Hello, On Mon, 2007-12-10 at 16:59 +0100, Benjamin Larsson wrote: > What's wrong with > > p->dsp.float_to_int16(&buf_out[subframe*AMR_SUBFRAME_SIZE] > , p->samples_in, AMR_SUBFRAME_SIZE); > > ? Is void *data aligned? Can I use it for the output of float_to_int16? If so, I'll change it. :) Rob