On Sat, Feb 24, 2024 at 11:33 AM Bouke / Videotoolshed < bo...@videotoolshed.com> wrote:
> Trying to split a 64 channel 48Khz 24 bits Wave file. > That is VERY slow. > > ffmpeg -i /Volumes/Data/magweg/64channel.wav -filter_complex > "[0]pan=mono|c0=c0[tr0]" -c:a pcm_s24le -map "[tr0]" > /Volumes/Data/magweg/64_test/out_1.wav > > Runs at about 15x > > But if I do all 64 in one pass, it slows down to about 0.7x > > I don’t get it. DiskIO should not be a factor, and there is no math > involved. (Granted, the ppm_s24le might involve math, but that should be > simple.) > > Is this a bug? > (Same happens vice versa, going from multiple non-interleaved to one big > interleaved Wave.) > > Bouke > For demuxing case, use -max_size 4000000 as input option for wav demuxer, order of this option matters. Issue is if single packets have small number of stored samples it hurts performance with ffmpeg utility, recently situation got even worse with threading work. I raised concerns and reported it to authorities multiple times without results, but got responses that is not relevant but very minor problem. Dunno what command you use for reverse direction so can not help you on that one. > > _______________________________________________ > ffmpeg-user mailing list > ffmpeg-user@ffmpeg.org > https://ffmpeg.org/mailman/listinfo/ffmpeg-user > > To unsubscribe, visit link above, or email > ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe". > _______________________________________________ ffmpeg-user mailing list ffmpeg-user@ffmpeg.org https://ffmpeg.org/mailman/listinfo/ffmpeg-user To unsubscribe, visit link above, or email ffmpeg-user-requ...@ffmpeg.org with subject "unsubscribe".
