools beyond standard IR convolution. We'll explore these unorthodox uses in depth later.
.PP
The interesting thing about convolution is that it can have a substantial effect on the timbre of a sound—just as a low-pass filter does on an FM synth. By attenuating one source through an IR, convolution can amplify hidden or subtle frequency content, revealing textures and resonances that weren’t previously audible. This makes it not just a spatial effect, but a powerful tool for timbral transformation.
.COLLATE