> > On 19 May 2025, at 03:01, George Welch via ffmpeg-user > <ffmpeg-user@ffmpeg.org> wrote: > > Howdy ffmpeg users. > > I have found a behavior that seems strange to me. The following two commands: > > $ ffmpeg -i input.mkv -an -c:v libx265 -preset slow -crf 21 output1.mkv > > and > > $ ffmpeg -i input.mkv -an -c:v libx265 -preset slow -crf 21 -vf > "setpts=PTS-STARTPTS" output2.mkv > > produce output files that are very different. The input file is 1080p h264 > recorded at 15Mbps. The output file in the first case is about 1.66 times > larger than the second output file. The video channel in the first case > averages 7342 kbps while the video channel in the second case averages 4423 > kbps. > > Unsurprisingly, dumping frames to png files and using imagemagick to compare > them shows that the first output file has much higher fidelity to the > original than the second. > > This is with ffmpeg 7.1.1 installed via homebrew on a fairly modern macbook > pro. Checking on a linux machine with a fairly old intel xeon processor (all > I have access to) also with ffmpeg installed via homebrew produces nearly > identical files. > > I wonder if this a bug in the mac version of ffmpeg? I don't see why the > setpts filter should produce a much lower bitrate, nor why it would be > different under linux.
The complete uncut console output in each case would have been a good starting point to explain the differences. Def > > Thank you for your time. > _______________________________ _______________________________________________ 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".
