[Libav-user] Criteria for h.264 key-frame detection

[Robert Krüger]

On Mon, Oct 21, 2013 at 8:26 PM, Carl Eugen Hoyos <cehoyos at ag.or.at> wrote:
> Robert Krüger <krueger at ...> writes:
>
>> I was interested to know how you found out that
>> ffmpeg "thinks" the 28th frame is a random access
>> point
>
> I concatenated the second until the 29th frame of the
> input sample and tested decoding with current FFmpeg.
> I also wanted to test with the reference decoder
> (that may have answered all our questions) but it
> crashes.

Thanks for the info.

I just cut away everything from the file until the start of the packet
containing the 27th frame (tail -c4329532 gh1_720p.mts >
gh1_720p_cat.mts) and the resulting file decodes without artifacts
using ffmpeg. The number of decoded frames is also as expected. So I
think this is at least an indication that the 27th frame is indeed a
valid random access point.

>
> I used the reference decoder but ffmpeg -vf showinfo
> allows you to see the type of frames.
> (I had the suspicion you patched the source to find
> out.)

I have but I am outputting the info on the parser/slice level. At
least for interlaced material I have cases where the information in
the frame is incorrect but it is still correct at the slice level.

Robert