updated ffmpeg (thanks to Aleksoid for finding the conflict in avcore\utils.c for Debug VS2010 builds)

git-svn-id: https://mpc-hc.svn.sourceforge.net/svnroot/mpc-hc/trunk@2180 10f7b99b-c216-0410-bff0-8a66a9350fd8

25 files changed, 1147 insertions, 1003 deletions

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/Makefile b/src/filters/transform/MPCVideoDec/ffmpeg/Makefile
index 1497e5ab7..f5b7c3b0b 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/Makefile
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/Makefile
@@ -1,4 +1,5 @@
 LAVC_DIR = libavcodec

+LAVCORE_DIR=libavcore

 LAVU_DIR = libavutil

 LSWS_DIR = libswscale

 PNG_DIR  = ../../../../thirdparty/libpng

@@ -17,6 +18,7 @@ OUT_DIRS = ../../../../../bin/obj/Release_x64/libavcodec_gcc/ \
     ../../../../../bin/obj/Release_x64/libavcodec_gcc/libavcodec \

     ../../../../../bin/obj/Release_x64/libavcodec_gcc/libavcodec/amr_float \

     ../../../../../bin/obj/Release_x64/libavcodec_gcc/libavcodec/x86 \

+    ../../../../../bin/obj/Release_x64/libavcodec_gcc/libavcore \

     ../../../../../bin/obj/Release_x64/libavcodec_gcc/libavutil \

     ../../../../../bin/obj/Release_x64/libavcodec_gcc/libswscale \

     $(SLIB_DIR)

@@ -25,12 +27,13 @@ OUT_DIRS = ../../../../../bin/obj/Release_Win32/libavcodec_gcc/ \
     ../../../../../bin/obj/Release_Win32/libavcodec_gcc/libavcodec \

     ../../../../../bin/obj/Release_Win32/libavcodec_gcc/libavcodec/amr_float \

     ../../../../../bin/obj/Release_Win32/libavcodec_gcc/libavcodec/x86 \

+    ../../../../../bin/obj/Release_Win32/libavcodec_gcc/libavcore \

     ../../../../../bin/obj/Release_Win32/libavcodec_gcc/libavutil \

     ../../../../../bin/obj/Release_Win32/libavcodec_gcc/libswscale \

     $(SLIB_DIR)

 endif

 

-CFLAGS+= -I. -I.. -I$(LAVC_DIR) -I$(LAVU_DIR) -I$(LSWS_DIR) -I$(ZLIB_DIR) -I$(PNG_DIR) \

+CFLAGS+= -I. -I.. -I$(LAVC_DIR) -I$(LAVCORE_DIR) -I$(LAVU_DIR) -I$(LSWS_DIR) -I$(ZLIB_DIR) -I$(PNG_DIR) \

          -DHAVE_AV_CONFIG_H -D_ISOC99_SOURCE -D_POSIX_C_SOURCE=200112 -std=gnu99

 

 SRCS_C=\

@@ -156,6 +159,9 @@ SRCS_C=\
     $(LAVC_DIR)/x86/vp6dsp_sse2.c \

     $(LAVC_DIR)/x86/vp8dsp-init.c \

 \

+    $(LAVCORE_DIR)/avcore_utils.c \

+    $(LAVCORE_DIR)/imgutils.c \

+\

     $(LAVU_DIR)/crc.c \

     $(LAVU_DIR)/intfloat_readwrite.c \

     $(LAVU_DIR)/inverse.c \

@@ -213,6 +219,7 @@ clean:
 	$(OUT_DIR)$(LAVC_DIR)/*.o $(OUT_DIR)$(LAVC_DIR)/*.d \

 	$(OUT_DIR)$(LAVC_DIR)/x86/*.o $(OUT_DIR)$(LAVC_DIR)/x86/*.d \

 	$(OUT_DIR)$(LAVC_DIR)/amr_float/*.o $(OUT_DIR)$(LAVC_DIR)/amr_float/*.d \

+	$(OUT_DIR)$(LAVCORE_DIR)/*.o $(OUT_DIR)$(LAVCORE_DIR)/*.d \

 	$(OUT_DIR)$(LAVU_DIR)/*.o $(OUT_DIR)$(LAVU_DIR)/*.d \

 	$(OUT_DIR)$(LSWS_DIR)/*.o $(OUT_DIR)$(LSWS_DIR)/*.d \

 	$(ZLIB_DIR)/*.o $(ZLIB_DIR)/*.d $(PNG_DIR)/*.o $(SLIB)

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/custom_code.txt b/src/filters/transform/MPCVideoDec/ffmpeg/custom_code.txt
index 3ad7a140e..74108e99b 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/custom_code.txt
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/custom_code.txt
@@ -13,5 +13,6 @@ The following files have MPC-specific custom code (compared to ffdshow):
 * libavcodec/mpc_helper.c

 * libavcodec/mpeg12.c

 * libavcodec/vp3.c

+* libavcore/avcore_utils.c (renamed from utils.c to avoid conflicts in MSVC2010)

 * libavutil/internal.h

 * libavutil/log.h

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcproj b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcproj
index 0d276b5ba..110c08cba 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcproj
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcproj
@@ -1591,6 +1591,26 @@
 				</FileConfiguration>

 			</File>

 		</Filter>

+		<Filter

+			Name="libavcore"

+			>

+			<File

+				RelativePath=".\libavcore\avcore.h"

+				>

+			</File>

+			<File

+				RelativePath=".\libavcore\avcore_utils.c"

+				>

+			</File>

+			<File

+				RelativePath=".\libavcore\imgutils.c"

+				>

+			</File>

+			<File

+				RelativePath=".\libavcore\imgutils.h"

+				>

+			</File>

+		</Filter>

 		<File

 			RelativePath=".\array_allocator.h"

 			>

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj
index fe19a7491..141c87d5c 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj
@@ -126,6 +126,8 @@
     <ClInclude Include="libavcodec\vp8dsp.h" />

     <ClInclude Include="libavcodec\wmv2.h" />

     <ClInclude Include="libavcodec\xiph.h" />

+    <ClInclude Include="libavcore\avcore.h" />

+    <ClInclude Include="libavcore\imgutils.h" />

     <ClInclude Include="libavutil\attributes.h" />

     <ClInclude Include="libavutil\avconfig.h" />

     <ClInclude Include="libavutil\avstring.h" />

@@ -276,6 +278,8 @@
     <ClCompile Include="libavcodec\wmv2.c" />

     <ClCompile Include="libavcodec\wmv2dec.c" />

     <ClCompile Include="libavcodec\xiph.c" />

+    <ClCompile Include="libavcore\imgutils.c" />

+    <ClCompile Include="libavcore\avcore_utils.c" />

     <ClCompile Include="libavutil\crc.c" />

     <ClCompile Include="libavutil\intfloat_readwrite.c">

       <ExcludedFromBuild Condition="'$(Configuration)|$(Platform)'=='Debug|Win32'">true</ExcludedFromBuild>

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj.filters b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj.filters
index 9cf688ad0..67bf63abb 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj.filters
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec.vcxproj.filters
@@ -13,6 +13,9 @@
     <Filter Include="libavcodec\amr_float">

       <UniqueIdentifier>{dcef6bb8-4262-415c-935a-c8dd0f056c4c}</UniqueIdentifier>

     </Filter>

+    <Filter Include="libavcore">

+      <UniqueIdentifier>{bb994511-43fc-42df-8ba2-b6186af844b2}</UniqueIdentifier>

+    </Filter>

   </ItemGroup>

   <ItemGroup>

     <ClInclude Include="libswscale\asmalign.h">

@@ -427,6 +430,12 @@
     <ClInclude Include="libavcodec\AVPaletteControl.h">

       <Filter>libavcodec</Filter>

     </ClInclude>

+    <ClInclude Include="libavcore\avcore.h">

+      <Filter>libavcore</Filter>

+    </ClInclude>

+    <ClInclude Include="libavcore\imgutils.h">

+      <Filter>libavcore</Filter>

+    </ClInclude>

   </ItemGroup>

   <ItemGroup>

     <ClCompile Include="libswscale\isP4HT.c">

@@ -813,5 +822,11 @@
     <ClCompile Include="libavcodec\avpacket.c">

       <Filter>libavcodec</Filter>

     </ClCompile>

+    <ClCompile Include="libavcore\imgutils.c">

+      <Filter>libavcore</Filter>

+    </ClCompile>

+    <ClCompile Include="libavcore\avcore_utils.c">

+      <Filter>libavcore</Filter>

+    </ClCompile>

   </ItemGroup>

 </Project>
\ No newline at end of file
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.c
index 3db5fb47d..293fa3c8a 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.c
@@ -35,9 +35,9 @@
 #include "internal.h"

 #include "imgconvert.h"

 #include "libavutil/pixdesc.h"

+#include "libavcore/imgutils.h"

 

 #if HAVE_MMX

-#include "x86/mmx.h"

 #include "x86/dsputil_mmx.h"

 #endif

 

@@ -748,144 +748,18 @@ int ff_set_systematic_pal(uint32_t pal[256], enum PixelFormat pix_fmt){
     return 0;

 }

 

+#if LIBAVCODEC_VERSION_MAJOR < 53

 int ff_fill_linesize(AVPicture *picture, enum PixelFormat pix_fmt, int width)

 {

-    int i;

-    const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[pix_fmt];

-    int max_plane_step     [4];

-    int max_plane_step_comp[4];

-

-    memset(picture->linesize, 0, sizeof(picture->linesize));

-

-    if (desc->flags & PIX_FMT_HWACCEL)

-        return -1;

-

-    if (desc->flags & PIX_FMT_BITSTREAM) {

-        picture->linesize[0] = (width * (desc->comp[0].step_minus1+1) + 7) >> 3;

-        return 0;

-    }

-

-    memset(max_plane_step     , 0, sizeof(max_plane_step     ));

-    memset(max_plane_step_comp, 0, sizeof(max_plane_step_comp));

-    for (i = 0; i < 4; i++) {

-        const AVComponentDescriptor *comp = &(desc->comp[i]);

-        if ((comp->step_minus1+1) > max_plane_step[comp->plane]) {

-            max_plane_step     [comp->plane] = comp->step_minus1+1;

-            max_plane_step_comp[comp->plane] = i;

-        }

-    }

-

-    for (i = 0; i < 4; i++) {

-        int s = (max_plane_step_comp[i] == 1 || max_plane_step_comp[i] == 2) ? desc->log2_chroma_w : 0;

-        picture->linesize[i] = max_plane_step[i] * (((width + (1 << s) - 1)) >> s);

-    }

-

-    return 0;

+    return av_fill_image_linesizes(picture->linesize, pix_fmt, width);

 }

 

 int ff_fill_pointer(AVPicture *picture, uint8_t *ptr, enum PixelFormat pix_fmt,

                     int height)

 {

-    int size, h2, size2;

-    const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[pix_fmt];

-

-    size = picture->linesize[0] * height;

-    switch(pix_fmt) {

-    case PIX_FMT_YUV420P:

-    case PIX_FMT_YUV422P:

-    case PIX_FMT_YUV444P:

-    case PIX_FMT_YUV410P:

-    case PIX_FMT_YUV411P:

-    case PIX_FMT_YUV440P:

-    case PIX_FMT_YUVJ420P:

-    case PIX_FMT_YUVJ422P:

-    case PIX_FMT_YUVJ444P:

-    case PIX_FMT_YUVJ440P:

-    case PIX_FMT_YUV420P16LE:

-    case PIX_FMT_YUV422P16LE:

-    case PIX_FMT_YUV444P16LE:

-    case PIX_FMT_YUV420P16BE:

-    case PIX_FMT_YUV422P16BE:

-    case PIX_FMT_YUV444P16BE:

-        h2 = (height + (1 << desc->log2_chroma_h) - 1) >> desc->log2_chroma_h;

-        size2 = picture->linesize[1] * h2;

-        picture->data[0] = ptr;

-        picture->data[1] = picture->data[0] + size;

-        picture->data[2] = picture->data[1] + size2;

-        picture->data[3] = NULL;

-        return size + 2 * size2;

-    case PIX_FMT_YUVA420P:

-        h2 = (height + (1 << desc->log2_chroma_h) - 1) >> desc->log2_chroma_h;

-        size2 = picture->linesize[1] * h2;

-        picture->data[0] = ptr;

-        picture->data[1] = picture->data[0] + size;

-        picture->data[2] = picture->data[1] + size2;

-        picture->data[3] = picture->data[1] + size2 + size2;

-        return 2 * size + 2 * size2;

-    case PIX_FMT_NV12:

-    case PIX_FMT_NV21:

-        h2 = (height + (1 << desc->log2_chroma_h) - 1) >> desc->log2_chroma_h;

-        size2 = picture->linesize[1] * h2;

-        picture->data[0] = ptr;

-        picture->data[1] = picture->data[0] + size;

-        picture->data[2] = NULL;

-        picture->data[3] = NULL;

-        return size + size2;

-    case PIX_FMT_RGB24:

-    case PIX_FMT_BGR24:

-    case PIX_FMT_ARGB:

-    case PIX_FMT_ABGR:

-    case PIX_FMT_RGBA:

-    case PIX_FMT_BGRA:

-    case PIX_FMT_RGB48BE:

-    case PIX_FMT_RGB48LE:

-    case PIX_FMT_GRAY16BE:

-    case PIX_FMT_GRAY16LE:

-    case PIX_FMT_BGR444BE:

-    case PIX_FMT_BGR444LE:

-    case PIX_FMT_BGR555BE:

-    case PIX_FMT_BGR555LE:

-    case PIX_FMT_BGR565BE:

-    case PIX_FMT_BGR565LE:

-    case PIX_FMT_RGB444BE:

-    case PIX_FMT_RGB444LE:

-    case PIX_FMT_RGB555BE:

-    case PIX_FMT_RGB555LE:

-    case PIX_FMT_RGB565BE:

-    case PIX_FMT_RGB565LE:

-    case PIX_FMT_YUYV422:

-    case PIX_FMT_UYVY422:

-    case PIX_FMT_UYYVYY411:

-    case PIX_FMT_RGB4:

-    case PIX_FMT_BGR4:

-    case PIX_FMT_MONOWHITE:

-    case PIX_FMT_MONOBLACK:

-    case PIX_FMT_Y400A:

-        picture->data[0] = ptr;

-        picture->data[1] = NULL;

-        picture->data[2] = NULL;

-        picture->data[3] = NULL;

-        return size;

-    case PIX_FMT_PAL8:

-    case PIX_FMT_RGB8:

-    case PIX_FMT_BGR8:

-    case PIX_FMT_RGB4_BYTE:

-    case PIX_FMT_BGR4_BYTE:

-    case PIX_FMT_GRAY8:

-        size2 = (size + 3) & ~3;

-        picture->data[0] = ptr;

-        picture->data[1] = ptr + size2; /* palette is stored here as 256 32 bit words */

-        picture->data[2] = NULL;

-        picture->data[3] = NULL;

-        return size2 + 256 * 4;

-    default:

-        picture->data[0] = NULL;

-        picture->data[1] = NULL;

-        picture->data[2] = NULL;

-        picture->data[3] = NULL;

-        return -1;

-    }

+    return av_fill_image_pointers(picture->data, pix_fmt, height, ptr, picture->linesize);

 }

+#endif

 

 void ff_img_copy_plane(uint8_t *dst, int dst_wrap,

                            const uint8_t *src, int src_wrap,

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.h
index 48e2f1271..f09fcbfc0 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.h
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/imgconvert.h
@@ -27,9 +27,13 @@
 #include <stdint.h>
 #include "avcodec.h"
 
+#if LIBAVCODEC_VERSION_MAJOR < 53
+attribute_deprecated
 int ff_fill_linesize(AVPicture *picture, enum PixelFormat pix_fmt, int width);
 
+attribute_deprecated
 int ff_fill_pointer(AVPicture *picture, uint8_t *ptr, enum PixelFormat pix_fmt, int height);
+#endif
 
 int ff_get_plane_bytewidth(enum PixelFormat pix_fmt, int width, int plane);
 
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/mpeg12.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/mpeg12.c
index 99ff3fa62..93268052a 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/mpeg12.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/mpeg12.c
@@ -1151,7 +1151,7 @@ typedef struct Mpeg1Context {
     MpegEncContext mpeg_enc_ctx;

     int mpeg_enc_ctx_allocated; /* true if decoding context allocated */

     int repeat_field; /* true if we must repeat the field */

-    AVPanScan pan_scan; /** some temporary storage for the panscan */

+    AVPanScan pan_scan;              /**< some temporary storage for the panscan */

     int slice_count;

     int swap_uv;//indicate VCR2

     int save_aspect_info;

