Closed GoogleCodeExporter closed 9 years ago
Hi Guys,
This is an issue about video call between libjingle client and GMail client, I
always receive packets with NAL type 18 from GMail client
Here is an example packet which I captured using wireshark:
[RTP Standard header: 12 bytes]
90 e1 4e de // Payload type is 97, so this is correct
71 46 f3 05
e2 28 97 85
[Extension header: 12 bytes, no problem] check this FYI:
http://tools.ietf.org/html/draft-ietf-avt-tfrc-profile-10
be de 00 03
66 00 00 00
02 c8 ea c5
[Actual payload: 0x72 indicates the NAL type is 18]
72 00 00 1b
21 e3 18 0b
...
"ITU-T Recommendation H.264" says:
----
NAL units having nal_unit_type equal to 13 and 19 may be discarded by decoders
without affecting the decoding process for NAL units having nal_unit_type not
equal to 13 or 19 and without affecting conformance to this Recommendation |
International Standard. (Page 57)
----
But I received a lot of NAL 18 packets, I think I cannot simply ignore them,
I'm not a H.264 expert, does anyone know to handle NAL 18 packets? Is it
specific to GMail client?
Kind Regards
Original comment by interfac...@gmail.com
on 18 Jul 2011 at 12:27
[deleted comment]
now I understand the extension header is:
be de 00 03
66 00 00 00
02 c8 ea c5
72 00 00 1b
So the payload starts with "0x21", but my decoder always returns error because
the first byte of my payload is always "0x21", why do I always get coded slice?
An idea?
Original comment by interfac...@gmail.com
on 20 Jul 2011 at 12:52
If the first byte after the extension header is 21, the NAL unit type is
1(non-IDR slice), but I think the first two NAL units you should receive are
7(SPS) and 8(PPS). Are you using wireshark to make the capture? Does it show
only non-IDR NAL units? Here I don't have this issue.
Original comment by diego.cd...@gmail.com
on 12 Aug 2011 at 4:09
I found I actually missed SPS and PPS frame because GMail started the stream
before I accepted, that's why I "always" got slice frames.
Thanks Diego.
Original comment by interfac...@gmail.com
on 18 Aug 2011 at 10:46
Original comment by juberti@google.com
on 9 Sep 2011 at 7:36
Original issue reported on code.google.com by
interfac...@gmail.com
on 18 Jul 2011 at 11:40