[FFmpeg-devel] [PATCH] lavc/videotoolboxenc: add MJPEG support

[Andreas Rheinhardt]

gnattu via ffmpeg-devel:
> For example:
> ./ffmpeg -hwaccel videotoolbox\
>    -i INPUT -c:v mjpeg_videotoolbox\
>     out.mp4
> 
> This encoder does not have many options and can only use ``-q:v` to control quality.
> 
> Signed-off-by: Gnattu OC <gnattuoc at me.com>
> ---
>  Changelog                    |  1 +
>  configure                    |  2 ++
>  libavcodec/Makefile          |  1 +
>  libavcodec/allcodecs.c       |  1 +
>  libavcodec/videotoolboxenc.c | 39 +++++++++++++++++++++++++++++++++---
>  5 files changed, 41 insertions(+), 3 deletions(-)
> 
> diff --git a/Changelog b/Changelog
> index e83a00e35c..b73a44f9ad 100644
> --- a/Changelog
> +++ b/Changelog
> @@ -38,6 +38,7 @@ version 7.0:
>  - ffplay with hwaccel decoding support (depends on vulkan renderer via libplacebo)
>  - dnn filter libtorch backend
>  - Android content URIs protocol
> +- VideoToolbox MJPEG encoder
>  
>  
>  version 6.1:
> diff --git a/configure b/configure
> index 2a1d22310b..52a0ee762f 100755
> --- a/configure
> +++ b/configure
> @@ -3474,6 +3474,8 @@ hevc_videotoolbox_encoder_deps="pthreads"
>  hevc_videotoolbox_encoder_select="atsc_a53 videotoolbox_encoder"
>  prores_videotoolbox_encoder_deps="pthreads"
>  prores_videotoolbox_encoder_select="videotoolbox_encoder"
> +mjpeg_videotoolbox_encoder_deps="pthreads"
> +mjpeg_videotoolbox_encoder_select="videotoolbox_encoder"
>  libaom_av1_decoder_deps="libaom"
>  libaom_av1_encoder_deps="libaom"
>  libaom_av1_encoder_select="extract_extradata_bsf"
> diff --git a/libavcodec/Makefile b/libavcodec/Makefile
> index 9ce6d445c1..f59b2ae691 100644
> --- a/libavcodec/Makefile
> +++ b/libavcodec/Makefile
> @@ -503,6 +503,7 @@ OBJS-$(CONFIG_MJPEGB_DECODER)          += mjpegbdec.o
>  OBJS-$(CONFIG_MJPEG_CUVID_DECODER)     += cuviddec.o
>  OBJS-$(CONFIG_MJPEG_QSV_ENCODER)       += qsvenc_jpeg.o
>  OBJS-$(CONFIG_MJPEG_VAAPI_ENCODER)     += vaapi_encode_mjpeg.o
> +OBJS-$(CONFIG_MJPEG_VIDEOTOOLBOX_ENCODER) += videotoolboxenc.o
>  OBJS-$(CONFIG_MLP_DECODER)             += mlpdec.o mlpdsp.o
>  OBJS-$(CONFIG_MLP_ENCODER)             += mlpenc.o mlp.o
>  OBJS-$(CONFIG_MMVIDEO_DECODER)         += mmvideo.o
> diff --git a/libavcodec/allcodecs.c b/libavcodec/allcodecs.c
> index 2386b450a6..f8a90fdca2 100644
> --- a/libavcodec/allcodecs.c
> +++ b/libavcodec/allcodecs.c
> @@ -879,6 +879,7 @@ extern const FFCodec ff_mpeg4_mediacodec_encoder;
>  extern const FFCodec ff_mpeg4_omx_encoder;
>  extern const FFCodec ff_mpeg4_v4l2m2m_encoder;
>  extern const FFCodec ff_prores_videotoolbox_encoder;
> +extern const FFCodec ff_mjpeg_videotoolbox_encoder;
>  extern const FFCodec ff_vc1_cuvid_decoder;
>  extern const FFCodec ff_vp8_cuvid_decoder;
>  extern const FFCodec ff_vp8_mediacodec_decoder;
> diff --git a/libavcodec/videotoolboxenc.c b/libavcodec/videotoolboxenc.c
> index 15c34d59c3..0574edab56 100644
> --- a/libavcodec/videotoolboxenc.c
> +++ b/libavcodec/videotoolboxenc.c
> @@ -547,6 +547,7 @@ static CMVideoCodecType get_cm_codec_type(AVCodecContext *avctx,
>              else
>                  return MKBETAG('a','p','c','n'); // kCMVideoCodecType_AppleProRes422
>          }
> +    case AV_CODEC_ID_MJPEG: return kCMVideoCodecType_JPEG;
>      default:               return 0;
