[FFmpeg-devel] [PATCH] Implement PAFF in H.264

[Jeff Downs]

On Fri, 28 Sep 2007, Neil Brown wrote:

> Why is the IDR frame missed?  Well with PAFF, it is an IDR field of
> course.  The first field of a complementary-field-pair is 'IDR', the
> second isn't.  So this line:
>             s->current_picture_ptr->key_frame = (hx->nal_unit_type == NAL_IDR_SLICE);

Yes, indeed. Thanks for tracking this down.

> [...]
> 
> My proposed fix (which "works for me") is:
> 
> Index: ffmpeg/libavcodec/h264.c
> ===================================================================
> --- ffmpeg.orig/libavcodec/h264.c	2007-09-23 13:37:14.000000000 +1000
> +++ ffmpeg/libavcodec/h264.c	2007-09-28 11:16:00.000000000 +1000
> @@ -7581,7 +7581,7 @@ static int decode_nal_units(H264Context 
>              if((err = decode_slice_header(hx, h)))
>                 break;
>  
> -            s->current_picture_ptr->key_frame= (hx->nal_unit_type == NAL_IDR_SLICE);
> +            s->current_picture_ptr->key_frame |= (hx->nal_unit_type == NAL_IDR_SLICE);
>              if(hx->redundant_pic_count==0 && hx->s.hurry_up < 5
>                 && (avctx->skip_frame < AVDISCARD_NONREF || hx->nal_ref_idc)
>                 && (avctx->skip_frame < AVDISCARD_BIDIR  || hx->slice_type!=B_TYPE)

This is half the solution. Problem here is that key_frame is wrong from 
the get-go (see MPV_frame_start and decode_slice_header() comments about 
s->pict_type).  It is based on the slice type being I Type or not, which 
is incorrect for H.264 because I slices can exist in non-IDR frames.

The original line overrides the bad value computed in MPV_frame_start.

I fixed this in my latest patches (coming soon) by combining this with an 
override (to zero) of the bad value from MPV_frame_start.  I would've 
corrected the setting of pict_type, but that's used in several places.

Thanks again for looking at this and providing the fix.

	-Jeff