On 6/21/2010 1:41 PM, Reimar D?ffinger wrote: > On Mon, Jun 21, 2010 at 12:40:30PM -0500, Rolf Ernst wrote: > >> On the other hand - the webm people released an mencoder patch to >> support VP in mencoder. Glancing at the svn code that patch has been >> (somewhat) integrated but apparently currently does not work or does >> not work right (you can say vcodec=libvpx but the resulting output >> is not playable). >> >> Do you know what the status of it is and what the plan? If you tell >> me, yes, we'll get it working within a few weeks it sounds like I >> might just hold off. If you say, well, we have no plans for it I may >> go the ffmpeg route but I'd rather not. >> > My point of view currently: I (and quite a few other FFmpeg developers) > seem unable to convince them to choose a bitstream format that works > with AVI or any other CFR-only format citing some as-of-yet-undiscovered > possible future case that might need their special handling of so-called ARFs. > So, if you provide more details I might look into it, but given the unwillingness > of making VP8 work properly with MEncoder, VirtualDub, VfW in general and any > other program that was never designed to do anything but CFR I'm not sure > I see much of a point, since MEncoder would always be "second-class", > never working fully with VP8. > Reimar, I am confused. I have used ffmpeg 0.6 which supports the webm container. Are you then saying it only provides CFR information? I am not sure how Handbrake provides VFR using ffmpeg but somehow they seem to introduce timestamping into the Matroska container. That all aside: Fact is that the Google guys have provided a patch for mencoder: http://code.google.com/p/webm/issues/detail?id=68 and http://code.google.com/p/webm/downloads/list This suggests to me that they have done some heavy lifting to integrate libvpx and ffmpeg (as executable) spits out webm. mplayer-vp8-encdec-support-r2.tar.bz2 <detail?name=mplayer-vp8-encdec-support-r2.tar.bz2&can=2&q=> Have you chosen not to flow this into the development path? Certainly some of these files can be found in svn. Has this not worked out. I think the 'second class' argument is not really applicable because it is doubtful that somehow magically some open source team will emerge which will pop out a tool with the vast amount of possibilities of mencoder/ffmpeg supporting vfr. I believe that the overwhelming community will be quite content to simply use VP8 with CFR (which it offers). If the webm guys (which I presume to be Google?) have delivered a patch - why not use it? /re -- /re Capital punishment is our society's recognition of the sanctity of human life. Orrin Hatch
