[Ffmpeg-devel] rtsp problem with ffplay

chris.2.dobbs at bt.com chris.2.dobbs
Wed Mar 8 15:59:50 CET 2006

Hi Luca,
Thanks again for info. Sorry to be a pain with this but I feel I am
getting closer!

I have totally misunderstood the meaning of Dynamic payloads - I thought
96 was specific for AMR!
I can see from the code that the sdp_parse_rtpmap() function succeeds in
extracting the RTSP session data for AMR i.e I get this from my debug

sdp_parse_line() picks this up now for each letter value (a, m and c)

a   -   'x-qt-text-nam:Session streamed by "NGC RTSP-Server V1.0e" on
port 3127 stream type:amr'
m   -   'audio 0 RTP/AVP 96'
c   -   'IN IP4'

And sdp_parse_rtpmap() gets the following buffer passed to it

'rtpmap:96 AMR/8000' 

So am I on the right track thinking I now need to add a entry to the
AVRtpDynamicPayloadTypes[] array to allow dynamic payload type 96 with
sdp map data equal to AMR to match this new entry and thus get the right
codec_id (0x12000 for AMR-NB I think...)  

Manay thanks for your input again

-----Original Message-----
From: ffmpeg-devel-bounces at mplayerhq.hu
[mailto:ffmpeg-devel-bounces at mplayerhq.hu] On Behalf Of Luca Barbato
Sent: 08 March 2006 13:30
To: FFMpeg development discussions and patches
Subject: Re: [Ffmpeg-devel] rtsp problem with ffplay

chris.2.dobbs at bt.com wrote:
> Many thanks for your reply Luca.
> I have not heard of libnms/nemesi ...is it part of the ffmpeg src


> BTW - I see what is going on with payload type 96 now, I suppose just 
> to test I could 'hard-code' the sdp_parser to pick up AMR NB codec 
> info for

DO NOT hardwire 96 with amr, the payload number is just a number now,
reading the sdp you discover what the server is sending with this
payload is the proper way to map number with codec.

> payload type 96....yuk I know but I am really interested to know if 
> ffplay can actually deal with AMR NB over rtsp

better check the rfc about it.

RTP is a pain because you need to write a separate framer for each
codec. Sometimes even write many framers if the people got too much
creative about it...


ffmpeg-devel mailing list
ffmpeg-devel at mplayerhq.hu

More information about the ffmpeg-devel mailing list