[FFmpeg-devel] [PATCH] h264_mp4toannexb_bsf: always set idr_sps_pps_seen when SPS/PPS is seen.

[Benoit Fouet]

Hi,

Le 26/09/2014 18:38, Michael Niedermayer a écrit :
> On Fri, Sep 26, 2014 at 09:27:01AM +0200, Benoit Fouet wrote:
>> Hi,
>>
>> ----- Mail original -----
>>> Michael Niedermayer <michaelni <at> gmx.at> writes:
>>>
>>>>
>>>> On Fri, Aug 01, 2014 at 01:54:14AM +0200, Michael Niedermayer
>>>> wrote:
>>>>> On Thu, Jul 31, 2014 at 03:40:51PM +0200, Benoit Fouet wrote:
>>>>>> In order not to break a sequence like "SPS IDR SPS IDR", the
>>> boolean
>>>>>> telling that the SPS/PPS has been seen should always be set.
>>>>>> ---
>>>>>>  libavcodec/h264_mp4toannexb_bsf.c | 2 +-
>>>>>>  1 file changed, 1 insertion(+), 1 deletion(-)
>>>>>
>>>>> LGTM
>>>>
>>>> applied
>>>>
>>>> thanks
>>>
>>> A GitHub user "@thomag" commented on this commit:
>>>
>>>> Consider an mp4 files with just pps in the h264 stream (no sps in
>>>> the
>>>> NAL units, but available from avcc).
>>>> In this case 'ctx->idr_sps_pps_seen' prevents inserting sps in the
>>>> extracted h264, which is then unusable.
>>>
>>> FYI. I cannot confirm if it is correct or not, and what the correct
>>> solution would be.
>>>
>>
>> I'm willing to try and fix this when/if we have a sample to test that.
>> If there is also an example with just sps in the avcc and only pps in
the stream, that'd be perfect...
>
> forwarding from github
>
>     Michael,
>     Here is a short clip @
https://s3.amazonaws.com/quickfire-public/dwclip2.mp4
>     which has just PPS in the h264 stream.
>     Please let me know if you need anything else.
>     Regards,
>     Thomas.
>

OK, cool. I'll have a look at this after the weekend then.

-- 
Ben