[FFmpeg-devel] Loop filter questions

Michael Niedermayer michaelni
Sun Jan 6 14:49:48 CET 2008


On Sun, Jan 06, 2008 at 02:21:10PM +0200, Kostya wrote:
> Here's what I was able to find out about loop filters employed in RV30/40.
> 
> RV30 loop filter:
>  if subblocks on both sides of edge are coded, then
>   deblock them with LIM=some parameter derived from top/left subblock
>   of this pair
>  I suspect it uses this table and quantizer to set strength:
>    0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 2 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5
>    0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 2 2 2 2 2 2 3 3 3 3 3 3 4 4 4
>    0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 2 2 2 2 2 3 3 3 3 4 4 4 4 5 5 5
>    0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 2 2 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6
>    0 0 0 0 0 0 0 0 0 0 0 1 1 1 2 2 2 3 3 3 4 4 4 5 5 5 6 6 6 7 7 7
>    0 0 0 0 0 0 0 0 0 0 0 1 1 2 2 3 3 4 4 4 5 5 5 6 6 6 7 7 7 8 8 8
>    0 0 0 0 0 0 0 0 0 0 0 1 2 3 4 4 4 5 5 5 6 6 6 7 7 7 8 8 8 9 9 9
> 
>  filter is weak loop filter taken from some H.264 draft:
>   t = ((SRC[-2] - SRC[1]) - (SRC[-1] - SRC[0])*4) >> 3;
>   d = clip(t, -LIM, LIM);
>   SRC[-1] = SRC[-1] + d;
>   SRC[ 0] = SRC[ 0] - d;
> 
> 
> RV40 loop filter:
>  this filter uses clip table, alpha and beta parameter tables from
>  early H.264 draft (A003r1),
>  loop filter is applied to the edges of coded blocks and depending
>  on where neighbouring block is coded or not, some limit parameters
>  are chosen (ClipTable[0][quant], ClipTable[1][quant] or ClipTable[2][quant])
> 
> filter code is in http://svn.mplayerhq.hu/soc/rv40/rv40.c?revision=1487&view=markup
>  it's adaptive and depending of pixel difference will use either
>  weak filtering similar to one employed in RV30, or stronger one, for up to
>  4 border pixels with coefficients [25 26 26 26 25] or [25 26 51 26]
> 
> 
> Can somebody point me to something similar? It's hard to me to search all
> H.264 drafts to find out where those things are proposed.  

well, i cant do anything else than looking through the drafts either, also
besides the drafts there were likely various proposals which never made it
in a draft but where rejected. Its possible one of them was used, and then
theres the reference software which may or may not be 100% identical to the
corresponding drafts
also the directories on the vceg/jvt ftp servers generally contain a file
which describes what the other files contain (in case you didnt notice)

you could also try to ask on the jvt-experts mailinglist, though iam not
sure how much they would appreciate a question about real video, its not
really on topic.

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

I wish the Xiph folks would stop pretending they've got something they
do not.  Somehow I fear this will remain a wish. -- M?ns Rullg?rd
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080106/d1cbff81/attachment.pgp>



More information about the ffmpeg-devel mailing list