[FFmpeg-devel] [PATCH 2/2] lavc/libx264: add mb_info option

[Anton Khirnov]

Quoting Carotti, Elias (2023-07-03 17:55:40)
> On Wed, 2023-06-21 at 15:57 +0000, Carotti, Elias wrote:
> > Hi all,
> > please find the second part of the patch set.
> > Best,
> > Elias
> 
> Hi all,
> please find the second part of the patch, updating libavcodec/libx264.c
> to use the AVVideoHint side data. This should use the block_size field
> to scan the AVVideoRect array.
> 
> Best,
> Elias
> 
> 
> 
> NICE SRL, viale Monte Grappa 3/5, 20124 Milano, Italia, Registro delle Imprese di Milano Monza Brianza Lodi REA n. 2096882, Capitale Sociale: 10.329,14 EUR i.v., Cod. Fisc. e P.IVA 01133050052, Societa con Socio Unico
> 
> 
> 
> From 5bdb624d8dbcff96493a63f02f68a3961cf72723 Mon Sep 17 00:00:00 2001
> From: Elias Carotti <eliascrt at amazon.it>
> Date: Tue, 20 Jun 2023 19:29:08 +0200
> Subject: [PATCH 2/2] lavc/libx264: add mb_info option
> 
> Pass the information about unchanged parts of the frame by means of the
> AVVideoHint side data.
> ---
>  Changelog            |  1 +
>  doc/APIchanges       |  3 ++
>  libavcodec/libx264.c | 94 ++++++++++++++++++++++++++++++++++++++++++++
>  libavcodec/version.h |  2 +-
>  4 files changed, 99 insertions(+), 1 deletion(-)
> 
> diff --git a/Changelog b/Changelog
> index 3876082844..70b0fe94a3 100644
> --- a/Changelog
> +++ b/Changelog
> @@ -25,6 +25,7 @@ version <next>:
>  - Raw VVC bitstream parser, muxer and demuxer
>  - Bitstream filter for editing metadata in VVC streams
>  - Bitstream filter for converting VVC from MP4 to Annex B
> +- support for the P_SKIP hinting to speed up libx264 encoding
>  
>  version 6.0:
>  - Radiance HDR image support
> diff --git a/doc/APIchanges b/doc/APIchanges
> index bfe04556d2..89ff88af15 100644
> --- a/doc/APIchanges
> +++ b/doc/APIchanges
> @@ -2,6 +2,9 @@ The last version increases of all libraries were on 2023-02-09
>  
>  API changes, most recent first:
>  
> +2023-06-21 - xxxxxxxxxx - lavc 60.23.100 - libx264.c
> +  Add mb_info option.

Do we actually need the option? If the encoder's caller bothered with
adding the side data, then I'd think it should always be used, otherwise
why is it even there?

-- 
Anton Khirnov