On 6 Sep 2009, at 21:37, Kassen wrote:
Have you discussed giving ChucK a 1-bit audio model?
I'm quite sure we could develop our own processing models in ChucK using single bits.
Sure.
I've been thinking about that but didn't yet get round to it. This man makes it out like it's a new thing but single bit audio is quite old, ...
The video says Yamasaki develop his stuff in the 1980's. Formats like AIFF do a tradeoff between data capacity and human hearing. So it is the high data capacity in recent years that makes it possible more widely.
...I have some sigma-delta based single bit delays here that are quite old, so old in fact that i got them for free. I think there are interesting opportunities here for non-conventional processing.
The 1-bit model tries avoiding the roundoff errors of fixed size sampling.
It'll work right now, but you will need to write it yourself; remember that this;
.001::samp => now;
is perfectly valid ChucK syntax. You can do sigma-delta or any other strategy you'd like. You may need to render in non-realtime though....
One problem is that one needs a much higher time resolution. The quoted device uses 2.8224 MHz, and a higher end model uses 5.6448 MHz. One might do a hybrid: there will be a timed stream, in which some data expresses a difference change to the current value, and some an absolute value. Hans