Merge remote-tracking branch 'qatar/master'

* qatar/master:
  swscale: remove misplaced comment.
  ffmpeg: fix streaming to ffserver.
  swscale: split out RGB48 output functions from yuv2packed[12X]_c().
  build: move vpath directives to main Makefile
  swscale: fix JPEG-range YUV scaling artifacts.
  build: move ALLFFLIBS to a more logical place
  ARM: factor some repetitive code into macros
  Fix SVQ3 after adding 4:4:4 H.264 support
  H.264: fix CODEC_FLAG_GRAY
  4:4:4 H.264 decoding support
  ac3enc: fix allocation of floating point samples.

Conflicts:
	ffmpeg.c
	libavcodec/dsputil_template.c
	libavcodec/h264.c
	libavcodec/mpegvideo.c
	libavcodec/snow.c
	libswscale/swscale.c
	libswscale/swscale_internal.h

Merged-by: Michael Niedermayer <michaelni@gmx.at>

24 files changed, 1793 insertions, 889 deletions

diff --git a/libavcodec/ac3enc.c b/libavcodec/ac3enc.c
index f61e7f8c4d..6ee8a7ac15 100644
--- a/libavcodec/ac3enc.c
+++ b/libavcodec/ac3enc.c
@@ -2215,15 +2215,9 @@ static av_cold int allocate_buffers(AVCodecContext *avctx)
     AC3EncodeContext *s = avctx->priv_data;
     int channels = s->channels + 1; /* includes coupling channel */
 
-    FF_ALLOC_OR_GOTO(avctx, s->windowed_samples, AC3_WINDOW_SIZE *
-                     sizeof(*s->windowed_samples), alloc_fail);
-    FF_ALLOC_OR_GOTO(avctx, s->planar_samples, s->channels * sizeof(*s->planar_samples),
-                     alloc_fail);
-    for (ch = 0; ch < s->channels; ch++) {
-        FF_ALLOCZ_OR_GOTO(avctx, s->planar_samples[ch],
-                          (AC3_FRAME_SIZE+AC3_BLOCK_SIZE) * sizeof(**s->planar_samples),
-                          alloc_fail);
-    }
+    if (s->allocate_sample_buffers(s))
+        goto alloc_fail;
+
     FF_ALLOC_OR_GOTO(avctx, s->bap_buffer,  AC3_MAX_BLOCKS * channels *
                      AC3_MAX_COEFS * sizeof(*s->bap_buffer),  alloc_fail);
     FF_ALLOC_OR_GOTO(avctx, s->bap1_buffer, AC3_MAX_BLOCKS * channels *
@@ -2323,6 +2317,8 @@ av_cold int ff_ac3_encode_init(AVCodecContext *avctx)
     AC3EncodeContext *s = avctx->priv_data;
     int ret, frame_size_58;
 
+    s->avctx = avctx;
+
     s->eac3 = avctx->codec_id == CODEC_ID_EAC3;
 
     avctx->frame_size = AC3_FRAME_SIZE;
@@ -2355,6 +2351,7 @@ av_cold int ff_ac3_encode_init(AVCodecContext *avctx)
         s->apply_window                 = ff_ac3_fixed_apply_window;
         s->normalize_samples            = ff_ac3_fixed_normalize_samples;
         s->scale_coefficients           = ff_ac3_fixed_scale_coefficients;
+        s->allocate_sample_buffers      = ff_ac3_fixed_allocate_sample_buffers;
         s->deinterleave_input_samples   = ff_ac3_fixed_deinterleave_input_samples;
         s->apply_mdct                   = ff_ac3_fixed_apply_mdct;
         s->apply_channel_coupling       = ff_ac3_fixed_apply_channel_coupling;
@@ -2364,6 +2361,7 @@ av_cold int ff_ac3_encode_init(AVCodecContext *avctx)
         s->mdct_init                    = ff_ac3_float_mdct_init;
         s->apply_window                 = ff_ac3_float_apply_window;
         s->scale_coefficients           = ff_ac3_float_scale_coefficients;
+        s->allocate_sample_buffers      = ff_ac3_float_allocate_sample_buffers;
         s->deinterleave_input_samples   = ff_ac3_float_deinterleave_input_samples;
         s->apply_mdct                   = ff_ac3_float_apply_mdct;
         s->apply_channel_coupling       = ff_ac3_float_apply_channel_coupling;
diff --git a/libavcodec/ac3enc.h b/libavcodec/ac3enc.h
index d5f662b4aa..563dcd532e 100644
--- a/libavcodec/ac3enc.h
+++ b/libavcodec/ac3enc.h
@@ -135,6 +135,7 @@ typedef struct AC3Block {
 typedef struct AC3EncodeContext {
     AVClass *av_class;                      ///< AVClass used for AVOption
     AC3EncOptions options;                  ///< encoding options
+    AVCodecContext *avctx;                  ///< parent AVCodecContext
     PutBitContext pb;                       ///< bitstream writer context
     DSPContext dsp;
     AC3DSPContext ac3dsp;                   ///< AC-3 optimized functions
@@ -230,6 +231,7 @@ typedef struct AC3EncodeContext {
     void (*scale_coefficients)(struct AC3EncodeContext *s);
 
     /* fixed vs. float templated function pointers */
+    int  (*allocate_sample_buffers)(struct AC3EncodeContext *s);
     void (*deinterleave_input_samples)(struct AC3EncodeContext *s,
                                        const SampleType *samples);
     void (*apply_mdct)(struct AC3EncodeContext *s);
@@ -276,6 +278,9 @@ void ff_ac3_float_scale_coefficients(AC3EncodeContext *s);
 
 /* prototypes for functions in ac3enc_template.c */
 
+int ff_ac3_fixed_allocate_sample_buffers(AC3EncodeContext *s);
+int ff_ac3_float_allocate_sample_buffers(AC3EncodeContext *s);
+
 void ff_ac3_fixed_deinterleave_input_samples(AC3EncodeContext *s,
                                              const SampleType *samples);
 void ff_ac3_float_deinterleave_input_samples(AC3EncodeContext *s,
diff --git a/libavcodec/ac3enc_template.c b/libavcodec/ac3enc_template.c
index d88fa225a1..0547165aaf 100644
--- a/libavcodec/ac3enc_template.c
+++ b/libavcodec/ac3enc_template.c
@@ -31,6 +31,26 @@
 #include "ac3enc.h"
 
 
+int AC3_NAME(allocate_sample_buffers)(AC3EncodeContext *s)
+{
+    int ch;
+
+    FF_ALLOC_OR_GOTO(s->avctx, s->windowed_samples, AC3_WINDOW_SIZE *
+                     sizeof(*s->windowed_samples), alloc_fail);
+    FF_ALLOC_OR_GOTO(s->avctx, s->planar_samples, s->channels * sizeof(*s->planar_samples),
+                     alloc_fail);
+    for (ch = 0; ch < s->channels; ch++) {
+        FF_ALLOCZ_OR_GOTO(s->avctx, s->planar_samples[ch],
+                          (AC3_FRAME_SIZE+AC3_BLOCK_SIZE) * sizeof(**s->planar_samples),
+                          alloc_fail);
+    }
+
+    return 0;
+alloc_fail:
+    return AVERROR(ENOMEM);
+}
+
+
 /**
  * Deinterleave input samples.
  * Channels are reordered from Libav's default order to AC-3 order.
diff --git a/libavcodec/arm/h264dsp_init_arm.c b/libavcodec/arm/h264dsp_init_arm.c
index e9146405c2..b344584799 100644
--- a/libavcodec/arm/h264dsp_init_arm.c
+++ b/libavcodec/arm/h264dsp_init_arm.c
@@ -122,7 +122,8 @@ static void ff_h264dsp_init_neon(H264DSPContext *c, const int bit_depth)
     c->h264_idct_dc_add     = ff_h264_idct_dc_add_neon;
     c->h264_idct_add16      = ff_h264_idct_add16_neon;
     c->h264_idct_add16intra = ff_h264_idct_add16intra_neon;
-    c->h264_idct_add8       = ff_h264_idct_add8_neon;
+    //FIXME: reenable when asm is updated.
+    //c->h264_idct_add8       = ff_h264_idct_add8_neon;
     c->h264_idct8_add       = ff_h264_idct8_add_neon;
     c->h264_idct8_dc_add    = ff_h264_idct8_dc_add_neon;
     c->h264_idct8_add4      = ff_h264_idct8_add4_neon;
diff --git a/libavcodec/arm/mpegvideo_armv5te_s.S b/libavcodec/arm/mpegvideo_armv5te_s.S
index 0606bb43b6..82095ab15d 100644
--- a/libavcodec/arm/mpegvideo_armv5te_s.S
+++ b/libavcodec/arm/mpegvideo_armv5te_s.S
@@ -35,6 +35,21 @@
  *
  * Inner loop should take 6 cycles per element on arm926ej-s (Nokia 770)
  */
+
+.macro  dequant_t       dst, src, mul, add, tmp
+        rsbs            \tmp, ip, \src, asr #16
+        addgt           \tmp, \add, #0
+        rsblt           \tmp, \add, #0
+        smlatbne        \dst, \src, \mul, \tmp
+.endm
+
+.macro  dequant_b       dst, src, mul, add, tmp
+        rsbs            \tmp, ip, \src, lsl #16
+        addgt           \tmp, \add, #0
+        rsblt           \tmp, \add, #0
+        smlabbne        \dst, \src, \mul, \tmp
+.endm
+
 function ff_dct_unquantize_h263_armv5te, export=1
         push            {r4-r9,lr}
         mov             ip, #0
@@ -44,50 +59,20 @@ function ff_dct_unquantize_h263_armv5te, export=1
 1:
         ldrd            r6, [r0, #8]
 
-        rsbs            r9, ip, r4, asr #16
-        addgt           r9, r2, #0
-        rsblt           r9, r2, #0
-        smlatbne        r9, r4, r1, r9
-
-        rsbs            lr, ip, r5, asr #16
-        addgt           lr, r2, #0
-        rsblt           lr, r2, #0
-        smlatbne        lr, r5, r1, lr
-
-        rsbs            r8, ip, r4, asl #16
-        addgt           r8, r2, #0
-        rsblt           r8, r2, #0
-        smlabbne        r4, r4, r1, r8
-
-        rsbs            r8, ip, r5, asl #16
-        addgt           r8, r2, #0
-        rsblt           r8, r2, #0
-        smlabbne        r5, r5, r1, r8
+        dequant_t       r9, r4, r1, r2, r9
+        dequant_t       lr, r5, r1, r2, lr
+        dequant_b       r4, r4, r1, r2, r8
+        dequant_b       r5, r5, r1, r2, r8
 
         strh            r4, [r0], #2
         strh            r9, [r0], #2
         strh            r5, [r0], #2
         strh            lr, [r0], #2
 
-        rsbs            r9, ip, r6, asr #16
-        addgt           r9, r2, #0
-        rsblt           r9, r2, #0
-        smlatbne        r9, r6, r1, r9
-
-        rsbs            lr, ip, r7, asr #16
-        addgt           lr, r2, #0
-        rsblt           lr, r2, #0
-        smlatbne        lr, r7, r1, lr
-
-        rsbs            r8, ip, r6, asl #16
-        addgt           r8, r2, #0
-        rsblt           r8, r2, #0
-        smlabbne        r6, r6, r1, r8
-
-        rsbs            r8, ip, r7, asl #16
-        addgt           r8, r2, #0
-        rsblt           r8, r2, #0
-        smlabbne        r7, r7, r1, r8
+        dequant_t       r9, r6, r1, r2, r9
+        dequant_t       lr, r7, r1, r2, lr
+        dequant_b       r6, r6, r1, r2, r8
+        dequant_b       r7, r7, r1, r2, r8
 
         strh            r6, [r0], #2
         strh            r9, [r0], #2
diff --git a/libavcodec/arm/simple_idct_armv5te.S b/libavcodec/arm/simple_idct_armv5te.S
index 27aeca4448..3c4b5c06d1 100644
--- a/libavcodec/arm/simple_idct_armv5te.S
+++ b/libavcodec/arm/simple_idct_armv5te.S
@@ -333,6 +333,20 @@ function idct_col_armv5te
         ldr    pc, [sp], #4
 endfunc
 
+.macro  clip   dst, src:vararg
+        movs   \dst, \src
+        movmi  \dst, #0
+        cmp    \dst, #255
+        movgt  \dst, #255
+.endm
+
+.macro  aclip  dst, src:vararg
+        adds   \dst, \src
+        movmi  \dst, #0
+        cmp    \dst, #255
+        movgt  \dst, #255
+.endm
+
 function idct_col_put_armv5te
         str    lr, [sp, #-4]!
 
@@ -341,27 +355,15 @@ function idct_col_put_armv5te
         ldmfd  sp!, {a3, a4}
         ldr    lr, [sp, #32]
         add    a2, a3, v1
-        movs   a2, a2, asr #20
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        clip   a2, a2, asr #20
         add    ip, a4, v2
-        movs   ip, ip, asr #20
-        movmi  ip, #0
-        cmp    ip, #255
-        movgt  ip, #255
+        clip   ip, ip, asr #20
         orr    a2, a2, ip, lsl #8
         sub    a3, a3, v1
-        movs   a3, a3, asr #20
-        movmi  a3, #0
-        cmp    a3, #255
-        movgt  a3, #255
+        clip   a3, a3, asr #20
         sub    a4, a4, v2
-        movs   a4, a4, asr #20
-        movmi  a4, #0
-        cmp    a4, #255
+        clip   a4, a4, asr #20
         ldr    v1, [sp, #28]
-        movgt  a4, #255
         strh   a2, [v1]
         add    a2, v1, #2
         str    a2, [sp, #28]
@@ -371,79 +373,43 @@ function idct_col_put_armv5te
         strh   a2, [v2, v1]!
 
         sub    a2, a3, v3
-        movs   a2, a2, asr #20
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        clip   a2, a2, asr #20
         sub    ip, a4, v4
-        movs   ip, ip, asr #20
-        movmi  ip, #0
-        cmp    ip, #255
-        movgt  ip, #255
+        clip   ip, ip, asr #20
         orr    a2, a2, ip, lsl #8
         strh   a2, [v1, lr]!
         add    a3, a3, v3
-        movs   a2, a3, asr #20
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        clip   a2, a3, asr #20
         add    a4, a4, v4
-        movs   a4, a4, asr #20
-        movmi  a4, #0
-        cmp    a4, #255
-        movgt  a4, #255
+        clip   a4, a4, asr #20
         orr    a2, a2, a4, lsl #8
         ldmfd  sp!, {a3, a4}
         strh   a2, [v2, -lr]!
 
         add    a2, a3, v5
-        movs   a2, a2, asr #20
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        clip   a2, a2, asr #20
         add    ip, a4, v6
-        movs   ip, ip, asr #20
-        movmi  ip, #0
-        cmp    ip, #255
-        movgt  ip, #255
+        clip   ip, ip, asr #20
         orr    a2, a2, ip, lsl #8
         strh   a2, [v1, lr]!
         sub    a3, a3, v5
-        movs   a2, a3, asr #20
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        clip   a2, a3, asr #20
         sub    a4, a4, v6
-        movs   a4, a4, asr #20
-        movmi  a4, #0
-        cmp    a4, #255
-        movgt  a4, #255
+        clip   a4, a4, asr #20
         orr    a2, a2, a4, lsl #8
         ldmfd  sp!, {a3, a4}
         strh   a2, [v2, -lr]!
 
         add    a2, a3, v7
-        movs   a2, a2, asr #20
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        clip   a2, a2, asr #20
         add    ip, a4, fp
-        movs   ip, ip, asr #20
-        movmi  ip, #0
-        cmp    ip, #255
-        movgt  ip, #255
+        clip   ip, ip, asr #20
         orr    a2, a2, ip, lsl #8
         strh   a2, [v1, lr]
         sub    a3, a3, v7
-        movs   a2, a3, asr #20
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        clip   a2, a3, asr #20
         sub    a4, a4, fp
-        movs   a4, a4, asr #20
-        movmi  a4, #0
-        cmp    a4, #255
-        movgt  a4, #255
+        clip   a4, a4, asr #20
         orr    a2, a2, a4, lsl #8
         strh   a2, [v2, -lr]
 
@@ -460,36 +426,22 @@ function idct_col_add_armv5te
         ldmfd  sp!, {a3, a4}
         ldrh   ip, [lr]
         add    a2, a3, v1
-        mov    a2, a2, asr #20
         sub    a3, a3, v1
         and    v1, ip, #255
-        adds   a2, a2, v1
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        aclip  a2, v1, a2, asr #20
         add    v1, a4, v2
         mov    v1, v1, asr #20
-        adds   v1, v1, ip, lsr #8
-        movmi  v1, #0
-        cmp    v1, #255
-        movgt  v1, #255
+        aclip  v1, v1, ip, lsr #8
         orr    a2, a2, v1, lsl #8
         ldr    v1, [sp, #32]
         sub    a4, a4, v2
         rsb    v2, v1, v1, lsl #3
         ldrh   ip, [v2, lr]!
         strh   a2, [lr]
-        mov    a3, a3, asr #20
         and    a2, ip, #255
-        adds   a3, a3, a2
-        movmi  a3, #0
-        cmp    a3, #255
-        movgt  a3, #255
+        aclip  a3, a2, a3, asr #20
         mov    a4, a4, asr #20
-        adds   a4, a4, ip, lsr #8
-        movmi  a4, #0
-        cmp    a4, #255
-        movgt  a4, #255
+        aclip  a4, a4, ip, lsr #8
         add    a2, lr, #2
         str    a2, [sp, #28]
         orr    a2, a3, a4, lsl #8
@@ -498,102 +450,60 @@ function idct_col_add_armv5te
         ldmfd  sp!, {a3, a4}
         ldrh   ip, [lr, v1]!
         sub    a2, a3, v3
-        mov    a2, a2, asr #20
         add    a3, a3, v3
         and    v3, ip, #255
-        adds   a2, a2, v3
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        aclip  a2, v3, a2, asr #20
         sub    v3, a4, v4
         mov    v3, v3, asr #20
-        adds   v3, v3, ip, lsr #8
-        movmi  v3, #0
-        cmp    v3, #255
-        movgt  v3, #255
+        aclip  v3, v3, ip, lsr #8
         orr    a2, a2, v3, lsl #8
         add    a4, a4, v4
         ldrh   ip, [v2, -v1]!
         strh   a2, [lr]
-        mov    a3, a3, asr #20
         and    a2, ip, #255
-        adds   a3, a3, a2
-        movmi  a3, #0
-        cmp    a3, #255
-        movgt  a3, #255
+        aclip  a3, a2, a3, asr #20
         mov    a4, a4, asr #20
-        adds   a4, a4, ip, lsr #8
-        movmi  a4, #0
-        cmp    a4, #255
-        movgt  a4, #255
+        aclip  a4, a4, ip, lsr #8
         orr    a2, a3, a4, lsl #8
         strh   a2, [v2]
 
         ldmfd  sp!, {a3, a4}
         ldrh   ip, [lr, v1]!
         add    a2, a3, v5
-        mov    a2, a2, asr #20
         sub    a3, a3, v5
         and    v3, ip, #255
-        adds   a2, a2, v3
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        aclip  a2, v3, a2, asr #20
         add    v3, a4, v6
         mov    v3, v3, asr #20
-        adds   v3, v3, ip, lsr #8
-        movmi  v3, #0
-        cmp    v3, #255
-        movgt  v3, #255
+        aclip  v3, v3, ip, lsr #8
         orr    a2, a2, v3, lsl #8
         sub    a4, a4, v6
         ldrh   ip, [v2, -v1]!
         strh   a2, [lr]
-        mov    a3, a3, asr #20
         and    a2, ip, #255
-        adds   a3, a3, a2
-        movmi  a3, #0
-        cmp    a3, #255
-        movgt  a3, #255
+        aclip  a3, a2, a3, asr #20
         mov    a4, a4, asr #20
-        adds   a4, a4, ip, lsr #8
-        movmi  a4, #0
-        cmp    a4, #255
-        movgt  a4, #255
+        aclip  a4, a4, ip, lsr #8
         orr    a2, a3, a4, lsl #8
         strh   a2, [v2]
 
         ldmfd  sp!, {a3, a4}
         ldrh   ip, [lr, v1]!
         add    a2, a3, v7
-        mov    a2, a2, asr #20
         sub    a3, a3, v7
         and    v3, ip, #255
-        adds   a2, a2, v3
-        movmi  a2, #0
-        cmp    a2, #255
-        movgt  a2, #255
+        aclip  a2, v3, a2, asr #20
         add    v3, a4, fp
         mov    v3, v3, asr #20
-        adds   v3, v3, ip, lsr #8
-        movmi  v3, #0
-        cmp    v3, #255
-        movgt  v3, #255
+        aclip  v3, v3, ip, lsr #8
         orr    a2, a2, v3, lsl #8
         sub    a4, a4, fp
         ldrh   ip, [v2, -v1]!
         strh   a2, [lr]
-        mov    a3, a3, asr #20
         and    a2, ip, #255
-        adds   a3, a3, a2
-        movmi  a3, #0
-        cmp    a3, #255
-        movgt  a3, #255
+        aclip  a3, a2, a3, asr #20
         mov    a4, a4, asr #20
-        adds   a4, a4, ip, lsr #8
-        movmi  a4, #0
-        cmp    a4, #255
-        movgt  a4, #255
+        aclip  a4, a4, ip, lsr #8
         orr    a2, a3, a4, lsl #8
         strh   a2, [v2]
 
diff --git a/libavcodec/dsputil.h b/libavcodec/dsputil.h
index f2c36b37e8..0143ed7d3c 100644
--- a/libavcodec/dsputil.h
+++ b/libavcodec/dsputil.h
@@ -505,7 +505,7 @@ typedef struct DSPContext {
 #define BASIS_SHIFT 16
 #define RECON_SHIFT 6
 
-    void (*draw_edges)(uint8_t *buf, int wrap, int width, int height, int w, int sides);
+    void (*draw_edges)(uint8_t *buf, int wrap, int width, int height, int w, int h, int sides);
 #define EDGE_WIDTH 16
 #define EDGE_TOP    1
 #define EDGE_BOTTOM 2
diff --git a/libavcodec/dsputil_template.c b/libavcodec/dsputil_template.c
index 1d73a7d858..58533d2ce9 100644
--- a/libavcodec/dsputil_template.c
+++ b/libavcodec/dsputil_template.c
@@ -79,7 +79,7 @@ static inline void FUNC(copy_block16)(uint8_t *dst, const uint8_t *src, int dstS
 
 /* draw the edges of width 'w' of an image of size width, height */
 //FIXME check that this is ok for mpeg4 interlaced
-static void FUNCC(draw_edges)(uint8_t *p_buf, int p_wrap, int width, int height, int w, int sides)
+static void FUNCC(draw_edges)(uint8_t *p_buf, int p_wrap, int width, int height, int w, int h, int sides)
 {
     pixel *buf = (pixel*)p_buf;
     int wrap = p_wrap / sizeof(pixel);
@@ -106,10 +106,10 @@ static void FUNCC(draw_edges)(uint8_t *p_buf, int p_wrap, int width, int height,
     buf -= w;
     last_line = buf + (height - 1) * wrap;
     if (sides & EDGE_TOP)
-        for(i = 0; i < w; i++)
+        for(i = 0; i < h; i++)
             memcpy(buf - (i + 1) * wrap, buf, (width + w + w) * sizeof(pixel)); // top
     if (sides & EDGE_BOTTOM)
-        for (i = 0; i < w; i++)
+        for (i = 0; i < h; i++)
             memcpy(last_line + (i + 1) * wrap, last_line, (width + w + w) * sizeof(pixel)); // bottom
 }
 
diff --git a/libavcodec/h264.c b/libavcodec/h264.c
index 5c812144c9..0919c12b92 100644
--- a/libavcodec/h264.c
+++ b/libavcodec/h264.c
@@ -451,12 +451,13 @@ static inline void mc_dir_part(H264Context *h, Picture *pic, int n, int square,
                            uint8_t *dest_y, uint8_t *dest_cb, uint8_t *dest_cr,
                            int src_x_offset, int src_y_offset,
                            qpel_mc_func *qpix_op, h264_chroma_mc_func chroma_op,
-                           int pixel_shift){
+                           int pixel_shift, int chroma444){
     MpegEncContext * const s = &h->s;
     const int mx= h->mv_cache[list][ scan8[n] ][0] + src_x_offset*8;
     int my=       h->mv_cache[list][ scan8[n] ][1] + src_y_offset*8;
     const int luma_xy= (mx&3) + ((my&3)<<2);
-    uint8_t * src_y = pic->data[0] + ((mx>>2) << pixel_shift) + (my>>2)*h->mb_linesize;
+    int offset = ((mx>>2) << pixel_shift) + (my>>2)*h->mb_linesize;
+    uint8_t * src_y = pic->data[0] + offset;
     uint8_t * src_cb, * src_cr;
     int extra_width= h->emu_edge_width;
     int extra_height= h->emu_edge_height;
@@ -485,6 +486,31 @@ static inline void mc_dir_part(H264Context *h, Picture *pic, int n, int square,
 
     if(CONFIG_GRAY && s->flags&CODEC_FLAG_GRAY) return;
 
+    if(chroma444){
+        src_cb = pic->data[1] + offset;
+        if(emu){
+            s->dsp.emulated_edge_mc(s->edge_emu_buffer, src_cb - (2 << pixel_shift) - 2*h->mb_linesize, h->mb_linesize,
+                                    16+5, 16+5/*FIXME*/, full_mx-2, full_my-2, pic_width, pic_height);
+            src_cb= s->edge_emu_buffer + (2 << pixel_shift) + 2*h->mb_linesize;
+        }
+        qpix_op[luma_xy](dest_cb, src_cb, h->mb_linesize); //FIXME try variable height perhaps?
+        if(!square){
+            qpix_op[luma_xy](dest_cb + delta, src_cb + delta, h->mb_linesize);
+        }
+
+        src_cr = pic->data[2] + offset;
+        if(emu){
+            s->dsp.emulated_edge_mc(s->edge_emu_buffer, src_cr - (2 << pixel_shift) - 2*h->mb_linesize, h->mb_linesize,
+                                    16+5, 16+5/*FIXME*/, full_mx-2, full_my-2, pic_width, pic_height);
+            src_cr= s->edge_emu_buffer + (2 << pixel_shift) + 2*h->mb_linesize;
+        }
+        qpix_op[luma_xy](dest_cr, src_cr, h->mb_linesize); //FIXME try variable height perhaps?
+        if(!square){
+            qpix_op[luma_xy](dest_cr + delta, src_cr + delta, h->mb_linesize);
+        }
+        return;
+    }
+
     if(MB_FIELD){
         // chroma offset when predicting from a field of opposite parity
         my += 2 * ((s->mb_y & 1) - (pic->reference - 1));
@@ -511,14 +537,19 @@ static inline void mc_part_std(H264Context *h, int n, int square, int chroma_hei
                            int x_offset, int y_offset,
                            qpel_mc_func *qpix_put, h264_chroma_mc_func chroma_put,
                            qpel_mc_func *qpix_avg, h264_chroma_mc_func chroma_avg,
-                           int list0, int list1, int pixel_shift){
+                           int list0, int list1, int pixel_shift, int chroma444){
     MpegEncContext * const s = &h->s;
     qpel_mc_func *qpix_op=  qpix_put;
     h264_chroma_mc_func chroma_op= chroma_put;
 
-    dest_y  += (2*x_offset << pixel_shift) + 2*y_offset*h->  mb_linesize;
-    dest_cb += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
-    dest_cr += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
+    dest_y  += (2*x_offset << pixel_shift) + 2*y_offset*h->mb_linesize;
+    if(chroma444){
+        dest_cb += (2*x_offset << pixel_shift) + 2*y_offset*h->mb_linesize;
+        dest_cr += (2*x_offset << pixel_shift) + 2*y_offset*h->mb_linesize;
+    }else{
+        dest_cb += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
+        dest_cr += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
+    }
     x_offset += 8*s->mb_x;
     y_offset += 8*(s->mb_y >> MB_FIELD);
 
@@ -526,7 +557,7 @@ static inline void mc_part_std(H264Context *h, int n, int square, int chroma_hei
         Picture *ref= &h->ref_list[0][ h->ref_cache[0][ scan8[n] ] ];
         mc_dir_part(h, ref, n, square, chroma_height, delta, 0,
                            dest_y, dest_cb, dest_cr, x_offset, y_offset,
-                           qpix_op, chroma_op, pixel_shift);
+                           qpix_op, chroma_op, pixel_shift, chroma444);
 
         qpix_op=  qpix_avg;
         chroma_op= chroma_avg;
@@ -536,7 +567,7 @@ static inline void mc_part_std(H264Context *h, int n, int square, int chroma_hei
         Picture *ref= &h->ref_list[1][ h->ref_cache[1][ scan8[n] ] ];
         mc_dir_part(h, ref, n, square, chroma_height, delta, 1,
                            dest_y, dest_cb, dest_cr, x_offset, y_offset,
-                           qpix_op, chroma_op, pixel_shift);
+                           qpix_op, chroma_op, pixel_shift, chroma444);
     }
 }
 
@@ -546,12 +577,19 @@ static inline void mc_part_weighted(H264Context *h, int n, int square, int chrom
                            qpel_mc_func *qpix_put, h264_chroma_mc_func chroma_put,
                            h264_weight_func luma_weight_op, h264_weight_func chroma_weight_op,
                            h264_biweight_func luma_weight_avg, h264_biweight_func chroma_weight_avg,
-                           int list0, int list1, int pixel_shift){
+                           int list0, int list1, int pixel_shift, int chroma444){
     MpegEncContext * const s = &h->s;
 
-    dest_y  += (2*x_offset << pixel_shift) + 2*y_offset*h->  mb_linesize;
-    dest_cb += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
-    dest_cr += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
+    dest_y += (2*x_offset << pixel_shift) + 2*y_offset*h->mb_linesize;
+    if(chroma444){
+        chroma_weight_avg = luma_weight_avg;
+        chroma_weight_op = luma_weight_op;
+        dest_cb += (2*x_offset << pixel_shift) + 2*y_offset*h->mb_linesize;
+        dest_cr += (2*x_offset << pixel_shift) + 2*y_offset*h->mb_linesize;
+    }else{
+        dest_cb += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
+        dest_cr += (  x_offset << pixel_shift) +   y_offset*h->mb_uvlinesize;
+    }
     x_offset += 8*s->mb_x;
     y_offset += 8*(s->mb_y >> MB_FIELD);
 
@@ -559,17 +597,17 @@ static inline void mc_part_weighted(H264Context *h, int n, int square, int chrom
         /* don't optimize for luma-only case, since B-frames usually
          * use implicit weights => chroma too. */
         uint8_t *tmp_cb = s->obmc_scratchpad;
-        uint8_t *tmp_cr = s->obmc_scratchpad + (8 << pixel_shift);
-        uint8_t *tmp_y  = s->obmc_scratchpad + 8*h->mb_uvlinesize;
+        uint8_t *tmp_cr = s->obmc_scratchpad + (16 << pixel_shift);
+        uint8_t *tmp_y  = s->obmc_scratchpad + 16*h->mb_uvlinesize;
         int refn0 = h->ref_cache[0][ scan8[n] ];
         int refn1 = h->ref_cache[1][ scan8[n] ];
 
         mc_dir_part(h, &h->ref_list[0][refn0], n, square, chroma_height, delta, 0,
                     dest_y, dest_cb, dest_cr,
-                    x_offset, y_offset, qpix_put, chroma_put, pixel_shift);
+                    x_offset, y_offset, qpix_put, chroma_put, pixel_shift, chroma444);
         mc_dir_part(h, &h->ref_list[1][refn1], n, square, chroma_height, delta, 1,
                     tmp_y, tmp_cb, tmp_cr,
-                    x_offset, y_offset, qpix_put, chroma_put, pixel_shift);
+                    x_offset, y_offset, qpix_put, chroma_put, pixel_shift, chroma444);
 
         if(h->use_weight == 2){
             int weight0 = h->implicit_weight[refn0][refn1][s->mb_y&1];
@@ -594,7 +632,7 @@ static inline void mc_part_weighted(H264Context *h, int n, int square, int chrom
         Picture *ref= &h->ref_list[list][refn];
         mc_dir_part(h, ref, n, square, chroma_height, delta, list,
                     dest_y, dest_cb, dest_cr, x_offset, y_offset,
-                    qpix_put, chroma_put, pixel_shift);
+                    qpix_put, chroma_put, pixel_shift, chroma444);
 
         luma_weight_op(dest_y, h->mb_linesize, h->luma_log2_weight_denom,
                        h->luma_weight[refn][list][0], h->luma_weight[refn][list][1]);
@@ -613,21 +651,21 @@ static inline void mc_part(H264Context *h, int n, int square, int chroma_height,
                            qpel_mc_func *qpix_put, h264_chroma_mc_func chroma_put,
                            qpel_mc_func *qpix_avg, h264_chroma_mc_func chroma_avg,
                            h264_weight_func *weight_op, h264_biweight_func *weight_avg,
-                           int list0, int list1, int pixel_shift){
+                           int list0, int list1, int pixel_shift, int chroma444){
     if((h->use_weight==2 && list0 && list1
         && (h->implicit_weight[ h->ref_cache[0][scan8[n]] ][ h->ref_cache[1][scan8[n]] ][h->s.mb_y&1] != 32))
        || h->use_weight==1)
         mc_part_weighted(h, n, square, chroma_height, delta, dest_y, dest_cb, dest_cr,
                          x_offset, y_offset, qpix_put, chroma_put,
                          weight_op[0], weight_op[3], weight_avg[0],
-                         weight_avg[3], list0, list1, pixel_shift);
+                         weight_avg[3], list0, list1, pixel_shift, chroma444);
     else
         mc_part_std(h, n, square, chroma_height, delta, dest_y, dest_cb, dest_cr,
                     x_offset, y_offset, qpix_put, chroma_put, qpix_avg,
-                    chroma_avg, list0, list1, pixel_shift);
+                    chroma_avg, list0, list1, pixel_shift, chroma444);
 }
 
