[FFmpeg-devel] Patch for High color and High bit-depth support

[James Zern]

On Fri, Apr 3, 2015 at 6:37 PM, Debargha Mukherjee <debargha at google.com> wrote:

>>
>> You can probably merge profile 1 like 2/3 to cut down on the duplication.
>> You
>> might want to set bitdepth to 8 by default reducing the ifdefs here too.
>>
>
> done.
>

> +#ifdef VPX_IMG_FMT_HIGHBITDEPTH
> +    case AV_PIX_FMT_YUV420P10LE:
> +    case AV_PIX_FMT_YUV420P12LE:
> +        if (codec_caps & VPX_CODEC_CAP_HIGHBITDEPTH) {
> +            enccfg->g_bit_depth = enccfg->g_input_bit_depth =
> +                avctx->pix_fmt == AV_PIX_FMT_YUV420P10LE ? 10 : 12;
> +            enccfg->g_profile = 2;
> +            *img_fmt = VPX_IMG_FMT_I42016;
> +            *flags |= VPX_CODEC_USE_HIGHBITDEPTH;
> +            return 0;
> +        } else {
> +            break;
> +        }
>

just drop the else, you return from the other block.

>
>>
>> > +#if CONFIG_LIBVPX_VP9_ENCODER
>> > +    if (avctx->codec_id == AV_CODEC_ID_VP9) {
>> > +        if (set_pix_fmt(avctx, codec_caps, &enccfg, &flags, &img_fmt))
>> > +            return AVERROR(EINVAL);
>> > +        av_log(avctx, AV_LOG_INFO, "Bit-depth: %d\n",
>> enccfg.g_bit_depth);
>> >
>>
>> Won't build with a non-high-bitdepth build.
>>
>
> done.
>

This should get output in dump_enc_cfg(), no? I think the log can be dropped.

>
>>
>> > -    .pix_fmts       = (const enum AVPixelFormat[]){ AV_PIX_FMT_YUV420P,
>> AV_PIX_FMT_NONE },
>> > +    .pix_fmts       = (const enum AVPixelFormat[]){ AV_PIX_FMT_YUV420P,
>> > +                                                    AV_PIX_FMT_YUV422P,
>> > +                                                    AV_PIX_FMT_YUV444P,
>> > +#ifdef VPX_IMG_FMT_HIGHBITDEPTH
>> > +
>> AV_PIX_FMT_YUV420P10LE,
>> > +
>> AV_PIX_FMT_YUV422P10LE,
>> > +
>> AV_PIX_FMT_YUV444P10LE,
>> > +
>> AV_PIX_FMT_YUV420P12LE,
>> > +
>> AV_PIX_FMT_YUV422P12LE,
>> > +
>> AV_PIX_FMT_YUV444P12LE,
>> > +#endif
>>
>
> done using ff_vp9_init_static().
>
> diff --git a/libavcodec/libvpx.c b/libavcodec/libvpx.c
> index 5d29893..5111eed 100644
> --- a/libavcodec/libvpx.c
> +++ b/libavcodec/libvpx.c
> @@ -19,12 +19,47 @@
>   */
>
>  #include <vpx/vpx_codec.h>
> -
> +#include <vpx/vpx_encoder.h>
> +#include <vpx/vp8cx.h>
>

You need to make sure the encoder is available. Keep the break between local
and system headers.

>      if (    vpx_codec_version_major() < 1
>          || (vpx_codec_version_major() == 1 && vpx_codec_version_minor() < 3))
>          codec->capabilities |= CODEC_CAP_EXPERIMENTAL;
> +    if (    vpx_codec_version_major() == 1 && vpx_codec_version_minor() >= 3) {
>

1.4.0 would be better for all high bitdepth since it wasn't working at 1.3.0
and identifying anything in between would need a configure change. Also check
for major > 1.