Many audio signal processing systems use floating-point sample buffers internally, even though the I/O is all in integer samples.
I added support for reading/writing normal sound file formats into floating-point sample buffers, and for supporting many-channel files. I did this about a month ago, and just noticed that the latest source version from SVN has some changes to these same methods, e.g., the new signature,
would be augmented with
which, in addition to writing normalized floating-point samples, would support numDestChannels > 2.
Is anyone else already doing this, or should I port my code to the latest API and distribute it?