Hi,

I want to add a processing step in dsp.c, but I must say I find it difficult to
fully understand how it works. A few questions and things I've noticed:

* Which format are the samples in during processing? I want to multiply them
with a fixed-point number - before clipping - so I need to know the exact
format.

* Which input sample depths are supported? The code implies 16 and 32 only,
but it isn't clear from the API.

* Samples seem to be scaled differently, depending on the stereo format (if
scaling is needed).

* When "rendering" the final samples, the scaling and clipping seems to do
things that aren't needed (unless GCC is smart enough to optimize away that).

Perhaps a cleanup would be in order? :)

  Magnus

_______________________________________________
http://cool.haxx.se/mailman/listinfo/rockbox
Received on Thu Jul 7 15:17:24 2005