-static inline void prefetch_motion(H264Context *h, int list, int pixel_shift){
+static inline void prefetch_motion(H264Context *h, int list, int pixel_shift, int chroma444){
     /* fetch pixels for estimated mv 4 macroblocks ahead
      * optimized for 64byte cache lines */
     MpegEncContext * const s = &h->s;
@@ -638,8 +676,13 @@ static inline void prefetch_motion(H264Context *h, int list, int pixel_shift){
         uint8_t **src= h->ref_list[list][refn].data;
         int off= ((mx+64)<<h->pixel_shift) + (my + (s->mb_x&3)*4)*h->mb_linesize;
         s->dsp.prefetch(src[0]+off, s->linesize, 4);
-        off= (((mx>>1)+64)<<h->pixel_shift) + ((my>>1) + (s->mb_x&7))*s->uvlinesize;
-        s->dsp.prefetch(src[1]+off, src[2]-src[1], 2);
+        if(chroma444){
+            s->dsp.prefetch(src[1]+off, s->linesize, 4);
+            s->dsp.prefetch(src[2]+off, s->linesize, 4);
+        }else{
+            off= (((mx>>1)+64)<<pixel_shift) + ((my>>1) + (s->mb_x&7))*s->uvlinesize;
+            s->dsp.prefetch(src[1]+off, src[2]-src[1], 2);
+        }
     }
 }
 
@@ -647,7 +690,7 @@ static av_always_inline void hl_motion(H264Context *h, uint8_t *dest_y, uint8_t
                       qpel_mc_func (*qpix_put)[16], h264_chroma_mc_func (*chroma_put),
                       qpel_mc_func (*qpix_avg)[16], h264_chroma_mc_func (*chroma_avg),
                       h264_weight_func *weight_op, h264_biweight_func *weight_avg,
-                      int pixel_shift){
+                      int pixel_shift, int chroma444){
     MpegEncContext * const s = &h->s;
     const int mb_xy= h->mb_xy;
     const int mb_type= s->current_picture.mb_type[mb_xy];
@@ -656,36 +699,36 @@ static av_always_inline void hl_motion(H264Context *h, uint8_t *dest_y, uint8_t
 
     if(HAVE_PTHREADS && (s->avctx->active_thread_type & FF_THREAD_FRAME))
         await_references(h);
-    prefetch_motion(h, 0, pixel_shift);
+    prefetch_motion(h, 0, pixel_shift, chroma444);
 
     if(IS_16X16(mb_type)){
         mc_part(h, 0, 1, 8, 0, dest_y, dest_cb, dest_cr, 0, 0,
                 qpix_put[0], chroma_put[0], qpix_avg[0], chroma_avg[0],
                 weight_op, weight_avg,
                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1),
-                pixel_shift);
+                pixel_shift, chroma444);
     }else if(IS_16X8(mb_type)){
         mc_part(h, 0, 0, 4, 8 << pixel_shift, dest_y, dest_cb, dest_cr, 0, 0,
                 qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
                 &weight_op[1], &weight_avg[1],
                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1),
-                pixel_shift);
+                pixel_shift, chroma444);
         mc_part(h, 8, 0, 4, 8 << pixel_shift, dest_y, dest_cb, dest_cr, 0, 4,
                 qpix_put[1], chroma_put[0], qpix_avg[1], chroma_avg[0],
                 &weight_op[1], &weight_avg[1],
                 IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1),
-                pixel_shift);
+                pixel_shift, chroma444);
     }else if(IS_8X16(mb_type)){
         mc_part(h, 0, 0, 8, 8*h->mb_linesize, dest_y, dest_cb, dest_cr, 0, 0,
                 qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
                 &weight_op[2], &weight_avg[2],
                 IS_DIR(mb_type, 0, 0), IS_DIR(mb_type, 0, 1),
-                pixel_shift);
+                pixel_shift, chroma444);
         mc_part(h, 4, 0, 8, 8*h->mb_linesize, dest_y, dest_cb, dest_cr, 4, 0,
                 qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
                 &weight_op[2], &weight_avg[2],
                 IS_DIR(mb_type, 1, 0), IS_DIR(mb_type, 1, 1),
-                pixel_shift);
+                pixel_shift, chroma444);
     }else{
         int i;
 
@@ -702,29 +745,29 @@ static av_always_inline void hl_motion(H264Context *h, uint8_t *dest_y, uint8_t
                     qpix_put[1], chroma_put[1], qpix_avg[1], chroma_avg[1],
                     &weight_op[3], &weight_avg[3],
                     IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1),
-                    pixel_shift);
+                    pixel_shift, chroma444);
             }else if(IS_SUB_8X4(sub_mb_type)){
                 mc_part(h, n  , 0, 2, 4 << pixel_shift, dest_y, dest_cb, dest_cr, x_offset, y_offset,
                     qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
                     &weight_op[4], &weight_avg[4],
                     IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1),
-                    pixel_shift);
+                    pixel_shift, chroma444);
                 mc_part(h, n+2, 0, 2, 4 << pixel_shift, dest_y, dest_cb, dest_cr, x_offset, y_offset+2,
                     qpix_put[2], chroma_put[1], qpix_avg[2], chroma_avg[1],
                     &weight_op[4], &weight_avg[4],
                     IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1),
-                    pixel_shift);
+                    pixel_shift, chroma444);
             }else if(IS_SUB_4X8(sub_mb_type)){
                 mc_part(h, n  , 0, 4, 4*h->mb_linesize, dest_y, dest_cb, dest_cr, x_offset, y_offset,
                     qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
                     &weight_op[5], &weight_avg[5],
                     IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1),
-                    pixel_shift);
+                    pixel_shift, chroma444);
                 mc_part(h, n+1, 0, 4, 4*h->mb_linesize, dest_y, dest_cb, dest_cr, x_offset+2, y_offset,
                     qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
                     &weight_op[5], &weight_avg[5],
                     IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1),
-                    pixel_shift);
+                    pixel_shift, chroma444);
             }else{
                 int j;
                 assert(IS_SUB_4X4(sub_mb_type));
@@ -735,13 +778,13 @@ static av_always_inline void hl_motion(H264Context *h, uint8_t *dest_y, uint8_t
                         qpix_put[2], chroma_put[2], qpix_avg[2], chroma_avg[2],
                         &weight_op[6], &weight_avg[6],
                         IS_DIR(sub_mb_type, 0, 0), IS_DIR(sub_mb_type, 0, 1),
-                        pixel_shift);
+                        pixel_shift, chroma444);
                 }
             }
         }
     }
 
-    prefetch_motion(h, 1, pixel_shift);
+    prefetch_motion(h, 1, pixel_shift, chroma444);
 }
 
 #define hl_motion_fn(sh, bits) \
@@ -753,10 +796,11 @@ static av_always_inline void hl_motion_ ## bits(H264Context *h, \
                                        qpel_mc_func (*qpix_avg)[16], \
                                        h264_chroma_mc_func (*chroma_avg), \
                                        h264_weight_func *weight_op, \
-                                       h264_biweight_func *weight_avg) \
+                                       h264_biweight_func *weight_avg, \
+                                       int chroma444) \
 { \
     hl_motion(h, dest_y, dest_cb, dest_cr, qpix_put, chroma_put, \
-              qpix_avg, chroma_avg, weight_op, weight_avg, sh); \
+              qpix_avg, chroma_avg, weight_op, weight_avg, sh, chroma444); \
 }
 hl_motion_fn(0, 8);
 hl_motion_fn(1, 16);
@@ -796,16 +840,19 @@ static void free_tables(H264Context *h, int free_rbsp){
 }
 
 static void init_dequant8_coeff_table(H264Context *h){
-    int i,q,x;
+    int i,j,q,x;
     const int max_qp = 51 + 6*(h->sps.bit_depth_luma-8);
-    h->dequant8_coeff[0] = h->dequant8_buffer[0];
-    h->dequant8_coeff[1] = h->dequant8_buffer[1];
 
-    for(i=0; i<2; i++ ){
-        if(i && !memcmp(h->pps.scaling_matrix8[0], h->pps.scaling_matrix8[1], 64*sizeof(uint8_t))){
-            h->dequant8_coeff[1] = h->dequant8_buffer[0];
-            break;
+    for(i=0; i<6; i++ ){
+        h->dequant8_coeff[i] = h->dequant8_buffer[i];
+        for(j=0; j<i; j++){
+            if(!memcmp(h->pps.scaling_matrix8[j], h->pps.scaling_matrix8[i], 64*sizeof(uint8_t))){
+                h->dequant8_coeff[i] = h->dequant8_buffer[j];
+                break;
+            }
         }
+        if(j<i)
+            continue;
 
         for(q=0; q<max_qp+1; q++){
             int shift = div6[q];
@@ -853,7 +900,7 @@ static void init_dequant_tables(H264Context *h){
             for(x=0; x<16; x++)
                 h->dequant4_coeff[i][0][x] = 1<<6;
         if(h->pps.transform_8x8_mode)
-            for(i=0; i<2; i++)
+            for(i=0; i<6; i++)
                 for(x=0; x<64; x++)
                     h->dequant8_coeff[i][0][x] = 1<<6;
     }
@@ -868,7 +915,7 @@ int ff_h264_alloc_tables(H264Context *h){
 
     FF_ALLOCZ_OR_GOTO(h->s.avctx, h->intra4x4_pred_mode, row_mb_num * 8  * sizeof(uint8_t), fail)
 
-    FF_ALLOCZ_OR_GOTO(h->s.avctx, h->non_zero_count    , big_mb_num * 32 * sizeof(uint8_t), fail)
+    FF_ALLOCZ_OR_GOTO(h->s.avctx, h->non_zero_count    , big_mb_num * 48 * sizeof(uint8_t), fail)
     FF_ALLOCZ_OR_GOTO(h->s.avctx, h->slice_table_base  , (big_mb_num+s->mb_stride) * sizeof(*h->slice_table_base), fail)
     FF_ALLOCZ_OR_GOTO(h->s.avctx, h->cbp_table, big_mb_num * sizeof(uint16_t), fail)
 
@@ -930,8 +977,8 @@ static void clone_tables(H264Context *dst, H264Context *src, int i){
  * Allocate buffers which are not shared amongst multiple threads.
  */
 static int context_init(H264Context *h){
-    FF_ALLOCZ_OR_GOTO(h->s.avctx, h->top_borders[0], h->s.mb_width * (16+8+8) * sizeof(uint8_t)*2, fail)
-    FF_ALLOCZ_OR_GOTO(h->s.avctx, h->top_borders[1], h->s.mb_width * (16+8+8) * sizeof(uint8_t)*2, fail)
+    FF_ALLOCZ_OR_GOTO(h->s.avctx, h->top_borders[0], h->s.mb_width * 16*3 * sizeof(uint8_t)*2, fail)
+    FF_ALLOCZ_OR_GOTO(h->s.avctx, h->top_borders[1], h->s.mb_width * 16*3 * sizeof(uint8_t)*2, fail)
 
     h->ref_cache[0][scan8[5 ]+1] = h->ref_cache[0][scan8[7 ]+1] = h->ref_cache[0][scan8[13]+1] =
     h->ref_cache[1][scan8[5 ]+1] = h->ref_cache[1][scan8[7 ]+1] = h->ref_cache[1][scan8[13]+1] = PART_NOT_AVAILABLE;
@@ -1130,9 +1177,10 @@ static int decode_update_thread_context(AVCodecContext *dst, const AVCodecContex
 
         // frame_start may not be called for the next thread (if it's decoding a bottom field)
         // so this has to be allocated here
-        h->s.obmc_scratchpad = av_malloc(16*2*s->linesize + 8*2*s->uvlinesize);
+        h->s.obmc_scratchpad = av_malloc(16*6*s->linesize);
 
         s->dsp.clear_blocks(h->mb);
+        s->dsp.clear_blocks(h->mb+(24*16<<h->pixel_shift));
     }
 
     //extradata/NAL handling
@@ -1151,7 +1199,7 @@ static int decode_update_thread_context(AVCodecContext *dst, const AVCodecContex
     for(i=0; i<6; i++)
         h->dequant4_coeff[i] = h->dequant4_buffer[0] + (h1->dequant4_coeff[i] - h1->dequant4_buffer[0]);
 
-    for(i=0; i<2; i++)
+    for(i=0; i<6; i++)
         h->dequant8_coeff[i] = h->dequant8_buffer[0] + (h1->dequant8_coeff[i] - h1->dequant8_buffer[0]);
 
     h->dequant_coeff_pps = h1->dequant_coeff_pps;
@@ -1206,20 +1254,20 @@ int ff_h264_frame_start(H264Context *h){
 
     for(i=0; i<16; i++){
         h->block_offset[i]= (4*((scan8[i] - scan8[0])&7) << pixel_shift) + 4*s->linesize*((scan8[i] - scan8[0])>>3);
-        h->block_offset[24+i]= (4*((scan8[i] - scan8[0])&7) << pixel_shift) + 8*s->linesize*((scan8[i] - scan8[0])>>3);
+        h->block_offset[48+i]= (4*((scan8[i] - scan8[0])&7) << pixel_shift) + 8*s->linesize*((scan8[i] - scan8[0])>>3);
     }
-    for(i=0; i<4; i++){
+    for(i=0; i<16; i++){
         h->block_offset[16+i]=
-        h->block_offset[20+i]= (4*((scan8[i] - scan8[0])&7) << pixel_shift) + 4*s->uvlinesize*((scan8[i] - scan8[0])>>3);
-        h->block_offset[24+16+i]=
-        h->block_offset[24+20+i]= (4*((scan8[i] - scan8[0])&7) << pixel_shift) + 8*s->uvlinesize*((scan8[i] - scan8[0])>>3);
+        h->block_offset[32+i]= (4*((scan8[i] - scan8[0])&7) << pixel_shift) + 4*s->uvlinesize*((scan8[i] - scan8[0])>>3);
+        h->block_offset[48+16+i]=
+        h->block_offset[48+32+i]= (4*((scan8[i] - scan8[0])&7) << pixel_shift) + 8*s->uvlinesize*((scan8[i] - scan8[0])>>3);
     }
 
     /* can't be in alloc_tables because linesize isn't known there.
      * FIXME: redo bipred weight to not require extra buffer? */
     for(i = 0; i < thread_count; i++)
         if(h->thread_context[i] && !h->thread_context[i]->s.obmc_scratchpad)
-            h->thread_context[i]->s.obmc_scratchpad = av_malloc(16*2*s->linesize + 8*2*s->uvlinesize);
+            h->thread_context[i]->s.obmc_scratchpad = av_malloc(16*6*s->linesize);
 
     /* some macroblocks can be accessed before they're available in case of lost slices, mbaff or threading*/
     memset(h->slice_table, -1, (s->mb_height*s->mb_stride-1) * sizeof(*h->slice_table));
@@ -1404,7 +1452,7 @@ static void decode_postinit(H264Context *h, int setup_finished){
         ff_thread_finish_setup(s->avctx);
 }
 
-static inline void backup_mb_border(H264Context *h, uint8_t *src_y, uint8_t *src_cb, uint8_t *src_cr, int linesize, int uvlinesize, int simple){
+static inline void backup_mb_border(H264Context *h, uint8_t *src_y, uint8_t *src_cb, uint8_t *src_cr, int linesize, int uvlinesize, int chroma444, int simple){
     MpegEncContext * const s = &h->s;
     uint8_t *top_border;
     int top_idx = 1;
@@ -1422,12 +1470,24 @@ static inline void backup_mb_border(H264Context *h, uint8_t *src_y, uint8_t *src
                 if (pixel_shift)
                     AV_COPY128(top_border+16, src_y+15*linesize+16);
                 if(simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)){
-                    if (pixel_shift) {
-                        AV_COPY128(top_border+32, src_cb+7*uvlinesize);
-                        AV_COPY128(top_border+48, src_cr+7*uvlinesize);
+                    if(chroma444){
+                        if (pixel_shift){
+                            AV_COPY128(top_border+32, src_cb + 15*uvlinesize);
+                            AV_COPY128(top_border+48, src_cb + 15*uvlinesize+16);
+                            AV_COPY128(top_border+64, src_cr + 15*uvlinesize);
+                            AV_COPY128(top_border+80, src_cr + 15*uvlinesize+16);
+                        } else {
+                            AV_COPY128(top_border+16, src_cb + 15*uvlinesize);
+                            AV_COPY128(top_border+32, src_cr + 15*uvlinesize);
+                        }
                     } else {
-                    AV_COPY64(top_border+16, src_cb+7*uvlinesize);
-                    AV_COPY64(top_border+24, src_cr+7*uvlinesize);
+                        if (pixel_shift) {
+                            AV_COPY128(top_border+32, src_cb+7*uvlinesize);
+                            AV_COPY128(top_border+48, src_cr+7*uvlinesize);
+                        } else {
+                            AV_COPY64(top_border+16, src_cb+7*uvlinesize);
+                            AV_COPY64(top_border+24, src_cr+7*uvlinesize);
+                        }
                     }
                 }
             }
@@ -1445,12 +1505,24 @@ static inline void backup_mb_border(H264Context *h, uint8_t *src_y, uint8_t *src
         AV_COPY128(top_border+16, src_y+16*linesize+16);
 
     if(simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)){
-        if (pixel_shift) {
-            AV_COPY128(top_border+32, src_cb+8*uvlinesize);
-            AV_COPY128(top_border+48, src_cr+8*uvlinesize);
+        if(chroma444){
+            if (pixel_shift){
+                AV_COPY128(top_border+32, src_cb + 16*linesize);
+                AV_COPY128(top_border+48, src_cb + 16*linesize+16);
+                AV_COPY128(top_border+64, src_cr + 16*linesize);
+                AV_COPY128(top_border+80, src_cr + 16*linesize+16);
+            } else {
+                AV_COPY128(top_border+16, src_cb + 16*linesize);
+                AV_COPY128(top_border+32, src_cr + 16*linesize);
+            }
         } else {
-        AV_COPY64(top_border+16, src_cb+8*uvlinesize);
-        AV_COPY64(top_border+24, src_cr+8*uvlinesize);
+            if (pixel_shift) {
+                AV_COPY128(top_border+32, src_cb+8*uvlinesize);
+                AV_COPY128(top_border+48, src_cr+8*uvlinesize);
+            } else {
+                AV_COPY64(top_border+16, src_cb+8*uvlinesize);
+                AV_COPY64(top_border+24, src_cr+8*uvlinesize);
+            }
         }
     }
 }
@@ -1458,7 +1530,8 @@ static inline void backup_mb_border(H264Context *h, uint8_t *src_y, uint8_t *src
 static inline void xchg_mb_border(H264Context *h, uint8_t *src_y,
                                   uint8_t *src_cb, uint8_t *src_cr,
                                   int linesize, int uvlinesize,
-                                  int xchg, int simple, int pixel_shift){
+                                  int xchg, int chroma444,
+                                  int simple, int pixel_shift){
     MpegEncContext * const s = &h->s;
     int deblock_topleft;
     int deblock_top;
@@ -1513,13 +1586,28 @@ else      AV_COPY64(b,a);
         }
     }
     if(simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)){
-        if(deblock_top){
+        if(chroma444){
             if(deblock_topleft){
-                XCHG(top_border_m1 + (16 << pixel_shift), src_cb - (7 << pixel_shift), 1);
-                XCHG(top_border_m1 + (24 << pixel_shift), src_cr - (7 << pixel_shift), 1);
+                XCHG(top_border_m1 + (24 << pixel_shift), src_cb - (7 << pixel_shift), 1);
+                XCHG(top_border_m1 + (40 << pixel_shift), src_cr - (7 << pixel_shift), 1);
+            }
+            XCHG(top_border + (16 << pixel_shift), src_cb + (1 << pixel_shift), xchg);
+            XCHG(top_border + (24 << pixel_shift), src_cb + (9 << pixel_shift), 1);
+            XCHG(top_border + (32 << pixel_shift), src_cr + (1 << pixel_shift), xchg);
+            XCHG(top_border + (40 << pixel_shift), src_cr + (9 << pixel_shift), 1);
+            if(s->mb_x+1 < s->mb_width){
+                XCHG(h->top_borders[top_idx][s->mb_x+1] + (16 << pixel_shift), src_cb + (17 << pixel_shift), 1);
+                XCHG(h->top_borders[top_idx][s->mb_x+1] + (32 << pixel_shift), src_cr + (17 << pixel_shift), 1);
+            }
+        } else {
+            if(deblock_top){
+                if(deblock_topleft){
+                    XCHG(top_border_m1 + (16 << pixel_shift), src_cb - (7 << pixel_shift), 1);
+                    XCHG(top_border_m1 + (24 << pixel_shift), src_cr - (7 << pixel_shift), 1);
+                }
+                XCHG(top_border + (16 << pixel_shift), src_cb+1+pixel_shift, 1);
+                XCHG(top_border + (24 << pixel_shift), src_cr+1+pixel_shift, 1);
             }
-            XCHG(top_border + (16 << pixel_shift), src_cb+1+pixel_shift, 1);
-            XCHG(top_border + (24 << pixel_shift), src_cr+1+pixel_shift, 1);
         }
     }
 }
@@ -1538,6 +1626,159 @@ static av_always_inline void dctcoef_set(DCTELEM *mb, int high_bit_depth, int in
         AV_WN16A(mb + index, value);
 }
 
+static av_always_inline void hl_decode_mb_predict_luma(H264Context *h, int mb_type, int is_h264, int simple, int transform_bypass,
+                                                       int pixel_shift, int *block_offset, int linesize, uint8_t *dest_y, int p)
+{
+    MpegEncContext * const s = &h->s;
+    void (*idct_add)(uint8_t *dst, DCTELEM *block, int stride);
+    void (*idct_dc_add)(uint8_t *dst, DCTELEM *block, int stride);
+    int i;
+    int qscale = p == 0 ? s->qscale : h->chroma_qp[p-1];
+    block_offset += 16*p;
+    if(IS_INTRA4x4(mb_type)){
+        if(simple || !s->encoding){
+            if(IS_8x8DCT(mb_type)){
+                if(transform_bypass){
+                    idct_dc_add =
+                    idct_add    = s->dsp.add_pixels8;
+                }else{
+                    idct_dc_add = h->h264dsp.h264_idct8_dc_add;
+                    idct_add    = h->h264dsp.h264_idct8_add;
+                }
+                for(i=0; i<16; i+=4){
+                    uint8_t * const ptr= dest_y + block_offset[i];
+                    const int dir= h->intra4x4_pred_mode_cache[ scan8[i] ];
+                    if(transform_bypass && h->sps.profile_idc==244 && dir<=1){
+                        h->hpc.pred8x8l_add[dir](ptr, h->mb + (i*16+p*256 << pixel_shift), linesize);
+                    }else{
+                        const int nnz = h->non_zero_count_cache[ scan8[i+p*16] ];
+                        h->hpc.pred8x8l[ dir ](ptr, (h->topleft_samples_available<<i)&0x8000,
+                                                    (h->topright_samples_available<<i)&0x4000, linesize);
+                        if(nnz){
+                            if(nnz == 1 && dctcoef_get(h->mb, pixel_shift, i*16+p*256))
+                                idct_dc_add(ptr, h->mb + (i*16+p*256 << pixel_shift), linesize);
+                            else
+                                idct_add   (ptr, h->mb + (i*16+p*256 << pixel_shift), linesize);
+                        }
+                    }
+                }
+            }else{
+                if(transform_bypass){
+                    idct_dc_add =
+                    idct_add    = s->dsp.add_pixels4;
+                }else{
+                    idct_dc_add = h->h264dsp.h264_idct_dc_add;
+                    idct_add    = h->h264dsp.h264_idct_add;
+                }
+                for(i=0; i<16; i++){
+                    uint8_t * const ptr= dest_y + block_offset[i];
+                    const int dir= h->intra4x4_pred_mode_cache[ scan8[i] ];
+
+                    if(transform_bypass && h->sps.profile_idc==244 && dir<=1){
+                        h->hpc.pred4x4_add[dir](ptr, h->mb + (i*16+p*256 << pixel_shift), linesize);
+                    }else{
+                        uint8_t *topright;
+                        int nnz, tr;
+                        uint64_t tr_high;
+                        if(dir == DIAG_DOWN_LEFT_PRED || dir == VERT_LEFT_PRED){
+                            const int topright_avail= (h->topright_samples_available<<i)&0x8000;
+                            assert(mb_y || linesize <= block_offset[i]);
+                            if(!topright_avail){
+                                if (pixel_shift) {
+                                    tr_high= ((uint16_t*)ptr)[3 - linesize/2]*0x0001000100010001ULL;
+                                    topright= (uint8_t*) &tr_high;
+                                } else {
+                                    tr= ptr[3 - linesize]*0x01010101;
+                                    topright= (uint8_t*) &tr;
+                                }
+                            }else
+                                topright= ptr + (4 << pixel_shift) - linesize;
+                        }else
+                            topright= NULL;
+
+                        h->hpc.pred4x4[ dir ](ptr, topright, linesize);
+                        nnz = h->non_zero_count_cache[ scan8[i+p*16] ];
+                        if(nnz){
+                            if(is_h264){
+                                if(nnz == 1 && dctcoef_get(h->mb, pixel_shift, i*16+p*256))
+                                    idct_dc_add(ptr, h->mb + (i*16+p*256 << pixel_shift), linesize);
+                                else
+                                    idct_add   (ptr, h->mb + (i*16+p*256 << pixel_shift), linesize);
+                            }else
+                                ff_svq3_add_idct_c(ptr, h->mb + i*16+p*256, linesize, qscale, 0);
+                        }
+                    }
+                }
+            }
+        }
+    }else{
+        h->hpc.pred16x16[ h->intra16x16_pred_mode ](dest_y , linesize);
+        if(is_h264){
+            if(h->non_zero_count_cache[ scan8[LUMA_DC_BLOCK_INDEX+p] ]){
+                if(!transform_bypass)
+                    h->h264dsp.h264_luma_dc_dequant_idct(h->mb+(p*256 << pixel_shift), h->mb_luma_dc[p], h->dequant4_coeff[p][qscale][0]);
+                else{
+                    static const uint8_t dc_mapping[16] = { 0*16, 1*16, 4*16, 5*16, 2*16, 3*16, 6*16, 7*16,
+                                                            8*16, 9*16,12*16,13*16,10*16,11*16,14*16,15*16};
+                    for(i = 0; i < 16; i++)
+                        dctcoef_set(h->mb+p*256, pixel_shift, dc_mapping[i], dctcoef_get(h->mb_luma_dc[p], pixel_shift, i));
+                }
+            }
+        }else
+            ff_svq3_luma_dc_dequant_idct_c(h->mb+p*256, h->mb_luma_dc[p], qscale);
+    }
+}
+
+static av_always_inline void hl_decode_mb_idct_luma(H264Context *h, int mb_type, int is_h264, int simple, int transform_bypass,
+                                                    int pixel_shift, int *block_offset, int linesize, uint8_t *dest_y, int p)
+{
+    MpegEncContext * const s = &h->s;
+    void (*idct_add)(uint8_t *dst, DCTELEM *block, int stride);
+    int i;
+    block_offset += 16*p;
+    if(!IS_INTRA4x4(mb_type)){
+        if(is_h264){
+            if(IS_INTRA16x16(mb_type)){
+                if(transform_bypass){
+                    if(h->sps.profile_idc==244 && (h->intra16x16_pred_mode==VERT_PRED8x8 || h->intra16x16_pred_mode==HOR_PRED8x8)){
+                        h->hpc.pred16x16_add[h->intra16x16_pred_mode](dest_y, block_offset, h->mb + (p*256 << pixel_shift), linesize);
+                    }else{
+                        for(i=0; i<16; i++){
+                            if(h->non_zero_count_cache[ scan8[i+p*16] ] || dctcoef_get(h->mb, pixel_shift, i*16))
+                                s->dsp.add_pixels4(dest_y + block_offset[i], h->mb + (i*16+p*256 << pixel_shift), linesize);
+                        }
+                    }
+                }else{
+                    h->h264dsp.h264_idct_add16intra(dest_y, block_offset, h->mb + (p*256 << pixel_shift), linesize, h->non_zero_count_cache+p*5*8);
+                }
+            }else if(h->cbp&15){
+                if(transform_bypass){
+                    const int di = IS_8x8DCT(mb_type) ? 4 : 1;
+                    idct_add= IS_8x8DCT(mb_type) ? s->dsp.add_pixels8 : s->dsp.add_pixels4;
+                    for(i=0; i<16; i+=di){
+                        if(h->non_zero_count_cache[ scan8[i+p*16] ]){
+                            idct_add(dest_y + block_offset[i], h->mb + (i*16+p*256 << pixel_shift), linesize);
+                        }
+                    }
+                }else{
+                    if(IS_8x8DCT(mb_type)){
+                        h->h264dsp.h264_idct8_add4(dest_y, block_offset, h->mb + (p*256 << pixel_shift), linesize, h->non_zero_count_cache+p*5*8);
+                    }else{
+                        h->h264dsp.h264_idct_add16(dest_y, block_offset, h->mb + (p*256 << pixel_shift), linesize, h->non_zero_count_cache+p*5*8);
+                    }
+                }
+            }
+        }else{
+            for(i=0; i<16; i++){
+                if(h->non_zero_count_cache[ scan8[i+p*16] ] || h->mb[i*16+p*256]){ //FIXME benchmark weird rule, & below
+                    uint8_t * const ptr= dest_y + block_offset[i];
+                    ff_svq3_add_idct_c(ptr, h->mb + i*16 + p*256, linesize, s->qscale, IS_INTRA(mb_type) ? 1 : 0);
+                }
+            }
+        }
+    }
+}
+
 static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple, int pixel_shift){
     MpegEncContext * const s = &h->s;
     const int mb_x= s->mb_x;
@@ -1546,13 +1787,12 @@ static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple, i
     const int mb_type= s->current_picture.mb_type[mb_xy];
     uint8_t  *dest_y, *dest_cb, *dest_cr;
     int linesize, uvlinesize /*dct_offset*/;
-    int i;
+    int i, j;
     int *block_offset = &h->block_offset[0];
     const int transform_bypass = !simple && (s->qscale == 0 && h->sps.transform_bypass);
     /* is_h264 should always be true if SVQ3 is disabled. */
     const int is_h264 = !CONFIG_SVQ3_DECODER || simple || s->codec_id == CODEC_ID_H264;
     void (*idct_add)(uint8_t *dst, DCTELEM *block, int stride);
-    void (*idct_dc_add)(uint8_t *dst, DCTELEM *block, int stride);
 
     dest_y  = s->current_picture.data[0] + ((mb_x << pixel_shift) + mb_y * s->linesize  ) * 16;
     dest_cb = s->current_picture.data[1] + ((mb_x << pixel_shift) + mb_y * s->uvlinesize) * 8;
@@ -1566,7 +1806,7 @@ static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple, i
     if (!simple && MB_FIELD) {
         linesize   = h->mb_linesize   = s->linesize * 2;
         uvlinesize = h->mb_uvlinesize = s->uvlinesize * 2;
-        block_offset = &h->block_offset[24];
+        block_offset = &h->block_offset[48];
         if(mb_y&1){ //FIXME move out of this function?
             dest_y -= s->linesize*15;
             dest_cb-= s->uvlinesize*7;
@@ -1607,227 +1847,95 @@ static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple, i
                 for (j = 0; j < 16; j++)
                     tmp_y[j] = get_bits(&gb, bit_depth);
             }
-            for (i = 0; i < 8; i++) {
-                uint16_t *tmp_cb = (uint16_t*)(dest_cb + i*uvlinesize);
-                for (j = 0; j < 8; j++)
-                    tmp_cb[j] = get_bits(&gb, bit_depth);
-            }
-            for (i = 0; i < 8; i++) {
-                uint16_t *tmp_cr = (uint16_t*)(dest_cr + i*uvlinesize);
-                for (j = 0; j < 8; j++)
-                    tmp_cr[j] = get_bits(&gb, bit_depth);
+            if(simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)){
+                for (i = 0; i < 8; i++) {
+                    uint16_t *tmp_cb = (uint16_t*)(dest_cb + i*uvlinesize);
+                    for (j = 0; j < 8; j++)
+                        tmp_cb[j] = get_bits(&gb, bit_depth);
+                }
+                for (i = 0; i < 8; i++) {
+                    uint16_t *tmp_cr = (uint16_t*)(dest_cr + i*uvlinesize);
+                    for (j = 0; j < 8; j++)
+                        tmp_cr[j] = get_bits(&gb, bit_depth);
+                }
             }
         } else {
-        for (i=0; i<16; i++) {
-            memcpy(dest_y + i*  linesize, h->mb       + i*8, 16);
-        }
-        for (i=0; i<8; i++) {
-            memcpy(dest_cb+ i*uvlinesize, h->mb + 128 + i*4,  8);
-            memcpy(dest_cr+ i*uvlinesize, h->mb + 160 + i*4,  8);
-        }
+            for (i=0; i<16; i++) {
+                memcpy(dest_y + i*  linesize, h->mb       + i*8, 16);
+            }
+            if(simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)){
+                for (i=0; i<8; i++) {
+                    memcpy(dest_cb+ i*uvlinesize, h->mb + 128 + i*4,  8);
+                    memcpy(dest_cr+ i*uvlinesize, h->mb + 160 + i*4,  8);
+                }
+            }
         }
     } else {
         if(IS_INTRA(mb_type)){
             if(h->deblocking_filter)
-                xchg_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 1, simple, pixel_shift);
+                xchg_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 1, 0, simple, pixel_shift);
 
             if(simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)){
                 h->hpc.pred8x8[ h->chroma_pred_mode ](dest_cb, uvlinesize);
                 h->hpc.pred8x8[ h->chroma_pred_mode ](dest_cr, uvlinesize);
             }
 
-            if(IS_INTRA4x4(mb_type)){
-                if(simple || !s->encoding){
-                    if(IS_8x8DCT(mb_type)){
-                        if(transform_bypass){
-                            idct_dc_add =
-                            idct_add    = s->dsp.add_pixels8;
-                        }else{
-                            idct_dc_add = h->h264dsp.h264_idct8_dc_add;
-                            idct_add    = h->h264dsp.h264_idct8_add;
-                        }
-                        for(i=0; i<16; i+=4){
-                            uint8_t * const ptr= dest_y + block_offset[i];
-                            const int dir= h->intra4x4_pred_mode_cache[ scan8[i] ];
-                            if(transform_bypass && h->sps.profile_idc==244 && dir<=1){
-                                h->hpc.pred8x8l_add[dir](ptr, h->mb + (i*16 << pixel_shift), linesize);
-                            }else{
-                                const int nnz = h->non_zero_count_cache[ scan8[i] ];
-                                h->hpc.pred8x8l[ dir ](ptr, (h->topleft_samples_available<<i)&0x8000,
-                                                            (h->topright_samples_available<<i)&0x4000, linesize);
-                                if(nnz){
-                                    if(nnz == 1 && dctcoef_get(h->mb, pixel_shift, i*16))
-                                        idct_dc_add(ptr, h->mb + (i*16 << pixel_shift), linesize);
-                                    else
-                                        idct_add   (ptr, h->mb + (i*16 << pixel_shift), linesize);
-                                }
-                            }
-                        }
-                    }else{
-                        if(transform_bypass){
-                            idct_dc_add =
-                            idct_add    = s->dsp.add_pixels4;
-                        }else{
-                            idct_dc_add = h->h264dsp.h264_idct_dc_add;
-                            idct_add    = h->h264dsp.h264_idct_add;
-                        }
-                        for(i=0; i<16; i++){
-                            uint8_t * const ptr= dest_y + block_offset[i];
-                            const int dir= h->intra4x4_pred_mode_cache[ scan8[i] ];
-
-                            if(transform_bypass && h->sps.profile_idc==244 && dir<=1){
-                                h->hpc.pred4x4_add[dir](ptr, h->mb + (i*16 << pixel_shift), linesize);
-                            }else{
-                                uint8_t *topright;
-                                int nnz, tr;
-                                uint64_t tr_high;
-                                if(dir == DIAG_DOWN_LEFT_PRED || dir == VERT_LEFT_PRED){
-                                    const int topright_avail= (h->topright_samples_available<<i)&0x8000;
-                                    assert(mb_y || linesize <= block_offset[i]);
-                                    if(!topright_avail){
-                                        if (pixel_shift) {
-                                            tr_high= ((uint16_t*)ptr)[3 - linesize/2]*0x0001000100010001ULL;
-                                            topright= (uint8_t*) &tr_high;
-                                        } else {
-                                        tr= ptr[3 - linesize]*0x01010101;
-                                        topright= (uint8_t*) &tr;
-                                        }
-                                    }else
-                                        topright= ptr + (4 << pixel_shift) - linesize;
-                                }else
-                                    topright= NULL;
-
-                                h->hpc.pred4x4[ dir ](ptr, topright, linesize);
-                                nnz = h->non_zero_count_cache[ scan8[i] ];
-                                if(nnz){
-                                    if(is_h264){
-                                        if(nnz == 1 && dctcoef_get(h->mb, pixel_shift, i*16))
-                                            idct_dc_add(ptr, h->mb + (i*16 << pixel_shift), linesize);
-                                        else
-                                            idct_add   (ptr, h->mb + (i*16<<pixel_shift), linesize);
-                                    }
-#if CONFIG_SVQ3_DECODER
-                                    else
-                                        ff_svq3_add_idct_c(ptr, h->mb + i*16, linesize, s->qscale, 0);
-#endif
-                                }
-                            }
-                        }
-                    }
-                }
-            }else{
-                h->hpc.pred16x16[ h->intra16x16_pred_mode ](dest_y , linesize);
-                if(is_h264){
-                    if(h->non_zero_count_cache[ scan8[LUMA_DC_BLOCK_INDEX] ]){
-                        if(!transform_bypass)
-                            h->h264dsp.h264_luma_dc_dequant_idct(h->mb, h->mb_luma_dc, h->dequant4_coeff[0][s->qscale][0]);
-                        else{
-                            static const uint8_t dc_mapping[16] = { 0*16, 1*16, 4*16, 5*16, 2*16, 3*16, 6*16, 7*16,
-                                                                    8*16, 9*16,12*16,13*16,10*16,11*16,14*16,15*16};
-                            for(i = 0; i < 16; i++)
-                                dctcoef_set(h->mb, pixel_shift, dc_mapping[i], dctcoef_get(h->mb_luma_dc, pixel_shift, i));
-                        }
-                    }
-                }
-#if CONFIG_SVQ3_DECODER
-                else
-                    ff_svq3_luma_dc_dequant_idct_c(h->mb, h->mb_luma_dc, s->qscale);
-#endif
-            }
+            hl_decode_mb_predict_luma(h, mb_type, is_h264, simple, transform_bypass, pixel_shift, block_offset, linesize, dest_y, 0);
+
             if(h->deblocking_filter)
-                xchg_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 0, simple, pixel_shift);
+                xchg_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 0, 0, simple, pixel_shift);
         }else if(is_h264){
             if (pixel_shift) {
                 hl_motion_16(h, dest_y, dest_cb, dest_cr,
                              s->me.qpel_put, s->dsp.put_h264_chroma_pixels_tab,
                              s->me.qpel_avg, s->dsp.avg_h264_chroma_pixels_tab,
                              h->h264dsp.weight_h264_pixels_tab,
-                             h->h264dsp.biweight_h264_pixels_tab);
+                             h->h264dsp.biweight_h264_pixels_tab, 0);
             } else
                 hl_motion_8(h, dest_y, dest_cb, dest_cr,
                             s->me.qpel_put, s->dsp.put_h264_chroma_pixels_tab,
                             s->me.qpel_avg, s->dsp.avg_h264_chroma_pixels_tab,
                             h->h264dsp.weight_h264_pixels_tab,
-                            h->h264dsp.biweight_h264_pixels_tab);
+                            h->h264dsp.biweight_h264_pixels_tab, 0);
         }
 
-
-        if(!IS_INTRA4x4(mb_type)){
-            if(is_h264){
-                if(IS_INTRA16x16(mb_type)){
-                    if(transform_bypass){
-                        if(h->sps.profile_idc==244 && (h->intra16x16_pred_mode==VERT_PRED8x8 || h->intra16x16_pred_mode==HOR_PRED8x8)){
-                            h->hpc.pred16x16_add[h->intra16x16_pred_mode](dest_y, block_offset, h->mb, linesize);
-                        }else{
-                            for(i=0; i<16; i++){
-                                if(h->non_zero_count_cache[ scan8[i] ] || dctcoef_get(h->mb, pixel_shift, i*16))
-                                    s->dsp.add_pixels4(dest_y + block_offset[i], h->mb + (i*16 << pixel_shift), linesize);
-                            }
-                        }
-                    }else{
-                         h->h264dsp.h264_idct_add16intra(dest_y, block_offset, h->mb, linesize, h->non_zero_count_cache);
-                    }
-                }else if(h->cbp&15){
-                    if(transform_bypass){
-                        const int di = IS_8x8DCT(mb_type) ? 4 : 1;
-                        idct_add= IS_8x8DCT(mb_type) ? s->dsp.add_pixels8 : s->dsp.add_pixels4;
-                        for(i=0; i<16; i+=di){
-                            if(h->non_zero_count_cache[ scan8[i] ]){
-                                idct_add(dest_y + block_offset[i], h->mb + (i*16 << pixel_shift), linesize);
-                            }
-                        }
-                    }else{
-                        if(IS_8x8DCT(mb_type)){
-                            h->h264dsp.h264_idct8_add4(dest_y, block_offset, h->mb, linesize, h->non_zero_count_cache);
-                        }else{
-                            h->h264dsp.h264_idct_add16(dest_y, block_offset, h->mb, linesize, h->non_zero_count_cache);
-                        }
-                    }
-                }
-            }
-#if CONFIG_SVQ3_DECODER
-            else{
-                for(i=0; i<16; i++){
-                    if(h->non_zero_count_cache[ scan8[i] ] || h->mb[i*16]){ //FIXME benchmark weird rule, & below
-                        uint8_t * const ptr= dest_y + block_offset[i];
-                        ff_svq3_add_idct_c(ptr, h->mb + i*16, linesize, s->qscale, IS_INTRA(mb_type) ? 1 : 0);
-                    }
-                }
-            }
-#endif
-        }
+        hl_decode_mb_idct_luma(h, mb_type, is_h264, simple, transform_bypass, pixel_shift, block_offset, linesize, dest_y, 0);
 
         if((simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)) && (h->cbp&0x30)){
             uint8_t *dest[2] = {dest_cb, dest_cr};
             if(transform_bypass){
                 if(IS_INTRA(mb_type) && h->sps.profile_idc==244 && (h->chroma_pred_mode==VERT_PRED8x8 || h->chroma_pred_mode==HOR_PRED8x8)){
-                    h->hpc.pred8x8_add[h->chroma_pred_mode](dest[0], block_offset + 16, h->mb + (16*16 << pixel_shift), uvlinesize);
-                    h->hpc.pred8x8_add[h->chroma_pred_mode](dest[1], block_offset + 20, h->mb + (20*16 << pixel_shift), uvlinesize);
+                    h->hpc.pred8x8_add[h->chroma_pred_mode](dest[0], block_offset + 16, h->mb + (16*16*1 << pixel_shift), uvlinesize);
+                    h->hpc.pred8x8_add[h->chroma_pred_mode](dest[1], block_offset + 32, h->mb + (16*16*2 << pixel_shift), uvlinesize);
                 }else{
                     idct_add = s->dsp.add_pixels4;
-                    for(i=16; i<16+8; i++){
-                        if(h->non_zero_count_cache[ scan8[i] ] || dctcoef_get(h->mb, pixel_shift, i*16))
-                            idct_add   (dest[(i&4)>>2] + block_offset[i], h->mb + (i*16 << pixel_shift), uvlinesize);
+                    for(j=1; j<3; j++){
+                        for(i=j*16; i<j*16+4; i++){
+                            if(h->non_zero_count_cache[ scan8[i] ] || dctcoef_get(h->mb, pixel_shift, i*16))
+                                idct_add   (dest[j-1] + block_offset[i], h->mb + (i*16 << pixel_shift), uvlinesize);
+                        }
                     }
                 }
             }else{
                 if(is_h264){
                     if(h->non_zero_count_cache[ scan8[CHROMA_DC_BLOCK_INDEX+0] ])
-                        h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + (16*16 << pixel_shift)       , h->dequant4_coeff[IS_INTRA(mb_type) ? 1:4][h->chroma_qp[0]][0]);
+                        h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + (16*16*1 << pixel_shift), h->dequant4_coeff[IS_INTRA(mb_type) ? 1:4][h->chroma_qp[0]][0]);
                     if(h->non_zero_count_cache[ scan8[CHROMA_DC_BLOCK_INDEX+1] ])
-                        h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + ((16*16+4*16) << pixel_shift), h->dequant4_coeff[IS_INTRA(mb_type) ? 2:5][h->chroma_qp[1]][0]);
+                        h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + (16*16*2 << pixel_shift), h->dequant4_coeff[IS_INTRA(mb_type) ? 2:5][h->chroma_qp[1]][0]);
                     h->h264dsp.h264_idct_add8(dest, block_offset,
                                               h->mb, uvlinesize,
                                               h->non_zero_count_cache);
                 }
 #if CONFIG_SVQ3_DECODER
                 else{
-                    h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + 16*16     , h->dequant4_coeff[IS_INTRA(mb_type) ? 1:4][h->chroma_qp[0]][0]);
-                    h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + 16*16+4*16, h->dequant4_coeff[IS_INTRA(mb_type) ? 2:5][h->chroma_qp[1]][0]);
-                    for(i=16; i<16+8; i++){
-                        if(h->non_zero_count_cache[ scan8[i] ] || h->mb[i*16]){
-                            uint8_t * const ptr= dest[(i&4)>>2] + block_offset[i];
-                            ff_svq3_add_idct_c(ptr, h->mb + i*16, uvlinesize, ff_h264_chroma_qp[0][s->qscale + 12] - 12, 2);
+                    h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + 16*16*1, h->dequant4_coeff[IS_INTRA(mb_type) ? 1:4][h->chroma_qp[0]][0]);
+                    h->h264dsp.h264_chroma_dc_dequant_idct(h->mb + 16*16*2, h->dequant4_coeff[IS_INTRA(mb_type) ? 2:5][h->chroma_qp[1]][0]);
+                    for(j=1; j<3; j++){
+                        for(i=j*16; i<j*16+4; i++){
+                            if(h->non_zero_count_cache[ scan8[i] ] || h->mb[i*16]){
+                                uint8_t * const ptr= dest[j-1] + block_offset[i];
+                                ff_svq3_add_idct_c(ptr, h->mb + i*16, uvlinesize, ff_h264_chroma_qp[0][s->qscale + 12] - 12, 2);
+                            }
                         }
                     }
                 }
@@ -1836,7 +1944,113 @@ static av_always_inline void hl_decode_mb_internal(H264Context *h, int simple, i
         }
     }
     if(h->cbp || IS_INTRA(mb_type))
+    {
+        s->dsp.clear_blocks(h->mb);
+        s->dsp.clear_blocks(h->mb+(24*16<<pixel_shift));
+    }
+}
+
+static av_always_inline void hl_decode_mb_444_internal(H264Context *h, int simple, int pixel_shift){
+    MpegEncContext * const s = &h->s;
+    const int mb_x= s->mb_x;
+    const int mb_y= s->mb_y;
+    const int mb_xy= h->mb_xy;
+    const int mb_type= s->current_picture.mb_type[mb_xy];
+    uint8_t  *dest[3];
+    int linesize;
+    int i, j, p;
+    int *block_offset = &h->block_offset[0];
+    const int transform_bypass = !simple && (s->qscale == 0 && h->sps.transform_bypass);
+    const int plane_count = (simple || !CONFIG_GRAY || !(s->flags&CODEC_FLAG_GRAY)) ? 3 : 1;
+
+    for (p = 0; p < plane_count; p++)
+    {
+        dest[p] = s->current_picture.data[p] + ((mb_x << pixel_shift) + mb_y * s->linesize) * 16;
+        s->dsp.prefetch(dest[p] + (s->mb_x&3)*4*s->linesize + (64 << pixel_shift), s->linesize, 4);
+    }
+
+    h->list_counts[mb_xy]= h->list_count;
+
+    if (!simple && MB_FIELD) {
+        linesize   = h->mb_linesize = h->mb_uvlinesize = s->linesize * 2;
+        block_offset = &h->block_offset[48];
+        if(mb_y&1) //FIXME move out of this function?
+            for (p = 0; p < 3; p++)
+                dest[p] -= s->linesize*15;
+        if(FRAME_MBAFF) {
+            int list;
+            for(list=0; list<h->list_count; list++){
+                if(!USES_LIST(mb_type, list))
+                    continue;
+                if(IS_16X16(mb_type)){
+                    int8_t *ref = &h->ref_cache[list][scan8[0]];
+                    fill_rectangle(ref, 4, 4, 8, (16+*ref)^(s->mb_y&1), 1);
+                }else{
+                    for(i=0; i<16; i+=4){
+                        int ref = h->ref_cache[list][scan8[i]];
+                        if(ref >= 0)
+                            fill_rectangle(&h->ref_cache[list][scan8[i]], 2, 2, 8, (16+ref)^(s->mb_y&1), 1);
+                    }
+                }
+            }
+        }
+    } else {
+        linesize   = h->mb_linesize = h->mb_uvlinesize = s->linesize;
+    }
+
+    if (!simple && IS_INTRA_PCM(mb_type)) {
+        if (pixel_shift) {
+            const int bit_depth = h->sps.bit_depth_luma;
+            GetBitContext gb;
+            init_get_bits(&gb, (uint8_t*)h->mb, 768*bit_depth);
+
+            for (p = 0; p < plane_count; p++) {
+                for (i = 0; i < 16; i++) {
+                    uint16_t *tmp = (uint16_t*)(dest[p] + i*linesize);
+                    for (j = 0; j < 16; j++)
+                        tmp[j] = get_bits(&gb, bit_depth);
+                }
+            }
+        } else {
+            for (p = 0; p < plane_count; p++) {
+                for (i = 0; i < 16; i++) {
+                    memcpy(dest[p] + i*linesize, h->mb + p*128 + i*8, 16);
+                }
+            }
+        }
+    } else {
+        if(IS_INTRA(mb_type)){
+            if(h->deblocking_filter)
+                xchg_mb_border(h, dest[0], dest[1], dest[2], linesize, linesize, 1, 1, simple, pixel_shift);
+
+            for (p = 0; p < plane_count; p++)
+                hl_decode_mb_predict_luma(h, mb_type, 1, simple, transform_bypass, pixel_shift, block_offset, linesize, dest[p], p);
+
+            if(h->deblocking_filter)
+                xchg_mb_border(h, dest[0], dest[1], dest[2], linesize, linesize, 0, 1, simple, pixel_shift);
+        }else{
+            if (pixel_shift) {
+                hl_motion_16(h, dest[0], dest[1], dest[2],
+                             s->me.qpel_put, s->dsp.put_h264_chroma_pixels_tab,
+                             s->me.qpel_avg, s->dsp.avg_h264_chroma_pixels_tab,
+                             h->h264dsp.weight_h264_pixels_tab,
+                             h->h264dsp.biweight_h264_pixels_tab, 1);
+            } else
+                hl_motion_8(h, dest[0], dest[1], dest[2],
+                            s->me.qpel_put, s->dsp.put_h264_chroma_pixels_tab,
+                            s->me.qpel_avg, s->dsp.avg_h264_chroma_pixels_tab,
+                            h->h264dsp.weight_h264_pixels_tab,
+                            h->h264dsp.biweight_h264_pixels_tab, 1);
+        }
+
+        for (p = 0; p < plane_count; p++)
+            hl_decode_mb_idct_luma(h, mb_type, 1, simple, transform_bypass, pixel_shift, block_offset, linesize, dest[p], p);
+    }
+    if(h->cbp || IS_INTRA(mb_type))
+    {
         s->dsp.clear_blocks(h->mb);
+        s->dsp.clear_blocks(h->mb+(24*16<<pixel_shift));
+    }
 }
 
 /**
@@ -1856,13 +2070,26 @@ static void av_noinline hl_decode_mb_complex(H264Context *h){
     hl_decode_mb_internal(h, 0, h->pixel_shift);
 }
 
+static void av_noinline hl_decode_mb_444_complex(H264Context *h){
+    hl_decode_mb_444_internal(h, 0, h->pixel_shift);
+}
+
+static void av_noinline hl_decode_mb_444_simple(H264Context *h){
+    hl_decode_mb_444_internal(h, 1, 0);
+}
+
 void ff_h264_hl_decode_mb(H264Context *h){
     MpegEncContext * const s = &h->s;
     const int mb_xy= h->mb_xy;
     const int mb_type= s->current_picture.mb_type[mb_xy];
     int is_complex = CONFIG_SMALL || h->is_complex || IS_INTRA_PCM(mb_type) || s->qscale == 0;
 
-    if (is_complex) {
+    if (CHROMA444) {
+        if(is_complex || h->pixel_shift)
+            hl_decode_mb_444_complex(h);
+        else
+            hl_decode_mb_444_simple(h);
+    } else if (is_complex) {
         hl_decode_mb_complex(h);
     } else if (h->pixel_shift) {
         hl_decode_mb_simple_16(h);
@@ -1878,7 +2105,7 @@ static int pred_weight_table(H264Context *h){
     h->use_weight= 0;
     h->use_weight_chroma= 0;
     h->luma_log2_weight_denom= get_ue_golomb(&s->gb);
-    if(CHROMA)
+    if(h->sps.chroma_format_idc)
         h->chroma_log2_weight_denom= get_ue_golomb(&s->gb);
     luma_def = 1<<h->luma_log2_weight_denom;
     chroma_def = 1<<h->chroma_log2_weight_denom;
@@ -1903,7 +2130,7 @@ static int pred_weight_table(H264Context *h){
                 h->luma_weight[i][list][1]= 0;
             }
 
-            if(CHROMA){
+            if(h->sps.chroma_format_idc){
                 chroma_weight_flag= get_bits1(&s->gb);
                 if(chroma_weight_flag){
                     int j;
@@ -2333,11 +2560,11 @@ static int decode_slice_header(H264Context *h, H264Context *h0){
 
     h->b_stride=  s->mb_width*4;
 
-    s->width = 16*s->mb_width - 2*FFMIN(h->sps.crop_right, 7);
+    s->width = 16*s->mb_width - (2>>CHROMA444)*FFMIN(h->sps.crop_right, (8<<CHROMA444)-1);
     if(h->sps.frame_mbs_only_flag)
-        s->height= 16*s->mb_height - 2*FFMIN(h->sps.crop_bottom, 7);
+        s->height= 16*s->mb_height - (2>>CHROMA444)*FFMIN(h->sps.crop_bottom, (8<<CHROMA444)-1);
     else
-        s->height= 16*s->mb_height - 4*FFMIN(h->sps.crop_bottom, 7);
+        s->height= 16*s->mb_height - (4>>CHROMA444)*FFMIN(h->sps.crop_bottom, (8<<CHROMA444)-1);
 
     if (s->context_initialized
         && (   s->width != s->avctx->width || s->height != s->avctx->height
@@ -2382,18 +2609,22 @@ static int decode_slice_header(H264Context *h, H264Context *h0){
 
         switch (h->sps.bit_depth_luma) {
             case 9 :
-                s->avctx->pix_fmt = PIX_FMT_YUV420P9;
+                s->avctx->pix_fmt = CHROMA444 ? PIX_FMT_YUV444P9 : PIX_FMT_YUV420P9;
                 break;
             case 10 :
-                s->avctx->pix_fmt = PIX_FMT_YUV420P10;
+                s->avctx->pix_fmt = CHROMA444 ? PIX_FMT_YUV444P10 : PIX_FMT_YUV420P10;
                 break;
             default:
-        s->avctx->pix_fmt = s->avctx->get_format(s->avctx,
-                                                 s->avctx->codec->pix_fmts ?
-                                                 s->avctx->codec->pix_fmts :
-                                                 s->avctx->color_range == AVCOL_RANGE_JPEG ?
-                                                 hwaccel_pixfmt_list_h264_jpeg_420 :
-                                                 ff_hwaccel_pixfmt_list_420);
+                if (CHROMA444){
+                    s->avctx->pix_fmt = s->avctx->color_range == AVCOL_RANGE_JPEG ? PIX_FMT_YUVJ444P : PIX_FMT_YUV444P;
+                }else{
+                    s->avctx->pix_fmt = s->avctx->get_format(s->avctx,
+                                                             s->avctx->codec->pix_fmts ?
+                                                             s->avctx->codec->pix_fmts :
+                                                             s->avctx->color_range == AVCOL_RANGE_JPEG ?
+                                                             hwaccel_pixfmt_list_h264_jpeg_420 :
+                                                             ff_hwaccel_pixfmt_list_420);
+                }
         }
 
         s->avctx->hwaccel = ff_find_hwaccel(s->avctx->codec->id, s->avctx->pix_fmt);
@@ -2896,10 +3127,10 @@ static int fill_filter_caches(H264Context *h, int mb_type){
     if(IS_INTRA(mb_type))
         return 0;
 
-    AV_COPY32(&h->non_zero_count_cache[4+8*1], &h->non_zero_count[mb_xy][ 4]);
-    AV_COPY32(&h->non_zero_count_cache[4+8*2], &h->non_zero_count[mb_xy][12]);
-    AV_COPY32(&h->non_zero_count_cache[4+8*3], &h->non_zero_count[mb_xy][20]);
-    AV_COPY32(&h->non_zero_count_cache[4+8*4], &h->non_zero_count[mb_xy][28]);
+    AV_COPY32(&h->non_zero_count_cache[4+8* 1], &h->non_zero_count[mb_xy][ 0]);
+    AV_COPY32(&h->non_zero_count_cache[4+8* 2], &h->non_zero_count[mb_xy][ 4]);
+    AV_COPY32(&h->non_zero_count_cache[4+8* 3], &h->non_zero_count[mb_xy][ 8]);
+    AV_COPY32(&h->non_zero_count_cache[4+8* 4], &h->non_zero_count[mb_xy][12]);
 
     h->cbp= h->cbp_table[mb_xy];
 
@@ -2951,45 +3182,45 @@ static int fill_filter_caches(H264Context *h, int mb_type){
 */
 //FIXME constraint_intra_pred & partitioning & nnz (let us hope this is just a typo in the spec)
     if(top_type){
-        AV_COPY32(&h->non_zero_count_cache[4+8*0], &h->non_zero_count[top_xy][4+3*8]);
+        AV_COPY32(&h->non_zero_count_cache[4+8*0], &h->non_zero_count[top_xy][3*4]);
     }
 
     if(left_type[0]){
-        h->non_zero_count_cache[3+8*1]= h->non_zero_count[left_xy[0]][7+0*8];
-        h->non_zero_count_cache[3+8*2]= h->non_zero_count[left_xy[0]][7+1*8];
-        h->non_zero_count_cache[3+8*3]= h->non_zero_count[left_xy[0]][7+2*8];
-        h->non_zero_count_cache[3+8*4]= h->non_zero_count[left_xy[0]][7+3*8];
+        h->non_zero_count_cache[3+8*1]= h->non_zero_count[left_xy[0]][3+0*4];
+        h->non_zero_count_cache[3+8*2]= h->non_zero_count[left_xy[0]][3+1*4];
+        h->non_zero_count_cache[3+8*3]= h->non_zero_count[left_xy[0]][3+2*4];
+        h->non_zero_count_cache[3+8*4]= h->non_zero_count[left_xy[0]][3+3*4];
     }
 
     // CAVLC 8x8dct requires NNZ values for residual decoding that differ from what the loop filter needs
     if(!CABAC && h->pps.transform_8x8_mode){
         if(IS_8x8DCT(top_type)){
             h->non_zero_count_cache[4+8*0]=
-            h->non_zero_count_cache[5+8*0]= h->cbp_table[top_xy] & 4;
+            h->non_zero_count_cache[5+8*0]= (h->cbp_table[top_xy] & 0x4000) >> 12;
             h->non_zero_count_cache[6+8*0]=
-            h->non_zero_count_cache[7+8*0]= h->cbp_table[top_xy] & 8;
+            h->non_zero_count_cache[7+8*0]= (h->cbp_table[top_xy] & 0x8000) >> 12;
         }
         if(IS_8x8DCT(left_type[0])){
             h->non_zero_count_cache[3+8*1]=
-            h->non_zero_count_cache[3+8*2]= h->cbp_table[left_xy[0]]&2; //FIXME check MBAFF
+            h->non_zero_count_cache[3+8*2]= (h->cbp_table[left_xy[0]]&0x2000) >> 12; //FIXME check MBAFF
         }
         if(IS_8x8DCT(left_type[1])){
             h->non_zero_count_cache[3+8*3]=
-            h->non_zero_count_cache[3+8*4]= h->cbp_table[left_xy[1]]&8; //FIXME check MBAFF
+            h->non_zero_count_cache[3+8*4]= (h->cbp_table[left_xy[1]]&0x8000) >> 12; //FIXME check MBAFF
         }
 
         if(IS_8x8DCT(mb_type)){
             h->non_zero_count_cache[scan8[0   ]]= h->non_zero_count_cache[scan8[1   ]]=
-            h->non_zero_count_cache[scan8[2   ]]= h->non_zero_count_cache[scan8[3   ]]= h->cbp & 1;
+            h->non_zero_count_cache[scan8[2   ]]= h->non_zero_count_cache[scan8[3   ]]= (h->cbp & 0x1000) >> 12;
 
             h->non_zero_count_cache[scan8[0+ 4]]= h->non_zero_count_cache[scan8[1+ 4]]=
-            h->non_zero_count_cache[scan8[2+ 4]]= h->non_zero_count_cache[scan8[3+ 4]]= h->cbp & 2;
+            h->non_zero_count_cache[scan8[2+ 4]]= h->non_zero_count_cache[scan8[3+ 4]]= (h->cbp & 0x2000) >> 12;
 
             h->non_zero_count_cache[scan8[0+ 8]]= h->non_zero_count_cache[scan8[1+ 8]]=
-            h->non_zero_count_cache[scan8[2+ 8]]= h->non_zero_count_cache[scan8[3+ 8]]= h->cbp & 4;
+            h->non_zero_count_cache[scan8[2+ 8]]= h->non_zero_count_cache[scan8[3+ 8]]= (h->cbp & 0x4000) >> 12;
 
             h->non_zero_count_cache[scan8[0+12]]= h->non_zero_count_cache[scan8[1+12]]=
-            h->non_zero_count_cache[scan8[2+12]]= h->non_zero_count_cache[scan8[3+12]]= h->cbp & 8;
+            h->non_zero_count_cache[scan8[2+12]]= h->non_zero_count_cache[scan8[3+12]]= (h->cbp & 0x8000) >> 12;
         }
     }
 
@@ -3063,8 +3294,8 @@ static void loop_filter(H264Context *h, int start_x, int end_x){
                 s->mb_x= mb_x;
                 s->mb_y= mb_y;
                 dest_y  = s->current_picture.data[0] + ((mb_x << pixel_shift) + mb_y * s->linesize  ) * 16;
-                dest_cb = s->current_picture.data[1] + ((mb_x << pixel_shift) + mb_y * s->uvlinesize) * 8;
-                dest_cr = s->current_picture.data[2] + ((mb_x << pixel_shift) + mb_y * s->uvlinesize) * 8;
+                dest_cb = s->current_picture.data[1] + ((mb_x << pixel_shift) + mb_y * s->uvlinesize) * (8 << CHROMA444);
+                dest_cr = s->current_picture.data[2] + ((mb_x << pixel_shift) + mb_y * s->uvlinesize) * (8 << CHROMA444);
                     //FIXME simplify above
 
                 if (MB_FIELD) {
@@ -3079,7 +3310,7 @@ static void loop_filter(H264Context *h, int start_x, int end_x){
                     linesize   = h->mb_linesize   = s->linesize;
                     uvlinesize = h->mb_uvlinesize = s->uvlinesize;
                 }
-                backup_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, 0);
+                backup_mb_border(h, dest_y, dest_cb, dest_cr, linesize, uvlinesize, CHROMA444, 0);
                 if(fill_filter_caches(h, mb_type))
                     continue;
                 h->chroma_qp[0] = get_chroma_qp(h, 0, s->current_picture.qscale_table[mb_xy]);
diff --git a/libavcodec/h264.h b/libavcodec/h264.h
index c075c4b48e..7bad91b215 100644
--- a/libavcodec/h264.h
+++ b/libavcodec/h264.h
@@ -39,9 +39,6 @@
 #define interlaced_dct interlaced_dct_is_a_bad_name
 #define mb_intra mb_intra_is_not_initialized_see_mb_type
 
-#define LUMA_DC_BLOCK_INDEX   24
-#define CHROMA_DC_BLOCK_INDEX 25
-
 #define CHROMA_DC_COEFF_TOKEN_VLC_BITS 8
 #define COEFF_TOKEN_VLC_BITS           8
 #define TOTAL_ZEROS_VLC_BITS           9
@@ -60,8 +57,6 @@
  * of progressive decoding by about 2%. */
 #define ALLOW_INTERLACE
 
-#define ALLOW_NOCHROMA
-
 #define FMO 0
 
 /**
@@ -85,16 +80,12 @@
 #endif
 #define FIELD_OR_MBAFF_PICTURE (FRAME_MBAFF || FIELD_PICTURE)
 
-#ifdef ALLOW_NOCHROMA
-#define CHROMA h->sps.chroma_format_idc
-#else
-#define CHROMA 1
-#endif
-
 #ifndef CABAC
 #define CABAC h->pps.cabac
 #endif
 
+#define CHROMA444 (h->sps.chroma_format_idc == 3)
+
 #define EXTENDED_SAR          255
 
 #define MB_TYPE_REF0       MB_TYPE_ACPRED //dirty but it fits in 16 bit
@@ -198,7 +189,7 @@ typedef struct SPS{
     int num_reorder_frames;
     int scaling_matrix_present;
     uint8_t scaling_matrix4[6][16];
-    uint8_t scaling_matrix8[2][64];
+    uint8_t scaling_matrix8[6][64];
     int nal_hrd_parameters_present_flag;
     int vcl_hrd_parameters_present_flag;
     int pic_struct_present_flag;
@@ -233,7 +224,7 @@ typedef struct PPS{
     int redundant_pic_cnt_present; ///< redundant_pic_cnt_present_flag
     int transform_8x8_mode;     ///< transform_8x8_mode_flag
     uint8_t scaling_matrix4[6][16];
-    uint8_t scaling_matrix8[2][64];
+    uint8_t scaling_matrix8[6][64];
     uint8_t chroma_qp_table[2][64];  ///< pre-scaled (with chroma_qp_index_offset) version of qp_table
     int chroma_qp_diff;
 }PPS;
@@ -298,21 +289,15 @@ typedef struct H264Context{
     unsigned int top_samples_available;
     unsigned int topright_samples_available;
     unsigned int left_samples_available;
-    uint8_t (*top_borders[2])[(16+2*8)*2];
+    uint8_t (*top_borders[2])[(16*3)*2];
 
     /**
      * non zero coeff count cache.
      * is 64 if not available.
      */
-    DECLARE_ALIGNED(8, uint8_t, non_zero_count_cache)[6*8];
+    DECLARE_ALIGNED(8, uint8_t, non_zero_count_cache)[15*8];
 
-    /*
-    .UU.YYYY
-    .UU.YYYY
-    .vv.YYYY
-    .VV.YYYY
-    */
-    uint8_t (*non_zero_count)[32];
+    uint8_t (*non_zero_count)[48];
 
     /**
      * Motion vector cache.
@@ -336,7 +321,7 @@ typedef struct H264Context{
      * block_offset[ 0..23] for frame macroblocks
      * block_offset[24..47] for field macroblocks
      */
-    int block_offset[2*(16+8)];
+    int block_offset[2*(16*3)];
 
     uint32_t *mb2b_xy; //FIXME are these 4 a good idea?
     uint32_t *mb2br_xy;
@@ -356,9 +341,9 @@ typedef struct H264Context{
     PPS pps; //FIXME move to Picture perhaps? (->no) do we need that?
 
     uint32_t dequant4_buffer[6][QP_MAX_NUM+1][16]; //FIXME should these be moved down?
-    uint32_t dequant8_buffer[2][QP_MAX_NUM+1][64];
+    uint32_t dequant8_buffer[6][QP_MAX_NUM+1][64];
     uint32_t (*dequant4_coeff[6])[16];
-    uint32_t (*dequant8_coeff[2])[64];
+    uint32_t (*dequant8_coeff[6])[64];
 
     int slice_num;
     uint16_t *slice_table;     ///< slice_table_base + 2*mb_stride + 1
@@ -408,15 +393,15 @@ typedef struct H264Context{
     GetBitContext *intra_gb_ptr;
     GetBitContext *inter_gb_ptr;
 
-    DECLARE_ALIGNED(16, DCTELEM, mb)[16*24*2]; ///< as a dct coeffecient is int32_t in high depth, we need to reserve twice the space.
-    DECLARE_ALIGNED(16, DCTELEM, mb_luma_dc)[16*2];
+    DECLARE_ALIGNED(16, DCTELEM, mb)[16*48*2]; ///< as a dct coeffecient is int32_t in high depth, we need to reserve twice the space.
+    DECLARE_ALIGNED(16, DCTELEM, mb_luma_dc)[3][16*2];
     DCTELEM mb_padding[256*2];        ///< as mb is addressed by scantable[i] and scantable is uint8_t we can either check that i is not too large or ensure that there is some unused stuff after mb
 
     /**
      * Cabac
      */
     CABACContext cabac;
-    uint8_t      cabac_state[460];
+    uint8_t      cabac_state[1024];
 
     /* 0x100 -> non null luma_dc, 0x80/0x40 -> non null chroma_dc (cb/cr), 0x?0 -> chroma_cbp(0,1,2), 0x0? luma_cbp */
     uint16_t     *cbp_table;
@@ -721,27 +706,43 @@ o-o o-o
 */
 
 /* Scan8 organization:
- *   0 1 2 3 4 5 6 7
- * 0   u u y y y y y
- * 1 u U U y Y Y Y Y
- * 2 u U U y Y Y Y Y
- * 3   v v y Y Y Y Y
- * 4 v V V y Y Y Y Y
- * 5 v V V   DYDUDV
+ *    0 1 2 3 4 5 6 7
+ * 0  DY    y y y y y
+ * 1        y Y Y Y Y
+ * 2        y Y Y Y Y
+ * 3        y Y Y Y Y
+ * 4        y Y Y Y Y
+ * 5  DU    u u u u u
+ * 6        u U U U U
+ * 7        u U U U U
+ * 8        u U U U U
+ * 9        u U U U U
+ * 10 DV    v v v v v
+ * 11       v V V V V
+ * 12       v V V V V
+ * 13       v V V V V
+ * 14       v V V V V
  * DY/DU/DV are for luma/chroma DC.
  */
 
+#define LUMA_DC_BLOCK_INDEX   48
+#define CHROMA_DC_BLOCK_INDEX 49
+
 //This table must be here because scan8[constant] must be known at compiletime
-static const uint8_t scan8[16 + 2*4 + 3]={
- 4+1*8, 5+1*8, 4+2*8, 5+2*8,
- 6+1*8, 7+1*8, 6+2*8, 7+2*8,
- 4+3*8, 5+3*8, 4+4*8, 5+4*8,
- 6+3*8, 7+3*8, 6+4*8, 7+4*8,
- 1+1*8, 2+1*8,
- 1+2*8, 2+2*8,
- 1+4*8, 2+4*8,
- 1+5*8, 2+5*8,
- 4+5*8, 5+5*8, 6+5*8
+static const uint8_t scan8[16*3 + 3]={
+ 4+ 1*8, 5+ 1*8, 4+ 2*8, 5+ 2*8,
+ 6+ 1*8, 7+ 1*8, 6+ 2*8, 7+ 2*8,
+ 4+ 3*8, 5+ 3*8, 4+ 4*8, 5+ 4*8,
+ 6+ 3*8, 7+ 3*8, 6+ 4*8, 7+ 4*8,
+ 4+ 6*8, 5+ 6*8, 4+ 7*8, 5+ 7*8,
+ 6+ 6*8, 7+ 6*8, 6+ 7*8, 7+ 7*8,
+ 4+ 8*8, 5+ 8*8, 4+ 9*8, 5+ 9*8,
+ 6+ 8*8, 7+ 8*8, 6+ 9*8, 7+ 9*8,
+ 4+11*8, 5+11*8, 4+12*8, 5+12*8,
+ 6+11*8, 7+11*8, 6+12*8, 7+12*8,
+ 4+13*8, 5+13*8, 4+14*8, 5+14*8,
+ 6+13*8, 7+13*8, 6+14*8, 7+14*8,
+ 0+ 0*8, 0+ 5*8, 0+10*8
 };
 
 static av_always_inline uint32_t pack16to32(int a, int b){
@@ -773,11 +774,11 @@ static void fill_decode_neighbors(H264Context *h, int mb_type){
     MpegEncContext * const s = &h->s;
     const int mb_xy= h->mb_xy;
     int topleft_xy, top_xy, topright_xy, left_xy[2];
-    static const uint8_t left_block_options[4][16]={
-        {0,1,2,3,7,10,8,11,7+0*8, 7+1*8, 7+2*8, 7+3*8, 2+0*8, 2+3*8, 2+1*8, 2+2*8},
-        {2,2,3,3,8,11,8,11,7+2*8, 7+2*8, 7+3*8, 7+3*8, 2+1*8, 2+2*8, 2+1*8, 2+2*8},
-        {0,0,1,1,7,10,7,10,7+0*8, 7+0*8, 7+1*8, 7+1*8, 2+0*8, 2+3*8, 2+0*8, 2+3*8},
-        {0,2,0,2,7,10,7,10,7+0*8, 7+2*8, 7+0*8, 7+2*8, 2+0*8, 2+3*8, 2+0*8, 2+3*8}
+    static const uint8_t left_block_options[4][32]={
+        {0,1,2,3,7,10,8,11,3+0*4, 3+1*4, 3+2*4, 3+3*4, 1+4*4, 1+8*4, 1+5*4, 1+9*4},
+        {2,2,3,3,8,11,8,11,3+2*4, 3+2*4, 3+3*4, 3+3*4, 1+5*4, 1+9*4, 1+5*4, 1+9*4},
+        {0,0,1,1,7,10,7,10,3+0*4, 3+0*4, 3+1*4, 3+1*4, 1+4*4, 1+8*4, 1+4*4, 1+8*4},
+        {0,2,0,2,7,10,7,10,3+0*4, 3+2*4, 3+0*4, 3+2*4, 1+4*4, 1+8*4, 1+4*4, 1+8*4}
     };
 
     h->topleft_partition= -1;
@@ -947,32 +948,41 @@ static void fill_decode_caches(H264Context *h, int mb_type){
 */
 //FIXME constraint_intra_pred & partitioning & nnz (let us hope this is just a typo in the spec)
     if(top_type){
-        AV_COPY32(&h->non_zero_count_cache[4+8*0], &h->non_zero_count[top_xy][4+3*8]);
-            h->non_zero_count_cache[1+8*0]= h->non_zero_count[top_xy][1+1*8];
-            h->non_zero_count_cache[2+8*0]= h->non_zero_count[top_xy][2+1*8];
-
-            h->non_zero_count_cache[1+8*3]= h->non_zero_count[top_xy][1+2*8];
-            h->non_zero_count_cache[2+8*3]= h->non_zero_count[top_xy][2+2*8];
-    }else {
-            h->non_zero_count_cache[1+8*0]=
-            h->non_zero_count_cache[2+8*0]=
-
-            h->non_zero_count_cache[1+8*3]=
-            h->non_zero_count_cache[2+8*3]=
-            AV_WN32A(&h->non_zero_count_cache[4+8*0], CABAC && !IS_INTRA(mb_type) ? 0 : 0x40404040);
+        AV_COPY32(&h->non_zero_count_cache[4+8* 0], &h->non_zero_count[top_xy][4*3]);
+        if(CHROMA444){
+            AV_COPY32(&h->non_zero_count_cache[4+8* 5], &h->non_zero_count[top_xy][4* 7]);
+            AV_COPY32(&h->non_zero_count_cache[4+8*10], &h->non_zero_count[top_xy][4*11]);
+        }else{
+            AV_COPY32(&h->non_zero_count_cache[4+8* 5], &h->non_zero_count[top_xy][4* 5]);
+            AV_COPY32(&h->non_zero_count_cache[4+8*10], &h->non_zero_count[top_xy][4* 9]);
+        }
+    }else{
+        uint32_t top_empty = CABAC && !IS_INTRA(mb_type) ? 0 : 0x40404040;
+        AV_WN32A(&h->non_zero_count_cache[4+8* 0], top_empty);
+        AV_WN32A(&h->non_zero_count_cache[4+8* 5], top_empty);
+        AV_WN32A(&h->non_zero_count_cache[4+8*10], top_empty);
     }
 
     for (i=0; i<2; i++) {
         if(left_type[i]){
-            h->non_zero_count_cache[3+8*1 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+0+2*i]];
-            h->non_zero_count_cache[3+8*2 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+1+2*i]];
-                h->non_zero_count_cache[0+8*1 +   8*i]= h->non_zero_count[left_xy[i]][left_block[8+4+2*i]];
-                h->non_zero_count_cache[0+8*4 +   8*i]= h->non_zero_count[left_xy[i]][left_block[8+5+2*i]];
+            h->non_zero_count_cache[3+8* 1 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+0+2*i]];
+            h->non_zero_count_cache[3+8* 2 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+1+2*i]];
+            if(CHROMA444){
+                h->non_zero_count_cache[3+8* 6 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+0+2*i]+4*4];
+                h->non_zero_count_cache[3+8* 7 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+1+2*i]+4*4];
+                h->non_zero_count_cache[3+8*11 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+0+2*i]+8*4];
+                h->non_zero_count_cache[3+8*12 + 2*8*i]= h->non_zero_count[left_xy[i]][left_block[8+1+2*i]+8*4];
+            }else{
+                h->non_zero_count_cache[3+8* 6 +   8*i]= h->non_zero_count[left_xy[i]][left_block[8+4+2*i]];
+                h->non_zero_count_cache[3+8*11 +   8*i]= h->non_zero_count[left_xy[i]][left_block[8+5+2*i]];
+            }
         }else{
-                h->non_zero_count_cache[3+8*1 + 2*8*i]=
-                h->non_zero_count_cache[3+8*2 + 2*8*i]=
-                h->non_zero_count_cache[0+8*1 +   8*i]=
-                h->non_zero_count_cache[0+8*4 +   8*i]= CABAC && !IS_INTRA(mb_type) ? 0 : 64;
+            h->non_zero_count_cache[3+8* 1 + 2*8*i]=
+            h->non_zero_count_cache[3+8* 2 + 2*8*i]=
+            h->non_zero_count_cache[3+8* 6 + 2*8*i]=
+            h->non_zero_count_cache[3+8* 7 + 2*8*i]=
+            h->non_zero_count_cache[3+8*11 + 2*8*i]=
+            h->non_zero_count_cache[3+8*12 + 2*8*i]= CABAC && !IS_INTRA(mb_type) ? 0 : 64;
         }
     }
 
@@ -981,15 +991,15 @@ static void fill_decode_caches(H264Context *h, int mb_type){
         if(top_type) {
             h->top_cbp = h->cbp_table[top_xy];
         } else {
-            h->top_cbp = IS_INTRA(mb_type) ? 0x1CF : 0x00F;
+            h->top_cbp = IS_INTRA(mb_type) ? 0x7CF : 0x00F;
         }
         // left_cbp
         if (left_type[0]) {
-            h->left_cbp = (h->cbp_table[left_xy[0]] & 0x1f0)
+            h->left_cbp =   (h->cbp_table[left_xy[0]] & 0x7F0)
                         |  ((h->cbp_table[left_xy[0]]>>(left_block[0]&(~1)))&2)
                         | (((h->cbp_table[left_xy[1]]>>(left_block[2]&(~1)))&2) << 2);
         } else {
-            h->left_cbp = IS_INTRA(mb_type) ? 0x1CF : 0x00F;
+            h->left_cbp = IS_INTRA(mb_type) ? 0x7CF : 0x00F;
         }
     }
     }
@@ -1190,11 +1200,21 @@ static inline int pred_intra_mode(H264Context *h, int n){
 static inline void write_back_non_zero_count(H264Context *h){
     const int mb_xy= h->mb_xy;
 
-    AV_COPY64(&h->non_zero_count[mb_xy][ 0], &h->non_zero_count_cache[0+8*1]);
-    AV_COPY64(&h->non_zero_count[mb_xy][ 8], &h->non_zero_count_cache[0+8*2]);
-    AV_COPY32(&h->non_zero_count[mb_xy][16], &h->non_zero_count_cache[0+8*5]);
-    AV_COPY32(&h->non_zero_count[mb_xy][20], &h->non_zero_count_cache[4+8*3]);
-    AV_COPY64(&h->non_zero_count[mb_xy][24], &h->non_zero_count_cache[0+8*4]);
+    AV_COPY32(&h->non_zero_count[mb_xy][ 0], &h->non_zero_count_cache[4+8* 1]);
+    AV_COPY32(&h->non_zero_count[mb_xy][ 4], &h->non_zero_count_cache[4+8* 2]);
+    AV_COPY32(&h->non_zero_count[mb_xy][ 8], &h->non_zero_count_cache[4+8* 3]);
+    AV_COPY32(&h->non_zero_count[mb_xy][12], &h->non_zero_count_cache[4+8* 4]);
+    AV_COPY32(&h->non_zero_count[mb_xy][16], &h->non_zero_count_cache[4+8* 6]);
+    AV_COPY32(&h->non_zero_count[mb_xy][20], &h->non_zero_count_cache[4+8* 7]);
+    AV_COPY32(&h->non_zero_count[mb_xy][32], &h->non_zero_count_cache[4+8*11]);
+    AV_COPY32(&h->non_zero_count[mb_xy][36], &h->non_zero_count_cache[4+8*12]);
+
+    if(CHROMA444){
+        AV_COPY32(&h->non_zero_count[mb_xy][24], &h->non_zero_count_cache[4+8* 8]);
+        AV_COPY32(&h->non_zero_count[mb_xy][28], &h->non_zero_count_cache[4+8* 9]);
+        AV_COPY32(&h->non_zero_count[mb_xy][40], &h->non_zero_count_cache[4+8*13]);
+        AV_COPY32(&h->non_zero_count[mb_xy][44], &h->non_zero_count_cache[4+8*14]);
+    }
 }
 
 static inline void write_back_motion(H264Context *h, int mb_type){
@@ -1267,8 +1287,7 @@ static void av_unused decode_mb_skip(H264Context *h){
     const int mb_xy= h->mb_xy;
     int mb_type=0;
 
-    memset(h->non_zero_count[mb_xy], 0, 32);
-    memset(h->non_zero_count_cache + 8, 0, 8*5); //FIXME ugly, remove pfui
+    memset(h->non_zero_count[mb_xy], 0, 48);
 
     if(MB_FIELD)
         mb_type|= MB_TYPE_INTERLACED;
diff --git a/libavcodec/h264_cabac.c b/libavcodec/h264_cabac.c
index 925ac44498..3975a61699 100644
--- a/libavcodec/h264_cabac.c
+++ b/libavcodec/h264_cabac.c
@@ -45,7 +45,7 @@
 
 /* Cabac pre state table */
 
-static const int8_t cabac_context_init_I[460][2] =
+static const int8_t cabac_context_init_I[1024][2] =
 {
     /* 0 - 10 */
     { 20, -15 }, {  2, 54 },  {  3,  74 }, { 20, -15 },
@@ -211,10 +211,153 @@ static const int8_t cabac_context_init_I[460][2] =
     { -10,  73 }, { -10,  70 }, { -10,  69 }, {  -5,  66 },
     {  -9,  64 }, {  -5,  58 }, {   2,  59 }, {  21, -10 },
     {  24, -11 }, {  28,  -8 }, {  28,  -1 }, {  29,   3 },
-    {  29,   9 }, {  35,  20 }, {  29,  36 }, {  14,  67 }
+    {  29,   9 }, {  35,  20 }, {  29,  36 }, {  14,  67 },
+
+    /* 460 -> 1024 */
+    { -17, 123 }, { -12, 115 }, { -16, 122 }, { -11, 115 },
+    { -12,  63 }, {  -2,  68 }, { -15,  84 }, { -13, 104 },
+    {  -3,  70 }, {  -8,  93 }, { -10,  90 }, { -30, 127 },
+    { -17, 123 }, { -12, 115 }, { -16, 122 }, { -11, 115 },
+    { -12,  63 }, {  -2,  68 }, { -15,  84 }, { -13, 104 },
+    {  -3,  70 }, {  -8,  93 }, { -10,  90 }, { -30, 127 },
+    {  -7,  93 }, { -11,  87 }, {  -3,  77 }, {  -5,  71 },
+    {  -4,  63 }, {  -4,  68 }, { -12,  84 }, {  -7,  62 },
+    {  -7,  65 }, {   8,  61 }, {   5,  56 }, {  -2,  66 },
+    {   1,  64 }, {   0,  61 }, {  -2,  78 }, {   1,  50 },
+    {   7,  52 }, {  10,  35 }, {   0,  44 }, {  11,  38 },
+    {   1,  45 }, {   0,  46 }, {   5,  44 }, {  31,  17 },
+    {   1,  51 }, {   7,  50 }, {  28,  19 }, {  16,  33 },
+    {  14,  62 }, { -13, 108 }, { -15, 100 }, { -13, 101 },
+    { -13,  91 }, { -12,  94 }, { -10,  88 }, { -16,  84 },
+    { -10,  86 }, {  -7,  83 }, { -13,  87 }, { -19,  94 },
+    {   1,  70 }, {   0,  72 }, {  -5,  74 }, {  18,  59 },
+    {  -7,  93 }, { -11,  87 }, {  -3,  77 }, {  -5,  71 },
+    {  -4,  63 }, {  -4,  68 }, { -12,  84 }, {  -7,  62 },
+    {  -7,  65 }, {   8,  61 }, {   5,  56 }, {  -2,  66 },
+    {   1,  64 }, {   0,  61 }, {  -2,  78 }, {   1,  50 },
+    {   7,  52 }, {  10,  35 }, {   0,  44 }, {  11,  38 },
+    {   1,  45 }, {   0,  46 }, {   5,  44 }, {  31,  17 },
+    {   1,  51 }, {   7,  50 }, {  28,  19 }, {  16,  33 },
+    {  14,  62 }, { -13, 108 }, { -15, 100 }, { -13, 101 },
+    { -13,  91 }, { -12,  94 }, { -10,  88 }, { -16,  84 },
+    { -10,  86 }, {  -7,  83 }, { -13,  87 }, { -19,  94 },
+    {   1,  70 }, {   0,  72 }, {  -5,  74 }, {  18,  59 },
+    {  24,   0 }, {  15,   9 }, {   8,  25 }, {  13,  18 },
+    {  15,   9 }, {  13,  19 }, {  10,  37 }, {  12,  18 },
+    {   6,  29 }, {  20,  33 }, {  15,  30 }, {   4,  45 },
+    {   1,  58 }, {   0,  62 }, {   7,  61 }, {  12,  38 },
+    {  11,  45 }, {  15,  39 }, {  11,  42 }, {  13,  44 },
+    {  16,  45 }, {  12,  41 }, {  10,  49 }, {  30,  34 },
+    {  18,  42 }, {  10,  55 }, {  17,  51 }, {  17,  46 },
+    {   0,  89 }, {  26, -19 }, {  22, -17 }, {  26, -17 },
+    {  30, -25 }, {  28, -20 }, {  33, -23 }, {  37, -27 },
+    {  33, -23 }, {  40, -28 }, {  38, -17 }, {  33, -11 },
+    {  40, -15 }, {  41,  -6 }, {  38,   1 }, {  41,  17 },
+    {  24,   0 }, {  15,   9 }, {   8,  25 }, {  13,  18 },
+    {  15,   9 }, {  13,  19 }, {  10,  37 }, {  12,  18 },
+    {   6,  29 }, {  20,  33 }, {  15,  30 }, {   4,  45 },
+    {   1,  58 }, {   0,  62 }, {   7,  61 }, {  12,  38 },
+    {  11,  45 }, {  15,  39 }, {  11,  42 }, {  13,  44 },
+    {  16,  45 }, {  12,  41 }, {  10,  49 }, {  30,  34 },
+    {  18,  42 }, {  10,  55 }, {  17,  51 }, {  17,  46 },
+    {   0,  89 }, {  26, -19 }, {  22, -17 }, {  26, -17 },
+    {  30, -25 }, {  28, -20 }, {  33, -23 }, {  37, -27 },
+    {  33, -23 }, {  40, -28 }, {  38, -17 }, {  33, -11 },
+    {  40, -15 }, {  41,  -6 }, {  38,   1 }, {  41,  17 },
+    { -17, 120 }, { -20, 112 }, { -18, 114 }, { -11,  85 },
+    { -15,  92 }, { -14,  89 }, { -26,  71 }, { -15,  81 },
+    { -14,  80 }, {   0,  68 }, { -14,  70 }, { -24,  56 },
+    { -23,  68 }, { -24,  50 }, { -11,  74 }, { -14, 106 },
+    { -13,  97 }, { -15,  90 }, { -12,  90 }, { -18,  88 },
+    { -10,  73 }, {  -9,  79 }, { -14,  86 }, { -10,  73 },
+    { -10,  70 }, { -10,  69 }, {  -5,  66 }, {  -9,  64 },
+    {  -5,  58 }, {   2,  59 }, {  23, -13 }, {  26, -13 },
+    {  40, -15 }, {  49, -14 }, {  44,   3 }, {  45,   6 },
+    {  44,  34 }, {  33,  54 }, {  19,  82 }, {  21, -10 },
+    {  24, -11 }, {  28,  -8 }, {  28,  -1 }, {  29,   3 },
+    {  29,   9 }, {  35,  20 }, {  29,  36 }, {  14,  67 },
+    {  -3,  75 }, {  -1,  23 }, {   1,  34 }, {   1,  43 },
+    {   0,  54 }, {  -2,  55 }, {   0,  61 }, {   1,  64 },
+    {   0,  68 }, {  -9,  92 }, { -17, 120 }, { -20, 112 },
+    { -18, 114 }, { -11,  85 }, { -15,  92 }, { -14,  89 },
+    { -26,  71 }, { -15,  81 }, { -14,  80 }, {   0,  68 },
+    { -14,  70 }, { -24,  56 }, { -23,  68 }, { -24,  50 },
+    { -11,  74 }, { -14, 106 }, { -13,  97 }, { -15,  90 },
+    { -12,  90 }, { -18,  88 }, { -10,  73 }, {  -9,  79 },
+    { -14,  86 }, { -10,  73 }, { -10,  70 }, { -10,  69 },
+    {  -5,  66 }, {  -9,  64 }, {  -5,  58 }, {   2,  59 },
+    {  23, -13 }, {  26, -13 }, {  40, -15 }, {  49, -14 },
+    {  44,   3 }, {  45,   6 }, {  44,  34 }, {  33,  54 },
+    {  19,  82 }, {  21, -10 }, {  24, -11 }, {  28,  -8 },
+    {  28,  -1 }, {  29,   3 }, {  29,   9 }, {  35,  20 },
+    {  29,  36 }, {  14,  67 }, {  -3,  75 }, {  -1,  23 },
+    {   1,  34 }, {   1,  43 }, {   0,  54 }, {  -2,  55 },
+    {   0,  61 }, {   1,  64 }, {   0,  68 }, {  -9,  92 },
+    {  -6,  93 }, {  -6,  84 }, {  -8,  79 }, {   0,  66 },
+    {  -1,  71 }, {   0,  62 }, {  -2,  60 }, {  -2,  59 },
+    {  -5,  75 }, {  -3,  62 }, {  -4,  58 }, {  -9,  66 },
+    {  -1,  79 }, {   0,  71 }, {   3,  68 }, {  10,  44 },
+    {  -7,  62 }, {  15,  36 }, {  14,  40 }, {  16,  27 },
+    {  12,  29 }, {   1,  44 }, {  20,  36 }, {  18,  32 },
+    {   5,  42 }, {   1,  48 }, {  10,  62 }, {  17,  46 },
+    {   9,  64 }, { -12, 104 }, { -11,  97 }, { -16,  96 },
+    {  -7,  88 }, {  -8,  85 }, {  -7,  85 }, {  -9,  85 },
+    { -13,  88 }, {   4,  66 }, {  -3,  77 }, {  -3,  76 },
+    {  -6,  76 }, {  10,  58 }, {  -1,  76 }, {  -1,  83 },
+    {  -6,  93 }, {  -6,  84 }, {  -8,  79 }, {   0,  66 },
+    {  -1,  71 }, {   0,  62 }, {  -2,  60 }, {  -2,  59 },
+    {  -5,  75 }, {  -3,  62 }, {  -4,  58 }, {  -9,  66 },
+    {  -1,  79 }, {   0,  71 }, {   3,  68 }, {  10,  44 },
+    {  -7,  62 }, {  15,  36 }, {  14,  40 }, {  16,  27 },
+    {  12,  29 }, {   1,  44 }, {  20,  36 }, {  18,  32 },
+    {   5,  42 }, {   1,  48 }, {  10,  62 }, {  17,  46 },
+    {   9,  64 }, { -12, 104 }, { -11,  97 }, { -16,  96 },
+    {  -7,  88 }, {  -8,  85 }, {  -7,  85 }, {  -9,  85 },
+    { -13,  88 }, {   4,  66 }, {  -3,  77 }, {  -3,  76 },
+    {  -6,  76 }, {  10,  58 }, {  -1,  76 }, {  -1,  83 },
+    {  15,   6 }, {   6,  19 }, {   7,  16 }, {  12,  14 },
+    {  18,  13 }, {  13,  11 }, {  13,  15 }, {  15,  16 },
+    {  12,  23 }, {  13,  23 }, {  15,  20 }, {  14,  26 },
+    {  14,  44 }, {  17,  40 }, {  17,  47 }, {  24,  17 },
+    {  21,  21 }, {  25,  22 }, {  31,  27 }, {  22,  29 },
+    {  19,  35 }, {  14,  50 }, {  10,  57 }, {   7,  63 },
+    {  -2,  77 }, {  -4,  82 }, {  -3,  94 }, {   9,  69 },
+    { -12, 109 }, {  36, -35 }, {  36, -34 }, {  32, -26 },
+    {  37, -30 }, {  44, -32 }, {  34, -18 }, {  34, -15 },
+    {  40, -15 }, {  33,  -7 }, {  35,  -5 }, {  33,   0 },
+    {  38,   2 }, {  33,  13 }, {  23,  35 }, {  13,  58 },
+    {  15,   6 }, {   6,  19 }, {   7,  16 }, {  12,  14 },
+    {  18,  13 }, {  13,  11 }, {  13,  15 }, {  15,  16 },
+    {  12,  23 }, {  13,  23 }, {  15,  20 }, {  14,  26 },
+    {  14,  44 }, {  17,  40 }, {  17,  47 }, {  24,  17 },
+    {  21,  21 }, {  25,  22 }, {  31,  27 }, {  22,  29 },
+    {  19,  35 }, {  14,  50 }, {  10,  57 }, {   7,  63 },
+    {  -2,  77 }, {  -4,  82 }, {  -3,  94 }, {   9,  69 },
+    { -12, 109 }, {  36, -35 }, {  36, -34 }, {  32, -26 },
+    {  37, -30 }, {  44, -32 }, {  34, -18 }, {  34, -15 },
+    {  40, -15 }, {  33,  -7 }, {  35,  -5 }, {  33,   0 },
+    {  38,   2 }, {  33,  13 }, {  23,  35 }, {  13,  58 },
+    {  -3,  71 }, {  -6,  42 }, {  -5,  50 }, {  -3,  54 },
+    {  -2,  62 }, {   0,  58 }, {   1,  63 }, {  -2,  72 },
+    {  -1,  74 }, {  -9,  91 }, {  -5,  67 }, {  -5,  27 },
+    {  -3,  39 }, {  -2,  44 }, {   0,  46 }, { -16,  64 },
+    {  -8,  68 }, { -10,  78 }, {  -6,  77 }, { -10,  86 },
+    { -12,  92 }, { -15,  55 }, { -10,  60 }, {  -6,  62 },
+    {  -4,  65 }, { -12,  73 }, {  -8,  76 }, {  -7,  80 },
+    {  -9,  88 }, { -17, 110 }, {  -3,  71 }, {  -6,  42 },
+    {  -5,  50 }, {  -3,  54 }, {  -2,  62 }, {   0,  58 },
+    {   1,  63 }, {  -2,  72 }, {  -1,  74 }, {  -9,  91 },
+    {  -5,  67 }, {  -5,  27 }, {  -3,  39 }, {  -2,  44 },
+    {   0,  46 }, { -16,  64 }, {  -8,  68 }, { -10,  78 },
+    {  -6,  77 }, { -10,  86 }, { -12,  92 }, { -15,  55 },
+    { -10,  60 }, {  -6,  62 }, {  -4,  65 }, { -12,  73 },
+    {  -8,  76 }, {  -7,  80 }, {  -9,  88 }, { -17, 110 },
+    {  -3,  70 }, {  -8,  93 }, { -10,  90 }, { -30, 127 },
+    {  -3,  70 }, {  -8,  93 }, { -10,  90 }, { -30, 127 },
+    {  -3,  70 }, {  -8,  93 }, { -10,  90 }, { -30, 127 }
 };
 
