[FFmpeg-devel] [Brainstorming] sound reproduction with a 3 voice soundchip

[François Revol]

Le 28 f?vr. 2011 ? 12:58, Michael Niedermayer a ?crit :

> On Mon, Feb 28, 2011 at 11:42:00AM +0100, Tobias Bindhammer wrote:
>> After playing around with video codecs to enable video playback on a
>> c64, it would be nice to have also sound for that. There would be some
>> options to reproduce sound via the soudnchip, for e.g. by putting 4 bit
[...]
> If you dont care about encoding speed, some form of iterative minimization
> of the difference seems like an interresting option.
> "difference" here can be some sum of squared differences or more advanced
> psyco acoustic thing. some FFT ignoring or downweighting phase could be tried
> too

FWIW I posted this on irc://ircnet/oric to some people doing music stuff on 8bit machines, here is what he said:

<twilighte_> ok, i am reminded of a similar technique dbug linked me to many years ago, something like ver.. something
memory not so good
its the idea of using the oscillators directly as pulse width generators, the idea that lower frequencies represent wider pulse widths and higher frequencies shorter pulse widths
in this way a pulse code modulation could be reproduced by directly feeding a pulse code modulated waveform into the oscilator frequency registers
It was theorised doing this could also raise the oppertunity to reproduce 3 channels samples to much greater accuracy than 4 bit samples
its a hack though since adjusting the oscillator frequency does not reset the waveform cycle. So sound reproduction quality may be lost
on the ay anyway. i don't know if the SID behaves in the same way
i also don't know if any time could be saved in doing it this way (which is the aim of the original post)


Fran?ois.