[FFmpeg-devel] "error, non monotone timestamps 45000 >= 45000" in ffserver

Michael Niedermayer michaelni
Fri Aug 24 00:42:58 CEST 2007


Hi

On Wed, Aug 22, 2007 at 03:21:02PM +0200, Luca Abeni wrote:
> Hi Luca,
>
> Luca Barbato wrote:
>> Luca Abeni wrote:
>> [seeking and rtp/rtsp]
>> You may try feng. ffplay seems to seek to a point if you use the mouse
>> clicks but sends wrong requests on ss.
>
> Well, I am not really interested in RTSP seeking... I was just trying to 
> have a look at the RTSP special case in ffplay, noticed by Michael. It 
> looked related to seeking, so I tried seeking and I noticed that it is 
> broken.
>
> Anyway, I think something like the attached patch is needed... It does not 
> completely fix seeking, but I think it makes it less broken than what we 
> currently have.
>
>
>
> 			Thanks,
> 				Luca

> Index: libavformat/rtsp.c
> ===================================================================
> --- libavformat/rtsp.c	(revision 10178)
> +++ libavformat/rtsp.c	(working copy)
> @@ -1267,7 +1267,7 @@
>  {
>      RTSPState *rt = s->priv_data;
>  
> -    rt->seek_timestamp = timestamp;
> +    rt->seek_timestamp = av_rescale(timestamp, AV_TIME_BASE * (int64_t)s->streams[stream_index]->time_base.num, s->streams[stream_index]->time_base.den);

av_rescale_q and AV_TIME_BASE_Q would be more readable
except that patch ok, the function gets a timestamp in stream timebase 
and its later divided by AV_TIME_BASE so the code apparently assumes
its in AV_TIME_BASE_Q timebase

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Complexity theory is the science of finding the exact solution to an
approximation. Benchmarking OTOH is finding an approximation of the exact
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20070824/d238d9a2/attachment.pgp>



More information about the ffmpeg-devel mailing list