[FFmpeg-devel] [PATCH] avcodec: Add librav1e encoder

[Derek Buitenhuis]

On 28/05/2019 20:49, Derek Buitenhuis wrote:
>>> +static const AVOption options[] = {
>>> +    { "quantizer", "use constant quantizer mode", OFFSET(quantizer), AV_OPT_TYPE_INT, { .i64 = -1 }, -1, 255, VE },
>>> +    { "max-quantizer", "max quantizer when using bitrate mode", OFFSET(max_quantizer), AV_OPT_TYPE_INT, { .i64 = 255 }, 1, 255, VE },
>> This should be mapped to qmax lavc option instead.
> OK. Is there a matching option for plain old quantizer?
> I couldn't find an obvious one.

Looking closer at this, it does not seem correct to use it.

As, per `ffmpeg -h full`:

      -qmax              <int>        E..V..... maximum video quantizer scale (VBR) (from -1 to 1024) (default 31)

The default of 31 is really bad for AV1. Further, rav1e's CLI default of 255
is probably the safest to use here.

- Derek