-static const int8_t cabac_context_init_PB[3][460][2] =
+static const int8_t cabac_context_init_PB[3][1024][2] =
 {
     /* i_cabac_init_idc == 0 */
     {
@@ -370,6 +513,149 @@ static const int8_t cabac_context_init_PB[3][460][2] =
         { -14,  66 }, {   0,  59 }, {   2,  59 }, {  21, -13 },
         {  33, -14 }, {  39,  -7 }, {  46,  -2 }, {  51,   2 },
         {  60,   6 }, {  61,  17 }, {  55,  34 }, {  42,  62 },
+
+        /* 460 - 1024 */
+        {  -7,  92 }, {  -5,  89 }, {  -7,  96 }, { -13, 108 },
+        {  -3,  46 }, {  -1,  65 }, {  -1,  57 }, {  -9,  93 },
+        {  -3,  74 }, {  -9,  92 }, {  -8,  87 }, { -23, 126 },
+        {  -7,  92 }, {  -5,  89 }, {  -7,  96 }, { -13, 108 },
+        {  -3,  46 }, {  -1,  65 }, {  -1,  57 }, {  -9,  93 },
+        {  -3,  74 }, {  -9,  92 }, {  -8,  87 }, { -23, 126 },
+        {  -2,  85 }, {  -6,  78 }, {  -1,  75 }, {  -7,  77 },
+        {   2,  54 }, {   5,  50 }, {  -3,  68 }, {   1,  50 },
+        {   6,  42 }, {  -4,  81 }, {   1,  63 }, {  -4,  70 },
+        {   0,  67 }, {   2,  57 }, {  -2,  76 }, {  11,  35 },
+        {   4,  64 }, {   1,  61 }, {  11,  35 }, {  18,  25 },
+        {  12,  24 }, {  13,  29 }, {  13,  36 }, { -10,  93 },
+        {  -7,  73 }, {  -2,  73 }, {  13,  46 }, {   9,  49 },
+        {  -7, 100 }, {   9,  53 }, {   2,  53 }, {   5,  53 },
+        {  -2,  61 }, {   0,  56 }, {   0,  56 }, { -13,  63 },
+        {  -5,  60 }, {  -1,  62 }, {   4,  57 }, {  -6,  69 },
+        {   4,  57 }, {  14,  39 }, {   4,  51 }, {  13,  68 },
+        {  -2,  85 }, {  -6,  78 }, {  -1,  75 }, {  -7,  77 },
+        {   2,  54 }, {   5,  50 }, {  -3,  68 }, {   1,  50 },
+        {   6,  42 }, {  -4,  81 }, {   1,  63 }, {  -4,  70 },
+        {   0,  67 }, {   2,  57 }, {  -2,  76 }, {  11,  35 },
+        {   4,  64 }, {   1,  61 }, {  11,  35 }, {  18,  25 },
+        {  12,  24 }, {  13,  29 }, {  13,  36 }, { -10,  93 },
+        {  -7,  73 }, {  -2,  73 }, {  13,  46 }, {   9,  49 },
+        {  -7, 100 }, {   9,  53 }, {   2,  53 }, {   5,  53 },
+        {  -2,  61 }, {   0,  56 }, {   0,  56 }, { -13,  63 },
+        {  -5,  60 }, {  -1,  62 }, {   4,  57 }, {  -6,  69 },
+        {   4,  57 }, {  14,  39 }, {   4,  51 }, {  13,  68 },
+        {  11,  28 }, {   2,  40 }, {   3,  44 }, {   0,  49 },
+        {   0,  46 }, {   2,  44 }, {   2,  51 }, {   0,  47 },
+        {   4,  39 }, {   2,  62 }, {   6,  46 }, {   0,  54 },
+        {   3,  54 }, {   2,  58 }, {   4,  63 }, {   6,  51 },
+        {   6,  57 }, {   7,  53 }, {   6,  52 }, {   6,  55 },
+        {  11,  45 }, {  14,  36 }, {   8,  53 }, {  -1,  82 },
+        {   7,  55 }, {  -3,  78 }, {  15,  46 }, {  22,  31 },
+        {  -1,  84 }, {  25,   7 }, {  30,  -7 }, {  28,   3 },
+        {  28,   4 }, {  32,   0 }, {  34,  -1 }, {  30,   6 },
+        {  30,   6 }, {  32,   9 }, {  31,  19 }, {  26,  27 },
+        {  26,  30 }, {  37,  20 }, {  28,  34 }, {  17,  70 },
+        {  11,  28 }, {   2,  40 }, {   3,  44 }, {   0,  49 },
+        {   0,  46 }, {   2,  44 }, {   2,  51 }, {   0,  47 },
+        {   4,  39 }, {   2,  62 }, {   6,  46 }, {   0,  54 },
+        {   3,  54 }, {   2,  58 }, {   4,  63 }, {   6,  51 },
+        {   6,  57 }, {   7,  53 }, {   6,  52 }, {   6,  55 },
+        {  11,  45 }, {  14,  36 }, {   8,  53 }, {  -1,  82 },
+        {   7,  55 }, {  -3,  78 }, {  15,  46 }, {  22,  31 },
+        {  -1,  84 }, {  25,   7 }, {  30,  -7 }, {  28,   3 },
+        {  28,   4 }, {  32,   0 }, {  34,  -1 }, {  30,   6 },
+        {  30,   6 }, {  32,   9 }, {  31,  19 }, {  26,  27 },
+        {  26,  30 }, {  37,  20 }, {  28,  34 }, {  17,  70 },
+        {  -4,  79 }, {  -7,  71 }, {  -5,  69 }, {  -9,  70 },
+        {  -8,  66 }, { -10,  68 }, { -19,  73 }, { -12,  69 },
+        { -16,  70 }, { -15,  67 }, { -20,  62 }, { -19,  70 },
+        { -16,  66 }, { -22,  65 }, { -20,  63 }, {  -5,  85 },
+        {  -6,  81 }, { -10,  77 }, {  -7,  81 }, { -17,  80 },
+        { -18,  73 }, {  -4,  74 }, { -10,  83 }, {  -9,  71 },
+        {  -9,  67 }, {  -1,  61 }, {  -8,  66 }, { -14,  66 },
+        {   0,  59 }, {   2,  59 }, {   9,  -2 }, {  26,  -9 },
+        {  33,  -9 }, {  39,  -7 }, {  41,  -2 }, {  45,   3 },
+        {  49,   9 }, {  45,  27 }, {  36,  59 }, {  21, -13 },
+        {  33, -14 }, {  39,  -7 }, {  46,  -2 }, {  51,   2 },
+        {  60,   6 }, {  61,  17 }, {  55,  34 }, {  42,  62 },
+        {  -6,  66 }, {  -7,  35 }, {  -7,  42 }, {  -8,  45 },
+        {  -5,  48 }, { -12,  56 }, {  -6,  60 }, {  -5,  62 },
+        {  -8,  66 }, {  -8,  76 }, {  -4,  79 }, {  -7,  71 },
+        {  -5,  69 }, {  -9,  70 }, {  -8,  66 }, { -10,  68 },
+        { -19,  73 }, { -12,  69 }, { -16,  70 }, { -15,  67 },
+        { -20,  62 }, { -19,  70 }, { -16,  66 }, { -22,  65 },
+        { -20,  63 }, {  -5,  85 }, {  -6,  81 }, { -10,  77 },
+        {  -7,  81 }, { -17,  80 }, { -18,  73 }, {  -4,  74 },
+        { -10,  83 }, {  -9,  71 }, {  -9,  67 }, {  -1,  61 },
+        {  -8,  66 }, { -14,  66 }, {   0,  59 }, {   2,  59 },
+        {   9,  -2 }, {  26,  -9 }, {  33,  -9 }, {  39,  -7 },
+        {  41,  -2 }, {  45,   3 }, {  49,   9 }, {  45,  27 },
+        {  36,  59 }, {  21, -13 }, {  33, -14 }, {  39,  -7 },
+        {  46,  -2 }, {  51,   2 }, {  60,   6 }, {  61,  17 },
+        {  55,  34 }, {  42,  62 }, {  -6,  66 }, {  -7,  35 },
+        {  -7,  42 }, {  -8,  45 }, {  -5,  48 }, { -12,  56 },
+        {  -6,  60 }, {  -5,  62 }, {  -8,  66 }, {  -8,  76 },
+        { -13, 106 }, { -16, 106 }, { -10,  87 }, { -21, 114 },
+        { -18, 110 }, { -14,  98 }, { -22, 110 }, { -21, 106 },
+        { -18, 103 }, { -21, 107 }, { -23, 108 }, { -26, 112 },
+        { -10,  96 }, { -12,  95 }, {  -5,  91 }, {  -9,  93 },
+        { -22,  94 }, {  -5,  86 }, {   9,  67 }, {  -4,  80 },
+        { -10,  85 }, {  -1,  70 }, {   7,  60 }, {   9,  58 },
+        {   5,  61 }, {  12,  50 }, {  15,  50 }, {  18,  49 },
+        {  17,  54 }, {  10,  41 }, {   7,  46 }, {  -1,  51 },
+        {   7,  49 }, {   8,  52 }, {   9,  41 }, {   6,  47 },
+        {   2,  55 }, {  13,  41 }, {  10,  44 }, {   6,  50 },
+        {   5,  53 }, {  13,  49 }, {   4,  63 }, {   6,  64 },
+        { -13, 106 }, { -16, 106 }, { -10,  87 }, { -21, 114 },
+        { -18, 110 }, { -14,  98 }, { -22, 110 }, { -21, 106 },
+        { -18, 103 }, { -21, 107 }, { -23, 108 }, { -26, 112 },
+        { -10,  96 }, { -12,  95 }, {  -5,  91 }, {  -9,  93 },
+        { -22,  94 }, {  -5,  86 }, {   9,  67 }, {  -4,  80 },
+        { -10,  85 }, {  -1,  70 }, {   7,  60 }, {   9,  58 },
+        {   5,  61 }, {  12,  50 }, {  15,  50 }, {  18,  49 },
+        {  17,  54 }, {  10,  41 }, {   7,  46 }, {  -1,  51 },
+        {   7,  49 }, {   8,  52 }, {   9,  41 }, {   6,  47 },
+        {   2,  55 }, {  13,  41 }, {  10,  44 }, {   6,  50 },
+        {   5,  53 }, {  13,  49 }, {   4,  63 }, {   6,  64 },
+        {  14,  11 }, {  11,  14 }, {   9,  11 }, {  18,  11 },
+        {  21,   9 }, {  23,  -2 }, {  32, -15 }, {  32, -15 },
+        {  34, -21 }, {  39, -23 }, {  42, -33 }, {  41, -31 },
+        {  46, -28 }, {  38, -12 }, {  21,  29 }, {  45, -24 },
+        {  53, -45 }, {  48, -26 }, {  65, -43 }, {  43, -19 },
+        {  39, -10 }, {  30,   9 }, {  18,  26 }, {  20,  27 },
+        {   0,  57 }, { -14,  82 }, {  -5,  75 }, { -19,  97 },
+        { -35, 125 }, {  27,   0 }, {  28,   0 }, {  31,  -4 },
+        {  27,   6 }, {  34,   8 }, {  30,  10 }, {  24,  22 },
+        {  33,  19 }, {  22,  32 }, {  26,  31 }, {  21,  41 },
+        {  26,  44 }, {  23,  47 }, {  16,  65 }, {  14,  71 },
+        {  14,  11 }, {  11,  14 }, {   9,  11 }, {  18,  11 },
+        {  21,   9 }, {  23,  -2 }, {  32, -15 }, {  32, -15 },
+        {  34, -21 }, {  39, -23 }, {  42, -33 }, {  41, -31 },
+        {  46, -28 }, {  38, -12 }, {  21,  29 }, {  45, -24 },
+        {  53, -45 }, {  48, -26 }, {  65, -43 }, {  43, -19 },
+        {  39, -10 }, {  30,   9 }, {  18,  26 }, {  20,  27 },
+        {   0,  57 }, { -14,  82 }, {  -5,  75 }, { -19,  97 },
+        { -35, 125 }, {  27,   0 }, {  28,   0 }, {  31,  -4 },
+        {  27,   6 }, {  34,   8 }, {  30,  10 }, {  24,  22 },
+        {  33,  19 }, {  22,  32 }, {  26,  31 }, {  21,  41 },
+        {  26,  44 }, {  23,  47 }, {  16,  65 }, {  14,  71 },
+        {  -6,  76 }, {  -2,  44 }, {   0,  45 }, {   0,  52 },
+        {  -3,  64 }, {  -2,  59 }, {  -4,  70 }, {  -4,  75 },
+        {  -8,  82 }, { -17, 102 }, {  -9,  77 }, {   3,  24 },
+        {   0,  42 }, {   0,  48 }, {   0,  55 }, {  -6,  59 },
+        {  -7,  71 }, { -12,  83 }, { -11,  87 }, { -30, 119 },
+        {   1,  58 }, {  -3,  29 }, {  -1,  36 }, {   1,  38 },
+        {   2,  43 }, {  -6,  55 }, {   0,  58 }, {   0,  64 },
+        {  -3,  74 }, { -10,  90 }, {  -6,  76 }, {  -2,  44 },
+        {   0,  45 }, {   0,  52 }, {  -3,  64 }, {  -2,  59 },
+        {  -4,  70 }, {  -4,  75 }, {  -8,  82 }, { -17, 102 },
+        {  -9,  77 }, {   3,  24 }, {   0,  42 }, {   0,  48 },
+        {   0,  55 }, {  -6,  59 }, {  -7,  71 }, { -12,  83 },
+        { -11,  87 }, { -30, 119 }, {   1,  58 }, {  -3,  29 },
+        {  -1,  36 }, {   1,  38 }, {   2,  43 }, {  -6,  55 },
+        {   0,  58 }, {   0,  64 }, {  -3,  74 }, { -10,  90 },
+        {  -3,  74 }, {  -9,  92 }, {  -8,  87 }, { -23, 126 },
+        {  -3,  74 }, {  -9,  92 }, {  -8,  87 }, { -23, 126 },
+        {  -3,  74 }, {  -9,  92 }, {  -8,  87 }, { -23, 126 }
     },
 
     /* i_cabac_init_idc == 1 */
@@ -526,6 +812,149 @@ static const int8_t cabac_context_init_PB[3][460][2] =
         {  -9,  60 }, {   1,  54 }, {   2,  58 }, {  17, -10 },
         {  32, -13 }, {  42,  -9 }, {  49,  -5 }, {  53,   0 },
         {  64,   3 }, {  68,  10 }, {  66,  27 }, {  47,  57 },
+
+        /* 460 - 1024 */
+        {   0,  80 }, {  -5,  89 }, {  -7,  94 }, {  -4,  92 },
+        {   0,  39 }, {   0,  65 }, { -15,  84 }, { -35, 127 },
+        {  -2,  73 }, { -12, 104 }, {  -9,  91 }, { -31, 127 },
+        {   0,  80 }, {  -5,  89 }, {  -7,  94 }, {  -4,  92 },
+        {   0,  39 }, {   0,  65 }, { -15,  84 }, { -35, 127 },
+        {  -2,  73 }, { -12, 104 }, {  -9,  91 }, { -31, 127 },
+        { -13, 103 }, { -13,  91 }, {  -9,  89 }, { -14,  92 },
+        {  -8,  76 }, { -12,  87 }, { -23, 110 }, { -24, 105 },
+        { -10,  78 }, { -20, 112 }, { -17,  99 }, { -78, 127 },
+        { -70, 127 }, { -50, 127 }, { -46, 127 }, {  -4,  66 },
+        {  -5,  78 }, {  -4,  71 }, {  -8,  72 }, {   2,  59 },
+        {  -1,  55 }, {  -7,  70 }, {  -6,  75 }, {  -8,  89 },
+        { -34, 119 }, {  -3,  75 }, {  32,  20 }, {  30,  22 },
+        { -44, 127 }, {   0,  54 }, {  -5,  61 }, {   0,  58 },
+        {  -1,  60 }, {  -3,  61 }, {  -8,  67 }, { -25,  84 },
+        { -14,  74 }, {  -5,  65 }, {   5,  52 }, {   2,  57 },
+        {   0,  61 }, {  -9,  69 }, { -11,  70 }, {  18,  55 },
+        { -13, 103 }, { -13,  91 }, {  -9,  89 }, { -14,  92 },
+        {  -8,  76 }, { -12,  87 }, { -23, 110 }, { -24, 105 },
+        { -10,  78 }, { -20, 112 }, { -17,  99 }, { -78, 127 },
+        { -70, 127 }, { -50, 127 }, { -46, 127 }, {  -4,  66 },
+        {  -5,  78 }, {  -4,  71 }, {  -8,  72 }, {   2,  59 },
+        {  -1,  55 }, {  -7,  70 }, {  -6,  75 }, {  -8,  89 },
+        { -34, 119 }, {  -3,  75 }, {  32,  20 }, {  30,  22 },
+        { -44, 127 }, {   0,  54 }, {  -5,  61 }, {   0,  58 },
+        {  -1,  60 }, {  -3,  61 }, {  -8,  67 }, { -25,  84 },
+        { -14,  74 }, {  -5,  65 }, {   5,  52 }, {   2,  57 },
+        {   0,  61 }, {  -9,  69 }, { -11,  70 }, {  18,  55 },
+        {   4,  45 }, {  10,  28 }, {  10,  31 }, {  33, -11 },
+        {  52, -43 }, {  18,  15 }, {  28,   0 }, {  35, -22 },
+        {  38, -25 }, {  34,   0 }, {  39, -18 }, {  32, -12 },
+        { 102, -94 }, {   0,   0 }, {  56, -15 }, {  33,  -4 },
+        {  29,  10 }, {  37,  -5 }, {  51, -29 }, {  39,  -9 },
+        {  52, -34 }, {  69, -58 }, {  67, -63 }, {  44,  -5 },
+        {  32,   7 }, {  55, -29 }, {  32,   1 }, {   0,   0 },
+        {  27,  36 }, {  33, -25 }, {  34, -30 }, {  36, -28 },
+        {  38, -28 }, {  38, -27 }, {  34, -18 }, {  35, -16 },
+        {  34, -14 }, {  32,  -8 }, {  37,  -6 }, {  35,   0 },
+        {  30,  10 }, {  28,  18 }, {  26,  25 }, {  29,  41 },
+        {   4,  45 }, {  10,  28 }, {  10,  31 }, {  33, -11 },
+        {  52, -43 }, {  18,  15 }, {  28,   0 }, {  35, -22 },
+        {  38, -25 }, {  34,   0 }, {  39, -18 }, {  32, -12 },
+        { 102, -94 }, {   0,   0 }, {  56, -15 }, {  33,  -4 },
+        {  29,  10 }, {  37,  -5 }, {  51, -29 }, {  39,  -9 },
+        {  52, -34 }, {  69, -58 }, {  67, -63 }, {  44,  -5 },
+        {  32,   7 }, {  55, -29 }, {  32,   1 }, {   0,   0 },
+        {  27,  36 }, {  33, -25 }, {  34, -30 }, {  36, -28 },
+        {  38, -28 }, {  38, -27 }, {  34, -18 }, {  35, -16 },
+        {  34, -14 }, {  32,  -8 }, {  37,  -6 }, {  35,   0 },
+        {  30,  10 }, {  28,  18 }, {  26,  25 }, {  29,  41 },
+        {  -5,  85 }, {  -6,  81 }, { -10,  77 }, {  -7,  81 },
+        { -17,  80 }, { -18,  73 }, {  -4,  74 }, { -10,  83 },
+        {  -9,  71 }, {  -9,  67 }, {  -1,  61 }, {  -8,  66 },
+        { -14,  66 }, {   0,  59 }, {   2,  59 }, {  -3,  81 },
+        {  -3,  76 }, {  -7,  72 }, {  -6,  78 }, { -12,  72 },
+        { -14,  68 }, {  -3,  70 }, {  -6,  76 }, {  -5,  66 },
+        {  -5,  62 }, {   0,  57 }, {  -4,  61 }, {  -9,  60 },
+        {   1,  54 }, {   2,  58 }, {  17, -10 }, {  32, -13 },
+        {  42,  -9 }, {  49,  -5 }, {  53,   0 }, {  64,   3 },
+        {  68,  10 }, {  66,  27 }, {  47,  57 }, {  17, -10 },
+        {  32, -13 }, {  42,  -9 }, {  49,  -5 }, {  53,   0 },
+        {  64,   3 }, {  68,  10 }, {  66,  27 }, {  47,  57 },
+        {  -5,  71 }, {   0,  24 }, {  -1,  36 }, {  -2,  42 },
+        {  -2,  52 }, {  -9,  57 }, {  -6,  63 }, {  -4,  65 },
+        {  -4,  67 }, {  -7,  82 }, {  -5,  85 }, {  -6,  81 },
+        { -10,  77 }, {  -7,  81 }, { -17,  80 }, { -18,  73 },
+        {  -4,  74 }, { -10,  83 }, {  -9,  71 }, {  -9,  67 },
+        {  -1,  61 }, {  -8,  66 }, { -14,  66 }, {   0,  59 },
+        {   2,  59 }, {  -3,  81 }, {  -3,  76 }, {  -7,  72 },
+        {  -6,  78 }, { -12,  72 }, { -14,  68 }, {  -3,  70 },
+        {  -6,  76 }, {  -5,  66 }, {  -5,  62 }, {   0,  57 },
+        {  -4,  61 }, {  -9,  60 }, {   1,  54 }, {   2,  58 },
+        {  17, -10 }, {  32, -13 }, {  42,  -9 }, {  49,  -5 },
+        {  53,   0 }, {  64,   3 }, {  68,  10 }, {  66,  27 },
+        {  47,  57 }, {  17, -10 }, {  32, -13 }, {  42,  -9 },
+        {  49,  -5 }, {  53,   0 }, {  64,   3 }, {  68,  10 },
+        {  66,  27 }, {  47,  57 }, {  -5,  71 }, {   0,  24 },
+        {  -1,  36 }, {  -2,  42 }, {  -2,  52 }, {  -9,  57 },
+        {  -6,  63 }, {  -4,  65 }, {  -4,  67 }, {  -7,  82 },
+        { -21, 126 }, { -23, 124 }, { -20, 110 }, { -26, 126 },
+        { -25, 124 }, { -17, 105 }, { -27, 121 }, { -27, 117 },
+        { -17, 102 }, { -26, 117 }, { -27, 116 }, { -33, 122 },
+        { -10,  95 }, { -14, 100 }, {  -8,  95 }, { -17, 111 },
+        { -28, 114 }, {  -6,  89 }, {  -2,  80 }, {  -4,  82 },
+        {  -9,  85 }, {  -8,  81 }, {  -1,  72 }, {   5,  64 },
+        {   1,  67 }, {   9,  56 }, {   0,  69 }, {   1,  69 },
+        {   7,  69 }, {  -7,  69 }, {  -6,  67 }, { -16,  77 },
+        {  -2,  64 }, {   2,  61 }, {  -6,  67 }, {  -3,  64 },
+        {   2,  57 }, {  -3,  65 }, {  -3,  66 }, {   0,  62 },
+        {   9,  51 }, {  -1,  66 }, {  -2,  71 }, {  -2,  75 },
+        { -21, 126 }, { -23, 124 }, { -20, 110 }, { -26, 126 },
+        { -25, 124 }, { -17, 105 }, { -27, 121 }, { -27, 117 },
+        { -17, 102 }, { -26, 117 }, { -27, 116 }, { -33, 122 },
+        { -10,  95 }, { -14, 100 }, {  -8,  95 }, { -17, 111 },
+        { -28, 114 }, {  -6,  89 }, {  -2,  80 }, {  -4,  82 },
+        {  -9,  85 }, {  -8,  81 }, {  -1,  72 }, {   5,  64 },
+        {   1,  67 }, {   9,  56 }, {   0,  69 }, {   1,  69 },
+        {   7,  69 }, {  -7,  69 }, {  -6,  67 }, { -16,  77 },
+        {  -2,  64 }, {   2,  61 }, {  -6,  67 }, {  -3,  64 },
+        {   2,  57 }, {  -3,  65 }, {  -3,  66 }, {   0,  62 },
+        {   9,  51 }, {  -1,  66 }, {  -2,  71 }, {  -2,  75 },
+        {  19,  -6 }, {  18,  -6 }, {  14,   0 }, {  26, -12 },
+        {  31, -16 }, {  33, -25 }, {  33, -22 }, {  37, -28 },
+        {  39, -30 }, {  42, -30 }, {  47, -42 }, {  45, -36 },
+        {  49, -34 }, {  41, -17 }, {  32,   9 }, {  69, -71 },
+        {  63, -63 }, {  66, -64 }, {  77, -74 }, {  54, -39 },
+        {  52, -35 }, {  41, -10 }, {  36,   0 }, {  40,  -1 },
+        {  30,  14 }, {  28,  26 }, {  23,  37 }, {  12,  55 },
+        {  11,  65 }, {  37, -33 }, {  39, -36 }, {  40, -37 },
+        {  38, -30 }, {  46, -33 }, {  42, -30 }, {  40, -24 },
+        {  49, -29 }, {  38, -12 }, {  40, -10 }, {  38,  -3 },
+        {  46,  -5 }, {  31,  20 }, {  29,  30 }, {  25,  44 },
+        {  19,  -6 }, {  18,  -6 }, {  14,   0 }, {  26, -12 },
+        {  31, -16 }, {  33, -25 }, {  33, -22 }, {  37, -28 },
+        {  39, -30 }, {  42, -30 }, {  47, -42 }, {  45, -36 },
+        {  49, -34 }, {  41, -17 }, {  32,   9 }, {  69, -71 },
+        {  63, -63 }, {  66, -64 }, {  77, -74 }, {  54, -39 },
+        {  52, -35 }, {  41, -10 }, {  36,   0 }, {  40,  -1 },
+        {  30,  14 }, {  28,  26 }, {  23,  37 }, {  12,  55 },
+        {  11,  65 }, {  37, -33 }, {  39, -36 }, {  40, -37 },
+        {  38, -30 }, {  46, -33 }, {  42, -30 }, {  40, -24 },
+        {  49, -29 }, {  38, -12 }, {  40, -10 }, {  38,  -3 },
+        {  46,  -5 }, {  31,  20 }, {  29,  30 }, {  25,  44 },
+        { -23, 112 }, { -15,  71 }, {  -7,  61 }, {   0,  53 },
+        {  -5,  66 }, { -11,  77 }, {  -9,  80 }, {  -9,  84 },
+        { -10,  87 }, { -34, 127 }, { -21, 101 }, {  -3,  39 },
+        {  -5,  53 }, {  -7,  61 }, { -11,  75 }, { -15,  77 },
+        { -17,  91 }, { -25, 107 }, { -25, 111 }, { -28, 122 },
+        { -11,  76 }, { -10,  44 }, { -10,  52 }, { -10,  57 },
+        {  -9,  58 }, { -16,  72 }, {  -7,  69 }, {  -4,  69 },
+        {  -5,  74 }, {  -9,  86 }, { -23, 112 }, { -15,  71 },
+        {  -7,  61 }, {   0,  53 }, {  -5,  66 }, { -11,  77 },
+        {  -9,  80 }, {  -9,  84 }, { -10,  87 }, { -34, 127 },
+        { -21, 101 }, {  -3,  39 }, {  -5,  53 }, {  -7,  61 },
+        { -11,  75 }, { -15,  77 }, { -17,  91 }, { -25, 107 },
+        { -25, 111 }, { -28, 122 }, { -11,  76 }, { -10,  44 },
+        { -10,  52 }, { -10,  57 }, {  -9,  58 }, { -16,  72 },
+        {  -7,  69 }, {  -4,  69 }, {  -5,  74 }, {  -9,  86 },
+        {  -2,  73 }, { -12, 104 }, {  -9,  91 }, { -31, 127 },
+        {  -2,  73 }, { -12, 104 }, {  -9,  91 }, { -31, 127 },
+        {  -2,  73 }, { -12, 104 }, {  -9,  91 }, { -31, 127 }
     },
 
     /* i_cabac_init_idc == 2 */
@@ -682,6 +1111,149 @@ static const int8_t cabac_context_init_PB[3][460][2] =
         { -14,  59 }, {  -9,  52 }, { -11,  68 }, {   9,  -2 },
         {  30, -10 }, {  31,  -4 }, {  33,  -1 }, {  33,   7 },
         {  31,  12 }, {  37,  23 }, {  31,  38 }, {  20,  64 },
+
+        /* 460 - 1024 */
+        {  11,  80 }, {   5,  76 }, {   2,  84 }, {   5,  78 },
+        {  -6,  55 }, {   4,  61 }, { -14,  83 }, { -37, 127 },
+        {  -5,  79 }, { -11, 104 }, { -11,  91 }, { -30, 127 },
+        {  11,  80 }, {   5,  76 }, {   2,  84 }, {   5,  78 },
+        {  -6,  55 }, {   4,  61 }, { -14,  83 }, { -37, 127 },
+        {  -5,  79 }, { -11, 104 }, { -11,  91 }, { -30, 127 },
+        {  -4,  86 }, { -12,  88 }, {  -5,  82 }, {  -3,  72 },
+        {  -4,  67 }, {  -8,  72 }, { -16,  89 }, {  -9,  69 },
+        {  -1,  59 }, {   5,  66 }, {   4,  57 }, {  -4,  71 },
+        {  -2,  71 }, {   2,  58 }, {  -1,  74 }, {  -4,  44 },
+        {  -1,  69 }, {   0,  62 }, {  -7,  51 }, {  -4,  47 },
+        {  -6,  42 }, {  -3,  41 }, {  -6,  53 }, {   8,  76 },
+        {  -9,  78 }, { -11,  83 }, {   9,  52 }, {   0,  67 },
+        {  -5,  90 }, {   1,  67 }, { -15,  72 }, {  -5,  75 },
+        {  -8,  80 }, { -21,  83 }, { -21,  64 }, { -13,  31 },
+        { -25,  64 }, { -29,  94 }, {   9,  75 }, {  17,  63 },
+        {  -8,  74 }, {  -5,  35 }, {  -2,  27 }, {  13,  91 },
+        {  -4,  86 }, { -12,  88 }, {  -5,  82 }, {  -3,  72 },
+        {  -4,  67 }, {  -8,  72 }, { -16,  89 }, {  -9,  69 },
+        {  -1,  59 }, {   5,  66 }, {   4,  57 }, {  -4,  71 },
+        {  -2,  71 }, {   2,  58 }, {  -1,  74 }, {  -4,  44 },
+        {  -1,  69 }, {   0,  62 }, {  -7,  51 }, {  -4,  47 },
+        {  -6,  42 }, {  -3,  41 }, {  -6,  53 }, {   8,  76 },
+        {  -9,  78 }, { -11,  83 }, {   9,  52 }, {   0,  67 },
+        {  -5,  90 }, {   1,  67 }, { -15,  72 }, {  -5,  75 },
+        {  -8,  80 }, { -21,  83 }, { -21,  64 }, { -13,  31 },
+        { -25,  64 }, { -29,  94 }, {   9,  75 }, {  17,  63 },
+        {  -8,  74 }, {  -5,  35 }, {  -2,  27 }, {  13,  91 },
+        {   4,  39 }, {   0,  42 }, {   7,  34 }, {  11,  29 },
+        {   8,  31 }, {   6,  37 }, {   7,  42 }, {   3,  40 },
+        {   8,  33 }, {  13,  43 }, {  13,  36 }, {   4,  47 },
+        {   3,  55 }, {   2,  58 }, {   6,  60 }, {   8,  44 },
+        {  11,  44 }, {  14,  42 }, {   7,  48 }, {   4,  56 },
+        {   4,  52 }, {  13,  37 }, {   9,  49 }, {  19,  58 },
+        {  10,  48 }, {  12,  45 }, {   0,  69 }, {  20,  33 },
+        {   8,  63 }, {  35, -18 }, {  33, -25 }, {  28,  -3 },
+        {  24,  10 }, {  27,   0 }, {  34, -14 }, {  52, -44 },
+        {  39, -24 }, {  19,  17 }, {  31,  25 }, {  36,  29 },
+        {  24,  33 }, {  34,  15 }, {  30,  20 }, {  22,  73 },
+        {   4,  39 }, {   0,  42 }, {   7,  34 }, {  11,  29 },
+        {   8,  31 }, {   6,  37 }, {   7,  42 }, {   3,  40 },
+        {   8,  33 }, {  13,  43 }, {  13,  36 }, {   4,  47 },
+        {   3,  55 }, {   2,  58 }, {   6,  60 }, {   8,  44 },
+        {  11,  44 }, {  14,  42 }, {   7,  48 }, {   4,  56 },
+        {   4,  52 }, {  13,  37 }, {   9,  49 }, {  19,  58 },
+        {  10,  48 }, {  12,  45 }, {   0,  69 }, {  20,  33 },
+        {   8,  63 }, {  35, -18 }, {  33, -25 }, {  28,  -3 },
+        {  24,  10 }, {  27,   0 }, {  34, -14 }, {  52, -44 },
+        {  39, -24 }, {  19,  17 }, {  31,  25 }, {  36,  29 },
+        {  24,  33 }, {  34,  15 }, {  30,  20 }, {  22,  73 },
+        {  -3,  78 }, {  -8,  74 }, {  -9,  72 }, { -10,  72 },
+        { -18,  75 }, { -12,  71 }, { -11,  63 }, {  -5,  70 },
+        { -17,  75 }, { -14,  72 }, { -16,  67 }, {  -8,  53 },
+        { -14,  59 }, {  -9,  52 }, { -11,  68 }, {  -3,  78 },
+        {  -8,  74 }, {  -9,  72 }, { -10,  72 }, { -18,  75 },
+        { -12,  71 }, { -11,  63 }, {  -5,  70 }, { -17,  75 },
+        { -14,  72 }, { -16,  67 }, {  -8,  53 }, { -14,  59 },
+        {  -9,  52 }, { -11,  68 }, {   9,  -2 }, {  30, -10 },
+        {  31,  -4 }, {  33,  -1 }, {  33,   7 }, {  31,  12 },
+        {  37,  23 }, {  31,  38 }, {  20,  64 }, {   9,  -2 },
+        {  30, -10 }, {  31,  -4 }, {  33,  -1 }, {  33,   7 },
+        {  31,  12 }, {  37,  23 }, {  31,  38 }, {  20,  64 },
+        {  -9,  71 }, {  -7,  37 }, {  -8,  44 }, { -11,  49 },
+        { -10,  56 }, { -12,  59 }, {  -8,  63 }, {  -9,  67 },
+        {  -6,  68 }, { -10,  79 }, {  -3,  78 }, {  -8,  74 },
+        {  -9,  72 }, { -10,  72 }, { -18,  75 }, { -12,  71 },
+        { -11,  63 }, {  -5,  70 }, { -17,  75 }, { -14,  72 },
+        { -16,  67 }, {  -8,  53 }, { -14,  59 }, {  -9,  52 },
+        { -11,  68 }, {  -3,  78 }, {  -8,  74 }, {  -9,  72 },
+        { -10,  72 }, { -18,  75 }, { -12,  71 }, { -11,  63 },
+        {  -5,  70 }, { -17,  75 }, { -14,  72 }, { -16,  67 },
+        {  -8,  53 }, { -14,  59 }, {  -9,  52 }, { -11,  68 },
+        {   9,  -2 }, {  30, -10 }, {  31,  -4 }, {  33,  -1 },
+        {  33,   7 }, {  31,  12 }, {  37,  23 }, {  31,  38 },
+        {  20,  64 }, {   9,  -2 }, {  30, -10 }, {  31,  -4 },
+        {  33,  -1 }, {  33,   7 }, {  31,  12 }, {  37,  23 },
+        {  31,  38 }, {  20,  64 }, {  -9,  71 }, {  -7,  37 },
+        {  -8,  44 }, { -11,  49 }, { -10,  56 }, { -12,  59 },
+        {  -8,  63 }, {  -9,  67 }, {  -6,  68 }, { -10,  79 },
+        { -22, 127 }, { -25, 127 }, { -25, 120 }, { -27, 127 },
+        { -19, 114 }, { -23, 117 }, { -25, 118 }, { -26, 117 },
+        { -24, 113 }, { -28, 118 }, { -31, 120 }, { -37, 124 },
+        { -10,  94 }, { -15, 102 }, { -10,  99 }, { -13, 106 },
+        { -50, 127 }, {  -5,  92 }, {  17,  57 }, {  -5,  86 },
+        { -13,  94 }, { -12,  91 }, {  -2,  77 }, {   0,  71 },
+        {  -1,  73 }, {   4,  64 }, {  -7,  81 }, {   5,  64 },
+        {  15,  57 }, {   1,  67 }, {   0,  68 }, { -10,  67 },
+        {   1,  68 }, {   0,  77 }, {   2,  64 }, {   0,  68 },
+        {  -5,  78 }, {   7,  55 }, {   5,  59 }, {   2,  65 },
+        {  14,  54 }, {  15,  44 }, {   5,  60 }, {   2,  70 },
+        { -22, 127 }, { -25, 127 }, { -25, 120 }, { -27, 127 },
+        { -19, 114 }, { -23, 117 }, { -25, 118 }, { -26, 117 },
+        { -24, 113 }, { -28, 118 }, { -31, 120 }, { -37, 124 },
+        { -10,  94 }, { -15, 102 }, { -10,  99 }, { -13, 106 },
+        { -50, 127 }, {  -5,  92 }, {  17,  57 }, {  -5,  86 },
+        { -13,  94 }, { -12,  91 }, {  -2,  77 }, {   0,  71 },
+        {  -1,  73 }, {   4,  64 }, {  -7,  81 }, {   5,  64 },
+        {  15,  57 }, {   1,  67 }, {   0,  68 }, { -10,  67 },
+        {   1,  68 }, {   0,  77 }, {   2,  64 }, {   0,  68 },
+        {  -5,  78 }, {   7,  55 }, {   5,  59 }, {   2,  65 },
+        {  14,  54 }, {  15,  44 }, {   5,  60 }, {   2,  70 },
+        {  17, -13 }, {  16,  -9 }, {  17, -12 }, {  27, -21 },
+        {  37, -30 }, {  41, -40 }, {  42, -41 }, {  48, -47 },
+        {  39, -32 }, {  46, -40 }, {  52, -51 }, {  46, -41 },
+        {  52, -39 }, {  43, -19 }, {  32,  11 }, {  61, -55 },
+        {  56, -46 }, {  62, -50 }, {  81, -67 }, {  45, -20 },
+        {  35,  -2 }, {  28,  15 }, {  34,   1 }, {  39,   1 },
+        {  30,  17 }, {  20,  38 }, {  18,  45 }, {  15,  54 },
+        {   0,  79 }, {  36, -16 }, {  37, -14 }, {  37, -17 },
+        {  32,   1 }, {  34,  15 }, {  29,  15 }, {  24,  25 },
+        {  34,  22 }, {  31,  16 }, {  35,  18 }, {  31,  28 },
+        {  33,  41 }, {  36,  28 }, {  27,  47 }, {  21,  62 },
+        {  17, -13 }, {  16,  -9 }, {  17, -12 }, {  27, -21 },
+        {  37, -30 }, {  41, -40 }, {  42, -41 }, {  48, -47 },
+        {  39, -32 }, {  46, -40 }, {  52, -51 }, {  46, -41 },
+        {  52, -39 }, {  43, -19 }, {  32,  11 }, {  61, -55 },
+        {  56, -46 }, {  62, -50 }, {  81, -67 }, {  45, -20 },
+        {  35,  -2 }, {  28,  15 }, {  34,   1 }, {  39,   1 },
+        {  30,  17 }, {  20,  38 }, {  18,  45 }, {  15,  54 },
+        {   0,  79 }, {  36, -16 }, {  37, -14 }, {  37, -17 },
+        {  32,   1 }, {  34,  15 }, {  29,  15 }, {  24,  25 },
+        {  34,  22 }, {  31,  16 }, {  35,  18 }, {  31,  28 },
+        {  33,  41 }, {  36,  28 }, {  27,  47 }, {  21,  62 },
+        { -24, 115 }, { -22,  82 }, {  -9,  62 }, {   0,  53 },
+        {   0,  59 }, { -14,  85 }, { -13,  89 }, { -13,  94 },
+        { -11,  92 }, { -29, 127 }, { -21, 100 }, { -14,  57 },
+        { -12,  67 }, { -11,  71 }, { -10,  77 }, { -21,  85 },
+        { -16,  88 }, { -23, 104 }, { -15,  98 }, { -37, 127 },
+        { -10,  82 }, {  -8,  48 }, {  -8,  61 }, {  -8,  66 },
+        {  -7,  70 }, { -14,  75 }, { -10,  79 }, {  -9,  83 },
+        { -12,  92 }, { -18, 108 }, { -24, 115 }, { -22,  82 },
+        {  -9,  62 }, {   0,  53 }, {   0,  59 }, { -14,  85 },
+        { -13,  89 }, { -13,  94 }, { -11,  92 }, { -29, 127 },
+        { -21, 100 }, { -14,  57 }, { -12,  67 }, { -11,  71 },
+        { -10,  77 }, { -21,  85 }, { -16,  88 }, { -23, 104 },
+        { -15,  98 }, { -37, 127 }, { -10,  82 }, {  -8,  48 },
+        {  -8,  61 }, {  -8,  66 }, {  -7,  70 }, { -14,  75 },
+        { -10,  79 }, {  -9,  83 }, { -12,  92 }, { -18, 108 },
+        {  -5,  79 }, { -11, 104 }, { -11,  91 }, { -30, 127 },
+        {  -5,  79 }, { -11, 104 }, { -11,  91 }, { -30, 127 },
+        {  -5,  79 }, { -11, 104 }, { -11,  91 }, { -30, 127 }
     }
 };
 
