[FFmpeg-devel] [PATCH v1] avformat/mxf: set stream frame rates for ST 422 essence containers

[Tomas Härdin]

sön 2022-08-28 klockan 08:31 -0700 skrev pal at sandflow.com:
> From: Pierre-Anthony Lemieux <pal at palemieux.com>
> 
> The MXF demuxer does not currently set AVStream::avg_frame_rate and
> ::r_frame_rate
> when J2K essence is wrapped according to SMPTE ST 422.
> 
> ---
>  libavformat/mxfdec.c | 25 +++++++++++++++++++++++++
>  1 file changed, 25 insertions(+)
> 
> diff --git a/libavformat/mxfdec.c b/libavformat/mxfdec.c
> index feebff67aa..043e2e06ec 100644
> --- a/libavformat/mxfdec.c
> +++ b/libavformat/mxfdec.c
> @@ -2140,6 +2140,13 @@ finish_decoding_index:
>      return ret;
>  }
>  
> +static int mxf_is_st_422(const UID *essence_container_ul) {
> +    static const uint8_t st_422_essence_container_ul[] = {
> 0x06,0x0e,0x2b,0x34,0x04,0x01,0x01,0x07,0x0d,0x01,0x03,0x01,0x02,0x0c
> };
> +
> +    return essence_container_ul &&
> mxf_match_uid(*essence_container_ul, st_422_essence_container_ul,
> +                                                
> sizeof(st_422_essence_container_ul));
> +}

Seems unnecessary to have a function for this

> +
>  static int mxf_is_intra_only(MXFDescriptor *descriptor)
>  {
>      return mxf_get_codec_ul(mxf_intra_only_essence_container_uls,
> @@ -2892,6 +2899,24 @@ static int
> mxf_parse_structural_metadata(MXFContext *mxf)
>                      av_log(mxf->fc, AV_LOG_INFO, "Unknown frame
> layout type: %d\n", descriptor->frame_layout);
>              }
>  
> +            if (mxf_is_st_422(essence_container_ul)) {
> +                switch ((*essence_container_ul)[14]) {
> +                case 2: /* Cn: Clip- wrapped Picture Element */
> +                case 3: /* I1: Interlaced Frame, 1 field/KLV */
> +                case 4: /* I2: Interlaced Frame, 2 fields/KLV */
> +                case 6: /* P1: Frame- wrapped Picture Element */
> +                    st->avg_frame_rate = source_track->edit_rate;
> +                    st->r_frame_rate = st->avg_frame_rate;
> +                    break;
> +                case 5: /* F1: Field-wrapped Picture Element */
> +                    st->avg_frame_rate = av_mul_q(av_make_q(2, 1),
> source_track->edit_rate);
> +                    st->r_frame_rate = st->avg_frame_rate;
> +                    break;
> +                default:
> +                    break;
> +                }
> +            }

Looks OK, but we should have a test for this as well I think

At some point we'll probably need to implement proper timecode
support..

/Tomas