On Sat, Nov 28, 2015 at 6:21 PM, Ganesh Ajjanagadde <gajja...@mit.edu> wrote: > On Sat, Nov 28, 2015 at 5:26 PM, Paul B Mahol <one...@gmail.com> wrote: >> Signed-off-by: Paul B Mahol <one...@gmail.com> > [...] > >> + case SQUARE: >> + val = (phs < 0.5) ? -1 : +1; > > Something I missed: consider using e.g FFSIGN(phs - 0.5) or FFSIGN(0.5 > - phase), can't recall.
This I still suggest for readability. > This is actually important because if one > feeds in 0.5 exactly (which is exactly representable), val becomes > unitialized and hence garbage. Sorry, this was complete garbage and should be ignored. What you choose to do at the discontinuity is practically irrelevant. The "purest" thing to do is take the midpoint of the jump, i.e 0 per Fourier theory, but I don't think it is good due to an extra bench for essentially zero gain. > > [...] _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel