Michael Niedermayer <mich...@niedermayer.cc> skrev: (26 februari 2016 22:56:42 CET) >On Fri, Feb 26, 2016 at 08:00:58AM +0100, Mats Peterson wrote: >> On 02/26/2016 07:41 AM, Mats Peterson wrote: >> >>Look at this snippet from libavformat/qtpalette.c that stores a >palette >> >>entry (palette[] is uint32_t): >> >> >> >>palette[i] = (a << 24 ) | (r << 16) | (g << 8) | (b); >> >> >> >>The way it is stored in memory is obviously dependent on the >endianness >> >>of the machine. For little-endian machines, it will be BGRA, and >for >> >>big-endian ones ARGB. >> >> >> >>Mats >> >> >> > >> >This shouldn't have any relevance, though. Something is possibly >wrong >> >with the palette storage endian-wise when converting that matrix >file >> >from yuv420p to pal8. >> > >> >Mats >> > >> >> My final spam for some time. The palette is stored in host byte >> order internally, isn't it? Then I don't understand whatsoever why >> v26 of my movenc patch will fail on a big-endian machine. No >> AV_RL32() should be needed. > >for normal AVPackets AVPacket.data should not be in some host >endian specific format > >[...]
That much I understand now. -- Mats Peterson http://matsp888.no-ip.org/~mats/ _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
