Christoph Gerstbauer <christophgerstbauer <at> gmail.com> writes:
> > (This is not the first time that I am slightly > > surprised people use FFmpeg to prove that an > > encoder of FFmpeg is lossless.) > Do you have any ideas for proving the > losslessness with other tools than ffmpeg? You could use gm to convert both the original and the converted file to a raw format and compare. But let me try it differently: You have an input file with a logarithmic colour space that is not supported by FFmpeg. If you are sure that you know what you are doing, you can of course use FFmpeg to compress your raw data (I believe ffv1 was tested much better already but this is not really related). But while the data will still be identical the colourspace will be linear because FFmpeg only supports linear colourspaces. So you will will have to "know" that the actual content is logarithmic when you use it again (or write the transfer information when creating the file). It appears to me that gm ignores the transfer information just like FFmpeg, I tested with: $ gm convert -define tiff:bits-per-sample=16 LOG_ORIGINAL.dpx out.tif Carl Eugen _______________________________________________ ffmpeg-user mailing list [email protected] http://ffmpeg.org/mailman/listinfo/ffmpeg-user