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/utils.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/utils.c
index 1391625c0..78925d915 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/utils.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/utils.c
@@ -28,6 +28,7 @@
 #include "libavutil/avstring.h"

 #include "libavutil/crc.h"

 #include "libavutil/pixdesc.h"

+#include "libavcore/imgutils.h"

 #include "avcodec.h"

 #include "dsputil.h"

 #include "imgconvert.h"

@@ -281,7 +282,7 @@ int avcodec_default_get_buffer(AVCodecContext *s, AVFrame *pic){
         do {

             // NOTE: do not align linesizes individually, this breaks e.g. assumptions

             // that linesize[0] == 2*linesize[1] in the MPEG-encoder for 4:2:2

-            ff_fill_linesize(&picture, s->pix_fmt, w);

+            av_fill_image_linesizes(picture.linesize, s->pix_fmt, w);

             // increase alignment of w for next try (rhs gives the lowest bit set in w)

             w += w & ~(w-1);

 

@@ -291,7 +292,7 @@ int avcodec_default_get_buffer(AVCodecContext *s, AVFrame *pic){
             }

         } while (unaligned);

 

-        tmpsize = ff_fill_pointer(&picture, NULL, s->pix_fmt, h);

+        tmpsize = av_fill_image_pointers(picture.data, s->pix_fmt, h, NULL, picture.linesize);

         if (tmpsize < 0)

             return -1;

 

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp5.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp5.c
index 3ebff017a..bb9b62e07 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp5.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp5.c
@@ -42,7 +42,7 @@ static int vp5_parse_header(VP56Context *s, const uint8_t *buf, int buf_size,
     vp56_init_range_decoder(&s->c, buf, buf_size);
     s->framep[VP56_FRAME_CURRENT]->key_frame = !vp56_rac_get(c);
     vp56_rac_get(c);
-    vp56_init_dequant(s, vp56_rac_gets(c, 6));
+    ff_vp56_init_dequant(s, vp56_rac_gets(c, 6));
     if (s->framep[VP56_FRAME_CURRENT]->key_frame)
     {
         vp56_rac_gets(c, 8);
@@ -254,7 +254,7 @@ static av_cold int vp5_decode_init(AVCodecContext *avctx)
 {
     VP56Context *s = avctx->priv_data;
 
-    vp56_init(avctx, 1, 0);
+    ff_vp56_init(avctx, 1, 0);
     s->vp56_coord_div = vp5_coord_div;
     s->parse_vector_adjustment = vp5_parse_vector_adjustment;
     s->parse_coeff = vp5_parse_coeff;
@@ -273,8 +273,8 @@ AVCodec vp5_decoder = {
     sizeof(VP56Context),
     vp5_decode_init,
     NULL,
-    vp56_free,
-    vp56_decode_frame,
+    ff_vp56_free,
+    ff_vp56_decode_frame,
     /*.capabilities = */CODEC_CAP_DR1,
     /*.next = */NULL,
     /*.flush = */NULL,
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.c
index 188bfcfbf..3f5569eb8 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.c
@@ -28,7 +28,7 @@
 #include "vp56data.h"
 
 
-void vp56_init_dequant(VP56Context *s, int quantizer)
+void ff_vp56_init_dequant(VP56Context *s, int quantizer)
 {
     s->quantizer = quantizer;
     s->dequant_dc = vp56_dc_dequant[quantizer] << 2;
@@ -481,7 +481,7 @@ static int vp56_size_changed(AVCodecContext *avctx)
     return 0;
 }
 
-int vp56_decode_frame(AVCodecContext *avctx, void *data, int *data_size,
+int ff_vp56_decode_frame(AVCodecContext *avctx, void *data, int *data_size,
                       const uint8_t *buf, int buf_size)
 {
     VP56Context *s = avctx->priv_data;
@@ -638,7 +638,7 @@ int vp56_decode_frame(AVCodecContext *avctx, void *data, int *data_size,
     return buf_size;
 }
 
-av_cold void vp56_init(AVCodecContext *avctx, int flip, int has_alpha)
+av_cold void ff_vp56_init(AVCodecContext *avctx, int flip, int has_alpha)
 {
     VP56Context *s = avctx->priv_data;
     int i;
@@ -677,7 +677,7 @@ av_cold void vp56_init(AVCodecContext *avctx, int flip, int has_alpha)
     }
 }
 
-av_cold int vp56_free(AVCodecContext *avctx)
+av_cold int ff_vp56_free(AVCodecContext *avctx)
 {
     VP56Context *s = avctx->priv_data;
 
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.h
index a4ef49ede..69518fa73 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.h
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp56.h
@@ -56,7 +56,7 @@ typedef struct {
                  bits left) in order to eliminate a negate in cache refilling */
     const uint8_t *buffer;
     const uint8_t *end;
-    unsigned long code_word;
+    unsigned int code_word;
 } VP56RangeCoder;
 
 typedef struct {
@@ -170,10 +170,10 @@ struct vp56_context {
 };
 
 
-void vp56_init(AVCodecContext *avctx, int flip, int has_alpha);
-int vp56_free(AVCodecContext *avctx);
-void vp56_init_dequant(VP56Context *s, int quantizer);
-int vp56_decode_frame(AVCodecContext *avctx, void *data, int *data_size,
+void ff_vp56_init(AVCodecContext *avctx, int flip, int has_alpha);
+int ff_vp56_free(AVCodecContext *avctx);
+void ff_vp56_init_dequant(VP56Context *s, int quantizer);
+int ff_vp56_decode_frame(AVCodecContext *avctx, void *data, int *data_size,
                       const uint8_t *buf, int buf_size);
 
 
@@ -191,25 +191,12 @@ static inline void vp56_init_range_decoder(VP56RangeCoder *c,
     c->code_word = bytestream_get_be16(&c->buffer);
 }
 
-static inline int vp56_rac_get_prob(VP56RangeCoder *c, uint8_t prob)
+static av_always_inline unsigned int vp56_rac_renorm(VP56RangeCoder *c)
 {
-    /* Don't put c->high in a local variable; if we do that, gcc gets
-     * the stupids and turns the code below into a branch again. */
+    int shift = ff_h264_norm_shift[c->high] - 1;
     int bits = c->bits;
-    unsigned long code_word = c->code_word;
-    unsigned int low = 1 + (((c->high - 1) * prob) >> 8);
-    unsigned int low_shift = low << 8;
-    int bit = code_word >= low_shift;
-    int shift;
+    unsigned int code_word = c->code_word;
 
-    /* Incantation to convince GCC to turn these into conditional moves
-     * instead of branches -- faster, as this branch is basically
-     * unpredictable. */
-    c->high = bit ? c->high - low : low;
-    code_word = bit ? code_word - low_shift : code_word;
-
-    /* normalize */
-    shift = ff_h264_norm_shift[c->high] - 1;
     c->high   <<= shift;
     code_word <<= shift;
     bits       += shift;
@@ -218,29 +205,62 @@ static inline int vp56_rac_get_prob(VP56RangeCoder *c, uint8_t prob)
         bits -= 8;
     }
     c->bits = bits;
-    c->code_word = code_word;
+    return code_word;
+}
+
+#if ARCH_X86
+#include "x86/vp56_arith.h"
+#endif
+
+#ifndef vp56_rac_get_prob
+#define vp56_rac_get_prob vp56_rac_get_prob
+static inline int vp56_rac_get_prob(VP56RangeCoder *c, uint8_t prob)
+{
+    unsigned int code_word = vp56_rac_renorm(c);
+    unsigned int low = 1 + (((c->high - 1) * prob) >> 8);
+    unsigned int low_shift = low << 8;
+    int bit = code_word >= low_shift;
+
+    c->high = bit ? c->high - low : low;
+    c->code_word = bit ? code_word - low_shift : code_word;
+
     return bit;
 }
+#endif
+
+// branchy variant, to be used where there's a branch based on the bit decoded
+static av_always_inline int vp56_rac_get_prob_branchy(VP56RangeCoder *c, int prob)
+{
+    unsigned long code_word = vp56_rac_renorm(c);
+    unsigned low = 1 + (((c->high - 1) * prob) >> 8);
+    unsigned low_shift = low << 8;
+
+    if (code_word >= low_shift) {
+        c->high     -= low;
+        c->code_word = code_word - low_shift;
+        return 1;
+    }
+
+    c->high = low;
+    c->code_word = code_word;
+    return 0;
+}
 
 static inline int vp56_rac_get(VP56RangeCoder *c)
 {
+    unsigned int code_word = vp56_rac_renorm(c);
     /* equiprobable */
     int low = (c->high + 1) >> 1;
     unsigned int low_shift = low << 8;
-    int bit = c->code_word >= low_shift;
+    int bit = code_word >= low_shift;
     if (bit) {
-        c->high = (c->high - low) << 1;
-        c->code_word -= low_shift;
+        c->high   -= low;
+        code_word -= low_shift;
     } else {
-        c->high = low << 1;
+        c->high = low;
     }
 
-    /* normalize */
-    c->code_word <<= 1;
-    if (++c->bits == 0 && c->buffer < c->end) {
-        c->bits = -8;
-        c->code_word |= *c->buffer++;
-    }
+    c->code_word = code_word;
     return bit;
 }
 
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp6.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp6.c
index ce0ec642a..4c2aec67a 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp6.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp6.c
@@ -54,7 +54,7 @@ static int vp6_parse_header(VP56Context *s, const uint8_t *buf, int buf_size,
     int separated_coeff = buf[0] & 1;
 
     s->framep[VP56_FRAME_CURRENT]->key_frame = !(buf[0] & 0x80);
-    vp56_init_dequant(s, (buf[0] >> 1) & 0x3F);
+    ff_vp56_init_dequant(s, (buf[0] >> 1) & 0x3F);
 
     if (s->framep[VP56_FRAME_CURRENT]->key_frame) {
         sub_version = buf[1] >> 3;
@@ -576,8 +576,8 @@ static av_cold int vp6_decode_init(AVCodecContext *avctx)
 {
     VP56Context *s = avctx->priv_data;
 
-    vp56_init(avctx, avctx->codec->id == CODEC_ID_VP6,
-                     avctx->codec->id == CODEC_ID_VP6A);
+    ff_vp56_init(avctx, avctx->codec->id == CODEC_ID_VP6,
+                        avctx->codec->id == CODEC_ID_VP6A);
     s->vp56_coord_div = vp6_coord_div;
     s->parse_vector_adjustment = vp6_parse_vector_adjustment;
     s->filter = vp6_filter;
@@ -594,7 +594,7 @@ static av_cold int vp6_decode_free(AVCodecContext *avctx)
     VP56Context *s = avctx->priv_data;
     int pt, ct, cg;
 
-    vp56_free(avctx);
+    ff_vp56_free(avctx);
 
     for (pt=0; pt<2; pt++) {
         free_vlc(&s->dccv_vlc[pt]);
@@ -614,7 +614,7 @@ AVCodec vp6_decoder = {
     vp6_decode_init,
     NULL,
     vp6_decode_free,
-    vp56_decode_frame,
+    ff_vp56_decode_frame,
     /*.capabilities = */CODEC_CAP_DR1,
     /*.next = */NULL,
     /*.flush = */NULL,
@@ -632,7 +632,7 @@ AVCodec vp6f_decoder = {
     vp6_decode_init,
     NULL,
     vp6_decode_free,
-    vp56_decode_frame,
+    ff_vp56_decode_frame,
     /*.capabilities = */CODEC_CAP_DR1,
     /*.next = */NULL,
     /*.flush = */NULL,
@@ -650,7 +650,7 @@ AVCodec vp6a_decoder = {
     vp6_decode_init,
     NULL,
     vp6_decode_free,
-    vp56_decode_frame,
+    ff_vp56_decode_frame,
     /*.capabilities = */CODEC_CAP_DR1,
     /*.next = */NULL,
     /*.flush = */NULL,
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8.c
index 2dd086b70..6524a7141 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8.c
@@ -3,6 +3,7 @@
  *
  * Copyright (C) 2010 David Conrad
  * Copyright (C) 2010 Ronald S. Bultje
+ * Copyright (C) 2010 Jason Garrett-Glaser
  *
  * This file is part of FFmpeg.
  *
@@ -198,8 +199,6 @@ typedef struct {
     } prob[2];
 } VP8Context;
 
-#define RL24(p) (AV_RL16(p) + ((p)[2] << 16))
-
 static void vp8_decode_flush(AVCodecContext *avctx)
 {
     VP8Context *s = avctx->priv_data;
@@ -211,6 +210,7 @@ static void vp8_decode_flush(AVCodecContext *avctx)
     memset(s->framep, 0, sizeof(s->framep));
 
     av_freep(&s->macroblocks_base);
+    av_freep(&s->filter_strength);
     av_freep(&s->intra4x4_pred_mode_base);
     av_freep(&s->top_nnz);
     av_freep(&s->edge_emu_buffer);
@@ -307,7 +307,7 @@ static int setup_partitions(VP8Context *s, const uint8_t *buf, int buf_size)
         return -1;
 
     for (i = 0; i < s->num_coeff_partitions-1; i++) {
-        int size = RL24(sizes + 3*i);
+        int size = AV_RL24(sizes + 3*i);
         if (buf_size - size < 0)
             return -1;
 
@@ -402,7 +402,7 @@ static int decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_size)
     s->keyframe  = !(buf[0] & 1);
     s->profile   =  (buf[0]>>1) & 7;
     s->invisible = !(buf[0] & 0x10);
-    header_size  = RL24(buf) >> 5;
+    header_size  = AV_RL24(buf) >> 5;
     buf      += 3;
     buf_size -= 3;
 
@@ -420,8 +420,8 @@ static int decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_size)
     }
 
     if (s->keyframe) {
-        if (RL24(buf) != 0x2a019d) {
-            av_log(s->avctx, AV_LOG_ERROR, "Invalid start code 0x%x\n", RL24(buf));
+        if (AV_RL24(buf) != 0x2a019d) {
+            av_log(s->avctx, AV_LOG_ERROR, "Invalid start code 0x%x\n", AV_RL24(buf));
             return AVERROR_INVALIDDATA;
         }
         width  = AV_RL16(buf+3) & 0x3fff;
@@ -495,7 +495,7 @@ static int decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_size)
         for (j = 0; j < 8; j++)
             for (k = 0; k < 3; k++)
                 for (l = 0; l < NUM_DCT_TOKENS-1; l++)
-                    if (vp56_rac_get_prob(c, vp8_token_update_probs[i][j][k][l]))
+                    if (vp56_rac_get_prob_branchy(c, vp8_token_update_probs[i][j][k][l]))
                         s->prob->token[i][j][k][l] = vp8_rac_get_uint(c, 8);
 
     if ((s->mbskip_enabled = vp8_rac_get(c)))
@@ -516,15 +516,15 @@ static int decode_frame_header(VP8Context *s, const uint8_t *buf, int buf_size)
         // 17.2 MV probability update
         for (i = 0; i < 2; i++)
             for (j = 0; j < 19; j++)
-                if (vp56_rac_get_prob(c, vp8_mv_update_prob[i][j]))
+                if (vp56_rac_get_prob_branchy(c, vp8_mv_update_prob[i][j]))
                     s->prob->mvc[i][j] = vp8_rac_get_nn(c);
     }
 
     return 0;
 }
 
-static inline void clamp_mv(VP8Context *s, VP56mv *dst, const VP56mv *src,
-                            int mb_x, int mb_y)
+static av_always_inline
+void clamp_mv(VP8Context *s, VP56mv *dst, const VP56mv *src, int mb_x, int mb_y)
 {
 #define MARGIN (16 << 2)
     dst->x = av_clip(src->x, -((mb_x << 6) + MARGIN),
@@ -533,8 +533,9 @@ static inline void clamp_mv(VP8Context *s, VP56mv *dst, const VP56mv *src,
                      ((s->mb_height - 1 - mb_y) << 6) + MARGIN);
 }
 
-static void find_near_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
-                          VP56mv near[2], VP56mv *best, uint8_t cnt[4])
+static av_always_inline
+void find_near_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
+                   VP56mv near[2], VP56mv *best, uint8_t cnt[4])
 {
     VP8Macroblock *mb_edge[3] = { mb + 2 /* top */,
                                   mb - 1 /* left */,
@@ -589,7 +590,7 @@ static void find_near_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
     if (cnt[CNT_NEAREST] >= cnt[CNT_ZERO])
         best_idx = CNT_NEAREST;
 
-    clamp_mv(s, best, &near_mv[best_idx], mb_x, mb_y);
+    mb->mv  = near_mv[best_idx];
     near[0] = near_mv[CNT_NEAREST];
     near[1] = near_mv[CNT_NEAR];
 }
@@ -599,9 +600,9 @@ static void find_near_mvs(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
  */
 static int read_mv_component(VP56RangeCoder *c, const uint8_t *p)
 {
-    int x = 0;
+    int bit, x = 0;
 
-    if (vp56_rac_get_prob(c, p[0])) {
+    if (vp56_rac_get_prob_branchy(c, p[0])) {
         int i;
 
         for (i = 0; i < 3; i++)
@@ -610,13 +611,23 @@ static int read_mv_component(VP56RangeCoder *c, const uint8_t *p)
             x += vp56_rac_get_prob(c, p[9 + i]) << i;
         if (!(x & 0xFFF0) || vp56_rac_get_prob(c, p[12]))
             x += 8;
-    } else
-        x = vp8_rac_get_tree(c, vp8_small_mvtree, &p[2]);
+    } else {
+        // small_mvtree
+        const uint8_t *ps = p+2;
+        bit = vp56_rac_get_prob(c, *ps);
+        ps += 1 + 3*bit;
+        x  += 4*bit;
+        bit = vp56_rac_get_prob(c, *ps);
+        ps += 1 + bit;
+        x  += 2*bit;
+        x  += vp56_rac_get_prob(c, *ps);
+    }
 
     return (x && vp56_rac_get_prob(c, p[1])) ? -x : x;
 }
 
-static const uint8_t *get_submv_prob(uint32_t left, uint32_t top)
+static av_always_inline
+const uint8_t *get_submv_prob(uint32_t left, uint32_t top)
 {
     if (left == top)
         return vp8_submv_prob[4-!!left];
@@ -629,8 +640,8 @@ static const uint8_t *get_submv_prob(uint32_t left, uint32_t top)
  * Split motion vector prediction, 16.4.
  * @returns the number of motion vectors parsed (2, 4 or 16)
  */
-static int decode_splitmvs(VP8Context    *s,  VP56RangeCoder *c,
-                            VP8Macroblock *mb, VP56mv         *base_mv)
+static av_always_inline
+int decode_splitmvs(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb)
 {
     int part_idx = mb->partitioning =
         vp8_rac_get_tree(c, vp8_mbsplit_tree, vp8_mbsplit_prob);
@@ -663,11 +674,11 @@ static int decode_splitmvs(VP8Context    *s,  VP56RangeCoder *c,
 
         switch (vp8_rac_get_tree(c, vp8_submv_ref_tree, submv_prob)) {
         case VP8_SUBMVMODE_NEW4X4:
-            mb->bmv[n].y = base_mv->y + read_mv_component(c, s->prob->mvc[0]);
-            mb->bmv[n].x = base_mv->x + read_mv_component(c, s->prob->mvc[1]);
+            mb->bmv[n].y = mb->mv.y + read_mv_component(c, s->prob->mvc[0]);
+            mb->bmv[n].x = mb->mv.x + read_mv_component(c, s->prob->mvc[1]);
             break;
         case VP8_SUBMVMODE_ZERO4X4:
-            AV_WN32A(&mb->bmv[n], 0);
+            AV_ZERO32(&mb->bmv[n]);
             break;
         case VP8_SUBMVMODE_LEFT4X4:
             AV_WN32A(&mb->bmv[n], left);
@@ -681,8 +692,9 @@ static int decode_splitmvs(VP8Context    *s,  VP56RangeCoder *c,
     return num;
 }
 
-static inline void decode_intra4x4_modes(VP56RangeCoder *c, uint8_t *intra4x4,
-                                         int stride, int keyframe)
+static av_always_inline
+void decode_intra4x4_modes(VP56RangeCoder *c, uint8_t *intra4x4,
+                           int stride, int keyframe)
 {
     int x, y, t, l, i;
 
@@ -703,8 +715,9 @@ static inline void decode_intra4x4_modes(VP56RangeCoder *c, uint8_t *intra4x4,
     }
 }
 
-static void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
-                           uint8_t *intra4x4, uint8_t *segment)
+static av_always_inline
+void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
+                    uint8_t *intra4x4, uint8_t *segment)
 {
     VP56RangeCoder *c = &s->c;
 
@@ -724,13 +737,13 @@ static void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
 
         s->chroma_pred_mode = vp8_rac_get_tree(c, vp8_pred8x8c_tree, vp8_pred8x8c_prob_intra);
         mb->ref_frame = VP56_FRAME_CURRENT;
-    } else if (vp56_rac_get_prob(c, s->prob->intra)) {
+    } else if (vp56_rac_get_prob_branchy(c, s->prob->intra)) {
         VP56mv near[2], best;
         uint8_t cnt[4] = { 0 };
         uint8_t p[4];
 
         // inter MB, 16.2
-        if (vp56_rac_get_prob(c, s->prob->last))
+        if (vp56_rac_get_prob_branchy(c, s->prob->last))
             mb->ref_frame = vp56_rac_get_prob(c, s->prob->golden) ?
                 VP56_FRAME_GOLDEN2 /* altref */ : VP56_FRAME_GOLDEN;
         else
@@ -746,10 +759,11 @@ static void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
         mb->mode = vp8_rac_get_tree(c, vp8_pred16x16_tree_mvinter, p);
         switch (mb->mode) {
         case VP8_MVMODE_SPLIT:
-            mb->mv = mb->bmv[decode_splitmvs(s, c, mb, &best) - 1];
+            clamp_mv(s, &mb->mv, &mb->mv, mb_x, mb_y);
+            mb->mv = mb->bmv[decode_splitmvs(s, c, mb) - 1];
             break;
         case VP8_MVMODE_ZERO:
-            AV_WN32A(&mb->mv, 0);
+            AV_ZERO32(&mb->mv);
             break;
         case VP8_MVMODE_NEAREST:
             clamp_mv(s, &mb->mv, &near[0], mb_x, mb_y);
@@ -758,8 +772,9 @@ static void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
             clamp_mv(s, &mb->mv, &near[1], mb_x, mb_y);
             break;
         case VP8_MVMODE_NEW:
-            mb->mv.y = best.y + read_mv_component(c, s->prob->mvc[0]);
-            mb->mv.x = best.x + read_mv_component(c, s->prob->mvc[1]);
+            clamp_mv(s, &mb->mv, &mb->mv, mb_x, mb_y);
+            mb->mv.y += + read_mv_component(c, s->prob->mvc[0]);
+            mb->mv.x += + read_mv_component(c, s->prob->mvc[1]);
             break;
         }
         if (mb->mode != VP8_MVMODE_SPLIT) {
@@ -776,7 +791,7 @@ static void decode_mb_mode(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y,
         s->chroma_pred_mode = vp8_rac_get_tree(c, vp8_pred8x8c_tree, s->prob->pred8x8c);
         mb->ref_frame = VP56_FRAME_CURRENT;
         mb->partitioning = VP8_SPLITMVMODE_NONE;
-        AV_WN32A(&mb->bmv[0], 0);
+        AV_ZERO32(&mb->bmv[0]);
     }
 }
 
@@ -795,41 +810,67 @@ static int decode_block_coeffs(VP56RangeCoder *c, DCTELEM block[16],
                                uint8_t probs[8][3][NUM_DCT_TOKENS-1],
                                int i, int zero_nhood, int16_t qmul[2])
 {
-    int token, nonzero = 0;
-    int offset = 0;
+    uint8_t *token_prob;
+    int nonzero = 0;
+    int coeff;
 
-    for (; i < 16; i++) {
-        token = vp8_rac_get_tree_with_offset(c, vp8_coeff_tree, probs[vp8_coeff_band[i]][zero_nhood], offset);
+    do {
+        token_prob = probs[vp8_coeff_band[i]][zero_nhood];
 
-        if (token == DCT_EOB)
-            break;
-        else if (token >= DCT_CAT1) {
-            int cat = token-DCT_CAT1;
-            token = vp8_rac_get_coeff(c, vp8_dct_cat_prob[cat]);
-            token += 3 + (2<<cat);
-        }
+        if (!vp56_rac_get_prob_branchy(c, token_prob[0]))   // DCT_EOB
+            return nonzero;
 
-        // after the first token, the non-zero prediction context becomes
-        // based on the last decoded coeff
-        if (!token) {
+skip_eob:
+        if (!vp56_rac_get_prob_branchy(c, token_prob[1])) { // DCT_0
             zero_nhood = 0;
-            offset = 1;
-            continue;
-        } else if (token == 1)
+            token_prob = probs[vp8_coeff_band[++i]][0];
+            if (i < 16)
+                goto skip_eob;
+            return nonzero; // invalid input; blocks should end with EOB
+        }
+
+        if (!vp56_rac_get_prob_branchy(c, token_prob[2])) { // DCT_1
+            coeff = 1;
             zero_nhood = 1;
-        else
+        } else {
             zero_nhood = 2;
 
+            if (!vp56_rac_get_prob_branchy(c, token_prob[3])) { // DCT 2,3,4
+                coeff = vp56_rac_get_prob(c, token_prob[4]);
+                if (coeff)
+                    coeff += vp56_rac_get_prob(c, token_prob[5]);
+                coeff += 2;
+            } else {
+                // DCT_CAT*
+                if (!vp56_rac_get_prob_branchy(c, token_prob[6])) {
+                    if (!vp56_rac_get_prob_branchy(c, token_prob[7])) { // DCT_CAT1
+                        coeff  = 5 + vp56_rac_get_prob(c, vp8_dct_cat1_prob[0]);
+                    } else {                                    // DCT_CAT2
+                        coeff  = 7;
+                        coeff += vp56_rac_get_prob(c, vp8_dct_cat2_prob[0]) << 1;
+                        coeff += vp56_rac_get_prob(c, vp8_dct_cat2_prob[1]);
+                    }
+                } else {    // DCT_CAT3 and up
+                    int a = vp56_rac_get_prob(c, token_prob[8]);
+                    int b = vp56_rac_get_prob(c, token_prob[9+a]);
+                    int cat = (a<<1) + b;
+                    coeff  = 3 + (8<<cat);
+                    coeff += vp8_rac_get_coeff(c, vp8_dct_cat_prob[cat]);
+                }
+            }
+        }
+
         // todo: full [16] qmat? load into register?
-        block[zigzag_scan[i]] = (vp8_rac_get(c) ? -token : token) * qmul[!!i];
-        nonzero = i+1;
-        offset = 0;
-    }
+        block[zigzag_scan[i]] = (vp8_rac_get(c) ? -coeff : coeff) * qmul[!!i];
+        nonzero = ++i;
+    } while (i < 16);
+
     return nonzero;
 }
 
-static void decode_mb_coeffs(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb,
-                             uint8_t t_nnz[9], uint8_t l_nnz[9])
+static av_always_inline
+void decode_mb_coeffs(VP8Context *s, VP56RangeCoder *c, VP8Macroblock *mb,
+                      uint8_t t_nnz[9], uint8_t l_nnz[9])
 {
     LOCAL_ALIGNED_16(DCTELEM, dc,[16]);
     int i, x, y, luma_start = 0, luma_ctx = 3;
@@ -926,21 +967,22 @@ void xchg_mb_border(uint8_t *top_border, uint8_t *src_y, uint8_t *src_cb, uint8_
     }
 }
 
-static int check_intra_pred_mode(int mode, int mb_x, int mb_y)
+static av_always_inline
+int check_intra_pred_mode(int mode, int mb_x, int mb_y)
 {
     if (mode == DC_PRED8x8) {
-        if (!(mb_x|mb_y))
-            mode = DC_128_PRED8x8;
-        else if (!mb_y)
+        if (!mb_x) {
+            mode = mb_y ? TOP_DC_PRED8x8 : DC_128_PRED8x8;
+        } else if (!mb_y) {
             mode = LEFT_DC_PRED8x8;
-        else if (!mb_x)
-            mode = TOP_DC_PRED8x8;
+        }
     }
     return mode;
 }
 
-static void intra_predict(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb,
-                          uint8_t *intra4x4, int mb_x, int mb_y)
+static av_always_inline
+void intra_predict(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb,
+                   uint8_t *intra4x4, int mb_x, int mb_y)
 {
     int x, y, mode, nnz, tr;
 
@@ -1022,11 +1064,12 @@ static void intra_predict(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb,
  * @param linesize size of a single line of plane data, including padding
  * @param mc_func motion compensation function pointers (bilinear or sixtap MC)
  */
-static inline void vp8_mc(VP8Context *s, int luma,
-                          uint8_t *dst, uint8_t *src, const VP56mv *mv,
-                          int x_off, int y_off, int block_w, int block_h,
-                          int width, int height, int linesize,
-                          vp8_mc_func mc_func[3][3])
+static av_always_inline
+void vp8_mc(VP8Context *s, int luma,
+            uint8_t *dst, uint8_t *src, const VP56mv *mv,
+            int x_off, int y_off, int block_w, int block_h,
+            int width, int height, int linesize,
+            vp8_mc_func mc_func[3][3])
 {
     if (AV_RN32A(mv)) {
         static const uint8_t idx[8] = { 0, 1, 2, 1, 2, 1, 2, 1 };
@@ -1050,11 +1093,12 @@ static inline void vp8_mc(VP8Context *s, int luma,
         mc_func[0][0](dst, linesize, src + y_off * linesize + x_off, linesize, block_h, 0, 0);
 }
 
-static inline void vp8_mc_part(VP8Context *s, uint8_t *dst[3],
-                               AVFrame *ref_frame, int x_off, int y_off,
-                               int bx_off, int by_off,
-                               int block_w, int block_h,
-                               int width, int height, VP56mv *mv)
+static av_always_inline
+void vp8_mc_part(VP8Context *s, uint8_t *dst[3],
+                 AVFrame *ref_frame, int x_off, int y_off,
+                 int bx_off, int by_off,
+                 int block_w, int block_h,
+                 int width, int height, VP56mv *mv)
 {
     VP56mv uvmv = *mv;
 
@@ -1085,7 +1129,7 @@ static inline void vp8_mc_part(VP8Context *s, uint8_t *dst[3],
 
 /* Fetch pixels for estimated mv 4 macroblocks ahead.
  * Optimized for 64-byte cache lines.  Inspired by ffh264 prefetch_motion. */
-static inline void prefetch_motion(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y, int mb_xy, int ref)
+static av_always_inline void prefetch_motion(VP8Context *s, VP8Macroblock *mb, int mb_x, int mb_y, int mb_xy, int ref)
 {
     /* Don't prefetch refs that haven't been used very often this frame. */
     if (s->ref_count[ref-1] > (mb_xy >> 5)) {
@@ -1103,8 +1147,9 @@ static inline void prefetch_motion(VP8Context *s, VP8Macroblock *mb, int mb_x, i
 /**
  * Apply motion vectors to prediction buffer, chapter 18.
  */
-static void inter_predict(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb,
-                          int mb_x, int mb_y)
+static av_always_inline
+void inter_predict(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb,
+                   int mb_x, int mb_y)
 {
     int x_off = mb_x << 4, y_off = mb_y << 4;
     int width = 16*s->mb_width, height = 16*s->mb_height;
@@ -1187,7 +1232,7 @@ static void inter_predict(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb,
     }
 }
 
-static void idct_mb(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb)
+static av_always_inline void idct_mb(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb)
 {
     int x, y, ch;
 
@@ -1238,7 +1283,7 @@ static void idct_mb(VP8Context *s, uint8_t *dst[3], VP8Macroblock *mb)
     }
 }
 
-static void filter_level_for_mb(VP8Context *s, VP8Macroblock *mb, VP8FilterStrength *f )
+static av_always_inline void filter_level_for_mb(VP8Context *s, VP8Macroblock *mb, VP8FilterStrength *f )
 {
     int interior_limit, filter_level;
 
@@ -1278,7 +1323,7 @@ static void filter_level_for_mb(VP8Context *s, VP8Macroblock *mb, VP8FilterStren
     f->inner_filter = !mb->skip || mb->mode == MODE_I4x4 || mb->mode == VP8_MVMODE_SPLIT;
 }
 
-static void filter_mb(VP8Context *s, uint8_t *dst[3], VP8FilterStrength *f, int mb_x, int mb_y)
+static av_always_inline void filter_mb(VP8Context *s, uint8_t *dst[3], VP8FilterStrength *f, int mb_x, int mb_y)
 {
     int mbedge_lim, bedge_lim, hev_thresh;
     int filter_level = f->filter_level;
@@ -1347,7 +1392,7 @@ static void filter_mb(VP8Context *s, uint8_t *dst[3], VP8FilterStrength *f, int
     }
 }
 
-static void filter_mb_simple(VP8Context *s, uint8_t *dst, VP8FilterStrength *f, int mb_x, int mb_y)
+static av_always_inline void filter_mb_simple(VP8Context *s, uint8_t *dst, VP8FilterStrength *f, int mb_x, int mb_y)
 {
     int mbedge_lim, bedge_lim;
     int filter_level = f->filter_level;
@@ -1416,7 +1461,7 @@ static int vp8_decode_frame(AVCodecContext *avctx, void *data, int *data_size,
     VP8Context *s = avctx->priv_data;
     int ret, mb_x, mb_y, i, y, referenced;
     enum AVDiscard skip_thresh;
-    AVFrame *curframe = NULL;
+    AVFrame *av_uninit(curframe);
 
     if ((ret = decode_frame_header(s, buf, buf_size)) < 0)
         return ret;
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8data.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8data.h
index 9f56ab63b..1bdac16be 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8data.h
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/vp8data.h
@@ -118,16 +118,6 @@ static const int8_t vp8_pred16x16_tree_mvinter[4][2] = {
        { -VP8_MVMODE_NEW, -VP8_MVMODE_SPLIT } // '1110', '1111'
 };
 
-static const int8_t vp8_small_mvtree[7][2] = {
-    {  1, 4 },
-     {  2, 3 },
-      { -0, -1 },                           // '000', '001'
-      { -2, -3 },                           // '010', '011'
-     {  5,  6 },
-      { -4, -5 },                           // '100', '101'
-      { -6, -7 }                            // '110', '111'
-};
-
 static const uint8_t vp8_mbsplits[5][16] = {
     {  0,  0,  0,  0,  0,  0,  0,  0,
        1,  1,  1,  1,  1,  1,  1,  1  },
@@ -337,21 +327,6 @@ static const uint8_t vp8_coeff_band[16] =
     0, 1, 2, 3, 6, 4, 5, 6, 6, 6, 6, 6, 6, 6, 6, 7
 };
 
-static const int8_t vp8_coeff_tree[NUM_DCT_TOKENS-1][2] =
-{
-    { -DCT_EOB, 1 },                // '0'
-     { -DCT_0, 2 },                 // '10'
-      { -DCT_1, 3 },                // '110'
-       { 4, 6 },
-        { -DCT_2, 5 },              // '11100'
-         { -DCT_3, -DCT_4 },        // '111010', '111011'
-        { 7, 8 },
-         { -DCT_CAT1, -DCT_CAT2 },  // '111100', '111101'
-         { 9, 10 },
-          { -DCT_CAT3, -DCT_CAT4 }, // '1111100', '1111101'
-          { -DCT_CAT5, -DCT_CAT6 }, // '1111110', '1111111'
-};
-
 static const uint8_t vp8_dct_cat1_prob[] = { 159, 0 };
 static const uint8_t vp8_dct_cat2_prob[] = { 165, 145, 0 };
 static const uint8_t vp8_dct_cat3_prob[] = { 173, 148, 140, 0 };
@@ -359,10 +334,9 @@ static const uint8_t vp8_dct_cat4_prob[] = { 176, 155, 140, 135, 0 };
 static const uint8_t vp8_dct_cat5_prob[] = { 180, 157, 141, 134, 130, 0 };
 static const uint8_t vp8_dct_cat6_prob[] = { 254, 254, 243, 230, 196, 177, 153, 140, 133, 130, 129, 0 };
 
-static const uint8_t * const vp8_dct_cat_prob[6] =
+// only used for cat3 and above; cat 1 and 2 are referenced directly
+static const uint8_t * const vp8_dct_cat_prob[] =
 {
-    vp8_dct_cat1_prob,
-    vp8_dct_cat2_prob,
     vp8_dct_cat3_prob,
     vp8_dct_cat4_prob,
     vp8_dct_cat5_prob,
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/idct_mmx.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/idct_mmx.c
index d1859080c..b158c71d2 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/idct_mmx.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/idct_mmx.c
@@ -23,8 +23,8 @@
 #include "libavutil/common.h"

 #include "libavcodec/dsputil.h"

 

+#include "libavutil/x86_cpu.h"

 #include "dsputil_mmx.h"

-#include "mmx.h"

 

 #define ROW_SHIFT 11

 #define COL_SHIFT 6

@@ -87,104 +87,115 @@ static inline void idct_row (int16_t * row, int offset,
 static inline void mmxext_row_head (int16_t * const row, const int offset,

                                     const int16_t * const table)

 {

-    movq_m2r (*(row+offset), mm2);      /* mm2 = x6 x4 x2 x0 */

+    __asm__ volatile(

+        "movq     (%0), %%mm2        \n\t"  /* mm2 = x6 x4 x2 x0 */

 

-    movq_m2r (*(row+offset+4), mm5);    /* mm5 = x7 x5 x3 x1 */

-    movq_r2r (mm2, mm0);                /* mm0 = x6 x4 x2 x0 */

+        "movq    8(%0), %%mm5        \n\t"  /* mm5 = x7 x5 x3 x1 */

+        "movq    %%mm2, %%mm0        \n\t"  /* mm0 = x6 x4 x2 x0 */

 

-    movq_m2r (*table, mm3);             /* mm3 = -C2 -C4 C2 C4 */

-    movq_r2r (mm5, mm6);                /* mm6 = x7 x5 x3 x1 */

+        "movq     (%1), %%mm3        \n\t"  /* mm3 = -C2 -C4 C2 C4 */

+        "movq    %%mm5, %%mm6        \n\t"  /* mm6 = x7 x5 x3 x1 */

 

-    movq_m2r (*(table+4), mm4);         /* mm4 = C6 C4 C6 C4 */

-    pmaddwd_r2r (mm0, mm3);             /* mm3 = -C4*x4-C2*x6 C4*x0+C2*x2 */

+        "movq    8(%1), %%mm4        \n\t"  /* mm4 = C6 C4 C6 C4 */

+        "pmaddwd %%mm0, %%mm3        \n\t"  /* mm3 = -C4*x4-C2*x6 C4*x0+C2*x2 */

 

-    pshufw_r2r (mm2, mm2, 0x4e);        /* mm2 = x2 x0 x6 x4 */

+        "pshufw  $0x4e, %%mm2, %%mm2 \n\t"  /* mm2 = x2 x0 x6 x4 */

+        :: "r" ((row+offset)), "r" (table)

+    );

 }

 

 static inline void mmxext_row (const int16_t * const table,

                                const int32_t * const rounder)

 {

-    movq_m2r (*(table+8), mm1);         /* mm1 = -C5 -C1 C3 C1 */

-    pmaddwd_r2r (mm2, mm4);             /* mm4 = C4*x0+C6*x2 C4*x4+C6*x6 */

+    __asm__ volatile (

+        "movq    16(%0), %%mm1         \n\t" /* mm1 = -C5 -C1 C3 C1 */

+        "pmaddwd  %%mm2, %%mm4         \n\t" /* mm4 = C4*x0+C6*x2 C4*x4+C6*x6 */

 

-    pmaddwd_m2r (*(table+16), mm0);     /* mm0 = C4*x4-C6*x6 C4*x0-C6*x2 */

-    pshufw_r2r (mm6, mm6, 0x4e);        /* mm6 = x3 x1 x7 x5 */

+        "pmaddwd 32(%0), %%mm0         \n\t" /* mm0 = C4*x4-C6*x6 C4*x0-C6*x2 */

+        "pshufw   $0x4e, %%mm6, %%mm6  \n\t" /* mm6 = x3 x1 x7 x5 */

 

-    movq_m2r (*(table+12), mm7);        /* mm7 = -C7 C3 C7 C5 */

-    pmaddwd_r2r (mm5, mm1);             /* mm1 = -C1*x5-C5*x7 C1*x1+C3*x3 */

+        "movq    24(%0), %%mm7         \n\t" /* mm7 = -C7 C3 C7 C5 */

+        "pmaddwd  %%mm5, %%mm1         \n\t" /* mm1= -C1*x5-C5*x7 C1*x1+C3*x3 */

 

-    paddd_m2r (*rounder, mm3);          /* mm3 += rounder */

-    pmaddwd_r2r (mm6, mm7);             /* mm7 = C3*x1-C7*x3 C5*x5+C7*x7 */

+        "paddd     (%1), %%mm3         \n\t" /* mm3 += rounder */

+        "pmaddwd  %%mm6, %%mm7         \n\t" /* mm7 = C3*x1-C7*x3 C5*x5+C7*x7 */

 

-    pmaddwd_m2r (*(table+20), mm2);     /* mm2 = C4*x0-C2*x2 -C4*x4+C2*x6 */

-    paddd_r2r (mm4, mm3);               /* mm3 = a1 a0 + rounder */

+        "pmaddwd 40(%0), %%mm2         \n\t" /* mm2= C4*x0-C2*x2 -C4*x4+C2*x6 */

+        "paddd    %%mm4, %%mm3         \n\t" /* mm3 = a1 a0 + rounder */

 

-    pmaddwd_m2r (*(table+24), mm5);     /* mm5 = C3*x5-C1*x7 C5*x1-C1*x3 */

-    movq_r2r (mm3, mm4);                /* mm4 = a1 a0 + rounder */

+        "pmaddwd 48(%0), %%mm5         \n\t" /* mm5 = C3*x5-C1*x7 C5*x1-C1*x3 */

+        "movq     %%mm3, %%mm4         \n\t" /* mm4 = a1 a0 + rounder */

 

-    pmaddwd_m2r (*(table+28), mm6);     /* mm6 = C7*x1-C5*x3 C7*x5+C3*x7 */

-    paddd_r2r (mm7, mm1);               /* mm1 = b1 b0 */

+        "pmaddwd 56(%0), %%mm6         \n\t" /* mm6 = C7*x1-C5*x3 C7*x5+C3*x7 */

+        "paddd    %%mm7, %%mm1         \n\t" /* mm1 = b1 b0 */

 

-    paddd_m2r (*rounder, mm0);          /* mm0 += rounder */

-    psubd_r2r (mm1, mm3);               /* mm3 = a1-b1 a0-b0 + rounder */

+        "paddd     (%1), %%mm0         \n\t" /* mm0 += rounder */

+        "psubd    %%mm1, %%mm3         \n\t" /* mm3 = a1-b1 a0-b0 + rounder */

 

-    psrad_i2r (ROW_SHIFT, mm3);         /* mm3 = y6 y7 */

-    paddd_r2r (mm4, mm1);               /* mm1 = a1+b1 a0+b0 + rounder */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm3         \n\t" /* mm3 = y6 y7 */

+        "paddd    %%mm4, %%mm1         \n\t" /* mm1 = a1+b1 a0+b0 + rounder */

 

-    paddd_r2r (mm2, mm0);               /* mm0 = a3 a2 + rounder */

-    psrad_i2r (ROW_SHIFT, mm1);         /* mm1 = y1 y0 */

+        "paddd    %%mm2, %%mm0         \n\t" /* mm0 = a3 a2 + rounder */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm1         \n\t" /* mm1 = y1 y0 */

 

-    paddd_r2r (mm6, mm5);               /* mm5 = b3 b2 */

-    movq_r2r (mm0, mm4);                /* mm4 = a3 a2 + rounder */

+        "paddd    %%mm6, %%mm5         \n\t" /* mm5 = b3 b2 */

+        "movq     %%mm0, %%mm4         \n\t" /* mm4 = a3 a2 + rounder */

 

-    paddd_r2r (mm5, mm0);               /* mm0 = a3+b3 a2+b2 + rounder */

-    psubd_r2r (mm5, mm4);               /* mm4 = a3-b3 a2-b2 + rounder */

+        "paddd    %%mm5, %%mm0         \n\t" /* mm0 = a3+b3 a2+b2 + rounder */

+        "psubd    %%mm5, %%mm4         \n\t" /* mm4 = a3-b3 a2-b2 + rounder */

+        : : "r" (table), "r" (rounder));

 }

 

 static inline void mmxext_row_tail (int16_t * const row, const int store)

 {

-    psrad_i2r (ROW_SHIFT, mm0);         /* mm0 = y3 y2 */

+    __asm__ volatile (

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm0        \n\t"  /* mm0 = y3 y2 */

 

-    psrad_i2r (ROW_SHIFT, mm4);         /* mm4 = y4 y5 */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm4  \n\t"  /* mm4 = y4 y5 */

 

-    packssdw_r2r (mm0, mm1);            /* mm1 = y3 y2 y1 y0 */

+        "packssdw %%mm0, %%mm1        \n\t"  /* mm1 = y3 y2 y1 y0 */

 

-    packssdw_r2r (mm3, mm4);            /* mm4 = y6 y7 y4 y5 */

+        "packssdw %%mm3, %%mm4        \n\t"  /* mm4 = y6 y7 y4 y5 */

 

-    movq_r2m (mm1, *(row+store));       /* save y3 y2 y1 y0 */

-    pshufw_r2r (mm4, mm4, 0xb1);        /* mm4 = y7 y6 y5 y4 */

+        "movq     %%mm1, (%0)         \n\t"  /* save y3 y2 y1 y0 */

+        "pshufw   $0xb1, %%mm4, %%mm4 \n\t"  /* mm4 = y7 y6 y5 y4 */

 

-    /* slot */

+        /* slot */

 

-    movq_r2m (mm4, *(row+store+4));     /* save y7 y6 y5 y4 */

+        "movq     %%mm4, 8(%0)        \n\t"  /* save y7 y6 y5 y4 */

+        :: "r" (row+store)

+        );

 }

 

 static inline void mmxext_row_mid (int16_t * const row, const int store,

                                    const int offset,

                                    const int16_t * const table)

 {

-    movq_m2r (*(row+offset), mm2);      /* mm2 = x6 x4 x2 x0 */

-    psrad_i2r (ROW_SHIFT, mm0);         /* mm0 = y3 y2 */

+    __asm__ volatile (

+        "movq     (%0,%1), %%mm2       \n\t" /* mm2 = x6 x4 x2 x0 */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm0 \n\t"   /* mm0 = y3 y2 */

 

-    movq_m2r (*(row+offset+4), mm5);    /* mm5 = x7 x5 x3 x1 */

-    psrad_i2r (ROW_SHIFT, mm4);         /* mm4 = y4 y5 */

+        "movq    8(%0,%1), %%mm5       \n\t" /* mm5 = x7 x5 x3 x1 */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm4 \n\t" /* mm4 = y4 y5 */

 

-    packssdw_r2r (mm0, mm1);            /* mm1 = y3 y2 y1 y0 */

-    movq_r2r (mm5, mm6);                /* mm6 = x7 x5 x3 x1 */

+        "packssdw   %%mm0, %%mm1       \n\t" /* mm1 = y3 y2 y1 y0 */

+        "movq       %%mm5, %%mm6       \n\t" /* mm6 = x7 x5 x3 x1 */

 

-    packssdw_r2r (mm3, mm4);            /* mm4 = y6 y7 y4 y5 */

-    movq_r2r (mm2, mm0);                /* mm0 = x6 x4 x2 x0 */

+        "packssdw   %%mm3, %%mm4       \n\t" /* mm4 = y6 y7 y4 y5 */

+        "movq       %%mm2, %%mm0       \n\t" /* mm0 = x6 x4 x2 x0 */

 

-    movq_r2m (mm1, *(row+store));       /* save y3 y2 y1 y0 */

-    pshufw_r2r (mm4, mm4, 0xb1);        /* mm4 = y7 y6 y5 y4 */

+        "movq       %%mm1, (%0,%2)     \n\t" /* save y3 y2 y1 y0 */

+        "pshufw     $0xb1, %%mm4, %%mm4\n\t" /* mm4 = y7 y6 y5 y4 */

 

-    movq_m2r (*table, mm3);             /* mm3 = -C2 -C4 C2 C4 */

-    movq_r2m (mm4, *(row+store+4));     /* save y7 y6 y5 y4 */

+        "movq        (%3), %%mm3       \n\t" /* mm3 = -C2 -C4 C2 C4 */

+        "movq       %%mm4, 8(%0,%2)    \n\t" /* save y7 y6 y5 y4 */

 

-    pmaddwd_r2r (mm0, mm3);             /* mm3 = -C4*x4-C2*x6 C4*x0+C2*x2 */

+        "pmaddwd    %%mm0, %%mm3       \n\t" /* mm3= -C4*x4-C2*x6 C4*x0+C2*x2 */

 

-    movq_m2r (*(table+4), mm4);         /* mm4 = C6 C4 C6 C4 */

-    pshufw_r2r (mm2, mm2, 0x4e);        /* mm2 = x2 x0 x6 x4 */

+        "movq       8(%3), %%mm4       \n\t" /* mm4 = C6 C4 C6 C4 */

+        "pshufw     $0x4e, %%mm2, %%mm2\n\t" /* mm2 = x2 x0 x6 x4 */

+        :: "r" (row), "r" ((x86_reg) (2*offset)), "r" ((x86_reg) (2*store)), "r" (table)

+        );

 }

 

 

@@ -202,119 +213,132 @@ static inline void mmxext_row_mid (int16_t * const row, const int store,
 static inline void mmx_row_head (int16_t * const row, const int offset,

                                  const int16_t * const table)

 {

-    movq_m2r (*(row+offset), mm2);      /* mm2 = x6 x4 x2 x0 */

+    __asm__ volatile (

+        "movq (%0), %%mm2       \n\t"    /* mm2 = x6 x4 x2 x0 */

 

-    movq_m2r (*(row+offset+4), mm5);    /* mm5 = x7 x5 x3 x1 */

-    movq_r2r (mm2, mm0);                /* mm0 = x6 x4 x2 x0 */

+        "movq 8(%0), %%mm5      \n\t"    /* mm5 = x7 x5 x3 x1 */

+        "movq %%mm2, %%mm0      \n\t"    /* mm0 = x6 x4 x2 x0 */

 

-    movq_m2r (*table, mm3);             /* mm3 = C6 C4 C2 C4 */

-    movq_r2r (mm5, mm6);                /* mm6 = x7 x5 x3 x1 */

+        "movq (%1), %%mm3       \n\t"    /* mm3 = C6 C4 C2 C4 */

+        "movq %%mm5, %%mm6      \n\t"    /* mm6 = x7 x5 x3 x1 */

 

-    punpckldq_r2r (mm0, mm0);           /* mm0 = x2 x0 x2 x0 */

+        "punpckldq %%mm0, %%mm0 \n\t"    /* mm0 = x2 x0 x2 x0 */

 

-    movq_m2r (*(table+4), mm4);         /* mm4 = -C2 -C4 C6 C4 */

-    pmaddwd_r2r (mm0, mm3);             /* mm3 = C4*x0+C6*x2 C4*x0+C2*x2 */

+        "movq 8(%1), %%mm4      \n\t"    /* mm4 = -C2 -C4 C6 C4 */

+        "pmaddwd %%mm0, %%mm3   \n\t"    /* mm3 = C4*x0+C6*x2 C4*x0+C2*x2 */

 

-    movq_m2r (*(table+8), mm1);         /* mm1 = -C7 C3 C3 C1 */

-    punpckhdq_r2r (mm2, mm2);           /* mm2 = x6 x4 x6 x4 */

+        "movq 16(%1), %%mm1     \n\t"    /* mm1 = -C7 C3 C3 C1 */

+        "punpckhdq %%mm2, %%mm2 \n\t"    /* mm2 = x6 x4 x6 x4 */

+        :: "r" ((row+offset)), "r" (table)

+        );

 }

 

 static inline void mmx_row (const int16_t * const table,

                             const int32_t * const rounder)

 {

-    pmaddwd_r2r (mm2, mm4);             /* mm4 = -C4*x4-C2*x6 C4*x4+C6*x6 */

-    punpckldq_r2r (mm5, mm5);           /* mm5 = x3 x1 x3 x1 */

+    __asm__ volatile (

+        "pmaddwd   %%mm2, %%mm4    \n\t"  /* mm4 = -C4*x4-C2*x6 C4*x4+C6*x6 */

+        "punpckldq %%mm5, %%mm5    \n\t"  /* mm5 = x3 x1 x3 x1 */

 

-    pmaddwd_m2r (*(table+16), mm0);     /* mm0 = C4*x0-C2*x2 C4*x0-C6*x2 */

-    punpckhdq_r2r (mm6, mm6);           /* mm6 = x7 x5 x7 x5 */

+        "pmaddwd  32(%0), %%mm0    \n\t"  /* mm0 = C4*x0-C2*x2 C4*x0-C6*x2 */

+        "punpckhdq %%mm6, %%mm6    \n\t"  /* mm6 = x7 x5 x7 x5 */

 

-    movq_m2r (*(table+12), mm7);        /* mm7 = -C5 -C1 C7 C5 */

-    pmaddwd_r2r (mm5, mm1);             /* mm1 = C3*x1-C7*x3 C1*x1+C3*x3 */

+        "movq     24(%0), %%mm7    \n\t"  /* mm7 = -C5 -C1 C7 C5 */

+        "pmaddwd   %%mm5, %%mm1    \n\t"  /* mm1 = C3*x1-C7*x3 C1*x1+C3*x3 */

 

-    paddd_m2r (*rounder, mm3);          /* mm3 += rounder */

-    pmaddwd_r2r (mm6, mm7);             /* mm7 = -C1*x5-C5*x7 C5*x5+C7*x7 */

+        "paddd      (%1), %%mm3    \n\t"  /* mm3 += rounder */

+        "pmaddwd   %%mm6, %%mm7    \n\t"  /* mm7 = -C1*x5-C5*x7 C5*x5+C7*x7 */

 

-    pmaddwd_m2r (*(table+20), mm2);     /* mm2 = C4*x4-C6*x6 -C4*x4+C2*x6 */

-    paddd_r2r (mm4, mm3);               /* mm3 = a1 a0 + rounder */

+        "pmaddwd  40(%0), %%mm2    \n\t"  /* mm2 = C4*x4-C6*x6 -C4*x4+C2*x6 */

+        "paddd     %%mm4, %%mm3    \n\t"  /* mm3 = a1 a0 + rounder */

 

-    pmaddwd_m2r (*(table+24), mm5);     /* mm5 = C7*x1-C5*x3 C5*x1-C1*x3 */

-    movq_r2r (mm3, mm4);                /* mm4 = a1 a0 + rounder */

+        "pmaddwd  48(%0), %%mm5    \n\t"  /* mm5 = C7*x1-C5*x3 C5*x1-C1*x3 */

+        "movq      %%mm3, %%mm4    \n\t"  /* mm4 = a1 a0 + rounder */

 

-    pmaddwd_m2r (*(table+28), mm6);     /* mm6 = C3*x5-C1*x7 C7*x5+C3*x7 */

-    paddd_r2r (mm7, mm1);               /* mm1 = b1 b0 */

+        "pmaddwd  56(%0), %%mm6    \n\t"  /* mm6 = C3*x5-C1*x7 C7*x5+C3*x7 */

+        "paddd     %%mm7, %%mm1    \n\t"  /* mm1 = b1 b0 */

 

-    paddd_m2r (*rounder, mm0);          /* mm0 += rounder */

-    psubd_r2r (mm1, mm3);               /* mm3 = a1-b1 a0-b0 + rounder */

+        "paddd      (%1), %%mm0    \n\t"  /* mm0 += rounder */

+        "psubd     %%mm1, %%mm3    \n\t"  /* mm3 = a1-b1 a0-b0 + rounder */

 

-    psrad_i2r (ROW_SHIFT, mm3);         /* mm3 = y6 y7 */

-    paddd_r2r (mm4, mm1);               /* mm1 = a1+b1 a0+b0 + rounder */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm3    \n\t"  /* mm3 = y6 y7 */

+        "paddd     %%mm4, %%mm1    \n\t"  /* mm1 = a1+b1 a0+b0 + rounder */

 

-    paddd_r2r (mm2, mm0);               /* mm0 = a3 a2 + rounder */

-    psrad_i2r (ROW_SHIFT, mm1);         /* mm1 = y1 y0 */

+        "paddd     %%mm2, %%mm0    \n\t"  /* mm0 = a3 a2 + rounder */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm1    \n\t"  /* mm1 = y1 y0 */

 

-    paddd_r2r (mm6, mm5);               /* mm5 = b3 b2 */

-    movq_r2r (mm0, mm7);                /* mm7 = a3 a2 + rounder */

+        "paddd     %%mm6, %%mm5    \n\t"  /* mm5 = b3 b2 */

+        "movq      %%mm0, %%mm7    \n\t"  /* mm7 = a3 a2 + rounder */

 

-    paddd_r2r (mm5, mm0);               /* mm0 = a3+b3 a2+b2 + rounder */

-    psubd_r2r (mm5, mm7);               /* mm7 = a3-b3 a2-b2 + rounder */

+        "paddd     %%mm5, %%mm0    \n\t"  /* mm0 = a3+b3 a2+b2 + rounder */

+        "psubd     %%mm5, %%mm7    \n\t"  /* mm7 = a3-b3 a2-b2 + rounder */

+        :: "r" (table), "r" (rounder)

+        );

 }

 

 static inline void mmx_row_tail (int16_t * const row, const int store)

 {

-    psrad_i2r (ROW_SHIFT, mm0);         /* mm0 = y3 y2 */

+    __asm__ volatile (

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm0      \n\t" /* mm0 = y3 y2 */

 

-    psrad_i2r (ROW_SHIFT, mm7);         /* mm7 = y4 y5 */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm7      \n\t" /* mm7 = y4 y5 */

 

-    packssdw_r2r (mm0, mm1);            /* mm1 = y3 y2 y1 y0 */

+        "packssdw %%mm0, %%mm1 \n\t" /* mm1 = y3 y2 y1 y0 */

 

-    packssdw_r2r (mm3, mm7);            /* mm7 = y6 y7 y4 y5 */

+        "packssdw %%mm3, %%mm7 \n\t" /* mm7 = y6 y7 y4 y5 */

 

-    movq_r2m (mm1, *(row+store));       /* save y3 y2 y1 y0 */

-    movq_r2r (mm7, mm4);                /* mm4 = y6 y7 y4 y5 */

+        "movq %%mm1, (%0)      \n\t" /* save y3 y2 y1 y0 */

+        "movq %%mm7, %%mm4     \n\t" /* mm4 = y6 y7 y4 y5 */

 

-    pslld_i2r (16, mm7);                /* mm7 = y7 0 y5 0 */

+        "pslld $16, %%mm7      \n\t" /* mm7 = y7 0 y5 0 */

 

-    psrld_i2r (16, mm4);                /* mm4 = 0 y6 0 y4 */

+        "psrld $16, %%mm4      \n\t" /* mm4 = 0 y6 0 y4 */

 

-    por_r2r (mm4, mm7);                 /* mm7 = y7 y6 y5 y4 */

+        "por %%mm4, %%mm7      \n\t" /* mm7 = y7 y6 y5 y4 */

 

-    /* slot */

+        /* slot */

 

-    movq_r2m (mm7, *(row+store+4));     /* save y7 y6 y5 y4 */

+        "movq %%mm7, 8(%0)     \n\t" /* save y7 y6 y5 y4 */

+        :: "r" (row+store)

+        );

 }

 

 static inline void mmx_row_mid (int16_t * const row, const int store,

                                 const int offset, const int16_t * const table)

 {

-    movq_m2r (*(row+offset), mm2);      /* mm2 = x6 x4 x2 x0 */

-    psrad_i2r (ROW_SHIFT, mm0);         /* mm0 = y3 y2 */

 

-    movq_m2r (*(row+offset+4), mm5);    /* mm5 = x7 x5 x3 x1 */

-    psrad_i2r (ROW_SHIFT, mm7);         /* mm7 = y4 y5 */

+    __asm__ volatile (

+        "movq    (%0,%1), %%mm2    \n\t" /* mm2 = x6 x4 x2 x0 */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm0 \n\t" /* mm0 = y3 y2 */

 

-    packssdw_r2r (mm0, mm1);            /* mm1 = y3 y2 y1 y0 */

-    movq_r2r (mm5, mm6);                /* mm6 = x7 x5 x3 x1 */

+        "movq   8(%0,%1), %%mm5    \n\t" /* mm5 = x7 x5 x3 x1 */

+        "psrad $" AV_STRINGIFY(ROW_SHIFT) ", %%mm7 \n\t" /* mm7 = y4 y5 */

 

-    packssdw_r2r (mm3, mm7);            /* mm7 = y6 y7 y4 y5 */

-    movq_r2r (mm2, mm0);                /* mm0 = x6 x4 x2 x0 */

+        "packssdw  %%mm0, %%mm1    \n\t" /* mm1 = y3 y2 y1 y0 */

+        "movq      %%mm5, %%mm6    \n\t" /* mm6 = x7 x5 x3 x1 */

 

-    movq_r2m (mm1, *(row+store));       /* save y3 y2 y1 y0 */

-    movq_r2r (mm7, mm1);                /* mm1 = y6 y7 y4 y5 */

+        "packssdw  %%mm3, %%mm7    \n\t" /* mm7 = y6 y7 y4 y5 */

+        "movq      %%mm2, %%mm0    \n\t" /* mm0 = x6 x4 x2 x0 */

 

-    punpckldq_r2r (mm0, mm0);           /* mm0 = x2 x0 x2 x0 */

-    psrld_i2r (16, mm7);                /* mm7 = 0 y6 0 y4 */

+        "movq      %%mm1, (%0,%2)  \n\t" /* save y3 y2 y1 y0 */

+        "movq      %%mm7, %%mm1    \n\t" /* mm1 = y6 y7 y4 y5 */

 

-    movq_m2r (*table, mm3);             /* mm3 = C6 C4 C2 C4 */

-    pslld_i2r (16, mm1);                /* mm1 = y7 0 y5 0 */

+        "punpckldq %%mm0, %%mm0    \n\t" /* mm0 = x2 x0 x2 x0 */

+        "psrld       $16, %%mm7    \n\t" /* mm7 = 0 y6 0 y4 */

 

-    movq_m2r (*(table+4), mm4);         /* mm4 = -C2 -C4 C6 C4 */

-    por_r2r (mm1, mm7);                 /* mm7 = y7 y6 y5 y4 */

+        "movq       (%3), %%mm3    \n\t" /* mm3 = C6 C4 C2 C4 */

+        "pslld       $16, %%mm1    \n\t" /* mm1 = y7 0 y5 0 */

 

-    movq_m2r (*(table+8), mm1);         /* mm1 = -C7 C3 C3 C1 */

-    punpckhdq_r2r (mm2, mm2);           /* mm2 = x6 x4 x6 x4 */

+        "movq      8(%3), %%mm4    \n\t" /* mm4 = -C2 -C4 C6 C4 */

+        "por       %%mm1, %%mm7    \n\t" /* mm7 = y7 y6 y5 y4 */

 

-    movq_r2m (mm7, *(row+store+4));     /* save y7 y6 y5 y4 */

-    pmaddwd_r2r (mm0, mm3);             /* mm3 = C4*x0+C6*x2 C4*x0+C2*x2 */

+        "movq     16(%3), %%mm1    \n\t" /* mm1 = -C7 C3 C3 C1 */

+        "punpckhdq %%mm2, %%mm2    \n\t" /* mm2 = x6 x4 x6 x4 */

+

+        "movq      %%mm7, 8(%0,%2) \n\t" /* save y7 y6 y5 y4 */

+        "pmaddwd   %%mm0, %%mm3    \n\t" /* mm3 = C4*x0+C6*x2 C4*x0+C2*x2 */

+        : : "r" (row), "r" ((x86_reg) (2*offset)), "r" ((x86_reg) (2*store)), "r" (table)

+        );

 }

 

 

@@ -398,140 +422,145 @@ static inline void idct_col (int16_t * const col, const int offset)
 #define T3 43790

 #define C4 23170

 

-    DECLARE_ALIGNED(8, static const short, t1_vector)[] = {T1,T1,T1,T1};

-    DECLARE_ALIGNED(8, static const short, t2_vector)[] = {T2,T2,T2,T2};

-    DECLARE_ALIGNED(8, static const short, t3_vector)[] = {T3,T3,T3,T3};

-    DECLARE_ALIGNED(8, static const short, c4_vector)[] = {C4,C4,C4,C4};

+    DECLARE_ALIGNED(8, static const short, t1_vector)[] = {

+        T1,T1,T1,T1,

+        T2,T2,T2,T2,

+        T3,T3,T3,T3,

+        C4,C4,C4,C4

+    };

 

     /* column code adapted from Peter Gubanov */

     /* http://www.elecard.com/peter/idct.shtml */

 

-    movq_m2r (*t1_vector, mm0);         /* mm0 = T1 */

+    __asm__ volatile (

+        "movq      (%0), %%mm0    \n\t" /* mm0 = T1 */

 

