On 02/13/2016 05:39 AM, Mats Peterson wrote:
On 02/13/2016 12:51 AM, Michael Niedermayer wrote:
+ if (context->has_pkt_pal) {
+ memcpy(context->palette->data,
+ avpkt->data + avpkt->size - AVPALETTE_SIZE,
AVPALETTE_SIZE);
This is incorrect, the 8bit palette in nut can have any number of
entries up to 256, this assumes its fixed 256
I had absolutely no idea, Michael. I'm doing my best ;) It would be much
easier and more consequent if it used AVPALETTE_SIZE, though.
ill try to split out some trivial parts from this patch and apply
them as seperate commits after some testing though
Alright.
again, if you could help fix the bugs (which are not caused by your
patch but which interact with these changes in nasty ways)
that would be great
The rawvideo muxing bugs (incorrect stride for AVI and QuickTime files
produced with "-vcodec rawvideo", and weird storage of the palette in
the frames for AVI and QuickTime) once again have nothing to do with my
patch. They've been there for a long time. It's up to someone else to
fix these, since I'm not confident on the muxer side.
These musing bugs don't interact anymore than before the patch, at that.
They were just as nasty back then (not being able to produce a standards
compliant file with "-vcodec rawvideo").
Mats
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
