On Sun, Jan 31, 2016 at 03:16:46PM +0100, Mats Peterson wrote: > On 01/31/2016 03:07 PM, Michael Niedermayer wrote: > >indeed, i should have clearly stated that i applied the pal8 > >patches with the plan to use pal8 only when neccessary. > >I had asked you to implement exactly that but it seemed you didnt > >know how to do that cleanly and simply so i thought "no problem, > >i know how to do that, ill do it" > >i had not expected this to be controversal > > > > Well, it would have been nice to know that you did it at least. I > had no idea until you mentioned it. > > >basically 100% of avis with monowhite are black&white, we do not have > >a single (no craftet) sample that contains paletted monochrome data > >in avi. > >IMHO its bizare to convert black and white 1 bpp data in the decoder > >to 8bit paletted data > >we reject colorspace/pixel format converting in the decoder normally > >unless its somehow a mess otherwise. > > > >and its a problem to for using the output because an encoder has to > >deal and encode 8bit paletted data from such avis even though they > >really are just black and white if pal8 is used > > > > It's not bizarre at all. Black & white is just a variant of > palettized data in this context, it's not monochrome. And just > because you don't have a non-b/w sample, it doesn't mean you should > use monow. A 1 bpp AVI is not monochrome once again, it's > palettized, even if it's only with black & white. The same semantics > as with 1 bpp depth in QuickTime.
well, avi is a "old" format and ive worked with avis and maintained the avi demuxer for a long time. So if i dont remember a paletted 1bpp avi thats some indication that these would be rather rare if they exist at all > > And since you don't have a "pal2" or "pal4" format, it suddenly > seems perfectly OK to convert this to pal8? THere's no logic in > this. the problem is what happens with the data after the rawvideo decoder if its to be stored in a file the encoder will store 8bit paletted data (if it supports that) vs. 1bit monochrome data (if it supports that) making generated files potentially several times bigger [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB Everything should be made as simple as possible, but not simpler. -- Albert Einstein
signature.asc
Description: Digital signature
_______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
