Thanks for the update! The sound format issue seems to be fixed.
However, I am wondering why the new build seems a bit slower than the old pvoc with the sound format bug? Can the speed difference be just because the old build wrote just 16 bit files whereas I am making the new build write 32 bit floating point files? Could it be that you accidentally posted a debug binary or some such for Windows now?
Actually, even the older pvoc program seems to run curiously slow compared to the CDP programs that actually process the analysed data...Might there be some chances for optimizations?