> passed by SDL_mixer to my sfx processing callback. 1880 samples at
The buffer size you get from an audio backend is normally difficult to
predict, and I wouldn't advice you to do so. Sometimes it may even
depend on the hardware you are using, so it is not guaranteed that it
will be the same on another machine. I don't know whether SDL
encapsulates that for you, that would be more a question for SDL
Most audio processing apps use ring buffers to account for that variability.
Just use any value you think provides you with a reasonable latency.
If you get more from the backend, just make several calls to the
LADSPA, it shouldn't mean much overhead. If you get less, just process
that bit or save it for the next iteration.
Linux-audio-dev mailing list