[FFmpeg-devel] Timestamp issues in live RTP->mpegts bridges

Alexandre FERRIEUX - FT/RD/SIRP/ASF/SOFTL alexandre.ferrieux
Tue Jan 20 15:59:02 CET 2009


Michael Niedermayer wrote:
> 
> what?
> you just have to use .h263 not .263 in the filename
> or -f h263 or so

OK sorry I took this extension from one of your examples ;-) (or was it 
Luca's ? never mind)

>> 29.97 tb(r)
>>
>> So it is wrongly guessing 29.97.
>> Maybe I could "force" the fps through some SDP parameter or command-line 
>> option ?
> 
> i have in the past renamed fps to tb, it seems this was not
> enough, how can i make it more clear that this value is NOT the framerate?

All my fault. I'm just beginning to dive into this. Just point me to the 
proper documentation and I won't do that mistake again.

>>> Try moving finalize_packet() out of the else block and see if it fixes 
>>> anything...
>> I did that, but it still detects 29.97.
> 
> H.263 allows exactly one timebase, and this timebase is 30000/1001, it does
> allow different framerates but not different timebases.
> H263+ allows different timebases

OK.

Talking about the h263/h263p differences, I noticed that though my 
sending device (a mobile phone) sends an H263+ flow, I could only read 
it correctly with an H263 SDP. That is, the following SDP works:

	m=video 6000 RTP/AVP 96
	a=rtpmap:96 H263

while the following doesn't:

	m=video 6000 RTP/AVP 96
	a=rtpmap:96 H263-1998/90000

(with or without the /90000)
By "doesn't work" I mean, keeps calling av_read_frame_internal, never 
ending the pre-scanning work.

Similarly, on the output side, with the patch posted earlier I am able 
to stream "-vcodec h263" (forcing the payload to 96 instead of 34) and 
the device displays it perfectly, while "-vcodec h263p" is not displayed.

-Alex






More information about the ffmpeg-devel mailing list