-    movq_m2r (*(col+offset+1*8), mm1);  /* mm1 = x1 */

-    movq_r2r (mm0, mm2);                /* mm2 = T1 */

+        "movq   2*8(%1), %%mm1    \n\t" /* mm1 = x1 */

+        "movq     %%mm0, %%mm2    \n\t" /* mm2 = T1 */

 

-    movq_m2r (*(col+offset+7*8), mm4);  /* mm4 = x7 */

-    pmulhw_r2r (mm1, mm0);              /* mm0 = T1*x1 */

+        "movq 7*2*8(%1), %%mm4    \n\t" /* mm4 = x7 */

+        "pmulhw   %%mm1, %%mm0    \n\t" /* mm0 = T1*x1 */

 

-    movq_m2r (*t3_vector, mm5);         /* mm5 = T3 */

-    pmulhw_r2r (mm4, mm2);              /* mm2 = T1*x7 */

+        "movq    16(%0), %%mm5    \n\t" /* mm5 = T3 */

+        "pmulhw   %%mm4, %%mm2    \n\t" /* mm2 = T1*x7 */

 

-    movq_m2r (*(col+offset+5*8), mm6);  /* mm6 = x5 */

-    movq_r2r (mm5, mm7);                /* mm7 = T3-1 */

+        "movq 2*5*8(%1), %%mm6    \n\t" /* mm6 = x5 */

+        "movq     %%mm5, %%mm7    \n\t" /* mm7 = T3-1 */

 

-    movq_m2r (*(col+offset+3*8), mm3);  /* mm3 = x3 */

-    psubsw_r2r (mm4, mm0);              /* mm0 = v17 */

+        "movq 3*8*2(%1), %%mm3    \n\t" /* mm3 = x3 */

+        "psubsw   %%mm4, %%mm0    \n\t" /* mm0 = v17 */

 

-    movq_m2r (*t2_vector, mm4);         /* mm4 = T2 */

-    pmulhw_r2r (mm3, mm5);              /* mm5 = (T3-1)*x3 */

+        "movq     8(%0), %%mm4    \n\t" /* mm4 = T2 */

+        "pmulhw   %%mm3, %%mm5    \n\t" /* mm5 = (T3-1)*x3 */

 

-    paddsw_r2r (mm2, mm1);              /* mm1 = u17 */

-    pmulhw_r2r (mm6, mm7);              /* mm7 = (T3-1)*x5 */

+        "paddsw   %%mm2, %%mm1    \n\t" /* mm1 = u17 */

+        "pmulhw   %%mm6, %%mm7    \n\t" /* mm7 = (T3-1)*x5 */

 

-    /* slot */

+        /* slot */

 

-    movq_r2r (mm4, mm2);                /* mm2 = T2 */

-    paddsw_r2r (mm3, mm5);              /* mm5 = T3*x3 */

+        "movq     %%mm4, %%mm2    \n\t" /* mm2 = T2 */

+        "paddsw   %%mm3, %%mm5    \n\t" /* mm5 = T3*x3 */

 

-    pmulhw_m2r (*(col+offset+2*8), mm4);/* mm4 = T2*x2 */

-    paddsw_r2r (mm6, mm7);              /* mm7 = T3*x5 */

+        "pmulhw 2*8*2(%1), %%mm4  \n\t" /* mm4 = T2*x2 */

+        "paddsw   %%mm6, %%mm7    \n\t" /* mm7 = T3*x5 */

 

-    psubsw_r2r (mm6, mm5);              /* mm5 = v35 */

-    paddsw_r2r (mm3, mm7);              /* mm7 = u35 */

+        "psubsw   %%mm6, %%mm5    \n\t" /* mm5 = v35 */

+        "paddsw   %%mm3, %%mm7    \n\t" /* mm7 = u35 */

 

-    movq_m2r (*(col+offset+6*8), mm3);  /* mm3 = x6 */

-    movq_r2r (mm0, mm6);                /* mm6 = v17 */

+        "movq 6*8*2(%1), %%mm3    \n\t" /* mm3 = x6 */

+        "movq     %%mm0, %%mm6    \n\t" /* mm6 = v17 */

 

-    pmulhw_r2r (mm3, mm2);              /* mm2 = T2*x6 */

-    psubsw_r2r (mm5, mm0);              /* mm0 = b3 */

+        "pmulhw   %%mm3, %%mm2    \n\t" /* mm2 = T2*x6 */

+        "psubsw   %%mm5, %%mm0    \n\t" /* mm0 = b3 */

 

-    psubsw_r2r (mm3, mm4);              /* mm4 = v26 */

-    paddsw_r2r (mm6, mm5);              /* mm5 = v12 */

+        "psubsw   %%mm3, %%mm4    \n\t" /* mm4 = v26 */

+        "paddsw   %%mm6, %%mm5    \n\t" /* mm5 = v12 */

 

-    movq_r2m (mm0, *(col+offset+3*8));  /* save b3 in scratch0 */

-    movq_r2r (mm1, mm6);                /* mm6 = u17 */

+        "movq     %%mm0, 3*8*2(%1)\n\t" /* save b3 in scratch0 */

+        "movq     %%mm1, %%mm6    \n\t" /* mm6 = u17 */

 

-    paddsw_m2r (*(col+offset+2*8), mm2);/* mm2 = u26 */

-    paddsw_r2r (mm7, mm6);              /* mm6 = b0 */

+        "paddsw 2*8*2(%1), %%mm2  \n\t" /* mm2 = u26 */

+        "paddsw   %%mm7, %%mm6    \n\t" /* mm6 = b0 */

 

-    psubsw_r2r (mm7, mm1);              /* mm1 = u12 */

-    movq_r2r (mm1, mm7);                /* mm7 = u12 */

+        "psubsw   %%mm7, %%mm1    \n\t" /* mm1 = u12 */

+        "movq     %%mm1, %%mm7    \n\t" /* mm7 = u12 */

 

-    movq_m2r (*(col+offset+0*8), mm3);  /* mm3 = x0 */

-    paddsw_r2r (mm5, mm1);              /* mm1 = u12+v12 */

+        "movq   0*8(%1), %%mm3    \n\t" /* mm3 = x0 */

+        "paddsw   %%mm5, %%mm1    \n\t" /* mm1 = u12+v12 */

 

-    movq_m2r (*c4_vector, mm0);         /* mm0 = C4/2 */

-    psubsw_r2r (mm5, mm7);              /* mm7 = u12-v12 */

+        "movq    24(%0), %%mm0    \n\t" /* mm0 = C4/2 */

+        "psubsw   %%mm5, %%mm7    \n\t" /* mm7 = u12-v12 */

 

-    movq_r2m (mm6, *(col+offset+5*8));  /* save b0 in scratch1 */

-    pmulhw_r2r (mm0, mm1);              /* mm1 = b1/2 */

+        "movq     %%mm6, 5*8*2(%1)\n\t" /* save b0 in scratch1 */

+        "pmulhw   %%mm0, %%mm1    \n\t" /* mm1 = b1/2 */

 

-    movq_r2r (mm4, mm6);                /* mm6 = v26 */

-    pmulhw_r2r (mm0, mm7);              /* mm7 = b2/2 */

+        "movq     %%mm4, %%mm6    \n\t" /* mm6 = v26 */

+        "pmulhw   %%mm0, %%mm7    \n\t" /* mm7 = b2/2 */

 

-    movq_m2r (*(col+offset+4*8), mm5);  /* mm5 = x4 */

-    movq_r2r (mm3, mm0);                /* mm0 = x0 */

+        "movq 4*8*2(%1), %%mm5    \n\t" /* mm5 = x4 */

+        "movq     %%mm3, %%mm0    \n\t" /* mm0 = x0 */

 

-    psubsw_r2r (mm5, mm3);              /* mm3 = v04 */

-    paddsw_r2r (mm5, mm0);              /* mm0 = u04 */

+        "psubsw   %%mm5, %%mm3    \n\t" /* mm3 = v04 */

+        "paddsw   %%mm5, %%mm0    \n\t" /* mm0 = u04 */

 

-    paddsw_r2r (mm3, mm4);              /* mm4 = a1 */

-    movq_r2r (mm0, mm5);                /* mm5 = u04 */

+        "paddsw   %%mm3, %%mm4    \n\t" /* mm4 = a1 */

+        "movq     %%mm0, %%mm5    \n\t" /* mm5 = u04 */

 

-    psubsw_r2r (mm6, mm3);              /* mm3 = a2 */

-    paddsw_r2r (mm2, mm5);              /* mm5 = a0 */

+        "psubsw   %%mm6, %%mm3    \n\t" /* mm3 = a2 */

+        "paddsw   %%mm2, %%mm5    \n\t" /* mm5 = a0 */

 

-    paddsw_r2r (mm1, mm1);              /* mm1 = b1 */

-    psubsw_r2r (mm2, mm0);              /* mm0 = a3 */

+        "paddsw   %%mm1, %%mm1    \n\t" /* mm1 = b1 */

+        "psubsw   %%mm2, %%mm0    \n\t" /* mm0 = a3 */

 

-    paddsw_r2r (mm7, mm7);              /* mm7 = b2 */

-    movq_r2r (mm3, mm2);                /* mm2 = a2 */

+        "paddsw   %%mm7, %%mm7    \n\t" /* mm7 = b2 */

+        "movq     %%mm3, %%mm2    \n\t" /* mm2 = a2 */

 

-    movq_r2r (mm4, mm6);                /* mm6 = a1 */

-    paddsw_r2r (mm7, mm3);              /* mm3 = a2+b2 */

+        "movq     %%mm4, %%mm6    \n\t" /* mm6 = a1 */

+        "paddsw   %%mm7, %%mm3    \n\t" /* mm3 = a2+b2 */

 

-    psraw_i2r (COL_SHIFT, mm3);         /* mm3 = y2 */

-    paddsw_r2r (mm1, mm4);              /* mm4 = a1+b1 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm3\n\t" /* mm3 = y2 */

+        "paddsw   %%mm1, %%mm4\n\t" /* mm4 = a1+b1 */

 

-    psraw_i2r (COL_SHIFT, mm4);         /* mm4 = y1 */

-    psubsw_r2r (mm1, mm6);              /* mm6 = a1-b1 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm4\n\t" /* mm4 = y1 */

+        "psubsw   %%mm1, %%mm6    \n\t" /* mm6 = a1-b1 */

 

-    movq_m2r (*(col+offset+5*8), mm1);  /* mm1 = b0 */

-    psubsw_r2r (mm7, mm2);              /* mm2 = a2-b2 */

+        "movq 5*8*2(%1), %%mm1    \n\t" /* mm1 = b0 */

+        "psubsw   %%mm7, %%mm2    \n\t" /* mm2 = a2-b2 */

 

-    psraw_i2r (COL_SHIFT, mm6);         /* mm6 = y6 */

-    movq_r2r (mm5, mm7);                /* mm7 = a0 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm6\n\t" /* mm6 = y6 */

+        "movq     %%mm5, %%mm7    \n\t" /* mm7 = a0 */

 

-    movq_r2m (mm4, *(col+offset+1*8));  /* save y1 */

-    psraw_i2r (COL_SHIFT, mm2);         /* mm2 = y5 */

+        "movq     %%mm4, 1*8*2(%1)\n\t" /* save y1 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm2\n\t" /* mm2 = y5 */

 

-    movq_r2m (mm3, *(col+offset+2*8));  /* save y2 */

-    paddsw_r2r (mm1, mm5);              /* mm5 = a0+b0 */

+        "movq     %%mm3, 2*8*2(%1)\n\t" /* save y2 */

+        "paddsw   %%mm1, %%mm5    \n\t" /* mm5 = a0+b0 */

 

-    movq_m2r (*(col+offset+3*8), mm4);  /* mm4 = b3 */

-    psubsw_r2r (mm1, mm7);              /* mm7 = a0-b0 */

+        "movq 3*8*2(%1), %%mm4    \n\t" /* mm4 = b3 */

+        "psubsw   %%mm1, %%mm7    \n\t" /* mm7 = a0-b0 */

 

-    psraw_i2r (COL_SHIFT, mm5);         /* mm5 = y0 */

-    movq_r2r (mm0, mm3);                /* mm3 = a3 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm5\n\t" /* mm5 = y0 */

+        "movq     %%mm0, %%mm3    \n\t" /* mm3 = a3 */

 

-    movq_r2m (mm2, *(col+offset+5*8));  /* save y5 */

-    psubsw_r2r (mm4, mm3);              /* mm3 = a3-b3 */

+        "movq     %%mm2, 5*8*2(%1)\n\t" /* save y5 */

+        "psubsw   %%mm4, %%mm3    \n\t" /* mm3 = a3-b3 */

 

-    psraw_i2r (COL_SHIFT, mm7);         /* mm7 = y7 */

-    paddsw_r2r (mm0, mm4);              /* mm4 = a3+b3 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm7\n\t" /* mm7 = y7 */

+        "paddsw   %%mm0, %%mm4    \n\t" /* mm4 = a3+b3 */

 

-    movq_r2m (mm5, *(col+offset+0*8));  /* save y0 */

-    psraw_i2r (COL_SHIFT, mm3);         /* mm3 = y4 */

+        "movq     %%mm5, 0*8*2(%1)\n\t" /* save y0 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm3\n\t" /* mm3 = y4 */

 

-    movq_r2m (mm6, *(col+offset+6*8));  /* save y6 */

-    psraw_i2r (COL_SHIFT, mm4);         /* mm4 = y3 */

+        "movq     %%mm6, 6*8*2(%1)\n\t" /* save y6 */

+        "psraw $" AV_STRINGIFY(COL_SHIFT) ", %%mm4\n\t" /* mm4 = y3 */

 

-    movq_r2m (mm7, *(col+offset+7*8));  /* save y7 */

+        "movq     %%mm7, 7*8*2(%1)\n\t" /* save y7 */

 

-    movq_r2m (mm3, *(col+offset+4*8));  /* save y4 */

+        "movq     %%mm3, 4*8*2(%1)\n\t" /* save y4 */

 

-    movq_r2m (mm4, *(col+offset+3*8));  /* save y3 */

+        "movq     %%mm4, 3*8*2(%1)\n\t" /* save y3 */

+        :: "r" (t1_vector), "r" (col+offset)

+        );

 

 #undef T1

 #undef T2

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/mmx.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/mmx.h
deleted file mode 100644
index ef064e3e3..000000000
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/mmx.h
+++ /dev/null
@@ -1,267 +0,0 @@
-/*

- * mmx.h

- * Copyright (C) 1997-2001 H. Dietz and R. Fisher

- *

- * This file is part of FFmpeg.

- *

- * FFmpeg is free software; you can redistribute it and/or

- * modify it under the terms of the GNU Lesser General Public

- * License as published by the Free Software Foundation; either

- * version 2.1 of the License, or (at your option) any later version.

- *

- * FFmpeg is distributed in the hope that it will be useful,

- * but WITHOUT ANY WARRANTY; without even the implied warranty of

- * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

- * Lesser General Public License for more details.

- *

- * You should have received a copy of the GNU Lesser General Public

- * License along with FFmpeg; if not, write to the Free Software

- * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

- */

-#ifndef AVCODEC_X86_MMX_H

-#define AVCODEC_X86_MMX_H

-

-#warning Everything in this header is deprecated, use plain __asm__()! New code using this header will be rejected.

