Quoting Mark Thompson (2017-01-08 00:05:46)
> This removes the additional factor of FF_QP2LAMBDA (118) in use of
> global_quality.  The new scale is entirely codec dependent but should
> resemble the H.26[345] quantisation parameter.
> ---
>  avconv.c                        | 5 ++---
>  avconv_opt.c                    | 2 +-
>  doc/APIchanges                  | 3 +++
>  doc/encoders.texi               | 2 +-
>  libavcodec/a64multienc.c        | 2 +-
>  libavcodec/aacpsy.c             | 2 +-
>  libavcodec/asvenc.c             | 5 ++---
>  libavcodec/avcodec.h            | 3 ++-
>  libavcodec/dnxhdenc.c           | 4 ++--
>  libavcodec/libfaac.c            | 2 +-
>  libavcodec/libmp3lame.c         | 2 +-
>  libavcodec/libschroedingerenc.c | 2 +-
>  libavcodec/libspeexenc.c        | 3 +--
>  libavcodec/libtheoraenc.c       | 2 +-
>  libavcodec/libtwolame.c         | 3 +--
>  libavcodec/libvorbis.c          | 2 +-
>  libavcodec/libwebpenc.c         | 5 ++---
>  libavcodec/libx264.c            | 4 ++--
>  libavcodec/libxavs.c            | 4 ++--
>  libavcodec/libxvid.c            | 6 +++---
>  libavcodec/mpegvideo_enc.c      | 4 ++--
>  libavcodec/proresenc.c          | 2 +-
>  libavcodec/qsvenc.c             | 2 +-
>  libavcodec/vorbisenc.c          | 2 +-
>  libavutil/avutil.h              | 2 --
>  libavutil/frame.h               | 2 +-
>  26 files changed, 37 insertions(+), 40 deletions(-)

In general very much ok, but I'd keep FF_QUALITY_SCALE for a while.
In theory FF-prefixed stuff is for internal use only, but this thing
hails from ye olde times when namespacing rules were rather loose.

> diff --git a/doc/APIchanges b/doc/APIchanges
> index 7633c9918..8cade923e 100644
> --- a/doc/APIchanges
> +++ b/doc/APIchanges
> @@ -13,6 +13,9 @@ libavutil:     2015-08-28
>  
>  API changes, most recent first:
>  
> +2017-xx-xx - xxxxxxx - lavc 58.xx.x - avcodec.h
> +  Change scale of AVCodecContext.global_quality.

This could be a bit more verbose.

-- 
Anton Khirnov
_______________________________________________
libav-devel mailing list
[email protected]
https://lists.libav.org/mailman/listinfo/libav-devel

Reply via email to