[Ffmpeg-cvslog] r7246 - trunk/libavcodec/avcodec.h

Panagiotis Issaris takis.issaris
Tue Dec 19 13:37:23 CET 2006


Hi,

On Tue, 2006-12-19 at 11:53 +0100, Baptiste Coudurier wrote:
> Panagiotis Issaris wrote:
> > Hi,
> > 
> > On Tue, 2006-12-19 at 10:54 +0100, Baptiste Coudurier wrote:
> >> Hi
> >> 
> >> takis wrote:
> >>> Author: takis Date: Thu Dec  7 16:59:42 2006 New Revision: 7246
> >>> 
> >>> Modified: trunk/libavcodec/avcodec.h
> >>> 
> >>> Log: Add codec ID for a native H.264 encoder.
> >>> 
> >>> 
> >>> Modified: trunk/libavcodec/avcodec.h 
> >>> ==============================================================================
> >>>  --- trunk/libavcodec/avcodec.h	(original) +++
> >>> trunk/libavcodec/avcodec.h	Thu Dec  7 16:59:42 2006 @@ -149,6
> >>> +149,7 @@ CODEC_ID_TIERTEXSEQVIDEO, CODEC_ID_TIFF, CODEC_ID_GIF, 
> >>> +    CODEC_ID_FFH264,
> >>> 
> >> We were discussing that issue, why did you add it ?
> > Well, frankly, because the maintainer of that file (Michael) had 
> > accepted that hunk of the patch and it seemed the following
> > discussion had halted without a clear decision.
> 
> Well, in that case, we should had a clear statement, asking again would
> not have hurt anything.
You are right. Sorry about that.

> 
> >> Codec id are for codecs, not a way to distinguish between 
> >> decoders/encoders IMHO.
> > How about CODEC_ID_MPEG2VIDEO and CODEC_ID_MPEG2VIDEO_XVMC or 
> > CODEC_ID_MPEG4 and the CODEC_ID_MSMPEG* variants of CODEC_ID_WMV3 and
> >  CODEC_ID_VC1?
> > 
> > Aren't they different implementations of the same standard? (MPEG2 in
> >  the first case, MPEG-4 in the second and VC1 in the third)
> > 
> 
> Don't know much about differences, but I guess bitstream contains
> differences according to lavc code (msmpeg4.c), so again I guess those
> are variants of one standard, not different implementations.
> Again maybe Im completely wrong.

But WMV3 vs VC1? I thought VC1 was the name of the standard and WMV3 the
name of Microsoft's implementation of that standard.

Note that I am just trying to illustrate that it seemed to me that
existing codec IDs did not strictly associate with codec standards.

In fact, the only thing I would fear for (if both x264 and ffh264 would
use the same CODEC_ID) is that it would introduce confusion. In the
sense that two runs of ffmpeg revision X with identical parameters would
result in different bitstreams&quality&performance.

> FFH264 is just a different implementation of CODEC_ID_H264, and should
> not have another id IMHO. Also since lavc API allows to choose
> decoder/encoder by name instead of id, though ffmpeg.c only uses id.


With friendly regards,
Takis

-- 
vCard: http://www.issaris.org/pi.vcf
Public key: http://www.issaris.org/pi.key





More information about the ffmpeg-cvslog mailing list