Jan 2, 2024, 10:48 by nuomi2...@gmail.com:

> On Mon, Jan 1, 2024 at 11:18 PM Lynne <d...@lynne.ee> wrote:
>
>> Jan 1, 2024, 15:15 by nuomi2...@gmail.com:
>>
>> > Co-authored-by: Xu Mu <tox...@outlook.com>
>> > Co-authored-by: Frank Plowman <p...@frankplowman.com>
>> > Co-authored-by: Shaun Loo <shaunlo...@gmail.com>
>> > Co-authored-by: Wu Jianhua <toq...@outlook.com>
>> > ---
>> >  libavcodec/vvc/Makefile              |    1 +
>> >  libavcodec/vvc/vvc_ctu.h             |    1 +
>> >  libavcodec/vvc/vvc_filter.c          | 1332 ++++++++++++++++++++++++++
>> >  libavcodec/vvc/vvc_filter.h          |   71 ++
>> >  libavcodec/vvc/vvc_filter_template.c | 1135 ++++++++++++++++++++++
>> >  5 files changed, 2540 insertions(+)
>> >  create mode 100644 libavcodec/vvc/vvc_filter.c
>> >  create mode 100644 libavcodec/vvc/vvc_filter.h
>> >  create mode 100644 libavcodec/vvc/vvc_filter_template.c
>> >
>>
>> Is there really absolutely nothing you can reuse from the HEVC decoder?
>>
> LMCS and ALF are new filters. SAO is entirely reusable, and deblock is
> partially reusable.
> However, all of them require some changes to the HEVC code.
> Currently, I prefer not to touch the HEVC code as there is no HEVC
> maintainer, and any HEVC code change requires a very long review time.
>

I disagree with this. There doesn't need to be an HEVC maintainer to merge
patches, many of us are experienced enough to review them.
If this does not get done now, it will never get done, and it'll be a nightmare
for someone trying to clean it up in 5 years.

Particularly, we've been trying to clean up and deduplicate code and tables
as much as possible in libavcodec, so I'd rather not have all our effort go to
waste.

Please, make them share the code. Put keep in in libavcodec/, since it's
common to both.
_______________________________________________
ffmpeg-devel mailing list
ffmpeg-devel@ffmpeg.org
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel

To unsubscribe, visit link above, or email
ffmpeg-devel-requ...@ffmpeg.org with subject "unsubscribe".

Reply via email to