[x264-devel] Multislicing support

[Etienne Bömcke]

On 07 Jan 2008, at 18:01, Etienne Bömcke wrote:

> There's nothing wrong with them I guess. I'm working on an application
> that needs to be able to replace a bottom slice by a completely
> different video sequence. So I cannot have motion vectors for
> macroblocks of the upper slice linking to any macroblock of the bottom
> one, or else when switching slices artefacts would appear in the upper
> part of the sequence.
Just a precision : I'm aware that this feature is not part of the  
standard H.264 definition, so I was planning on an additional command- 
line parameter alowing the user to choose between standard slices  
(independant in term of intra-prediction only) and fully independant  
slices (both intra and inter-prediction).

Etienne

>
> I tried to modify the x264_mb_analyse_init function. It seems that
> adjusting the values of mb.mv_min[] and mb.mv_max[] relatively to the
> slice being encoded is sufficient to obtain perfectly independent
> slices. Is there anything wrong with this method?
>
> Etienne
>
> On 07 Jan 2008, at 17:46, Loren Merritt wrote:
>
>> On Mon, 7 Jan 2008, Etienne Bömcke wrote:
>>
>>> I still have to define additional constraints to disable usage of
>>> motion vectors liking two different slices.
>>
>> What's wrong with the mv constraints in x264_macroblock_cache_load?
>>
>> --Loren Merritt_______________________________________________
>> x264-devel mailing list
>> x264-devel at videolan.org
>> http://mailman.videolan.org/listinfo/x264-devel
>
> --
> Etienne Bömcke
> UCL - Laboratoire TELE
> etienne.bomcke at uclouvain.be
>
>
>
> _______________________________________________
> x264-devel mailing list
> x264-devel at videolan.org
> http://mailman.videolan.org/listinfo/x264-devel

--
Etienne Bömcke
UCL - Laboratoire TELE
etienne.bomcke at uclouvain.be