@@ -695,7 +1267,7 @@ void ff_h264_init_cabac_states(H264Context *h) {
     else                                 tab = cabac_context_init_PB[h->cabac_init_idc];
 
     /* calculate pre-state */
-    for( i= 0; i < 460; i++ ) {
+    for( i= 0; i < 1024; i++ ) {
         int pre = 2*(((tab[i][0] * slice_qp) >>4 ) + tab[i][1]) - 127;
 
         pre^= pre>>31;
@@ -957,21 +1529,22 @@ static int decode_cabac_mb_mvd( H264Context *h, int ctxbase, int amvd, int *mvda
     my += decode_cabac_mb_mvd( h, 47, amvd1, &mpy );\
 }
 
-static av_always_inline int get_cabac_cbf_ctx( H264Context *h, int cat, int idx, int is_dc ) {
+static av_always_inline int get_cabac_cbf_ctx( H264Context *h, int cat, int idx, int max_coeff, int is_dc ) {
     int nza, nzb;
     int ctx = 0;
+    static const uint16_t base_ctx[14] = {85,89,93,97,101,1012,460,464,468,1016,472,476,480,1020};
 
     if( is_dc ) {
-        if( cat == 0 ) {
-            nza = h->left_cbp&0x100;
-            nzb = h-> top_cbp&0x100;
-        } else {
+        if( cat == 3 ) {
             idx -= CHROMA_DC_BLOCK_INDEX;
             nza = (h->left_cbp>>(6+idx))&0x01;
             nzb = (h-> top_cbp>>(6+idx))&0x01;
+        } else {
+            idx -= LUMA_DC_BLOCK_INDEX;
+            nza = h->left_cbp&(0x100<<idx);
+            nzb = h-> top_cbp&(0x100<<idx);
         }
     } else {
-        assert(cat == 1 || cat == 2 || cat == 4);
         nza = h->non_zero_count_cache[scan8[idx] - 1];
         nzb = h->non_zero_count_cache[scan8[idx] - 8];
     }
@@ -982,7 +1555,7 @@ static av_always_inline int get_cabac_cbf_ctx( H264Context *h, int cat, int idx,
     if( nzb > 0 )
         ctx += 2;
 
-    return ctx + 4 * cat;
+    return base_ctx[cat] + ctx;
 }
 
 DECLARE_ASM_CONST(1, uint8_t, last_coeff_flag_offset_8x8)[63] = {
@@ -993,16 +1566,16 @@ DECLARE_ASM_CONST(1, uint8_t, last_coeff_flag_offset_8x8)[63] = {
 };
 
 static av_always_inline void decode_cabac_residual_internal( H264Context *h, DCTELEM *block, int cat, int n, const uint8_t *scantable, const uint32_t *qmul, int max_coeff, int is_dc ) {
-    static const int significant_coeff_flag_offset[2][6] = {
-      { 105+0, 105+15, 105+29, 105+44, 105+47, 402 },
-      { 277+0, 277+15, 277+29, 277+44, 277+47, 436 }
+    static const int significant_coeff_flag_offset[2][14] = {
+      { 105+0, 105+15, 105+29, 105+44, 105+47, 402, 484+0, 484+15, 484+29, 660, 528+0, 528+15, 528+29, 718 },
+      { 277+0, 277+15, 277+29, 277+44, 277+47, 436, 776+0, 776+15, 776+29, 675, 820+0, 820+15, 820+29, 733 }
     };
-    static const int last_coeff_flag_offset[2][6] = {
-      { 166+0, 166+15, 166+29, 166+44, 166+47, 417 },
-      { 338+0, 338+15, 338+29, 338+44, 338+47, 451 }
+    static const int last_coeff_flag_offset[2][14] = {
+      { 166+0, 166+15, 166+29, 166+44, 166+47, 417, 572+0, 572+15, 572+29, 690, 616+0, 616+15, 616+29, 748 },
+      { 338+0, 338+15, 338+29, 338+44, 338+47, 451, 864+0, 864+15, 864+29, 699, 908+0, 908+15, 908+29, 757 }
     };
-    static const int coeff_abs_level_m1_offset[6] = {
-        227+0, 227+10, 227+20, 227+30, 227+39, 426
+    static const int coeff_abs_level_m1_offset[14] = {
+        227+0, 227+10, 227+20, 227+30, 227+39, 426, 952+0, 952+10, 952+20, 708, 982+0, 982+10, 982+20, 766
     };
     static const uint8_t significant_coeff_flag_offset_8x8[2][63] = {
       { 0, 1, 2, 3, 4, 5, 5, 4, 4, 3, 3, 4, 4, 4, 5, 5,
@@ -1057,7 +1630,7 @@ static av_always_inline void decode_cabac_residual_internal( H264Context *h, DCT
     abs_level_m1_ctx_base = h->cabac_state
         + coeff_abs_level_m1_offset[cat];
 
-    if( !is_dc && cat == 5 ) {
+    if( !is_dc && max_coeff == 64 ) {
 #define DECODE_SIGNIFICANCE( coefs, sig_off, last_off ) \
         for(last= 0; last < coefs; last++) { \
             uint8_t *sig_ctx = significant_coeff_ctx_base + sig_off; \
@@ -1075,9 +1648,11 @@ static av_always_inline void decode_cabac_residual_internal( H264Context *h, DCT
         }
         const uint8_t *sig_off = significant_coeff_flag_offset_8x8[MB_FIELD];
 #if ARCH_X86 && HAVE_7REGS && HAVE_EBX_AVAILABLE && !defined(BROKEN_RELOCATIONS)
-        coeff_count= decode_significance_8x8_x86(CC, significant_coeff_ctx_base, index, sig_off);
+        coeff_count= decode_significance_8x8_x86(CC, significant_coeff_ctx_base, index,
+                                                 last_coeff_ctx_base-significant_coeff_ctx_base, sig_off);
     } else {
-        coeff_count= decode_significance_x86(CC, max_coeff, significant_coeff_ctx_base, index);
+        coeff_count= decode_significance_x86(CC, max_coeff, significant_coeff_ctx_base, index,
+                                             last_coeff_ctx_base-significant_coeff_ctx_base);
 #else
         DECODE_SIGNIFICANCE( 63, sig_off[last], last_coeff_flag_offset_8x8[last] );
     } else {
@@ -1087,16 +1662,16 @@ static av_always_inline void decode_cabac_residual_internal( H264Context *h, DCT
     assert(coeff_count > 0);
 
     if( is_dc ) {
-        if( cat == 0 )
-            h->cbp_table[h->mb_xy] |= 0x100;
-        else
+        if( cat == 3 )
             h->cbp_table[h->mb_xy] |= 0x40 << (n - CHROMA_DC_BLOCK_INDEX);
+        else
+            h->cbp_table[h->mb_xy] |= 0x100 << (n - LUMA_DC_BLOCK_INDEX);
         h->non_zero_count_cache[scan8[n]] = coeff_count;
     } else {
-        if( cat == 5 )
+        if( max_coeff == 64 )
             fill_rectangle(&h->non_zero_count_cache[scan8[n]], 2, 2, 8, coeff_count, 1);
         else {
-            assert( cat == 1 || cat == 2 || cat == 4 );
+            assert( cat == 1 || cat ==  2 || cat ==  4 || cat == 7 || cat == 8 || cat == 11 || cat == 12 );
             h->non_zero_count_cache[scan8[n]] = coeff_count;
         }
     }
@@ -1180,7 +1755,7 @@ static void decode_cabac_residual_nondc_internal( H264Context *h, DCTELEM *block
 
 static av_always_inline void decode_cabac_residual_dc( H264Context *h, DCTELEM *block, int cat, int n, const uint8_t *scantable, int max_coeff ) {
     /* read coded block flag */
-    if( get_cabac( &h->cabac, &h->cabac_state[85 + get_cabac_cbf_ctx( h, cat, n, 1 ) ] ) == 0 ) {
+    if( get_cabac( &h->cabac, &h->cabac_state[get_cabac_cbf_ctx( h, cat, n, max_coeff, 1 ) ] ) == 0 ) {
         h->non_zero_count_cache[scan8[n]] = 0;
         return;
     }
@@ -1189,13 +1764,68 @@ static av_always_inline void decode_cabac_residual_dc( H264Context *h, DCTELEM *
 
 static av_always_inline void decode_cabac_residual_nondc( H264Context *h, DCTELEM *block, int cat, int n, const uint8_t *scantable, const uint32_t *qmul, int max_coeff ) {
     /* read coded block flag */
-    if( cat != 5 && get_cabac( &h->cabac, &h->cabac_state[85 + get_cabac_cbf_ctx( h, cat, n, 0 ) ] ) == 0 ) {
-        h->non_zero_count_cache[scan8[n]] = 0;
+    if( (cat != 5 || CHROMA444) && get_cabac( &h->cabac, &h->cabac_state[get_cabac_cbf_ctx( h, cat, n, max_coeff, 0 ) ] ) == 0 ) {
+        if( max_coeff == 64 ) {
+            fill_rectangle(&h->non_zero_count_cache[scan8[n]], 2, 2, 8, 0, 1);
+        } else {
+            h->non_zero_count_cache[scan8[n]] = 0;
+        }
         return;
     }
     decode_cabac_residual_nondc_internal( h, block, cat, n, scantable, qmul, max_coeff );
 }
 
+static av_always_inline void decode_cabac_luma_residual( H264Context *h, const uint8_t *scan, const uint8_t *scan8x8, int pixel_shift, int mb_type, int cbp, int p )
+{
+    static const uint8_t ctx_cat[4][3] = {{0,6,10},{1,7,11},{2,8,12},{5,9,13}};
+    const uint32_t *qmul;
+    int i8x8, i4x4;
+    MpegEncContext * const s = &h->s;
+    int qscale = p == 0 ? s->qscale : h->chroma_qp[p-1];
+    if( IS_INTRA16x16( mb_type ) ) {
+        //av_log( s->avctx, AV_LOG_ERROR, "INTRA16x16 DC\n" );
+        AV_ZERO128(h->mb_luma_dc[p]+0);
+        AV_ZERO128(h->mb_luma_dc[p]+8);
+        AV_ZERO128(h->mb_luma_dc[p]+16);
+        AV_ZERO128(h->mb_luma_dc[p]+24);
+        decode_cabac_residual_dc(h, h->mb_luma_dc[p], ctx_cat[0][p], LUMA_DC_BLOCK_INDEX+p, scan, 16);
+
+        if( cbp&15 ) {
+            qmul = h->dequant4_coeff[p][qscale];
+            for( i4x4 = 0; i4x4 < 16; i4x4++ ) {
+                const int index = 16*p + i4x4;
+                //av_log( s->avctx, AV_LOG_ERROR, "INTRA16x16 AC:%d\n", index );
+                decode_cabac_residual_nondc(h, h->mb + (16*index << pixel_shift), ctx_cat[1][p], index, scan + 1, qmul, 15);
+            }
+        } else {
+            fill_rectangle(&h->non_zero_count_cache[scan8[16*p]], 4, 4, 8, 0, 1);
+        }
+    } else {
+        int cqm = (IS_INTRA( mb_type ) ? 0:3) + p;
+        for( i8x8 = 0; i8x8 < 4; i8x8++ ) {
+            if( cbp & (1<<i8x8) ) {
+                if( IS_8x8DCT(mb_type) ) {
+                    const int index = 16*p + 4*i8x8;
+                    decode_cabac_residual_nondc(h, h->mb + (16*index << pixel_shift), ctx_cat[3][p], index,
+                                                scan8x8, h->dequant8_coeff[cqm][qscale], 64);
+                } else {
+                    qmul = h->dequant4_coeff[cqm][qscale];
+                    for( i4x4 = 0; i4x4 < 4; i4x4++ ) {
+                        const int index = 16*p + 4*i8x8 + i4x4;
+                        //av_log( s->avctx, AV_LOG_ERROR, "Luma4x4: %d\n", index );
+//START_TIMER
+                        decode_cabac_residual_nondc(h, h->mb + (16*index << pixel_shift), ctx_cat[2][p], index, scan, qmul, 16);
+//STOP_TIMER("decode_residual")
+                    }
+                }
+            } else {
+                uint8_t * const nnz= &h->non_zero_count_cache[ scan8[4*i8x8+16*p] ];
+                nnz[0] = nnz[1] = nnz[8] = nnz[9] = 0;
+            }
+        }
+    }
+}
+
 /**
  * decodes a macroblock
  * @return 0 if OK, AC_ERROR / DC_ERROR / MV_ERROR if an error is noticed
@@ -1205,6 +1835,7 @@ int ff_h264_decode_mb_cabac(H264Context *h) {
     int mb_xy;
     int mb_type, partition_count, cbp = 0;
     int dct8x8_allowed= h->pps.transform_8x8_mode;
+    int decode_chroma = h->sps.chroma_format_idc == 1 || h->sps.chroma_format_idc == 2;
     const int pixel_shift = h->pixel_shift;
 
     mb_xy = h->mb_xy = s->mb_x + s->mb_y*s->mb_stride;
@@ -1314,7 +1945,8 @@ decode_intra_mb:
     h->slice_table[ mb_xy ]= h->slice_num;
 
     if(IS_INTRA_PCM(mb_type)) {
-        const int mb_size = (384*h->sps.bit_depth_luma) >> 3;
+        static const uint16_t mb_sizes[4] = {256,384,512,768};
+        const int mb_size = mb_sizes[h->sps.chroma_format_idc]*h->sps.bit_depth_luma >> 3;
         const uint8_t *ptr;
 
         // We assume these blocks are very rare so we do not optimize it.
@@ -1327,20 +1959,17 @@ decode_intra_mb:
         }
 
         // The pixels are stored in the same order as levels in h->mb array.
-        memcpy(h->mb, ptr, 2*mb_size/3); ptr+=2*mb_size/3;
-        if(CHROMA){
-            memcpy(h->mb+mb_size/3, ptr, mb_size/3); ptr+=mb_size/3;
-        }
+        memcpy(h->mb, ptr, mb_size); ptr+=mb_size;
 
         ff_init_cabac_decoder(&h->cabac, ptr, h->cabac.bytestream_end - ptr);
 
         // All blocks are present
-        h->cbp_table[mb_xy] = 0x1ef;
+        h->cbp_table[mb_xy] = 0xf7ef;
         h->chroma_pred_mode_table[mb_xy] = 0;
         // In deblocking, the quantizer is 0
         s->current_picture.qscale_table[mb_xy]= 0;
         // All coeffs are present
-        memset(h->non_zero_count[mb_xy], 16, 32);
+        memset(h->non_zero_count[mb_xy], 16, 48);
         s->current_picture.mb_type[mb_xy]= mb_type;
         h->last_qscale_diff = 0;
         return 0;
@@ -1377,7 +2006,7 @@ decode_intra_mb:
             h->intra16x16_pred_mode= ff_h264_check_intra_pred_mode( h, h->intra16x16_pred_mode );
             if( h->intra16x16_pred_mode < 0 ) return -1;
         }
-        if(CHROMA){
+        if(decode_chroma){
             h->chroma_pred_mode_table[mb_xy] =
             pred_mode                        = decode_cabac_mb_chroma_pre_mode( h );
 
@@ -1606,7 +2235,7 @@ decode_intra_mb:
 
     if( !IS_INTRA16x16( mb_type ) ) {
         cbp  = decode_cabac_mb_cbp_luma( h );
-        if(CHROMA)
+        if(decode_chroma)
             cbp |= decode_cabac_mb_cbp_chroma( h ) << 4;
     }
 
@@ -1615,6 +2244,28 @@ decode_intra_mb:
     if( dct8x8_allowed && (cbp&15) && !IS_INTRA( mb_type ) ) {
         mb_type |= MB_TYPE_8x8DCT * get_cabac_noinline( &h->cabac, &h->cabac_state[399 + h->neighbor_transform_size] );
     }
+
+    /* It would be better to do this in fill_decode_caches, but we don't know
+     * the transform mode of the current macroblock there. */
+    if (CHROMA444 && IS_8x8DCT(mb_type)){
+        int i;
+        for (i = 0; i < 2; i++){
+            if (h->left_type[i] && !IS_8x8DCT(h->left_type[i])){
+                h->non_zero_count_cache[3+8* 1 + 2*8*i]=
+                h->non_zero_count_cache[3+8* 2 + 2*8*i]=
+                h->non_zero_count_cache[3+8* 6 + 2*8*i]=
+                h->non_zero_count_cache[3+8* 7 + 2*8*i]=
+                h->non_zero_count_cache[3+8*11 + 2*8*i]=
+                h->non_zero_count_cache[3+8*12 + 2*8*i]= IS_INTRA(mb_type) ? 64 : 0;
+            }
+        }
+        if (h->top_type && !IS_8x8DCT(h->top_type)){
+            uint32_t top_empty = CABAC && !IS_INTRA(mb_type) ? 0 : 0x40404040;
+            AV_WN32A(&h->non_zero_count_cache[4+8* 0], top_empty);
+            AV_WN32A(&h->non_zero_count_cache[4+8* 5], top_empty);
+            AV_WN32A(&h->non_zero_count_cache[4+8*10], top_empty);
+        }
+    }
     s->current_picture.mb_type[mb_xy]= mb_type;
 
     if( cbp || IS_INTRA16x16( mb_type ) ) {
@@ -1659,76 +2310,38 @@ decode_intra_mb:
         }else
             h->last_qscale_diff=0;
 
-        if( IS_INTRA16x16( mb_type ) ) {
-            int i;
-            //av_log( s->avctx, AV_LOG_ERROR, "INTRA16x16 DC\n" );
-            AV_ZERO128(h->mb_luma_dc+0);
-            AV_ZERO128(h->mb_luma_dc+8);
-            AV_ZERO128(h->mb_luma_dc+16);
-            AV_ZERO128(h->mb_luma_dc+24);
-            decode_cabac_residual_dc( h, h->mb_luma_dc, 0, LUMA_DC_BLOCK_INDEX, scan, 16);
-
-            if( cbp&15 ) {
-                qmul = h->dequant4_coeff[0][s->qscale];
-                for( i = 0; i < 16; i++ ) {
-                    //av_log( s->avctx, AV_LOG_ERROR, "INTRA16x16 AC:%d\n", i );
-                    decode_cabac_residual_nondc(h, h->mb + (16*i << pixel_shift), 1, i, scan + 1, qmul, 15);
-                }
-            } else {
-                fill_rectangle(&h->non_zero_count_cache[scan8[0]], 4, 4, 8, 0, 1);
-            }
+        decode_cabac_luma_residual(h, scan, scan8x8, pixel_shift, mb_type, cbp, 0);
+        if(CHROMA444){
+            decode_cabac_luma_residual(h, scan, scan8x8, pixel_shift, mb_type, cbp, 1);
+            decode_cabac_luma_residual(h, scan, scan8x8, pixel_shift, mb_type, cbp, 2);
         } else {
-            int i8x8, i4x4;
-            for( i8x8 = 0; i8x8 < 4; i8x8++ ) {
-                if( cbp & (1<<i8x8) ) {
-                    if( IS_8x8DCT(mb_type) ) {
-                        decode_cabac_residual_nondc(h, h->mb + (64*i8x8 << pixel_shift), 5, 4*i8x8,
-                            scan8x8, h->dequant8_coeff[IS_INTRA( mb_type ) ? 0:1][s->qscale], 64);
-                    } else {
-                        qmul = h->dequant4_coeff[IS_INTRA( mb_type ) ? 0:3][s->qscale];
-                        for( i4x4 = 0; i4x4 < 4; i4x4++ ) {
-                            const int index = 4*i8x8 + i4x4;
-                            //av_log( s->avctx, AV_LOG_ERROR, "Luma4x4: %d\n", index );
-//START_TIMER
-                            decode_cabac_residual_nondc(h, h->mb + (16*index << pixel_shift), 2, index, scan, qmul, 16);
-//STOP_TIMER("decode_residual")
-                        }
-                    }
-                } else {
-                    uint8_t * const nnz= &h->non_zero_count_cache[ scan8[4*i8x8] ];
-                    nnz[0] = nnz[1] = nnz[8] = nnz[9] = 0;
+            if( cbp&0x30 ){
+                int c;
+                for( c = 0; c < 2; c++ ) {
+                    //av_log( s->avctx, AV_LOG_ERROR, "INTRA C%d-DC\n",c );
+                    decode_cabac_residual_dc(h, h->mb + ((256 + 16*16*c) << pixel_shift), 3, CHROMA_DC_BLOCK_INDEX+c, chroma_dc_scan, 4);
                 }
             }
-        }
 
-        if( cbp&0x30 ){
-            int c;
-            for( c = 0; c < 2; c++ ) {
-                //av_log( s->avctx, AV_LOG_ERROR, "INTRA C%d-DC\n",c );
-                decode_cabac_residual_dc(h, h->mb + ((256 + 16*4*c) << pixel_shift), 3, CHROMA_DC_BLOCK_INDEX+c, chroma_dc_scan, 4);
-            }
-        }
-
-        if( cbp&0x20 ) {
-            int c, i;
-            for( c = 0; c < 2; c++ ) {
-                qmul = h->dequant4_coeff[c+1+(IS_INTRA( mb_type ) ? 0:3)][h->chroma_qp[c]];
-                for( i = 0; i < 4; i++ ) {
-                    const int index = 16 + 4 * c + i;
-                    //av_log( s->avctx, AV_LOG_ERROR, "INTRA C%d-AC %d\n",c, index - 16 );
-                    decode_cabac_residual_nondc(h, h->mb + (16*index << pixel_shift), 4, index, scan + 1, qmul, 15);
+            if( cbp&0x20 ) {
+                int c, i;
+                for( c = 0; c < 2; c++ ) {
+                    qmul = h->dequant4_coeff[c+1+(IS_INTRA( mb_type ) ? 0:3)][h->chroma_qp[c]];
+                    for( i = 0; i < 4; i++ ) {
+                        const int index = 16 + 16 * c + i;
+                        //av_log( s->avctx, AV_LOG_ERROR, "INTRA C%d-AC %d\n",c, index - 16 );
+                        decode_cabac_residual_nondc(h, h->mb + (16*index << pixel_shift), 4, index, scan + 1, qmul, 15);
+                    }
                 }
+            } else {
+                fill_rectangle(&h->non_zero_count_cache[scan8[16]], 4, 4, 8, 0, 1);
+                fill_rectangle(&h->non_zero_count_cache[scan8[32]], 4, 4, 8, 0, 1);
             }
-        } else {
-            uint8_t * const nnz= &h->non_zero_count_cache[0];
-            nnz[ scan8[16]+0 ] = nnz[ scan8[16]+1 ] =nnz[ scan8[16]+8 ] =nnz[ scan8[16]+9 ] =
-            nnz[ scan8[20]+0 ] = nnz[ scan8[20]+1 ] =nnz[ scan8[20]+8 ] =nnz[ scan8[20]+9 ] = 0;
         }
     } else {
-        uint8_t * const nnz= &h->non_zero_count_cache[0];
-        fill_rectangle(&nnz[scan8[0]], 4, 4, 8, 0, 1);
-        nnz[ scan8[16]+0 ] = nnz[ scan8[16]+1 ] =nnz[ scan8[16]+8 ] =nnz[ scan8[16]+9 ] =
-        nnz[ scan8[20]+0 ] = nnz[ scan8[20]+1 ] =nnz[ scan8[20]+8 ] =nnz[ scan8[20]+9 ] = 0;
+        fill_rectangle(&h->non_zero_count_cache[scan8[ 0]], 4, 4, 8, 0, 1);
+        fill_rectangle(&h->non_zero_count_cache[scan8[16]], 4, 4, 8, 0, 1);
+        fill_rectangle(&h->non_zero_count_cache[scan8[32]], 4, 4, 8, 0, 1);
         h->last_qscale_diff = 0;
     }
 
diff --git a/libavcodec/h264_cavlc.c b/libavcodec/h264_cavlc.c
index 62e30f1311..0ddc430661 100644
--- a/libavcodec/h264_cavlc.c
+++ b/libavcodec/h264_cavlc.c
@@ -371,12 +371,12 @@ static int decode_residual(H264Context *h, GetBitContext *gb, DCTELEM *block, in
 
     //FIXME put trailing_onex into the context
 
-    if(n >= CHROMA_DC_BLOCK_INDEX){
+    if(max_coeff <= 8){
         coeff_token= get_vlc2(gb, chroma_dc_coeff_token_vlc.table, CHROMA_DC_COEFF_TOKEN_VLC_BITS, 1);
         total_coeff= coeff_token>>2;
     }else{
-        if(n == LUMA_DC_BLOCK_INDEX){
-            total_coeff= pred_non_zero_count(h, 0);
+        if(n >= LUMA_DC_BLOCK_INDEX){
+            total_coeff= pred_non_zero_count(h, (n - LUMA_DC_BLOCK_INDEX)*16);
             coeff_token= get_vlc2(gb, coeff_token_vlc[ coeff_token_table_index[total_coeff] ].table, COEFF_TOKEN_VLC_BITS, 2);
             total_coeff= coeff_token>>2;
         }else{
@@ -482,7 +482,8 @@ static int decode_residual(H264Context *h, GetBitContext *gb, DCTELEM *block, in
     if(total_coeff == max_coeff)
         zeros_left=0;
     else{
-        if(n >= CHROMA_DC_BLOCK_INDEX)
+        /* FIXME: we don't actually support 4:2:2 yet. */
+        if(max_coeff <= 8)
             zeros_left= get_vlc2(gb, (chroma_dc_total_zeros_vlc-1)[ total_coeff ].table, CHROMA_DC_TOTAL_ZEROS_VLC_BITS, 1);
         else
             zeros_left= get_vlc2(gb, (total_zeros_vlc-1)[ total_coeff ].table, TOTAL_ZEROS_VLC_BITS, 1);
@@ -536,12 +537,80 @@ static int decode_residual(H264Context *h, GetBitContext *gb, DCTELEM *block, in
     return 0;
 }
 
+static av_always_inline int decode_luma_residual(H264Context *h, GetBitContext *gb, const uint8_t *scan, const uint8_t *scan8x8, int pixel_shift, int mb_type, int cbp, int p){
+    int i4x4, i8x8;
+    MpegEncContext * const s = &h->s;
+    int qscale = p == 0 ? s->qscale : h->chroma_qp[p-1];
+    if(IS_INTRA16x16(mb_type)){
+        AV_ZERO128(h->mb_luma_dc[p]+0);
+        AV_ZERO128(h->mb_luma_dc[p]+8);
+        AV_ZERO128(h->mb_luma_dc[p]+16);
+        AV_ZERO128(h->mb_luma_dc[p]+24);
+        if( decode_residual(h, h->intra_gb_ptr, h->mb_luma_dc[p], LUMA_DC_BLOCK_INDEX+p, scan, NULL, 16) < 0){
+            return -1; //FIXME continue if partitioned and other return -1 too
+        }
+
+        assert((cbp&15) == 0 || (cbp&15) == 15);
+
+        if(cbp&15){
+            for(i8x8=0; i8x8<4; i8x8++){
+                for(i4x4=0; i4x4<4; i4x4++){
+                    const int index= i4x4 + 4*i8x8 + p*16;
+                    if( decode_residual(h, h->intra_gb_ptr, h->mb + (16*index << pixel_shift),
+                        index, scan + 1, h->dequant4_coeff[p][qscale], 15) < 0 ){
+                        return -1;
+                    }
+                }
+            }
+            return 0xf;
+        }else{
+            fill_rectangle(&h->non_zero_count_cache[scan8[p*16]], 4, 4, 8, 0, 1);
+            return 0;
+        }
+    }else{
+        int cqm = (IS_INTRA( mb_type ) ? 0:3)+p;
+        /* For CAVLC 4:4:4, we need to keep track of the luma 8x8 CBP for deblocking nnz purposes. */
+        int new_cbp = 0;
+        for(i8x8=0; i8x8<4; i8x8++){
+            if(cbp & (1<<i8x8)){
+                if(IS_8x8DCT(mb_type)){
+                    DCTELEM *buf = &h->mb[64*i8x8+256*p << pixel_shift];
+                    uint8_t *nnz;
+                    for(i4x4=0; i4x4<4; i4x4++){
+                        const int index= i4x4 + 4*i8x8 + p*16;
+                        if( decode_residual(h, gb, buf, index, scan8x8+16*i4x4,
+                                            h->dequant8_coeff[cqm][qscale], 16) < 0 )
+                            return -1;
+                    }
+                    nnz= &h->non_zero_count_cache[ scan8[4*i8x8+p*16] ];
+                    nnz[0] += nnz[1] + nnz[8] + nnz[9];
+                    new_cbp |= !!nnz[0] << i8x8;
+                }else{
+                    for(i4x4=0; i4x4<4; i4x4++){
+                        const int index= i4x4 + 4*i8x8 + p*16;
+                        if( decode_residual(h, gb, h->mb + (16*index << pixel_shift), index,
+                                            scan, h->dequant4_coeff[cqm][qscale], 16) < 0 ){
+                            return -1;
+                        }
+                        new_cbp |= h->non_zero_count_cache[ scan8[index] ] << i8x8;
+                    }
+                }
+            }else{
+                uint8_t * const nnz= &h->non_zero_count_cache[ scan8[4*i8x8+p*16] ];
+                nnz[0] = nnz[1] = nnz[8] = nnz[9] = 0;
+            }
+        }
+        return new_cbp;
+    }
+}
+
 int ff_h264_decode_mb_cavlc(H264Context *h){
     MpegEncContext * const s = &h->s;
     int mb_xy;
     int partition_count;
     unsigned int mb_type, cbp;
     int dct8x8_allowed= h->pps.transform_8x8_mode;
+    int decode_chroma = h->sps.chroma_format_idc == 1 || h->sps.chroma_format_idc == 2;
     const int pixel_shift = h->pixel_shift;
 
     mb_xy = h->mb_xy = s->mb_x + s->mb_y*s->mb_stride;
@@ -608,19 +677,21 @@ decode_intra_mb:
 
     if(IS_INTRA_PCM(mb_type)){
         unsigned int x;
+        static const uint16_t mb_sizes[4] = {256,384,512,768};
+        const int mb_size = mb_sizes[h->sps.chroma_format_idc]*h->sps.bit_depth_luma >> 3;
 
         // We assume these blocks are very rare so we do not optimize it.
         align_get_bits(&s->gb);
 
         // The pixels are stored in the same order as levels in h->mb array.
-        for(x=0; x < (CHROMA ? 384 : 256)*h->sps.bit_depth_luma/8; x++){
+        for(x=0; x < mb_size; x++){
             ((uint8_t*)h->mb)[x]= get_bits(&s->gb, 8);
         }
 
         // In deblocking, the quantizer is 0
         s->current_picture.qscale_table[mb_xy]= 0;
         // All coeffs are present
-        memset(h->non_zero_count[mb_xy], 16, 32);
+        memset(h->non_zero_count[mb_xy], 16, 48);
 
         s->current_picture.mb_type[mb_xy]= mb_type;
         return 0;
@@ -668,7 +739,7 @@ decode_intra_mb:
             if(h->intra16x16_pred_mode < 0)
                 return -1;
         }
-        if(CHROMA){
+        if(decode_chroma){
             pred_mode= ff_h264_check_intra_pred_mode(h, get_ue_golomb_31(&s->gb));
             if(pred_mode < 0)
                 return -1;
@@ -896,15 +967,19 @@ decode_intra_mb:
 
     if(!IS_INTRA16x16(mb_type)){
         cbp= get_ue_golomb(&s->gb);
-        if(cbp > 47){
-            av_log(h->s.avctx, AV_LOG_ERROR, "cbp too large (%u) at %d %d\n", cbp, s->mb_x, s->mb_y);
-            return -1;
-        }
 
-        if(CHROMA){
+        if(decode_chroma){
+            if(cbp > 47){
+                av_log(h->s.avctx, AV_LOG_ERROR, "cbp too large (%u) at %d %d\n", cbp, s->mb_x, s->mb_y);
+                return -1;
+            }
             if(IS_INTRA4x4(mb_type)) cbp= golomb_to_intra4x4_cbp[cbp];
             else                     cbp= golomb_to_inter_cbp   [cbp];
         }else{
+            if(cbp > 15){
+                av_log(h->s.avctx, AV_LOG_ERROR, "cbp too large (%u) at %d %d\n", cbp, s->mb_x, s->mb_y);
+                return -1;
+            }
             if(IS_INTRA4x4(mb_type)) cbp= golomb_to_intra4x4_cbp_gray[cbp];
             else                     cbp= golomb_to_inter_cbp_gray[cbp];
         }
@@ -918,8 +993,9 @@ decode_intra_mb:
     s->current_picture.mb_type[mb_xy]= mb_type;
 
     if(cbp || IS_INTRA16x16(mb_type)){
-        int i8x8, i4x4, chroma_idx;
+        int i4x4, chroma_idx;
         int dquant;
+        int ret;
         GetBitContext *gb= IS_INTRA(mb_type) ? h->intra_gb_ptr : h->inter_gb_ptr;
         const uint8_t *scan, *scan8x8;
         const int max_qp = 51 + 6*(h->sps.bit_depth_luma-8);
@@ -947,85 +1023,45 @@ decode_intra_mb:
 
         h->chroma_qp[0]= get_chroma_qp(h, 0, s->qscale);
         h->chroma_qp[1]= get_chroma_qp(h, 1, s->qscale);
-        if(IS_INTRA16x16(mb_type)){
-            AV_ZERO128(h->mb_luma_dc+0);
-            AV_ZERO128(h->mb_luma_dc+8);
-            AV_ZERO128(h->mb_luma_dc+16);
-            AV_ZERO128(h->mb_luma_dc+24);
-            if( decode_residual(h, h->intra_gb_ptr, h->mb_luma_dc, LUMA_DC_BLOCK_INDEX, scan, h->dequant4_coeff[0][s->qscale], 16) < 0){
-                return -1; //FIXME continue if partitioned and other return -1 too
-            }
 
-            assert((cbp&15) == 0 || (cbp&15) == 15);
+        if( (ret = decode_luma_residual(h, gb, scan, scan8x8, pixel_shift, mb_type, cbp, 0)) < 0 ){
+            return -1;
+        }
+        h->cbp_table[mb_xy] |= ret << 12;
+        if(CHROMA444){
+            if( decode_luma_residual(h, gb, scan, scan8x8, pixel_shift, mb_type, cbp, 1) < 0 ){
+                return -1;
+            }
+            if( decode_luma_residual(h, gb, scan, scan8x8, pixel_shift, mb_type, cbp, 2) < 0 ){
+                return -1;
+            }
+        } else {
+            if(cbp&0x30){
+                for(chroma_idx=0; chroma_idx<2; chroma_idx++)
+                    if( decode_residual(h, gb, h->mb + ((256 + 16*16*chroma_idx) << pixel_shift), CHROMA_DC_BLOCK_INDEX+chroma_idx, chroma_dc_scan, NULL, 4) < 0){
+                        return -1;
+                    }
+            }
 
-            if(cbp&15){
-                for(i8x8=0; i8x8<4; i8x8++){
+            if(cbp&0x20){
+                for(chroma_idx=0; chroma_idx<2; chroma_idx++){
+                    const uint32_t *qmul = h->dequant4_coeff[chroma_idx+1+(IS_INTRA( mb_type ) ? 0:3)][h->chroma_qp[chroma_idx]];
                     for(i4x4=0; i4x4<4; i4x4++){
-                        const int index= i4x4 + 4*i8x8;
-                        if( decode_residual(h, h->intra_gb_ptr, h->mb + (16*index << pixel_shift), index, scan + 1, h->dequant4_coeff[0][s->qscale], 15) < 0 ){
+                        const int index= 16 + 16*chroma_idx + i4x4;
+                        if( decode_residual(h, gb, h->mb + (16*index << pixel_shift), index, scan + 1, qmul, 15) < 0){
                             return -1;
                         }
                     }
                 }
             }else{
-                fill_rectangle(&h->non_zero_count_cache[scan8[0]], 4, 4, 8, 0, 1);
+                fill_rectangle(&h->non_zero_count_cache[scan8[16]], 4, 4, 8, 0, 1);
+                fill_rectangle(&h->non_zero_count_cache[scan8[32]], 4, 4, 8, 0, 1);
             }
-        }else{
-            for(i8x8=0; i8x8<4; i8x8++){
-                if(cbp & (1<<i8x8)){
-                    if(IS_8x8DCT(mb_type)){
-                        DCTELEM *buf = &h->mb[64*i8x8 << pixel_shift];
-                        uint8_t *nnz;
-                        for(i4x4=0; i4x4<4; i4x4++){
-                            if( decode_residual(h, gb, buf, i4x4+4*i8x8, scan8x8+16*i4x4,
-                                                h->dequant8_coeff[IS_INTRA( mb_type ) ? 0:1][s->qscale], 16) <0 )
-                                return -1;
-                        }
-                        nnz= &h->non_zero_count_cache[ scan8[4*i8x8] ];
-                        nnz[0] += nnz[1] + nnz[8] + nnz[9];
-                    }else{
-                        for(i4x4=0; i4x4<4; i4x4++){
-                            const int index= i4x4 + 4*i8x8;
-
-                            if( decode_residual(h, gb, h->mb + (16*index << pixel_shift), index, scan, h->dequant4_coeff[IS_INTRA( mb_type ) ? 0:3][s->qscale], 16) <0 ){
-                                return -1;
-                            }
-                        }
-                    }
-                }else{
-                    uint8_t * const nnz= &h->non_zero_count_cache[ scan8[4*i8x8] ];
-                    nnz[0] = nnz[1] = nnz[8] = nnz[9] = 0;
-                }
-            }
-        }
-
-        if(cbp&0x30){
-            for(chroma_idx=0; chroma_idx<2; chroma_idx++)
-                if( decode_residual(h, gb, h->mb + ((256 + 16*4*chroma_idx) << pixel_shift), CHROMA_DC_BLOCK_INDEX+chroma_idx, chroma_dc_scan, NULL, 4) < 0){
-                    return -1;
-                }
-        }
-
-        if(cbp&0x20){
-            for(chroma_idx=0; chroma_idx<2; chroma_idx++){
-                const uint32_t *qmul = h->dequant4_coeff[chroma_idx+1+(IS_INTRA( mb_type ) ? 0:3)][h->chroma_qp[chroma_idx]];
-                for(i4x4=0; i4x4<4; i4x4++){
-                    const int index= 16 + 4*chroma_idx + i4x4;
-                    if( decode_residual(h, gb, h->mb + (16*index << pixel_shift), index, scan + 1, qmul, 15) < 0){
-                        return -1;
-                    }
-                }
-            }
-        }else{
-            uint8_t * const nnz= &h->non_zero_count_cache[0];
-            nnz[ scan8[16]+0 ] = nnz[ scan8[16]+1 ] =nnz[ scan8[16]+8 ] =nnz[ scan8[16]+9 ] =
-            nnz[ scan8[20]+0 ] = nnz[ scan8[20]+1 ] =nnz[ scan8[20]+8 ] =nnz[ scan8[20]+9 ] = 0;
         }
     }else{
-        uint8_t * const nnz= &h->non_zero_count_cache[0];
-        fill_rectangle(&nnz[scan8[0]], 4, 4, 8, 0, 1);
-        nnz[ scan8[16]+0 ] = nnz[ scan8[16]+1 ] =nnz[ scan8[16]+8 ] =nnz[ scan8[16]+9 ] =
-        nnz[ scan8[20]+0 ] = nnz[ scan8[20]+1 ] =nnz[ scan8[20]+8 ] =nnz[ scan8[20]+9 ] = 0;
+        fill_rectangle(&h->non_zero_count_cache[scan8[ 0]], 4, 4, 8, 0, 1);
+        fill_rectangle(&h->non_zero_count_cache[scan8[16]], 4, 4, 8, 0, 1);
+        fill_rectangle(&h->non_zero_count_cache[scan8[32]], 4, 4, 8, 0, 1);
     }
     s->current_picture.qscale_table[mb_xy]= s->qscale;
     write_back_non_zero_count(h);
diff --git a/libavcodec/h264_loopfilter.c b/libavcodec/h264_loopfilter.c
index 2e61a3110a..fd1fb5f662 100644
--- a/libavcodec/h264_loopfilter.c
+++ b/libavcodec/h264_loopfilter.c
@@ -217,10 +217,11 @@ void ff_h264_filter_mb_fast( H264Context *h, int mb_x, int mb_y, uint8_t *img_y,
     int mb_xy;
     int mb_type, left_type;
     int qp, qp0, qp1, qpc, qpc0, qpc1, qp_thresh;
+    int chroma = !(CONFIG_GRAY && (s->flags&CODEC_FLAG_GRAY));
 
     mb_xy = h->mb_xy;
 
-    if(!h->top_type || !h->h264dsp.h264_loop_filter_strength || h->pps.chroma_qp_diff) {
+    if(!h->top_type || !h->h264dsp.h264_loop_filter_strength || h->pps.chroma_qp_diff || CHROMA444) {
         ff_h264_filter_mb(h, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize);
         return;
     }
@@ -262,16 +263,18 @@ void ff_h264_filter_mb_fast( H264Context *h, int mb_x, int mb_y, uint8_t *img_y,
             filter_mb_edgeh( &img_y[4*2*linesize], linesize, bS3, qp, h);
             filter_mb_edgeh( &img_y[4*3*linesize], linesize, bS3, qp, h);
         }
-        if(left_type){
-            filter_mb_edgecv( &img_cb[2*0], uvlinesize, bS4, qpc0, h);
-            filter_mb_edgecv( &img_cr[2*0], uvlinesize, bS4, qpc0, h);
+        if(chroma){
+            if(left_type){
+                filter_mb_edgecv( &img_cb[2*0], uvlinesize, bS4, qpc0, h);
+                filter_mb_edgecv( &img_cr[2*0], uvlinesize, bS4, qpc0, h);
+            }
+            filter_mb_edgecv( &img_cb[2*2], uvlinesize, bS3, qpc, h);
+            filter_mb_edgecv( &img_cr[2*2], uvlinesize, bS3, qpc, h);
+            filter_mb_edgech( &img_cb[2*0*uvlinesize], uvlinesize, bSH, qpc1, h);
+            filter_mb_edgech( &img_cb[2*2*uvlinesize], uvlinesize, bS3, qpc, h);
+            filter_mb_edgech( &img_cr[2*0*uvlinesize], uvlinesize, bSH, qpc1, h);
+            filter_mb_edgech( &img_cr[2*2*uvlinesize], uvlinesize, bS3, qpc, h);
         }
-        filter_mb_edgecv( &img_cb[2*2], uvlinesize, bS3, qpc, h);
-        filter_mb_edgecv( &img_cr[2*2], uvlinesize, bS3, qpc, h);
-        filter_mb_edgech( &img_cb[2*0*uvlinesize], uvlinesize, bSH, qpc1, h);
-        filter_mb_edgech( &img_cb[2*2*uvlinesize], uvlinesize, bS3, qpc, h);
-        filter_mb_edgech( &img_cr[2*0*uvlinesize], uvlinesize, bSH, qpc1, h);
-        filter_mb_edgech( &img_cr[2*2*uvlinesize], uvlinesize, bS3, qpc, h);
         return;
     } else {
         LOCAL_ALIGNED_8(int16_t, bS, [2], [4][4]);
@@ -298,7 +301,7 @@ void ff_h264_filter_mb_fast( H264Context *h, int mb_x, int mb_y, uint8_t *img_y,
 #define FILTER(hv,dir,edge)\
         if(AV_RN64A(bS[dir][edge])) {                                   \
             filter_mb_edge##hv( &img_y[4*edge*(dir?linesize:1)], linesize, bS[dir][edge], edge ? qp : qp##dir, h );\
-            if(!(edge&1)) {\
+            if(chroma && !(edge&1)) {\
                 filter_mb_edgec##hv( &img_cb[2*edge*(dir?uvlinesize:1)], uvlinesize, bS[dir][edge], edge ? qpc : qpc##dir, h );\
                 filter_mb_edgec##hv( &img_cr[2*edge*(dir?uvlinesize:1)], uvlinesize, bS[dir][edge], edge ? qpc : qpc##dir, h );\
             }\
@@ -353,9 +356,10 @@ static int check_mv(H264Context *h, long b_idx, long bn_idx, int mvy_limit){
     return v;
 }
 
-static av_always_inline void filter_mb_dir(H264Context *h, int mb_x, int mb_y, uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr, unsigned int linesize, unsigned int uvlinesize, int mb_xy, int mb_type, int mvy_limit, int first_vertical_edge_done, int dir) {
+static av_always_inline void filter_mb_dir(H264Context *h, int mb_x, int mb_y, uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr, unsigned int linesize, unsigned int uvlinesize, int mb_xy, int mb_type, int mvy_limit, int first_vertical_edge_done, int chroma, int chroma444, int dir) {
     MpegEncContext * const s = &h->s;
     int edge;
+    int chroma_qp_avg[2];
     const int mbm_xy = dir == 0 ? mb_xy -1 : h->top_mb_xy;
     const int mbm_type = dir == 0 ? h->left_type[0] : h->top_type;
 
@@ -394,7 +398,7 @@ static av_always_inline void filter_mb_dir(H264Context *h, int mb_x, int mb_y, u
                         bS[2]= 1+((h->cbp_table[mbn_xy] & 8)||h->non_zero_count_cache[scan8[0]+2]);
                         bS[3]= 1+((h->cbp_table[mbn_xy] & 8)||h->non_zero_count_cache[scan8[0]+3]);
                     }else{
-                    const uint8_t *mbn_nnz = h->non_zero_count[mbn_xy] + 4+3*8;
+                    const uint8_t *mbn_nnz = h->non_zero_count[mbn_xy] + 3*4;
                     int i;
                     for( i = 0; i < 4; i++ ) {
                         bS[i] = 1 + !!(h->non_zero_count_cache[scan8[0]+i] | mbn_nnz[i]);
@@ -407,10 +411,17 @@ static av_always_inline void filter_mb_dir(H264Context *h, int mb_x, int mb_y, u
                 tprintf(s->avctx, "filter mb:%d/%d dir:%d edge:%d, QPy:%d ls:%d uvls:%d", mb_x, mb_y, dir, edge, qp, tmp_linesize, tmp_uvlinesize);
                 { int i; for (i = 0; i < 4; i++) tprintf(s->avctx, " bS[%d]:%d", i, bS[i]); tprintf(s->avctx, "\n"); }
                 filter_mb_edgeh( &img_y[j*linesize], tmp_linesize, bS, qp, h );
-                filter_mb_edgech( &img_cb[j*uvlinesize], tmp_uvlinesize, bS,
-                                ( h->chroma_qp[0] + get_chroma_qp( h, 0, s->current_picture.qscale_table[mbn_xy] ) + 1 ) >> 1, h);
-                filter_mb_edgech( &img_cr[j*uvlinesize], tmp_uvlinesize, bS,
-                                ( h->chroma_qp[1] + get_chroma_qp( h, 1, s->current_picture.qscale_table[mbn_xy] ) + 1 ) >> 1, h);
+                chroma_qp_avg[0] = (h->chroma_qp[0] + get_chroma_qp( h, 0, s->current_picture.qscale_table[mbn_xy] ) + 1 ) >> 1;
+                chroma_qp_avg[1] = (h->chroma_qp[1] + get_chroma_qp( h, 1, s->current_picture.qscale_table[mbn_xy] ) + 1 ) >> 1;
+                if (chroma) {
+                    if (chroma444) {
+                        filter_mb_edgeh (&img_cb[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[0], h);
+                        filter_mb_edgeh (&img_cr[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[1], h);
+                    } else {
+                        filter_mb_edgech(&img_cb[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[0], h);
+                        filter_mb_edgech(&img_cr[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[1], h);
+                    }
+                }
             }
         }else{
             DECLARE_ALIGNED(8, int16_t, bS)[4];
@@ -465,23 +476,29 @@ static av_always_inline void filter_mb_dir(H264Context *h, int mb_x, int mb_y, u
                 //tprintf(s->avctx, "filter mb:%d/%d dir:%d edge:%d, QPy:%d, QPc:%d, QPcn:%d\n", mb_x, mb_y, dir, edge, qp, h->chroma_qp[0], s->current_picture.qscale_table[mbn_xy]);
                 tprintf(s->avctx, "filter mb:%d/%d dir:%d edge:%d, QPy:%d ls:%d uvls:%d", mb_x, mb_y, dir, edge, qp, linesize, uvlinesize);
                 //{ int i; for (i = 0; i < 4; i++) tprintf(s->avctx, " bS[%d]:%d", i, bS[i]); tprintf(s->avctx, "\n"); }
+                chroma_qp_avg[0] = (h->chroma_qp[0] + get_chroma_qp( h, 0, s->current_picture.qscale_table[mbm_xy] ) + 1 ) >> 1;
+                chroma_qp_avg[1] = (h->chroma_qp[1] + get_chroma_qp( h, 1, s->current_picture.qscale_table[mbm_xy] ) + 1 ) >> 1;
                 if( dir == 0 ) {
                     filter_mb_edgev( &img_y[0], linesize, bS, qp, h );
-                    {
-                        int qp= ( h->chroma_qp[0] + get_chroma_qp( h, 0, s->current_picture.qscale_table[mbm_xy] ) + 1 ) >> 1;
-                        filter_mb_edgecv( &img_cb[0], uvlinesize, bS, qp, h);
-                        if(h->pps.chroma_qp_diff)
-                            qp= ( h->chroma_qp[1] + get_chroma_qp( h, 1, s->current_picture.qscale_table[mbm_xy] ) + 1 ) >> 1;
-                        filter_mb_edgecv( &img_cr[0], uvlinesize, bS, qp, h);
+                    if (chroma) {
+                        if (chroma444) {
+                            filter_mb_edgev ( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], h);
+                            filter_mb_edgev ( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], h);
+                        } else {
+                            filter_mb_edgecv( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], h);
+                            filter_mb_edgecv( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], h);
+                        }
                     }
                 } else {
                     filter_mb_edgeh( &img_y[0], linesize, bS, qp, h );
-                    {
-                        int qp= ( h->chroma_qp[0] + get_chroma_qp( h, 0, s->current_picture.qscale_table[mbm_xy] ) + 1 ) >> 1;
-                        filter_mb_edgech( &img_cb[0], uvlinesize, bS, qp, h);
-                        if(h->pps.chroma_qp_diff)
-                            qp= ( h->chroma_qp[1] + get_chroma_qp( h, 1, s->current_picture.qscale_table[mbm_xy] ) + 1 ) >> 1;
-                        filter_mb_edgech( &img_cr[0], uvlinesize, bS, qp, h);
+                    if (chroma) {
+                        if (chroma444) {
+                            filter_mb_edgeh ( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], h);
+                            filter_mb_edgeh ( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], h);
+                        } else {
+                            filter_mb_edgech( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], h);
+                            filter_mb_edgech( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], h);
+                        }
                     }
                 }
             }
@@ -545,15 +562,25 @@ static av_always_inline void filter_mb_dir(H264Context *h, int mb_x, int mb_y, u
         //{ int i; for (i = 0; i < 4; i++) tprintf(s->avctx, " bS[%d]:%d", i, bS[i]); tprintf(s->avctx, "\n"); }
         if( dir == 0 ) {
             filter_mb_edgev( &img_y[4*edge << h->pixel_shift], linesize, bS, qp, h );
-            if( (edge&1) == 0 ) {
-                filter_mb_edgecv( &img_cb[2*edge << h->pixel_shift], uvlinesize, bS, h->chroma_qp[0], h);
-                filter_mb_edgecv( &img_cr[2*edge << h->pixel_shift], uvlinesize, bS, h->chroma_qp[1], h);
+            if (chroma) {
+                if (chroma444) {
+                    filter_mb_edgev ( &img_cb[4*edge << h->pixel_shift], uvlinesize, bS, h->chroma_qp[0], h);
+                    filter_mb_edgev ( &img_cr[4*edge << h->pixel_shift], uvlinesize, bS, h->chroma_qp[1], h);
+                } else if( (edge&1) == 0 ) {
+                    filter_mb_edgecv( &img_cb[2*edge << h->pixel_shift], uvlinesize, bS, h->chroma_qp[0], h);
+                    filter_mb_edgecv( &img_cr[2*edge << h->pixel_shift], uvlinesize, bS, h->chroma_qp[1], h);
+                }
             }
         } else {
             filter_mb_edgeh( &img_y[4*edge*linesize], linesize, bS, qp, h );
-            if( (edge&1) == 0 ) {
-                filter_mb_edgech( &img_cb[2*edge*uvlinesize], uvlinesize, bS, h->chroma_qp[0], h);
-                filter_mb_edgech( &img_cr[2*edge*uvlinesize], uvlinesize, bS, h->chroma_qp[1], h);
+            if (chroma) {
+                if (chroma444) {
+                    filter_mb_edgeh ( &img_cb[4*edge*uvlinesize], uvlinesize, bS, h->chroma_qp[0], h);
+                    filter_mb_edgeh ( &img_cr[4*edge*uvlinesize], uvlinesize, bS, h->chroma_qp[1], h);
+                } else if( (edge&1) == 0 ) {
+                    filter_mb_edgech( &img_cb[2*edge*uvlinesize], uvlinesize, bS, h->chroma_qp[0], h);
+                    filter_mb_edgech( &img_cr[2*edge*uvlinesize], uvlinesize, bS, h->chroma_qp[1], h);
+                }
             }
         }
     }
@@ -566,6 +593,7 @@ void ff_h264_filter_mb( H264Context *h, int mb_x, int mb_y, uint8_t *img_y, uint
     const int mvy_limit = IS_INTERLACED(mb_type) ? 2 : 4;
     int first_vertical_edge_done = 0;
     av_unused int dir;
+    int chroma = !(CONFIG_GRAY && (s->flags&CODEC_FLAG_GRAY));
 
     if (FRAME_MBAFF
             // and current and left pair do not have the same interlaced type
@@ -589,11 +617,11 @@ void ff_h264_filter_mb( H264Context *h, int mb_x, int mb_y, uint8_t *img_y, uint
         } else {
             static const uint8_t offset[2][2][8]={
                 {
-                    {7+8*0, 7+8*0, 7+8*0, 7+8*0, 7+8*1, 7+8*1, 7+8*1, 7+8*1},
-                    {7+8*2, 7+8*2, 7+8*2, 7+8*2, 7+8*3, 7+8*3, 7+8*3, 7+8*3},
+                    {3+4*0, 3+4*0, 3+4*0, 3+4*0, 3+4*1, 3+4*1, 3+4*1, 3+4*1},
+                    {3+4*2, 3+4*2, 3+4*2, 3+4*2, 3+4*3, 3+4*3, 3+4*3, 3+4*3},
                 },{
-                    {7+8*0, 7+8*1, 7+8*2, 7+8*3, 7+8*0, 7+8*1, 7+8*2, 7+8*3},
-                    {7+8*0, 7+8*1, 7+8*2, 7+8*3, 7+8*0, 7+8*1, 7+8*2, 7+8*3},
+                    {3+4*0, 3+4*1, 3+4*2, 3+4*3, 3+4*0, 3+4*1, 3+4*2, 3+4*3},
+                    {3+4*0, 3+4*1, 3+4*2, 3+4*3, 3+4*0, 3+4*1, 3+4*2, 3+4*3},
                 }
             };
             const uint8_t *off= offset[MB_FIELD][mb_y&1];
@@ -634,25 +662,29 @@ void ff_h264_filter_mb( H264Context *h, int mb_x, int mb_y, uint8_t *img_y, uint
         if(MB_FIELD){
             filter_mb_mbaff_edgev ( h, img_y                ,   linesize, bS  , 1, qp [0] );
             filter_mb_mbaff_edgev ( h, img_y  + 8*  linesize,   linesize, bS+4, 1, qp [1] );
-            filter_mb_mbaff_edgecv( h, img_cb,                uvlinesize, bS  , 1, bqp[0] );
-            filter_mb_mbaff_edgecv( h, img_cb + 4*uvlinesize, uvlinesize, bS+4, 1, bqp[1] );
-            filter_mb_mbaff_edgecv( h, img_cr,                uvlinesize, bS  , 1, rqp[0] );
-            filter_mb_mbaff_edgecv( h, img_cr + 4*uvlinesize, uvlinesize, bS+4, 1, rqp[1] );
+            if (chroma){
+                filter_mb_mbaff_edgecv( h, img_cb,                uvlinesize, bS  , 1, bqp[0] );
+                filter_mb_mbaff_edgecv( h, img_cb + 4*uvlinesize, uvlinesize, bS+4, 1, bqp[1] );
+                filter_mb_mbaff_edgecv( h, img_cr,                uvlinesize, bS  , 1, rqp[0] );
+                filter_mb_mbaff_edgecv( h, img_cr + 4*uvlinesize, uvlinesize, bS+4, 1, rqp[1] );
+            }
         }else{
             filter_mb_mbaff_edgev ( h, img_y              , 2*  linesize, bS  , 2, qp [0] );
             filter_mb_mbaff_edgev ( h, img_y  +   linesize, 2*  linesize, bS+1, 2, qp [1] );
-            filter_mb_mbaff_edgecv( h, img_cb,              2*uvlinesize, bS  , 2, bqp[0] );
-            filter_mb_mbaff_edgecv( h, img_cb + uvlinesize, 2*uvlinesize, bS+1, 2, bqp[1] );
-            filter_mb_mbaff_edgecv( h, img_cr,              2*uvlinesize, bS  , 2, rqp[0] );
-            filter_mb_mbaff_edgecv( h, img_cr + uvlinesize, 2*uvlinesize, bS+1, 2, rqp[1] );
+            if (chroma){
+                filter_mb_mbaff_edgecv( h, img_cb,              2*uvlinesize, bS  , 2, bqp[0] );
+                filter_mb_mbaff_edgecv( h, img_cb + uvlinesize, 2*uvlinesize, bS+1, 2, bqp[1] );
+                filter_mb_mbaff_edgecv( h, img_cr,              2*uvlinesize, bS  , 2, rqp[0] );
+                filter_mb_mbaff_edgecv( h, img_cr + uvlinesize, 2*uvlinesize, bS+1, 2, rqp[1] );
+            }
         }
     }
 
