[x265] fix cbf context

[Satoshi Nakagawa]

> Does this fix coding errors? or just compression loss?

Coding error.

context for trDepthC = 3 is used --tu-inter-depth=4.
context for trDepthC = 4 is used in 4:4:4.


> -----Original Message-----
> From: x265-devel [mailto:x265-devel-bounces at videolan.org] On Behalf Of
> Steve Borho
> Sent: Friday, September 05, 2014 1:00 AM
> To: Development for x265
> Subject: Re: [x265] fix cbf context
> 
> On 09/05, Satoshi Nakagawa wrote:
> > # HG changeset patch
> > # User Satoshi Nakagawa <nakagawa424 at oki.com> # Date 1409843113 -32400
> > #      Fri Sep 05 00:05:13 2014 +0900
> > # Node ID 85a4327fe163a91b4725891515234c87e1153289
> > # Parent  b686cb0abd713f6fefcc75d00725232d12e36089
> > fix cbf context
> 
> Does this fix coding errors? or just compression loss?
> 
> > diff -r b686cb0abd71 -r 85a4327fe163
> source/Lib/TLibCommon/ContextTables.h
> > --- a/source/Lib/TLibCommon/ContextTables.h	Thu Sep 04 17:06:16 2014
> +0530
> > +++ b/source/Lib/TLibCommon/ContextTables.h	Fri Sep 05 00:05:13 2014
> +0900
> > @@ -65,7 +65,7 @@
> >
> >  #define NUM_REF_NO_CTX                2       ///< number of
> context models for reference index
> >  #define NUM_TRANS_SUBDIV_FLAG_CTX     3       ///< number of
> context models for transform subdivision flags
> > -#define NUM_QT_CBF_CTX                6       ///< number of context
> models for QT CBF
> > +#define NUM_QT_CBF_CTX                7       ///< number of context
> models for QT CBF
> >  #define NUM_QT_ROOT_CBF_CTX           1       ///< number of
> context models for QT ROOT CBF
> >  #define NUM_DELTA_QP_CTX              3       ///< number of
> context models for dQP
> >
> > @@ -144,7 +144,7 @@
> >  #define NEIGHBOURHOOD_00_CONTEXT_1_THRESHOLD_4x4  3  #define
> > NEIGHBOURHOOD_00_CONTEXT_2_THRESHOLD_4x4  1
> >
> > -static const uint32_t ctxCbf[3][3] = { { 1, 0, 0 }, { 2, 3, 4 }, { 2,
> > 3, 4} };
> > +static const uint32_t ctxCbf[3][5] = { { 1, 0, 0, 0, 0 }, { 2, 3, 4,
> > +5, 6 }, { 2, 3, 4, 5, 6 } };
> >  static const uint32_t
> > significanceMapContextSetStart[MAX_NUM_CHANNEL_TYPE][3] = { { 0,  9,
> 21 }, { 0,  9, 12 } };  static const uint32_t
> significanceMapContextSetSize[MAX_NUM_CHANNEL_TYPE][3]  = { { 9, 12,
> 6 }, { 9,  3,  3 } };
> >  static const uint32_t
> nonDiagonalScan8x8ContextOffset[MAX_NUM_CHANNEL_TYPE]   = {  6, 0  };
> > @@ -259,9 +259,9 @@
> >  static const uint8_t
> >      INIT_QT_CBF[3][NUM_QT_CBF_CTX] =
> >  {
> > -    { 153,  111,  149,   92,  167,  154, },
> > -    { 153,  111,  149,  107,  167,  154, },
> > -    { 111,  141,   94,  138,  182,  154, },
> > +    { 153,  111,  149,   92,  167,  154,  154 },
> > +    { 153,  111,  149,  107,  167,  154,  154 },
> > +    { 111,  141,   94,  138,  182,  154,  154 },
> >  };
> >
> >  static const uint8_t
> > _______________________________________________
> > x265-devel mailing list
> > x265-devel at videolan.org
> > https://mailman.videolan.org/listinfo/x265-devel
> 
> --
> Steve Borho
> _______________________________________________
> x265-devel mailing list
> x265-devel at videolan.org
> https://mailman.videolan.org/listinfo/x265-devel