[FFmpeg-devel] [PATCH] avcodec/libdav1d: parse sequence headers in extradata if available

[James Almer]

On 7/11/2021 8:40 AM, Derek Buitenhuis wrote:
> On 7/9/2021 3:53 PM, James Almer wrote:
>> +    res = dav1d_parse_sequence_header(&seq, c->extradata + offset,
>> +                                      c->extradata_size  - offset);
>> +    if (res < 0) {
>> +        av_log(c, explode ? AV_LOG_ERROR : AV_LOG_INFO,
>> +               "Error decoding extradata\n");
>> +        return explode ? AVERROR_INVALIDDATA : 0;
>> +    }
>> +
> 
> I don't think it is a good idea to fail like this, even in explode
> mode.
> 
> Both the AV1-in-ISOBMFF and AV1-in-Matroska specs do specify that
> the sequence header OBU must be first if there is one in the
> configOBUs, however, they do not require on actually be present.
> 
> For example, configOBUs may contain a single metadata OBU, and
> nothing else - and it would be entirely valid - so failing here
> would actually be wrong, even in explode mode. This can happen,
> if, for example, you have a file with HDR metadata in the av1c box,
> but in-band sequence headers.

I can amend this patch locally to check for DAV1D_ERR(ENOENT), which was 
implemented to signal "No seqhdr is present", and return 0 in that case, 
but it's a very recent addition, so older libdav1d builds will never 
emit that error code.

Guess just returning 0 on all scenarios here until we increase the 
minimum required library version is better, so I'll do that.

> 
> - Derek
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
> 
> To unsubscribe, visit link above, or email
> ffmpeg-devel-request at ffmpeg.org with subject "unsubscribe".
>