[FFmpeg-devel] [PATCH] Change libaom default to crf=28.

[Elliott Karpilovsky]

On Wed, Aug 14, 2019 at 5:35 PM James Zern
<jzern-at-google.com at ffmpeg.org> wrote:
>
> Hi,
>
> On Tue, Aug 13, 2019 at 8:23 PM elliottk
> <elliottk-at-google.com at ffmpeg.org> wrote:
> >
> > Current default is 256kbps, which produces inconsistent
> > results (too high for low-res, too low for hi-res).
> > Use CRF instead, which will adapt.
> > ---
> >  libavcodec/libaomenc.c | 8 ++++----
> >  1 file changed, 4 insertions(+), 4 deletions(-)
> >
>
> I think this is OK as it's similar to what is done for x264/5 [1].
>
> > diff --git a/libavcodec/libaomenc.c b/libavcodec/libaomenc.c
> > index 9b4fb3b4eb..a18d11c8aa 100644
> > --- a/libavcodec/libaomenc.c
> > +++ b/libavcodec/libaomenc.c
> > @@ -575,10 +575,10 @@ static av_cold int aom_init(AVCodecContext *avctx,
> >          if (enccfg.rc_end_usage == AOM_CQ) {
> >              enccfg.rc_target_bitrate = 1000000;
> >          } else {
> > -            avctx->bit_rate = enccfg.rc_target_bitrate * 1000;
> > +            enccfg.rc_end_usage = AOM_Q;
> > +            ctx->crf = 28;
> >
>
> Can we take a library default here or does it default to bitrate and
> have cq_level cleared?
I believe the library defaults to bitrate without a CQ level.
>
> >              av_log(avctx, AV_LOG_WARNING,
> > -                   "Neither bitrate nor constrained quality specified, using default bitrate of %dkbit/sec\n",
> > -                   enccfg.rc_target_bitrate);
> > +                   "Neither bitrate nor constrained quality specified, using default CRF of 28\n");
> >
>
> You may want to use the variable to produce the output in case it changes.
Good suggestion. I will update the patch.
>
> [1] https://bugs.chromium.org/p/aomedia/issues/detail?id=2219#c9
> _______________________________________________
> 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".