-

-

-#define         mmx_i2r(op,imm,reg) \

-        __asm__ volatile (#op " %0, %%" #reg \

-                              : /* nothing */ \

-                              : "i" (imm) )

-

-#define         mmx_m2r(op,mem,reg) \

-        __asm__ volatile (#op " %0, %%" #reg \

-                              : /* nothing */ \

-                              : "m" (mem))

-

-#define         mmx_r2m(op,reg,mem) \

-        __asm__ volatile (#op " %%" #reg ", %0" \

-                              : "=m" (mem) \

-                              : /* nothing */ )

-

-#define         mmx_r2r(op,regs,regd) \

-        __asm__ volatile (#op " %" #regs ", %" #regd)

-

-

-#define         emms() __asm__ volatile ("emms")

-

-#define         movd_m2r(var,reg)           mmx_m2r (movd, var, reg)

-#define         movd_r2m(reg,var)           mmx_r2m (movd, reg, var)

-#define         movd_r2r(regs,regd)         mmx_r2r (movd, regs, regd)

-

-#define         movq_m2r(var,reg)           mmx_m2r (movq, var, reg)

-#define         movq_r2m(reg,var)           mmx_r2m (movq, reg, var)

-#define         movq_r2r(regs,regd)         mmx_r2r (movq, regs, regd)

-

-#define         packssdw_m2r(var,reg)       mmx_m2r (packssdw, var, reg)

-#define         packssdw_r2r(regs,regd)     mmx_r2r (packssdw, regs, regd)

-#define         packsswb_m2r(var,reg)       mmx_m2r (packsswb, var, reg)

-#define         packsswb_r2r(regs,regd)     mmx_r2r (packsswb, regs, regd)

-

-#define         packuswb_m2r(var,reg)       mmx_m2r (packuswb, var, reg)

-#define         packuswb_r2r(regs,regd)     mmx_r2r (packuswb, regs, regd)

-

-#define         paddb_m2r(var,reg)          mmx_m2r (paddb, var, reg)

-#define         paddb_r2r(regs,regd)        mmx_r2r (paddb, regs, regd)

-#define         paddd_m2r(var,reg)          mmx_m2r (paddd, var, reg)

-#define         paddd_r2r(regs,regd)        mmx_r2r (paddd, regs, regd)

-#define         paddw_m2r(var,reg)          mmx_m2r (paddw, var, reg)

-#define         paddw_r2r(regs,regd)        mmx_r2r (paddw, regs, regd)

-

-#define         paddsb_m2r(var,reg)         mmx_m2r (paddsb, var, reg)

-#define         paddsb_r2r(regs,regd)       mmx_r2r (paddsb, regs, regd)

-#define         paddsw_m2r(var,reg)         mmx_m2r (paddsw, var, reg)

-#define         paddsw_r2r(regs,regd)       mmx_r2r (paddsw, regs, regd)

-

-#define         paddusb_m2r(var,reg)        mmx_m2r (paddusb, var, reg)

-#define         paddusb_r2r(regs,regd)      mmx_r2r (paddusb, regs, regd)

-#define         paddusw_m2r(var,reg)        mmx_m2r (paddusw, var, reg)

-#define         paddusw_r2r(regs,regd)      mmx_r2r (paddusw, regs, regd)

-

-#define         pand_m2r(var,reg)           mmx_m2r (pand, var, reg)

-#define         pand_r2r(regs,regd)         mmx_r2r (pand, regs, regd)

-

-#define         pandn_m2r(var,reg)          mmx_m2r (pandn, var, reg)

-#define         pandn_r2r(regs,regd)        mmx_r2r (pandn, regs, regd)

-

-#define         pcmpeqb_m2r(var,reg)        mmx_m2r (pcmpeqb, var, reg)

-#define         pcmpeqb_r2r(regs,regd)      mmx_r2r (pcmpeqb, regs, regd)

-#define         pcmpeqd_m2r(var,reg)        mmx_m2r (pcmpeqd, var, reg)

-#define         pcmpeqd_r2r(regs,regd)      mmx_r2r (pcmpeqd, regs, regd)

-#define         pcmpeqw_m2r(var,reg)        mmx_m2r (pcmpeqw, var, reg)

-#define         pcmpeqw_r2r(regs,regd)      mmx_r2r (pcmpeqw, regs, regd)

-

-#define         pcmpgtb_m2r(var,reg)        mmx_m2r (pcmpgtb, var, reg)

-#define         pcmpgtb_r2r(regs,regd)      mmx_r2r (pcmpgtb, regs, regd)

-#define         pcmpgtd_m2r(var,reg)        mmx_m2r (pcmpgtd, var, reg)

-#define         pcmpgtd_r2r(regs,regd)      mmx_r2r (pcmpgtd, regs, regd)

-#define         pcmpgtw_m2r(var,reg)        mmx_m2r (pcmpgtw, var, reg)

-#define         pcmpgtw_r2r(regs,regd)      mmx_r2r (pcmpgtw, regs, regd)

-

-#define         pmaddwd_m2r(var,reg)        mmx_m2r (pmaddwd, var, reg)

-#define         pmaddwd_r2r(regs,regd)      mmx_r2r (pmaddwd, regs, regd)

-

-#define         pmulhw_m2r(var,reg)         mmx_m2r (pmulhw, var, reg)

-#define         pmulhw_r2r(regs,regd)       mmx_r2r (pmulhw, regs, regd)

-

-#define         pmullw_m2r(var,reg)         mmx_m2r (pmullw, var, reg)

-#define         pmullw_r2r(regs,regd)       mmx_r2r (pmullw, regs, regd)

-

-#define         por_m2r(var,reg)            mmx_m2r (por, var, reg)

-#define         por_r2r(regs,regd)          mmx_r2r (por, regs, regd)

-

-#define         pslld_i2r(imm,reg)          mmx_i2r (pslld, imm, reg)

-#define         pslld_m2r(var,reg)          mmx_m2r (pslld, var, reg)

-#define         pslld_r2r(regs,regd)        mmx_r2r (pslld, regs, regd)

-#define         psllq_i2r(imm,reg)          mmx_i2r (psllq, imm, reg)

-#define         psllq_m2r(var,reg)          mmx_m2r (psllq, var, reg)

-#define         psllq_r2r(regs,regd)        mmx_r2r (psllq, regs, regd)

-#define         psllw_i2r(imm,reg)          mmx_i2r (psllw, imm, reg)

-#define         psllw_m2r(var,reg)          mmx_m2r (psllw, var, reg)

-#define         psllw_r2r(regs,regd)        mmx_r2r (psllw, regs, regd)

-

-#define         psrad_i2r(imm,reg)          mmx_i2r (psrad, imm, reg)

-#define         psrad_m2r(var,reg)          mmx_m2r (psrad, var, reg)

-#define         psrad_r2r(regs,regd)        mmx_r2r (psrad, regs, regd)

-#define         psraw_i2r(imm,reg)          mmx_i2r (psraw, imm, reg)

-#define         psraw_m2r(var,reg)          mmx_m2r (psraw, var, reg)

-#define         psraw_r2r(regs,regd)        mmx_r2r (psraw, regs, regd)

-

-#define         psrld_i2r(imm,reg)          mmx_i2r (psrld, imm, reg)

-#define         psrld_m2r(var,reg)          mmx_m2r (psrld, var, reg)

-#define         psrld_r2r(regs,regd)        mmx_r2r (psrld, regs, regd)

-#define         psrlq_i2r(imm,reg)          mmx_i2r (psrlq, imm, reg)

-#define         psrlq_m2r(var,reg)          mmx_m2r (psrlq, var, reg)

-#define         psrlq_r2r(regs,regd)        mmx_r2r (psrlq, regs, regd)

-#define         psrlw_i2r(imm,reg)          mmx_i2r (psrlw, imm, reg)

-#define         psrlw_m2r(var,reg)          mmx_m2r (psrlw, var, reg)

-#define         psrlw_r2r(regs,regd)        mmx_r2r (psrlw, regs, regd)

-

-#define         psubb_m2r(var,reg)          mmx_m2r (psubb, var, reg)

-#define         psubb_r2r(regs,regd)        mmx_r2r (psubb, regs, regd)

-#define         psubd_m2r(var,reg)          mmx_m2r (psubd, var, reg)

-#define         psubd_r2r(regs,regd)        mmx_r2r (psubd, regs, regd)

-#define         psubw_m2r(var,reg)          mmx_m2r (psubw, var, reg)

-#define         psubw_r2r(regs,regd)        mmx_r2r (psubw, regs, regd)

-

-#define         psubsb_m2r(var,reg)         mmx_m2r (psubsb, var, reg)

-#define         psubsb_r2r(regs,regd)       mmx_r2r (psubsb, regs, regd)

-#define         psubsw_m2r(var,reg)         mmx_m2r (psubsw, var, reg)

-#define         psubsw_r2r(regs,regd)       mmx_r2r (psubsw, regs, regd)

-

-#define         psubusb_m2r(var,reg)        mmx_m2r (psubusb, var, reg)

-#define         psubusb_r2r(regs,regd)      mmx_r2r (psubusb, regs, regd)

-#define         psubusw_m2r(var,reg)        mmx_m2r (psubusw, var, reg)

-#define         psubusw_r2r(regs,regd)      mmx_r2r (psubusw, regs, regd)

-

-#define         punpckhbw_m2r(var,reg)      mmx_m2r (punpckhbw, var, reg)

-#define         punpckhbw_r2r(regs,regd)    mmx_r2r (punpckhbw, regs, regd)

-#define         punpckhdq_m2r(var,reg)      mmx_m2r (punpckhdq, var, reg)

-#define         punpckhdq_r2r(regs,regd)    mmx_r2r (punpckhdq, regs, regd)

-#define         punpckhwd_m2r(var,reg)      mmx_m2r (punpckhwd, var, reg)

-#define         punpckhwd_r2r(regs,regd)    mmx_r2r (punpckhwd, regs, regd)

-

-#define         punpcklbw_m2r(var,reg)      mmx_m2r (punpcklbw, var, reg)

-#define         punpcklbw_r2r(regs,regd)    mmx_r2r (punpcklbw, regs, regd)

-#define         punpckldq_m2r(var,reg)      mmx_m2r (punpckldq, var, reg)

-#define         punpckldq_r2r(regs,regd)    mmx_r2r (punpckldq, regs, regd)

-#define         punpcklwd_m2r(var,reg)      mmx_m2r (punpcklwd, var, reg)

-#define         punpcklwd_r2r(regs,regd)    mmx_r2r (punpcklwd, regs, regd)

-

-#define         pxor_m2r(var,reg)           mmx_m2r (pxor, var, reg)

-#define         pxor_r2r(regs,regd)         mmx_r2r (pxor, regs, regd)

-

-

-/* 3DNOW extensions */

-

-#define         pavgusb_m2r(var,reg)        mmx_m2r (pavgusb, var, reg)

-#define         pavgusb_r2r(regs,regd)      mmx_r2r (pavgusb, regs, regd)

-

-

-/* AMD MMX extensions - also available in intel SSE */

-

-

-#define         mmx_m2ri(op,mem,reg,imm) \

-        __asm__ volatile (#op " %1, %0, %%" #reg \

-                              : /* nothing */ \

-                              : "m" (mem), "i" (imm))

-#define         mmx_r2ri(op,regs,regd,imm) \

-        __asm__ volatile (#op " %0, %%" #regs ", %%" #regd \

-                              : /* nothing */ \

-                              : "i" (imm) )

-

-#define         mmx_fetch(mem,hint) \

-        __asm__ volatile ("prefetch" #hint " %0" \

-                              : /* nothing */ \

-                              : "m" (mem))

-

-

-#define         maskmovq(regs,maskreg)      mmx_r2ri (maskmovq, regs, maskreg)

-

-#define         movntq_r2m(mmreg,var)       mmx_r2m (movntq, mmreg, var)

-

-#define         pavgb_m2r(var,reg)          mmx_m2r (pavgb, var, reg)

-#define         pavgb_r2r(regs,regd)        mmx_r2r (pavgb, regs, regd)

-#define         pavgw_m2r(var,reg)          mmx_m2r (pavgw, var, reg)

-#define         pavgw_r2r(regs,regd)        mmx_r2r (pavgw, regs, regd)

-

-#define         pextrw_r2r(mmreg,reg,imm)   mmx_r2ri (pextrw, mmreg, reg, imm)

-

-#define         pinsrw_r2r(reg,mmreg,imm)   mmx_r2ri (pinsrw, reg, mmreg, imm)

-

-#define         pmaxsw_m2r(var,reg)         mmx_m2r (pmaxsw, var, reg)

-#define         pmaxsw_r2r(regs,regd)       mmx_r2r (pmaxsw, regs, regd)

-

-#define         pmaxub_m2r(var,reg)         mmx_m2r (pmaxub, var, reg)

-#define         pmaxub_r2r(regs,regd)       mmx_r2r (pmaxub, regs, regd)

-

-#define         pminsw_m2r(var,reg)         mmx_m2r (pminsw, var, reg)

-#define         pminsw_r2r(regs,regd)       mmx_r2r (pminsw, regs, regd)

-

-#define         pminub_m2r(var,reg)         mmx_m2r (pminub, var, reg)

-#define         pminub_r2r(regs,regd)       mmx_r2r (pminub, regs, regd)

-

-#define         pmovmskb(mmreg,reg) \

-        __asm__ volatile ("movmskps %" #mmreg ", %" #reg)

-

-#define         pmulhuw_m2r(var,reg)        mmx_m2r (pmulhuw, var, reg)

-#define         pmulhuw_r2r(regs,regd)      mmx_r2r (pmulhuw, regs, regd)

-

-#define         prefetcht0(mem)             mmx_fetch (mem, t0)

-#define         prefetcht1(mem)             mmx_fetch (mem, t1)

-#define         prefetcht2(mem)             mmx_fetch (mem, t2)

-#define         prefetchnta(mem)            mmx_fetch (mem, nta)

-

-#define         psadbw_m2r(var,reg)         mmx_m2r (psadbw, var, reg)

-#define         psadbw_r2r(regs,regd)       mmx_r2r (psadbw, regs, regd)

-

-#define         pshufw_m2r(var,reg,imm)     mmx_m2ri(pshufw, var, reg, imm)

-#define         pshufw_r2r(regs,regd,imm)   mmx_r2ri(pshufw, regs, regd, imm)

-

-#define         sfence() __asm__ volatile ("sfence\n\t")

-

-/* SSE2 */

-#define         pshufhw_m2r(var,reg,imm)    mmx_m2ri(pshufhw, var, reg, imm)

-#define         pshufhw_r2r(regs,regd,imm)  mmx_r2ri(pshufhw, regs, regd, imm)

-#define         pshuflw_m2r(var,reg,imm)    mmx_m2ri(pshuflw, var, reg, imm)

-#define         pshuflw_r2r(regs,regd,imm)  mmx_r2ri(pshuflw, regs, regd, imm)

-

-#define         pshufd_r2r(regs,regd,imm)   mmx_r2ri(pshufd, regs, regd, imm)

-

-#define         movdqa_m2r(var,reg)         mmx_m2r (movdqa, var, reg)

-#define         movdqa_r2m(reg,var)         mmx_r2m (movdqa, reg, var)

-#define         movdqa_r2r(regs,regd)       mmx_r2r (movdqa, regs, regd)

-#define         movdqu_m2r(var,reg)         mmx_m2r (movdqu, var, reg)

-#define         movdqu_r2m(reg,var)         mmx_r2m (movdqu, reg, var)

-#define         movdqu_r2r(regs,regd)       mmx_r2r (movdqu, regs, regd)

-

-#define         pmullw_r2m(reg,var)         mmx_r2m (pmullw, reg, var)

-

-#define         pslldq_i2r(imm,reg)         mmx_i2r (pslldq, imm, reg)

-#define         psrldq_i2r(imm,reg)         mmx_i2r (psrldq, imm, reg)

-

-#define         punpcklqdq_r2r(regs,regd)   mmx_r2r (punpcklqdq, regs, regd)

-#define         punpckhqdq_r2r(regs,regd)   mmx_r2r (punpckhqdq, regs, regd)

-

-

-#endif /* AVCODEC_X86_MMX_H */

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp56_arith.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp56_arith.h
new file mode 100644
index 000000000..95f96e365
--- /dev/null
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp56_arith.h
@@ -0,0 +1,54 @@
+/**

+ * VP5 and VP6 compatible video decoder (arith decoder)

+ *

+ * Copyright (C) 2006  Aurelien Jacobs <aurel@gnuage.org>

+ * Copyright (C) 2010  Eli Friedman

+ *

+ * This file is part of FFmpeg.

+ *

+ * FFmpeg is free software; you can redistribute it and/or

+ * modify it under the terms of the GNU Lesser General Public

+ * License as published by the Free Software Foundation; either

+ * version 2.1 of the License, or (at your option) any later version.

+ *

+ * FFmpeg is distributed in the hope that it will be useful,

+ * but WITHOUT ANY WARRANTY; without even the implied warranty of

+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

+ * Lesser General Public License for more details.

