[FFmpeg-devel] [WIP] Seek / subfile support in concat demuxer

Michael Niedermayer michaelni at gmx.at
Sun Mar 2 18:38:01 CET 2014


On Sun, Mar 02, 2014 at 04:12:39PM +0100, Nicolas George wrote:
> Le quintidi 5 ventôse, an CCXXII, Michael Niedermayer a écrit :
> > i neither have a strong oppinon but side data would be more consistent
> 
> That is true. SKIP_SAMPLES can not be reused, though, because it has SAMPLES
> in it.

IMHO, a video frame strictly speaking is a sample


> 
> > also side data would allow conveyng the information earlier
> > that is a keyframe could already contain the exact duration covering
> > multiple frames that should be discarded
> 
> I am not sure what you mean here. Is it: if frames #42 #43 and #44 must be
> skipped, set SKIP_FRAME=3 on packet #42 instead of FLAG=SKIP on all three
> packets?

yes, i think thats what i meant


> 
> > and in theory side data would allow cutting a frame in half
> > if for example in case of a slide show frames sra displayed for
> > significat time (iam not suggesting to implement this ATM though)
> 
> This case can already be handled by adjusting the timestamps on the packet,
> unless I am mistaken.

that assumes you have timestamps on the packets and
unless i miss something it wont work in some cases even with
timestamps
Consider there are multiple frames encoded in 1 packet
H263 PB frames come to mind here or field pictures in mpeg2

also theres frame reordering related to B frames


[...]


-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

The worst form of inequality is to try to make unequal things equal.
-- Aristotle
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20140302/639d640d/attachment.asc>


More information about the ffmpeg-devel mailing list