>      }
>  }
> @@ -1233,7 +1234,7 @@ static int vtenc_create_encoder(AVCodecContext   *avctx,
>                                        kVTCompressionPropertyKey_Quality,
>                                        quality_num);
>          CFRelease(quality_num);
> -    } else if (avctx->codec_id != AV_CODEC_ID_PRORES) {
> +    } else if (avctx->codec_id != AV_CODEC_ID_PRORES && avctx->codec_id != AV_CODEC_ID_MJPEG) {
>          bit_rate_num = CFNumberCreate(kCFAllocatorDefault,
>                                        kCFNumberSInt32Type,
>                                        &bit_rate);
> @@ -1347,7 +1348,7 @@ static int vtenc_create_encoder(AVCodecContext   *avctx,
>          }
>      }
>  
> -    if (avctx->gop_size > 0 && avctx->codec_id != AV_CODEC_ID_PRORES) {
> +    if (avctx->gop_size > 0 && avctx->codec_id != AV_CODEC_ID_PRORES && avctx->codec_id != AV_CODEC_ID_MJPEG) {
>          CFNumberRef interval = CFNumberCreate(kCFAllocatorDefault,
>                                                kCFNumberIntType,
>                                                &avctx->gop_size);
> @@ -1496,7 +1497,7 @@ static int vtenc_create_encoder(AVCodecContext   *avctx,
>          }
>      }
>  
> -    if (!vtctx->has_b_frames && avctx->codec_id != AV_CODEC_ID_PRORES) {
> +    if (!vtctx->has_b_frames && avctx->codec_id != AV_CODEC_ID_PRORES && avctx->codec_id != AV_CODEC_ID_MJPEG) {
>          status = VTSessionSetProperty(vtctx->session,
>                                        kVTCompressionPropertyKey_AllowFrameReordering,
>                                        kCFBooleanFalse);
> @@ -2844,6 +2845,13 @@ static const enum AVPixelFormat prores_pix_fmts[] = {
>      AV_PIX_FMT_NONE
>  };
>  
> +static const enum AVPixelFormat mjpeg_pix_fmts[] = {
> +    AV_PIX_FMT_VIDEOTOOLBOX,
> +    AV_PIX_FMT_NV12,
> +    AV_PIX_FMT_YUV420P,
> +    AV_PIX_FMT_NONE
> +};
> +
>  #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
>  #define COMMON_OPTIONS \
>      { "allow_sw", "Allow software encoding", OFFSET(allow_sw), AV_OPT_TYPE_BOOL, \
> @@ -3004,3 +3012,28 @@ const FFCodec ff_prores_videotoolbox_encoder = {
>      .p.wrapper_name   = "videotoolbox",
>      .hw_configs       = vt_encode_hw_configs,
>  };
> +
> +static const AVClass mjpeg_videotoolbox_class = {
> +    .class_name = "mjpeg_videotoolbox",
> +    .item_name  = av_default_item_name,
> +    .option     = NULL,
> +    .version    = LIBAVUTIL_VERSION_INT,
> +};

If you have no options, it makes no sense to use an AVClass.

> +
> +const FFCodec ff_mjpeg_videotoolbox_encoder = {
> +    .p.name           = "mjpeg_videotoolbox",
> +    CODEC_LONG_NAME("VideoToolbox MJPEG Encoder"),
> +    .p.type           = AVMEDIA_TYPE_VIDEO,
> +    .p.id             = AV_CODEC_ID_MJPEG,
> +    .p.capabilities   = AV_CODEC_CAP_DR1 | AV_CODEC_CAP_DELAY |
> +                        AV_CODEC_CAP_HARDWARE,
> +    .priv_data_size   = sizeof(VTEncContext),
> +    .p.pix_fmts       = prores_pix_fmts,
> +    .init             = vtenc_init,
> +    FF_CODEC_ENCODE_CB(vtenc_frame),
> +    .close            = vtenc_close,
> +    .p.priv_class     = &mjpeg_videotoolbox_class,
> +    .caps_internal    = FF_CODEC_CAP_INIT_CLEANUP,
> +    .p.wrapper_name   = "videotoolbox",
> +    .hw_configs       = vt_encode_hw_configs,
> +};