[FFmpeg-devel] [PATCH v4 4/9] avcodec: add cbs for h266/vvc

[Nuo Mi]

On Tue, Jan 26, 2021 at 2:35 AM James Almer <jamrial at gmail.com> wrote:

> On 1/25/2021 2:12 PM, James Almer wrote:
> > On 1/25/2021 11:15 AM, Nuo Mi wrote:
> >> +static int FUNC(ref_pic_list_struct)(CodedBitstreamContext *ctx,
> >> RWContext *rw,
> >> +                                     H266RefPicListStruct *current,
> >> +                                     uint8_t list_idx, uint8_t
> rpls_idx,
> >> +                                     const H266RawSPS *sps)
> >> +{
> >> +    CodedBitstreamH266Context *h266 = ctx->priv_data;
> >> +    int err, i, j, general_layer_idx = -1, num_direct_ref_layers = 0;
> >> +    const H266RawVPS *vps = h266->vps[sps->sps_video_parameter_set_id];
> >> +
> >> +    if (!vps) {
> >> +        av_log(ctx->log_ctx, AV_LOG_ERROR,
> >> +               "VPS id %d not available.\n",
> >> sps->sps_video_parameter_set_id);
> >> +        return AVERROR_INVALIDDATA;
> >
> > Wont this break all the samples that lack a VPS?
> >
> > (I thought for that matter that all those samples having this field but
> > then no VPS at all was odd).
>
> Right, i see now that vps_video_parameter_set_id in a VPS is meant to
> never be 0, so apparently a VPS with id 0 and default values is implicit
> to always exist while not being coded in the bitstream, and that's what
> these SPS with sps_video_parameter_set_id == 0 reference.
>
> I don't know if allocating it when parsing the first SPS is the best
> idea, but i can't think if a better alternative right now.
>
Yes. You are right.

> _______________________________________________
> 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".