Actually, here's another test that might be a bit more meaningful. This is using graphicksmagic and the identify/signature command in order to generate sha256 hashes for the image data only.
It is producing identical hashes for: 1. The original 12-bit DPX from Da Vinci Resolve (not packed to 16-bit) 2. The graphicsmagick 12-bit DPX to 12-bit DPX (padded to 16-bit DPX) file 3. The ffmpeg convert 12-bit DPX to 12-bit DPX (padded to 16-bit DPX) which uses the ffmpeg decoder with Jerome's patch. $ gm identify -verbose 6e0edc3b-7d89-4710-8b9a-66c8c990f9dc_1_00094787.dpx |grep -i signature Signature: 3fb67059dca860c483b1deac912973b76d48fa9f0114d63e8116cff69d55bb4c $ gm identify -verbose gm_dpx_to_dpx.dpx |grep -i signature Signature: 3fb67059dca860c483b1deac912973b76d48fa9f0114d63e8116cff69d55bb4c $ gm identify -verbose ffmpeg_dpx_to_dpx.dpx |grep -i signature Signature: 3fb67059dca860c483b1deac912973b76d48fa9f0114d63e8116cff69d55bb4c Hopefully this is a bit better than just using ffmpeg? Best, Kieran. _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel