[FFmpeg-devel] H.264 encoding/decoding process

Sergio Fadda superatropo
Tue Oct 16 10:38:07 CEST 2007

I'm working on a project of video conference.
I have to implement encoding/RTP encapsulation/RTP
decapsulation/decoding for a H.264 video stream.
The RTP (enc/dec)apsulation works well (I have compared the stream
before encapsulation and after decapsulation); the problem is that
when I try to decode the transmitted video frames ffmpeg print out
these errors:

[h264 @ ...] non existing PPS reference
[h264 @ ...] decode_slice_header error
[h264 @ ...] no frame!

The strange is that these errors appear every two frames I try to decode.

In encoding I set AVCodecContext::flags to CODEC_FLAG_TRUNCATED.
Should I use flags2 too? Which flags in this case?
The CODEC_ID_x I use is CODEC_ID_H264; I see a CODEC_ID_FFH264...
should I use this instead? What are the differences?
The procedure I use to encode/decode is:

initialization(); // codec, context, parser
encoding(); // just a avcodec_encode_video() call
// encapsulation and decapsulation omitted
decoding(); // just a avcodec_decode_video() call

Where I make the mistake?

P.S. I use a trunk version downloaded a few days ago.

More information about the ffmpeg-devel mailing list