 #if CONFIG_SMALL
     for( dir = 0; dir < 2; dir++ )
-        filter_mb_dir(h, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, dir ? 0 : first_vertical_edge_done, dir);
+        filter_mb_dir(h, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, dir ? 0 : first_vertical_edge_done, chroma, CHROMA444, dir);
 #else
-    filter_mb_dir(h, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, first_vertical_edge_done, 0);
-    filter_mb_dir(h, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, 0, 1);
+    filter_mb_dir(h, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, first_vertical_edge_done, chroma, CHROMA444, 0);
+    filter_mb_dir(h, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, 0, chroma, CHROMA444, 1);
 #endif
 }
diff --git a/libavcodec/h264_ps.c b/libavcodec/h264_ps.c
index f77a013112..ddfe1d2c64 100644
--- a/libavcodec/h264_ps.c
+++ b/libavcodec/h264_ps.c
@@ -269,7 +269,7 @@ static void decode_scaling_matrices(H264Context *h, SPS *sps, PPS *pps, int is_s
         fallback_sps ? sps->scaling_matrix4[0] : default_scaling4[0],
         fallback_sps ? sps->scaling_matrix4[3] : default_scaling4[1],
         fallback_sps ? sps->scaling_matrix8[0] : default_scaling8[0],
-        fallback_sps ? sps->scaling_matrix8[1] : default_scaling8[1]
+        fallback_sps ? sps->scaling_matrix8[3] : default_scaling8[1]
     };
     if(get_bits1(&s->gb)){
         sps->scaling_matrix_present |= is_sps;
@@ -281,7 +281,15 @@ static void decode_scaling_matrices(H264Context *h, SPS *sps, PPS *pps, int is_s
         decode_scaling_list(h,scaling_matrix4[5],16,default_scaling4[1],scaling_matrix4[4]); // Inter, Cb
         if(is_sps || pps->transform_8x8_mode){
             decode_scaling_list(h,scaling_matrix8[0],64,default_scaling8[0],fallback[2]);  // Intra, Y
-            decode_scaling_list(h,scaling_matrix8[1],64,default_scaling8[1],fallback[3]);  // Inter, Y
+            if(h->sps.chroma_format_idc == 3){
+                decode_scaling_list(h,scaling_matrix8[1],64,default_scaling8[0],scaling_matrix8[0]);  // Intra, Cr
+                decode_scaling_list(h,scaling_matrix8[2],64,default_scaling8[0],scaling_matrix8[1]);  // Intra, Cb
+            }
+            decode_scaling_list(h,scaling_matrix8[3],64,default_scaling8[1],fallback[3]);  // Inter, Y
+            if(h->sps.chroma_format_idc == 3){
+                decode_scaling_list(h,scaling_matrix8[4],64,default_scaling8[1],scaling_matrix8[3]);  // Inter, Cr
+                decode_scaling_list(h,scaling_matrix8[5],64,default_scaling8[1],scaling_matrix8[4]);  // Inter, Cb
+            }
         }
     }
 }
@@ -395,7 +403,7 @@ int ff_h264_decode_seq_parameter_set(H264Context *h){
         if(sps->crop_left || sps->crop_top){
             av_log(h->s.avctx, AV_LOG_ERROR, "insane cropping not completely supported, this could look slightly wrong ...\n");
         }
-        if(sps->crop_right >= 8 || sps->crop_bottom >= 8){
+        if(sps->crop_right >= (8<<CHROMA444) || sps->crop_bottom >= (8<<CHROMA444)){
             av_log(h->s.avctx, AV_LOG_ERROR, "brainfart cropping not supported, this could look slightly wrong ...\n");
         }
     }else{
diff --git a/libavcodec/h264dsp.h b/libavcodec/h264dsp.h
index 6ed0b537cd..b61967adfc 100644
--- a/libavcodec/h264dsp.h
+++ b/libavcodec/h264dsp.h
@@ -66,10 +66,10 @@ typedef struct H264DSPContext{
     void (*h264_idct_dc_add)(uint8_t *dst/*align 4*/, DCTELEM *block/*align 16*/, int stride);
     void (*h264_idct8_dc_add)(uint8_t *dst/*align 8*/, DCTELEM *block/*align 16*/, int stride);
 
-    void (*h264_idct_add16)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
-    void (*h264_idct8_add4)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
-    void (*h264_idct_add8)(uint8_t **dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
-    void (*h264_idct_add16intra)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[6*8]);
+    void (*h264_idct_add16)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[15*8]);
+    void (*h264_idct8_add4)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[15*8]);
+    void (*h264_idct_add8)(uint8_t **dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[15*8]);
+    void (*h264_idct_add16intra)(uint8_t *dst/*align 16*/, const int *blockoffset, DCTELEM *block/*align 16*/, int stride, const uint8_t nnzc[15*8]);
     void (*h264_luma_dc_dequant_idct)(DCTELEM *output, DCTELEM *input/*align 16*/, int qmul);
     void (*h264_chroma_dc_dequant_idct)(DCTELEM *block, int qmul);
 }H264DSPContext;
diff --git a/libavcodec/h264idct_template.c b/libavcodec/h264idct_template.c
index 1b3c635319..83f6f38691 100644
--- a/libavcodec/h264idct_template.c
+++ b/libavcodec/h264idct_template.c
@@ -30,15 +30,19 @@
 #ifndef AVCODEC_H264IDCT_INTERNAL_H
 #define AVCODEC_H264IDCT_INTERNAL_H
 //FIXME this table is a duplicate from h264data.h, and will be removed once the tables from, h264 have been split
-static const uint8_t scan8[16 + 2*4]={
- 4+1*8, 5+1*8, 4+2*8, 5+2*8,
- 6+1*8, 7+1*8, 6+2*8, 7+2*8,
- 4+3*8, 5+3*8, 4+4*8, 5+4*8,
- 6+3*8, 7+3*8, 6+4*8, 7+4*8,
- 1+1*8, 2+1*8,
- 1+2*8, 2+2*8,
- 1+4*8, 2+4*8,
- 1+5*8, 2+5*8,
+static const uint8_t scan8[16*3]={
+ 4+ 1*8, 5+ 1*8, 4+ 2*8, 5+ 2*8,
+ 6+ 1*8, 7+ 1*8, 6+ 2*8, 7+ 2*8,
+ 4+ 3*8, 5+ 3*8, 4+ 4*8, 5+ 4*8,
+ 6+ 3*8, 7+ 3*8, 6+ 4*8, 7+ 4*8,
+ 4+ 6*8, 5+ 6*8, 4+ 7*8, 5+ 7*8,
+ 6+ 6*8, 7+ 6*8, 6+ 7*8, 7+ 7*8,
+ 4+ 8*8, 5+ 8*8, 4+ 9*8, 5+ 9*8,
+ 6+ 8*8, 7+ 8*8, 6+ 9*8, 7+ 9*8,
+ 4+11*8, 5+11*8, 4+12*8, 5+12*8,
+ 6+11*8, 7+11*8, 6+12*8, 7+12*8,
+ 4+13*8, 5+13*8, 4+14*8, 5+14*8,
+ 6+13*8, 7+13*8, 6+14*8, 7+14*8
 };
 #endif
 
@@ -190,7 +194,7 @@ void FUNCC(ff_h264_idct8_dc_add)(uint8_t *p_dst, DCTELEM *block, int stride){
     }
 }
 
-void FUNCC(ff_h264_idct_add16)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+void FUNCC(ff_h264_idct_add16)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
     int i;
     for(i=0; i<16; i++){
         int nnz = nnzc[ scan8[i] ];
@@ -201,7 +205,7 @@ void FUNCC(ff_h264_idct_add16)(uint8_t *dst, const int *block_offset, DCTELEM *b
     }
 }
 
-void FUNCC(ff_h264_idct_add16intra)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+void FUNCC(ff_h264_idct_add16intra)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
     int i;
     for(i=0; i<16; i++){
         if(nnzc[ scan8[i] ])             FUNCC(idct_internal      )(dst + block_offset[i], block + i*16*sizeof(pixel), stride, 4, 6, 1);
@@ -209,7 +213,7 @@ void FUNCC(ff_h264_idct_add16intra)(uint8_t *dst, const int *block_offset, DCTEL
     }
 }
 
-void FUNCC(ff_h264_idct8_add4)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+void FUNCC(ff_h264_idct8_add4)(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
     int i;
     for(i=0; i<16; i+=4){
         int nnz = nnzc[ scan8[i] ];
@@ -220,13 +224,15 @@ void FUNCC(ff_h264_idct8_add4)(uint8_t *dst, const int *block_offset, DCTELEM *b
     }
 }
 
-void FUNCC(ff_h264_idct_add8)(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
-    int i;
-    for(i=16; i<16+8; i++){
-        if(nnzc[ scan8[i] ])
-            FUNCC(ff_h264_idct_add   )(dest[(i&4)>>2] + block_offset[i], block + i*16*sizeof(pixel), stride);
-        else if(((dctcoef*)block)[i*16])
-            FUNCC(ff_h264_idct_dc_add)(dest[(i&4)>>2] + block_offset[i], block + i*16*sizeof(pixel), stride);
+void FUNCC(ff_h264_idct_add8)(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
+    int i, j;
+    for(j=1; j<3; j++){
+        for(i=j*16; i<j*16+4; i++){
+            if(nnzc[ scan8[i] ])
+                FUNCC(ff_h264_idct_add   )(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
+            else if(((dctcoef*)block)[i*16])
+                FUNCC(ff_h264_idct_dc_add)(dest[j-1] + block_offset[i], block + i*16*sizeof(pixel), stride);
+        }
     }
 }
 /**
diff --git a/libavcodec/mpegvideo.c b/libavcodec/mpegvideo.c
index 8e9a79167d..c3c3e15033 100644
--- a/libavcodec/mpegvideo.c
+++ b/libavcodec/mpegvideo.c
@@ -1180,12 +1180,17 @@ void MPV_frame_end(MpegEncContext *s)
        && s->current_picture.reference
        && !s->intra_only
        && !(s->flags&CODEC_FLAG_EMU_EDGE)) {
-        int edges = EDGE_BOTTOM | EDGE_TOP, h = s->v_edge_pos;
-
-            s->dsp.draw_edges(s->current_picture.data[0], s->linesize  , s->h_edge_pos   , h   , EDGE_WIDTH  , edges);
-            s->dsp.draw_edges(s->current_picture.data[1], s->uvlinesize, s->h_edge_pos>>1, h>>1, EDGE_WIDTH/2, edges);
-            s->dsp.draw_edges(s->current_picture.data[2], s->uvlinesize, s->h_edge_pos>>1, h>>1, EDGE_WIDTH/2, edges);
-
+            int hshift = av_pix_fmt_descriptors[s->avctx->pix_fmt].log2_chroma_w;
+            int vshift = av_pix_fmt_descriptors[s->avctx->pix_fmt].log2_chroma_h;
+            s->dsp.draw_edges(s->current_picture.data[0], s->linesize  ,
+                              s->h_edge_pos             , s->v_edge_pos,
+                              EDGE_WIDTH        , EDGE_WIDTH        , EDGE_TOP | EDGE_BOTTOM);
+            s->dsp.draw_edges(s->current_picture.data[1], s->uvlinesize,
+                              s->h_edge_pos>>hshift, s->v_edge_pos>>vshift,
+                              EDGE_WIDTH>>hshift, EDGE_WIDTH>>vshift, EDGE_TOP | EDGE_BOTTOM);
+            s->dsp.draw_edges(s->current_picture.data[2], s->uvlinesize,
+                              s->h_edge_pos>>hshift, s->v_edge_pos>>vshift,
+                              EDGE_WIDTH>>hshift, EDGE_WIDTH>>vshift, EDGE_TOP | EDGE_BOTTOM);
     }
 
     emms_c();
@@ -2289,14 +2294,19 @@ void ff_draw_horiz_band(MpegEncContext *s, int y, int h){
        && !s->intra_only
        && !(s->flags&CODEC_FLAG_EMU_EDGE)) {
         int sides = 0, edge_h;
+        int hshift = av_pix_fmt_descriptors[s->avctx->pix_fmt].log2_chroma_w;
+        int vshift = av_pix_fmt_descriptors[s->avctx->pix_fmt].log2_chroma_h;
         if (y==0) sides |= EDGE_TOP;
         if (y + h >= s->v_edge_pos) sides |= EDGE_BOTTOM;
 
         edge_h= FFMIN(h, s->v_edge_pos - y);
 
-        s->dsp.draw_edges(s->current_picture_ptr->data[0] +  y    *s->linesize  , s->linesize  , s->h_edge_pos   , edge_h   , EDGE_WIDTH  , sides);
-        s->dsp.draw_edges(s->current_picture_ptr->data[1] + (y>>1)*s->uvlinesize, s->uvlinesize, s->h_edge_pos>>1, edge_h>>1, EDGE_WIDTH/2, sides);
-        s->dsp.draw_edges(s->current_picture_ptr->data[2] + (y>>1)*s->uvlinesize, s->uvlinesize, s->h_edge_pos>>1, edge_h>>1, EDGE_WIDTH/2, sides);
+        s->dsp.draw_edges(s->current_picture_ptr->data[0] +  y         *s->linesize  , s->linesize,
+                          s->h_edge_pos        , edge_h        , EDGE_WIDTH        , EDGE_WIDTH        , sides);
+        s->dsp.draw_edges(s->current_picture_ptr->data[1] + (y>>vshift)*s->uvlinesize, s->uvlinesize,
+                          s->h_edge_pos>>hshift, edge_h>>hshift, EDGE_WIDTH>>hshift, EDGE_WIDTH>>vshift, sides);
+        s->dsp.draw_edges(s->current_picture_ptr->data[2] + (y>>vshift)*s->uvlinesize, s->uvlinesize,
+                          s->h_edge_pos>>hshift, edge_h>>hshift, EDGE_WIDTH>>hshift, EDGE_WIDTH>>vshift, sides);
     }
 
     h= FFMIN(h, s->avctx->height - y);
diff --git a/libavcodec/ppc/h264_altivec.c b/libavcodec/ppc/h264_altivec.c
index 9df18888ad..9ba6bbaf2e 100644
--- a/libavcodec/ppc/h264_altivec.c
+++ b/libavcodec/ppc/h264_altivec.c
@@ -527,7 +527,7 @@ static void ff_h264_idct8_dc_add_altivec(uint8_t *dst, DCTELEM *block, int strid
     h264_idct_dc_add_internal(dst, block, stride, 8);
 }
 
-static void ff_h264_idct_add16_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+static void ff_h264_idct_add16_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
     int i;
     for(i=0; i<16; i++){
         int nnz = nnzc[ scan8[i] ];
@@ -538,7 +538,7 @@ static void ff_h264_idct_add16_altivec(uint8_t *dst, const int *block_offset, DC
     }
 }
 
-static void ff_h264_idct_add16intra_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+static void ff_h264_idct_add16intra_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
     int i;
     for(i=0; i<16; i++){
         if(nnzc[ scan8[i] ]) ff_h264_idct_add_altivec(dst + block_offset[i], block + i*16, stride);
@@ -546,7 +546,7 @@ static void ff_h264_idct_add16intra_altivec(uint8_t *dst, const int *block_offse
     }
 }
 
-static void ff_h264_idct8_add4_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
+static void ff_h264_idct8_add4_altivec(uint8_t *dst, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
     int i;
     for(i=0; i<16; i+=4){
         int nnz = nnzc[ scan8[i] ];
@@ -557,13 +557,15 @@ static void ff_h264_idct8_add4_altivec(uint8_t *dst, const int *block_offset, DC
     }
 }
 
-static void ff_h264_idct_add8_altivec(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[6*8]){
-    int i;
-    for(i=16; i<16+8; i++){
-        if(nnzc[ scan8[i] ])
-            ff_h264_idct_add_altivec(dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
-        else if(block[i*16])
-            h264_idct_dc_add_altivec(dest[(i&4)>>2] + block_offset[i], block + i*16, stride);
+static void ff_h264_idct_add8_altivec(uint8_t **dest, const int *block_offset, DCTELEM *block, int stride, const uint8_t nnzc[15*8]){
+    int i, j;
+    for (j = 1; j < 3; j++) {
+        for(i = j * 16; i < j * 16 + 4; i++){
+            if(nnzc[ scan8[i] ])
+                ff_h264_idct_add_altivec(dest[j-1] + block_offset[i], block + i*16, stride);
+            else if(block[i*16])
+                h264_idct_dc_add_altivec(dest[j-1] + block_offset[i], block + i*16, stride);
+        }
     }
 }
 
diff --git a/libavcodec/snow.c b/libavcodec/snow.c
index 4ee1722e4c..43a71904eb 100644
--- a/libavcodec/snow.c
+++ b/libavcodec/snow.c
@@ -1976,9 +1976,15 @@ static int frame_start(SnowContext *s){
    int h= s->avctx->height;
 
     if(s->current_picture.data[0]){
-        s->dsp.draw_edges(s->current_picture.data[0], s->current_picture.linesize[0], w   , h   , EDGE_WIDTH  , EDGE_TOP|EDGE_BOTTOM);
-        s->dsp.draw_edges(s->current_picture.data[1], s->current_picture.linesize[1], w>>1, h>>1, EDGE_WIDTH/2, EDGE_TOP|EDGE_BOTTOM);
-        s->dsp.draw_edges(s->current_picture.data[2], s->current_picture.linesize[2], w>>1, h>>1, EDGE_WIDTH/2, EDGE_TOP|EDGE_BOTTOM);
+        s->dsp.draw_edges(s->current_picture.data[0],
+                          s->current_picture.linesize[0], w   , h   ,
+                          EDGE_WIDTH  , EDGE_WIDTH  , EDGE_TOP | EDGE_BOTTOM);
+        s->dsp.draw_edges(s->current_picture.data[1],
+                          s->current_picture.linesize[1], w>>1, h>>1,
+                          EDGE_WIDTH/2, EDGE_WIDTH/2, EDGE_TOP | EDGE_BOTTOM);
+        s->dsp.draw_edges(s->current_picture.data[2],
+                          s->current_picture.linesize[2], w>>1, h>>1,
+                          EDGE_WIDTH/2, EDGE_WIDTH/2, EDGE_TOP | EDGE_BOTTOM);
     }
 
     release_buffer(s->avctx);
diff --git a/libavcodec/svq3.c b/libavcodec/svq3.c
index 9fec0350d2..30cd0f102f 100644
--- a/libavcodec/svq3.c
+++ b/libavcodec/svq3.c
@@ -635,8 +635,9 @@ static int svq3_decode_mb(SVQ3Context *svq3, unsigned int mb_type)
         memset(h->intra4x4_pred_mode+h->mb2br_xy[mb_xy], DC_PRED, 8);
     }
     if (!IS_SKIP(mb_type) || s->pict_type == AV_PICTURE_TYPE_B) {
-        memset(h->non_zero_count_cache + 8, 0, 4*9*sizeof(uint8_t));
-        s->dsp.clear_blocks(h->mb);
+        memset(h->non_zero_count_cache + 8, 0, 14*8*sizeof(uint8_t));
+        s->dsp.clear_blocks(h->mb+  0);
+        s->dsp.clear_blocks(h->mb+384);
     }
 
     if (!IS_INTRA16x16(mb_type) && (!IS_SKIP(mb_type) || s->pict_type == AV_PICTURE_TYPE_B)) {
@@ -656,8 +657,8 @@ static int svq3_decode_mb(SVQ3Context *svq3, unsigned int mb_type)
         }
     }
     if (IS_INTRA16x16(mb_type)) {
-        AV_ZERO128(h->mb_luma_dc+0);
-        AV_ZERO128(h->mb_luma_dc+8);
+        AV_ZERO128(h->mb_luma_dc[0]+0);
+        AV_ZERO128(h->mb_luma_dc[0]+8);
         if (svq3_decode_block(&s->gb, h->mb_luma_dc, 0, 1)){
             av_log(h->s.avctx, AV_LOG_ERROR, "error while decoding intra luma dc\n");
             return -1;
@@ -683,20 +684,23 @@ static int svq3_decode_mb(SVQ3Context *svq3, unsigned int mb_type)
         }
 
         if ((cbp & 0x30)) {
-            for (i = 0; i < 2; ++i) {
-              if (svq3_decode_block(&s->gb, &h->mb[16*(16 + 4*i)], 0, 3)){
+            for (i = 1; i < 3; ++i) {
+              if (svq3_decode_block(&s->gb, &h->mb[16*16*i], 0, 3)){
                 av_log(h->s.avctx, AV_LOG_ERROR, "error while decoding chroma dc block\n");
                 return -1;
               }
             }
 
             if ((cbp & 0x20)) {
-                for (i = 0; i < 8; i++) {
-                    h->non_zero_count_cache[ scan8[16+i] ] = 1;
-
-                    if (svq3_decode_block(&s->gb, &h->mb[16*(16 + i)], 1, 1)){
-                        av_log(h->s.avctx, AV_LOG_ERROR, "error while decoding chroma ac block\n");
-                        return -1;
+                for (i = 1; i < 3; i++) {
+                    for (j = 0; j < 4; j++) {
+                        k = 16*i + j;
+                        h->non_zero_count_cache[ scan8[k] ] = 1;
+
+                        if (svq3_decode_block(&s->gb, &h->mb[16*k], 1, 1)){
+                            av_log(h->s.avctx, AV_LOG_ERROR, "error while decoding chroma ac block\n");
+                            return -1;
+                        }
                     }
                 }
             }
diff --git a/libavcodec/x86/dsputil_mmx.c b/libavcodec/x86/dsputil_mmx.c
index d867dc3e6a..71bbd11741 100644
--- a/libavcodec/x86/dsputil_mmx.c
+++ b/libavcodec/x86/dsputil_mmx.c
@@ -762,7 +762,7 @@ static void h263_h_loop_filter_mmx(uint8_t *src, int stride, int qscale){
 
 /* draw the edges of width 'w' of an image of size width, height
    this mmx version can only handle w==8 || w==16 */
-static void draw_edges_mmx(uint8_t *buf, int wrap, int width, int height, int w, int sides)
+static void draw_edges_mmx(uint8_t *buf, int wrap, int width, int height, int w, int h, int sides)
 {
     uint8_t *ptr, *last_line;
     int i;
@@ -817,7 +817,7 @@ static void draw_edges_mmx(uint8_t *buf, int wrap, int width, int height, int w,
 
     /* top and bottom (and hopefully also the corners) */
     if (sides&EDGE_TOP) {
-        for(i = 0; i < w; i += 4) {
+        for(i = 0; i < h; i += 4) {
             ptr= buf - (i + 1) * wrap - w;
             __asm__ volatile(
                     "1:                             \n\t"
diff --git a/libavcodec/x86/h264_i386.h b/libavcodec/x86/h264_i386.h
index 26c163be3c..2f55879ce1 100644
--- a/libavcodec/x86/h264_i386.h
+++ b/libavcodec/x86/h264_i386.h
@@ -36,7 +36,7 @@
 #if ARCH_X86 && HAVE_7REGS && HAVE_EBX_AVAILABLE && !defined(BROKEN_RELOCATIONS)
 static int decode_significance_x86(CABACContext *c, int max_coeff,
                                    uint8_t *significant_coeff_ctx_base,
-                                   int *index){
+                                   int *index, x86_reg last_off){
     void *end= significant_coeff_ctx_base + max_coeff - 1;
     int minusstart= -(int)significant_coeff_ctx_base;
     int minusindex= 4-(int)index;
@@ -52,10 +52,12 @@ static int decode_significance_x86(CABACContext *c, int max_coeff,
 
         "test $1, %%edx                         \n\t"
         " jz 3f                                 \n\t"
+        "add  %7, %1                            \n\t"
 
-        BRANCHLESS_GET_CABAC("%%edx", "%3", "61(%1)", "%%ebx",
+        BRANCHLESS_GET_CABAC("%%edx", "%3", "(%1)", "%%ebx",
                              "%%bx", "%%esi", "%%eax", "%%al")
 
+        "sub  %7, %1                            \n\t"
         "mov  %2, %%"REG_a"                     \n\t"
         "movl %4, %%ecx                         \n\t"
         "add  %1, %%"REG_c"                     \n\t"
@@ -82,7 +84,7 @@ static int decode_significance_x86(CABACContext *c, int max_coeff,
         "movl %%esi, "RANGE    "(%3)            \n\t"
         "movl %%ebx, "LOW      "(%3)            \n\t"
         :"=&a"(coeff_count), "+r"(significant_coeff_ctx_base), "+m"(index)
-        :"r"(c), "m"(minusstart), "m"(end), "m"(minusindex)
+        :"r"(c), "m"(minusstart), "m"(end), "m"(minusindex), "m"(last_off)
         : "%"REG_c, "%ebx", "%edx", "%esi", "memory"
     );
     return coeff_count;
@@ -90,7 +92,7 @@ static int decode_significance_x86(CABACContext *c, int max_coeff,
 
 static int decode_significance_8x8_x86(CABACContext *c,
                                        uint8_t *significant_coeff_ctx_base,
-                                       int *index, const uint8_t *sig_off){
+                                       int *index, x86_reg last_off, const uint8_t *sig_off){
     int minusindex= 4-(int)index;
     int coeff_count;
     x86_reg last=0;
@@ -114,8 +116,9 @@ static int decode_significance_8x8_x86(CABACContext *c,
 
         "movzbl "MANGLE(last_coeff_flag_offset_8x8)"(%%edi), %%edi\n\t"
         "add %5, %%"REG_D"                      \n\t"
+        "add %7, %%"REG_D"                      \n\t"
 
-        BRANCHLESS_GET_CABAC("%%edx", "%3", "15(%%"REG_D")", "%%ebx",
+        BRANCHLESS_GET_CABAC("%%edx", "%3", "(%%"REG_D")", "%%ebx",
                              "%%bx", "%%esi", "%%eax", "%%al")
 
         "mov %2, %%"REG_a"                      \n\t"
@@ -142,7 +145,7 @@ static int decode_significance_8x8_x86(CABACContext *c,
         "movl %%esi, "RANGE    "(%3)            \n\t"
         "movl %%ebx, "LOW      "(%3)            \n\t"
         :"=&a"(coeff_count),"+m"(last), "+m"(index)
-        :"r"(c), "m"(minusindex), "m"(significant_coeff_ctx_base), "m"(sig_off)
+        :"r"(c), "m"(minusindex), "m"(significant_coeff_ctx_base), "m"(sig_off), "m"(last_off)
         : "%"REG_c, "%ebx", "%edx", "%esi", "%"REG_D, "memory"
     );
     return coeff_count;
diff --git a/libavcodec/x86/h264_idct.asm b/libavcodec/x86/h264_idct.asm
index 87fbbf2fd5..04dabc3a2d 100644
--- a/libavcodec/x86/h264_idct.asm
+++ b/libavcodec/x86/h264_idct.asm
@@ -32,14 +32,18 @@
 SECTION_RODATA
 
 ; FIXME this table is a duplicate from h264data.h, and will be removed once the tables from, h264 have been split
-scan8_mem: db 4+1*8, 5+1*8, 4+2*8, 5+2*8
-           db 6+1*8, 7+1*8, 6+2*8, 7+2*8
-           db 4+3*8, 5+3*8, 4+4*8, 5+4*8
-           db 6+3*8, 7+3*8, 6+4*8, 7+4*8
-           db 1+1*8, 2+1*8
-           db 1+2*8, 2+2*8
-           db 1+4*8, 2+4*8
-           db 1+5*8, 2+5*8
+scan8_mem: db  4+ 1*8, 5+ 1*8, 4+ 2*8, 5+ 2*8
+           db  6+ 1*8, 7+ 1*8, 6+ 2*8, 7+ 2*8
+           db  4+ 3*8, 5+ 3*8, 4+ 4*8, 5+ 4*8
+           db  6+ 3*8, 7+ 3*8, 6+ 4*8, 7+ 4*8
+           db  4+ 6*8, 5+ 6*8, 4+ 7*8, 5+ 7*8
+           db  6+ 6*8, 7+ 6*8, 6+ 7*8, 7+ 7*8
+           db  4+ 8*8, 5+ 8*8, 4+ 9*8, 5+ 9*8
+           db  6+ 8*8, 7+ 8*8, 6+ 9*8, 7+ 9*8
+           db  4+11*8, 5+11*8, 4+12*8, 5+12*8
+           db  6+11*8, 7+11*8, 6+12*8, 7+12*8
+           db  4+13*8, 5+13*8, 4+14*8, 5+14*8
+           db  6+13*8, 7+13*8, 6+14*8, 7+14*8
 %ifdef PIC
 %define scan8 r11
 %else
@@ -617,6 +621,8 @@ cglobal h264_idct_add8_8_mmx, 5, 7, 0
     mov         r10, r0
 %endif
     call         h264_idct_add8_mmx_plane
+    mov          r5, 32
+    add          r2, 384
 %ifdef ARCH_X86_64
     add         r10, gprsize
 %else
@@ -678,6 +684,8 @@ cglobal h264_idct_add8_8_mmx2, 5, 7, 0
     lea         r11, [scan8_mem]
 %endif
     call h264_idct_add8_mmx2_plane
+    mov          r5, 32
+    add          r2, 384
 %ifdef ARCH_X86_64
     add         r10, gprsize
 %else
@@ -810,12 +818,12 @@ cglobal h264_idct_add16intra_8_sse2, 5, 7, 8
     test        r0, r0
     jz .try%1dc
 %ifdef ARCH_X86_64
-    mov        r0d, dword [r1+%1*8+64]
+    mov        r0d, dword [r1+(%1&1)*8+64*(1+(%1>>1))]
     add         r0, [r10]
 %else
     mov         r0, r0m
     mov         r0, [r0]
-    add         r0, dword [r1+%1*8+64]
+    add         r0, dword [r1+(%1&1)*8+64*(1+(%1>>1))]
 %endif
     call        x264_add8x4_idct_sse2
     jmp .cycle%1end
@@ -824,16 +832,18 @@ cglobal h264_idct_add16intra_8_sse2, 5, 7, 8
     or         r0w, word [r2+32]
     jz .cycle%1end
 %ifdef ARCH_X86_64
-    mov        r0d, dword [r1+%1*8+64]
+    mov        r0d, dword [r1+(%1&1)*8+64*(1+(%1>>1))]
     add         r0, [r10]
 %else
     mov         r0, r0m
     mov         r0, [r0]
-    add         r0, dword [r1+%1*8+64]
+    add         r0, dword [r1+(%1&1)*8+64*(1+(%1>>1))]
 %endif
     call        h264_idct_dc_add8_mmx2
 .cycle%1end
-%if %1 < 3
+%if %1 == 1
+    add         r2, 384+64
+%elif %1 < 3
     add         r2, 64
 %endif
 %endmacro
@@ -845,15 +855,15 @@ cglobal h264_idct_add8_8_sse2, 5, 7, 8
 %ifdef ARCH_X86_64
     mov         r10, r0
 %endif
-    add8_sse2_cycle 0, 0x09
-    add8_sse2_cycle 1, 0x11
+    add8_sse2_cycle 0, 0x34
+    add8_sse2_cycle 1, 0x3c
 %ifdef ARCH_X86_64
     add         r10, gprsize
 %else
     add        r0mp, gprsize
 %endif
-    add8_sse2_cycle 2, 0x21
-    add8_sse2_cycle 3, 0x29
+    add8_sse2_cycle 2, 0x5c
+    add8_sse2_cycle 3, 0x64
     RET
 
 ;void ff_h264_luma_dc_dequant_idct_mmx(DCTELEM *output, DCTELEM *input, int qmul)
diff --git a/libavcodec/x86/h264_idct_10bit.asm b/libavcodec/x86/h264_idct_10bit.asm
index 3f7cf4cefc..54636a95d0 100644
--- a/libavcodec/x86/h264_idct_10bit.asm
+++ b/libavcodec/x86/h264_idct_10bit.asm
@@ -29,14 +29,18 @@ SECTION_RODATA
 
 pw_pixel_max: times 8 dw ((1 << 10)-1)
 pd_32:        times 4 dd 32
-scan8_mem: db 4+1*8, 5+1*8, 4+2*8, 5+2*8
-           db 6+1*8, 7+1*8, 6+2*8, 7+2*8
-           db 4+3*8, 5+3*8, 4+4*8, 5+4*8
-           db 6+3*8, 7+3*8, 6+4*8, 7+4*8
-           db 1+1*8, 2+1*8
-           db 1+2*8, 2+2*8
-           db 1+4*8, 2+4*8
-           db 1+5*8, 2+5*8
+scan8_mem: db  4+ 1*8, 5+ 1*8, 4+ 2*8, 5+ 2*8
+           db  6+ 1*8, 7+ 1*8, 6+ 2*8, 7+ 2*8
+           db  4+ 3*8, 5+ 3*8, 4+ 4*8, 5+ 4*8
+           db  6+ 3*8, 7+ 3*8, 6+ 4*8, 7+ 4*8
+           db  4+ 6*8, 5+ 6*8, 4+ 7*8, 5+ 7*8
+           db  6+ 6*8, 7+ 6*8, 6+ 7*8, 7+ 7*8
+           db  4+ 8*8, 5+ 8*8, 4+ 9*8, 5+ 9*8
+           db  6+ 8*8, 7+ 8*8, 6+ 9*8, 7+ 9*8
+           db  4+11*8, 5+11*8, 4+12*8, 5+12*8
+           db  6+11*8, 7+11*8, 6+12*8, 7+12*8
+           db  4+13*8, 5+13*8, 4+14*8, 5+14*8
+           db  6+13*8, 7+13*8, 6+14*8, 7+14*8
 
 %ifdef PIC
 %define scan8 r11
@@ -306,7 +310,7 @@ INIT_AVX
 IDCT_ADD16INTRA_10 avx
 %endif
 
-%assign last_block 24
+%assign last_block 36
 ;-----------------------------------------------------------------------------
 ; h264_idct_add8(pixel **dst, const int *block_offset, dctcoef *block, int stride, const uint8_t nnzc[6*8])
 ;-----------------------------------------------------------------------------
@@ -317,21 +321,22 @@ cglobal h264_idct_add8_10_%1,5,7
 %endif
     add      r2, 1024
     mov      r0, [r0]
-    ADD16_OP_INTRA %1, 16, 1+1*8
-    ADD16_OP_INTRA %1, 18, 1+2*8
+    ADD16_OP_INTRA %1, 16, 4+ 6*8
+    ADD16_OP_INTRA %1, 18, 4+ 7*8
+    add      r2, 1024-128*2
 %ifdef ARCH_X86_64
     mov      r0, [r10+gprsize]
 %else
     mov      r0, r0m
     mov      r0, [r0+gprsize]
 %endif
-    ADD16_OP_INTRA %1, 20, 1+4*8
-    ADD16_OP_INTRA %1, 22, 1+5*8
+    ADD16_OP_INTRA %1, 32, 4+11*8
+    ADD16_OP_INTRA %1, 34, 4+12*8
     REP_RET
     AC %1, 16
     AC %1, 18
-    AC %1, 20
-    AC %1, 22
+    AC %1, 32
+    AC %1, 34
 
 %endmacro ; IDCT_ADD8