Hello,

I'm developing a DSP plugin and at this point I have some doubts:

1) How should I deal with audio chunks with a sample count of zero? Discard them or just pass them through?

2) How can I neatly discard an audio chunk? Maybe setting the sample count to zero?

3) What really happens if I return false in on_chunk?

4) Is the main architecture of foobar explained anywhere? Is it push or pull flow-control based? From the DSP point of view looks like push, but...

Thanks

If you don't discard them, the playback engine will.

If on_chunk() returns false, the chunk will be removed from the chunk list passed to you by the playback engine, so unless you store the data in a private buffer, the chunk is effectively discarded.

Decoders, DSPs and outputs are all driven by the playback engine and process or produce data on its request.

Thanks for your reply, foosion . I needed to be sure, since I'm working on a DSP that its input:output ratio is not 1:1. You're right, part of the data will be held in a private buffer.

Regarding the null sized chunks, actually it was my fault. I thought I was receiving them, but I was generating them instead . An ugly console error is raised each time a null chunk is detected so it cannot pass unnoticed!