On Feb 25, 2013, at 18:57, Brad O'Hearne wrote: > On Feb 25, 2013, at 10:43 AM, "René J.V. Bertin" <rjvber...@gmail.com> wrote: > > I'm building for 64-bit. But anyway, I'm not sure that changing rounding > necessarily solves the problem, as any approach could result in 0 -- perhaps > not for a selected sample, but eventually, for other random samples, I would > think it possible.
True enough, I should have seen that too of course. Another thought - the divisor is *probably* the input range (multiplied by something like SHORT_MAX or SHORT_MAX/2). I don't see any specification of the input range in your sample code, I presume you have checked that it corresponds to what swr_convert expects? R _______________________________________________ Libav-user mailing list Libav-user@ffmpeg.org http://ffmpeg.org/mailman/listinfo/libav-user