[FFmpeg-devel] [PATCH] librav1e: Don't make users explicitly set -qp -1 to use bit rate mode

James Almer jamrial at gmail.com
Mon Nov 11 22:00:39 EET 2019


On 11/11/2019 11:16 AM, Derek Buitenhuis wrote:
> Signed-off-by: Derek Buitenhuis <derek.buitenhuis at gmail.com>
> ---
>  libavcodec/librav1e.c | 6 +++++-
>  1 file changed, 5 insertions(+), 1 deletion(-)
> 
> diff --git a/libavcodec/librav1e.c b/libavcodec/librav1e.c
> index 5052cac896..11c1d09e36 100644
> --- a/libavcodec/librav1e.c
> +++ b/libavcodec/librav1e.c
> @@ -329,6 +329,10 @@ static av_cold int librav1e_encode_init(AVCodecContext *avctx)
>          }
>      }
>  
> +    /* We can't easily represent this behavior with AVOptions, so do it here. */
> +    if (!avctx->bit_rate && ctx->quantizer < 0)
> +        ctx->quantizer = 100;

Looks like the default of 100 is done in the library itself, if
rav1e_config_parse_int() isn't called for either quantizer or bitrate.

If you remove this chunk and leave the one below, the output bitstreams
when calling this encoder with no arguments vs using -qp 100 are the same.

> +
>      if (avctx->bit_rate && ctx->quantizer < 0) {
>          int max_quantizer = avctx->qmax >= 0 ? avctx->qmax : 255;
>  
> @@ -533,7 +537,7 @@ retry:
>  #define VE AV_OPT_FLAG_VIDEO_PARAM | AV_OPT_FLAG_ENCODING_PARAM
>  
>  static const AVOption options[] = {
> -    { "qp", "use constant quantizer mode", OFFSET(quantizer), AV_OPT_TYPE_INT, { .i64 = 100 }, -1, 255, VE },
> +    { "qp", "use constant quantizer mode (defaults to 100 if no bit rate is set)", OFFSET(quantizer), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 255, VE },
>      { "speed", "what speed preset to use", OFFSET(speed), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 10, VE },
>      { "tiles", "number of tiles encode with", OFFSET(tiles), AV_OPT_TYPE_INT, { .i64 = 0 }, -1, INT64_MAX, VE },
>      { "tile-rows", "number of tiles rows to encode with", OFFSET(tile_rows), AV_OPT_TYPE_INT, { .i64 = 0 }, -1, INT64_MAX, VE },


More information about the ffmpeg-devel mailing list