[Ffmpeg-devel] Re: Bug - Framerate limitation for H.264 in .mp4

Michael Niedermayer michaelni
Wed Mar 22 19:22:05 CET 2006


Hi

On Wed, Mar 22, 2006 at 06:15:53PM +0100, Baptiste COUDURIER wrote:
> Hi
> 
> Michael Niedermayer wrote:
> > Hi
> > 
> > On Wed, Mar 22, 2006 at 12:34:02AM +0100, Baptiste COUDURIER wrote:
> >> Tyler Loch wrote:
> >>> Michael Niedermayer <michaelni <at> gmx.at> writes:
> >>>> could you provide such a file or _exact_ instructions to create one
> >>>> (low resolution like 16x16 constant black could be used to reduce size
> >>> Here are two video-only clips: one 2,127 frames long (working)
> >>> and one 2,180 frames long (duration error).
> >>>
> >>> http://www.isquint.org/2127sxsw.mp4
> >>> http://www.isquint.org/2180sxsw.mp4
> >>>
> >> First one is fine. Second has really weird stts atom :
> >>
> >> first stts duration is set to 1001000 while timescale is set to 15015000
> >> Duration overflow.
> > 
> > isnt it possible to store the duration(s) in 64bit? (yeah iam too lazy
> > to look at the spec ATM ...)
> > 
> > 
> > [...]
> 
> Patch attached, acceptable ?

yes


> 
> Btw, Michael did you consider my last request ? have a CODEC_ID_OTHER,
> CODEC_ID_UNKNOWN, or something similar to use when 'mp4s', 'tcmd' is
> encoutered in stsd atom ?

iam have a somehow unhappy feeling with this, somehow i think CODEC_ID_NONE
should be used but then this doesnt fit well into the current code on mov.c
[...]

-- 
Michael





More information about the ffmpeg-devel mailing list