On Wed, Jan 21, 2015 at 09:33:21 +0100, Andrea Rastelli wrote: > Actually the problem is not in the creation of the JPEG itself, but in > the resulting chroma subsampling that (apparently) is not stored in the > FFmpeg resulting image.
You did not mention this so far within this thread, if I may point this out! > With ImageMagick, when I set -sampling-factor 4:4:4 or 4:2:2 a tool like > "mediainfo" shows the exact subsampling. With ffmpeg this information is > not set nor visible. Unfortunately, I don't know much about this topic. > Also, ImageMagick produce a bigger file (1.4Mb) than ffmpeg (0.1Mb). [...] > If, with FFMpeg, exists some parameter that produce the same result that > I have obtained with ImageMagick will be prefect (because, now, I have > two different tools to manage in my scripts) I can't manage to tickle extremely high quality out of ffmpeg, but the resulting image with -q:v 0..2 is very good. I converted a big JPEG, and the resulting file never reached more than half the size of the original, despite it being a recoding. Either my original had too much redundancy, or ffmpeg wasn't investing enough. ;-) But I think my method and my argumentation are totally wrong. Yet either way, the resulting image can't be considered bad. You may want to give us samples - an input TIFF, your ffmpeg results, and your ImageMagick results. Moritz _______________________________________________ ffmpeg-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/ffmpeg-user