+ *

+ * You should have received a copy of the GNU Lesser General Public

+ * License along with FFmpeg; if not, write to the Free Software

+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

+ */

+

+#ifndef AVCODEC_X86_VP56_ARITH_H

+#define AVCODEC_X86_VP56_ARITH_H

+

+#if HAVE_FAST_CMOV

+#define vp56_rac_get_prob vp56_rac_get_prob

+static inline int vp56_rac_get_prob(VP56RangeCoder *c, uint8_t prob)

+{

+    unsigned int code_word = vp56_rac_renorm(c);

+    unsigned int high = c->high;

+    unsigned int low = 1 + (((high - 1) * prob) >> 8);

+    unsigned int low_shift = low << 8;

+    int bit = 0;

+

+    __asm__(

+        "subl  %4, %1      \n\t"

+        "subl  %3, %2      \n\t"

+        "leal (%2, %3), %3 \n\t"

+        "setae %b0         \n\t"

+        "cmovb %4, %1      \n\t"

+        "cmovb %3, %2      \n\t"

+        : "+q"(bit), "+r"(high), "+r"(code_word), "+r"(low_shift)

+        : "r"(low)

+    );

+

+    c->high      = high;

+    c->code_word = code_word;

+    return bit;

+}

+#endif

+

+#endif /* AVCODEC_X86_VP56_ARITH_H */

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp-init.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp-init.c
index e06da5e42..dd7dc696e 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp-init.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp-init.c
@@ -350,7 +350,6 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c)
 #endif
 
         c->vp8_v_loop_filter_simple = ff_vp8_v_loop_filter_simple_sse2;
-        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse2;
 
         c->vp8_v_loop_filter16y_inner = ff_vp8_v_loop_filter16y_inner_sse2;
         c->vp8_v_loop_filter8uv_inner = ff_vp8_v_loop_filter8uv_inner_sse2;
@@ -362,6 +361,8 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c)
     if (mm_flags & FF_MM_SSE2) {
         c->vp8_idct_dc_add4y          = ff_vp8_idct_dc_add4y_sse2;
 
+        c->vp8_h_loop_filter_simple = ff_vp8_h_loop_filter_simple_sse2;
+
         c->vp8_h_loop_filter16y_inner = ff_vp8_h_loop_filter16y_inner_sse2;
         c->vp8_h_loop_filter8uv_inner = ff_vp8_h_loop_filter8uv_inner_sse2;
 
@@ -396,6 +397,7 @@ av_cold void ff_vp8dsp_init_x86(VP8DSPContext* c)
     if (mm_flags & FF_MM_SSE4) {
         c->vp8_idct_dc_add                  = ff_vp8_idct_dc_add_sse4;
 
+        c->vp8_h_loop_filter_simple   = ff_vp8_h_loop_filter_simple_sse4;
         c->vp8_h_loop_filter16y       = ff_vp8_h_loop_filter16y_mbedge_sse4;
         c->vp8_h_loop_filter8uv       = ff_vp8_h_loop_filter8uv_mbedge_sse4;
     }
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp.asm b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp.asm
index 4aa901e27..4f430d80c 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp.asm
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcodec/x86/vp8dsp.asm
@@ -145,6 +145,10 @@ filter_h6_shuf3: db 3, 4, 4, 5, 5, 6, 6, 7, 7, 8, 8,  9, 9, 10, 10, 11
 pw_20091: times 4 dw 20091
 pw_17734: times 4 dw 17734
 
+pb_27_63: times 8 db 27, 63
+pb_18_63: times 8 db 18, 63
+pb_9_63:  times 8 db  9, 63
+
 cextern pb_1
 cextern pw_3
 cextern pb_3
@@ -438,48 +442,43 @@ cglobal put_vp8_epel4_h6_mmxext, 6, 6
     jg .nextrow
     REP_RET
 
-; 4x4 block, H-only 4-tap filter
 INIT_XMM
-cglobal put_vp8_epel8_h4_sse2, 6, 6, 8
-    shl      r5d, 4
+cglobal put_vp8_epel8_h4_sse2, 6, 6, 10
+    shl      r5d, 5
 %ifdef PIC
-    lea      r11, [fourtap_filter_hw_m]
+    lea      r11, [fourtap_filter_v_m]
 %endif
-    mova      m5, [fourtap_filter_hw+r5-16] ; set up 4tap filter in words
-    mova      m6, [fourtap_filter_hw+r5]
+    lea       r5, [fourtap_filter_v+r5-32]
     pxor      m7, m7
-
+    mova      m4, [pw_64]
+    mova      m5, [r5+ 0]
+    mova      m6, [r5+16]
+%ifdef m8
+    mova      m8, [r5+32]
+    mova      m9, [r5+48]
+%endif
 .nextrow
-    movh      m0, [r2-1]
-    punpcklbw m0, m7        ; ABCDEFGH
-    mova      m1, m0
-    mova      m2, m0
-    mova      m3, m0
-    psrldq    m1, 2         ; BCDEFGH
-    psrldq    m2, 4         ; CDEFGH
-    psrldq    m3, 6         ; DEFGH
-    punpcklwd m0, m1        ; ABBCCDDE
-    punpcklwd m2, m3        ; CDDEEFFG
-    pmaddwd   m0, m5
-    pmaddwd   m2, m6
-    paddd     m0, m2
-
-    movh      m1, [r2+3]
-    punpcklbw m1, m7        ; ABCDEFGH
-    mova      m2, m1
-    mova      m3, m1
-    mova      m4, m1
-    psrldq    m2, 2         ; BCDEFGH
-    psrldq    m3, 4         ; CDEFGH
-    psrldq    m4, 6         ; DEFGH
-    punpcklwd m1, m2        ; ABBCCDDE
-    punpcklwd m3, m4        ; CDDEEFFG
-    pmaddwd   m1, m5
-    pmaddwd   m3, m6
-    paddd     m1, m3
-
-    packssdw  m0, m1
-    paddsw    m0, [pw_64]
+    movq      m0, [r2-1]
+    movq      m1, [r2-0]
+    movq      m2, [r2+1]
+    movq      m3, [r2+2]
+    punpcklbw m0, m7
+    punpcklbw m1, m7
+    punpcklbw m2, m7
+    punpcklbw m3, m7
+    pmullw    m0, m5
+    pmullw    m1, m6
+%ifdef m8
+    pmullw    m2, m8
+    pmullw    m3, m9
+%else
+    pmullw    m2, [r5+32]
+    pmullw    m3, [r5+48]
+%endif
+    paddsw    m0, m1
+    paddsw    m2, m3
+    paddsw    m0, m2
+    paddsw    m0, m4
     psraw     m0, 7
     packuswb  m0, m7
     movh    [r0], m0        ; store
@@ -491,62 +490,57 @@ cglobal put_vp8_epel8_h4_sse2, 6, 6, 8
     jg .nextrow
     REP_RET
 
-cglobal put_vp8_epel8_h6_sse2, 6, 6, 8
+cglobal put_vp8_epel8_h6_sse2, 6, 6, 14
     lea      r5d, [r5*3]
+    shl      r5d, 4
 %ifdef PIC
-    lea      r11, [sixtap_filter_hw_m]
+    lea      r11, [sixtap_filter_v_m]
 %endif
-    lea       r5, [sixtap_filter_hw+r5*8]
+    lea       r5, [sixtap_filter_v+r5-96]
     pxor      m7, m7
-
+    mova      m6, [pw_64]
+%ifdef m8
+    mova      m8, [r5+ 0]
+    mova      m9, [r5+16]
+    mova     m10, [r5+32]
+    mova     m11, [r5+48]
+    mova     m12, [r5+64]
+    mova     m13, [r5+80]
+%endif
 .nextrow
-    movu      m0, [r2-2]
-    mova      m6, m0
-    mova      m4, m0
-    punpcklbw m0, m7        ; ABCDEFGHI
-    mova      m1, m0
-    mova      m2, m0
-    mova      m3, m0
-    psrldq    m1, 2         ; BCDEFGH
-    psrldq    m2, 4         ; CDEFGH
-    psrldq    m3, 6         ; DEFGH
-    psrldq    m4, 4
-    punpcklbw m4, m7        ; EFGH
-    mova      m5, m4
-    psrldq    m5, 2         ; FGH
-    punpcklwd m0, m1        ; ABBCCDDE
-    punpcklwd m2, m3        ; CDDEEFFG
-    punpcklwd m4, m5        ; EFFGGHHI
-    pmaddwd   m0, [r5-48]
-    pmaddwd   m2, [r5-32]
-    pmaddwd   m4, [r5-16]
-    paddd     m0, m2
-    paddd     m0, m4
-
-    psrldq    m6, 4
-    mova      m4, m6
-    punpcklbw m6, m7        ; ABCDEFGHI
-    mova      m1, m6
-    mova      m2, m6
-    mova      m3, m6
-    psrldq    m1, 2         ; BCDEFGH
-    psrldq    m2, 4         ; CDEFGH
-    psrldq    m3, 6         ; DEFGH
-    psrldq    m4, 4
-    punpcklbw m4, m7        ; EFGH
-    mova      m5, m4
-    psrldq    m5, 2         ; FGH
-    punpcklwd m6, m1        ; ABBCCDDE
-    punpcklwd m2, m3        ; CDDEEFFG
-    punpcklwd m4, m5        ; EFFGGHHI
-    pmaddwd   m6, [r5-48]
-    pmaddwd   m2, [r5-32]
-    pmaddwd   m4, [r5-16]
-    paddd     m6, m2
-    paddd     m6, m4
-
-    packssdw  m0, m6
-    paddsw    m0, [pw_64]
+    movq      m0, [r2-2]
+    movq      m1, [r2-1]
+    movq      m2, [r2-0]
+    movq      m3, [r2+1]
+    movq      m4, [r2+2]
+    movq      m5, [r2+3]
+    punpcklbw m0, m7
+    punpcklbw m1, m7
+    punpcklbw m2, m7
+    punpcklbw m3, m7
+    punpcklbw m4, m7
+    punpcklbw m5, m7
+%ifdef m8
+    pmullw    m0, m8
+    pmullw    m1, m9
+    pmullw    m2, m10
+    pmullw    m3, m11
+    pmullw    m4, m12
+    pmullw    m5, m13
+%else
+    pmullw    m0, [r5+ 0]
+    pmullw    m1, [r5+16]
+    pmullw    m2, [r5+32]
+    pmullw    m3, [r5+48]
+    pmullw    m4, [r5+64]
+    pmullw    m5, [r5+80]
+%endif
+    paddsw    m1, m4
+    paddsw    m0, m5
+    paddsw    m1, m2
+    paddsw    m0, m3
+    paddsw    m0, m1
+    paddsw    m0, m6
     psraw     m0, 7
     packuswb  m0, m7
     movh    [r0], m0        ; store
@@ -1360,44 +1354,123 @@ cglobal vp8_luma_dc_wht_mmx, 2,3
     movd    [%7+%9*2], m%4
 %endmacro
 
-%macro SPLATB_REG 3-4
+; write 4 or 8 words in the mmx/xmm registers as 8 lines
+; 1 and 2 are the registers to write, this can be the same (for SSE2)
+; for pre-SSE4:
+; 3 is a general-purpose register that we will clobber
+; for SSE4:
+; 3 is a pointer to the destination's 5th line
+; 4 is a pointer to the destination's 4th line
+; 5/6 is -stride and +stride
+%macro WRITE_2x4W 6
+    movd             %3, %1
+    punpckhdq        %1, %1
+    mov       [%4+%5*4], %3w
+    shr              %3, 16
+    add              %4, %6
+    mov       [%4+%5*4], %3w
+
+    movd             %3, %1
+    add              %4, %5
+    mov       [%4+%5*2], %3w
+    shr              %3, 16
+    mov       [%4+%5  ], %3w
+
+    movd             %3, %2
+    punpckhdq        %2, %2
+    mov       [%4     ], %3w
+    shr              %3, 16
+    mov       [%4+%6  ], %3w
+
+    movd             %3, %2
+    add              %4, %6
+    mov       [%4+%6  ], %3w
+    shr              %3, 16
+    mov       [%4+%6*2], %3w
+    add              %4, %5
+%endmacro
+
+%macro WRITE_8W_SSE2 5
+    movd             %2, %1
+    psrldq           %1, 4
+    mov       [%3+%4*4], %2w
+    shr              %2, 16
+    add              %3, %5
+    mov       [%3+%4*4], %2w
+
+    movd             %2, %1
+    psrldq           %1, 4
+    add              %3, %4
+    mov       [%3+%4*2], %2w
+    shr              %2, 16
+    mov       [%3+%4  ], %2w
+
+    movd             %2, %1
+    psrldq           %1, 4
+    mov       [%3     ], %2w
+    shr              %2, 16
+    mov       [%3+%5  ], %2w
+
+    movd             %2, %1
+    add              %3, %5
+    mov       [%3+%5  ], %2w
+    shr              %2, 16
+    mov       [%3+%5*2], %2w
+%endmacro
+
+%macro WRITE_8W_SSE4 5
+    pextrw    [%3+%4*4], %1, 0
+    pextrw    [%2+%4*4], %1, 1
+    pextrw    [%3+%4*2], %1, 2
+    pextrw    [%3+%4  ], %1, 3
+    pextrw    [%3     ], %1, 4
+    pextrw    [%2     ], %1, 5
+    pextrw    [%2+%5  ], %1, 6
+    pextrw    [%2+%5*2], %1, 7
+%endmacro
+
+%macro SPLATB_REG_MMX 2-3
     movd           %1, %2
-%ifidn %3, ssse3
-    pshufb         %1, %4
-%else
     punpcklbw      %1, %1
-%if mmsize == 16 ; sse2
-    pshuflw        %1, %1, 0x0
-    punpcklqdq     %1, %1
-%elifidn %3, mmx
     punpcklwd      %1, %1
     punpckldq      %1, %1
-%else ; mmxext
+%endmacro
+
+%macro SPLATB_REG_MMXEXT 2-3
+    movd           %1, %2
+    punpcklbw      %1, %1
     pshufw         %1, %1, 0x0
-%endif
-%endif
+%endmacro
+
+%macro SPLATB_REG_SSE2 2-3
+    movd           %1, %2
+    punpcklbw      %1, %1
+    pshuflw        %1, %1, 0x0
+    punpcklqdq     %1, %1
+%endmacro
+
+%macro SPLATB_REG_SSSE3 3
+    movd           %1, %2
+    pshufb         %1, %3
 %endmacro
 
 %macro SIMPLE_LOOPFILTER 3
 cglobal vp8_%2_loop_filter_simple_%1, 3, %3
-%ifidn %2, h
-    mov            r5, rsp          ; backup stack pointer
-    and           rsp, ~(mmsize-1)  ; align stack
-%endif
 %if mmsize == 8 ; mmx/mmxext
     mov            r3, 2
 %endif
-%ifidn %1, ssse3
+%ifnidn %1, sse2
+%if mmsize == 16
     pxor           m0, m0
 %endif
-    SPLATB_REG     m7, r2, %1, m0   ; splat "flim" into register
+%endif
+    SPLATB_REG     m7, r2, m0       ; splat "flim" into register
 
     ; set up indexes to address 4 rows
     mov            r2, r1
     neg            r1
 %ifidn %2, h
     lea            r0, [r0+4*r2-2]
-    sub           rsp, mmsize*2     ; (aligned) storage space for saving p1/q1
 %endif
 
 %if mmsize == 8 ; mmx / mmxext
@@ -1418,9 +1491,6 @@ cglobal vp8_%2_loop_filter_simple_%1, 3, %3
     READ_16x4_INTERLEAVED 0, 1, 2, 3, 4, 5, 6, r0, r4, r1, r2, r3
 %endif
     TRANSPOSE4x4W         0, 1, 2, 3, 4
-
-    mova        [rsp], m0           ; store p1
-    mova [rsp+mmsize], m3           ; store q1
 %endif
 
     ; simple_limit
@@ -1491,17 +1561,21 @@ cglobal vp8_%2_loop_filter_simple_%1, 3, %3
     mova         [r0], m4
     mova      [r0+r1], m6
 %else ; h
-    mova           m0, [rsp]        ; p1
-    SWAP            2, 4            ; p0
-    SWAP            1, 6            ; q0
-    mova           m3, [rsp+mmsize] ; q1
+    inc           r0
+    SBUTTERFLY    bw, 6, 4, 0
 
-    TRANSPOSE4x4B  0, 1, 2, 3, 4
 %if mmsize == 16 ; sse2
-    add            r3, r1           ; change from r4*8*stride to r0+8*stride
-    WRITE_4x4D 0, 1, 2, 3, r0, r4, r3, r1, r2, 16
+%ifidn %1, sse4
+    inc            r4
+%endif
+    WRITE_8W       m6, r4, r0, r1, r2
+    lea            r4, [r3+r1+1]
+%ifidn %1, sse4
+    inc            r3
+%endif
+    WRITE_8W       m4, r3, r4, r1, r2
 %else ; mmx/mmxext
-    WRITE_4x2D 0, 1, 2, 3, r0, r4, r1, r2
+    WRITE_2x4W     m6, m4, r4, r0, r1, r2
 %endif
 %endif
 
@@ -1510,34 +1584,33 @@ cglobal vp8_%2_loop_filter_simple_%1, 3, %3
 %ifidn %2, v
     add            r0, 8            ; advance 8 cols = pixels
 %else ; h
-    lea            r0, [r0+r2*8]    ; advance 8 rows = lines
+    lea            r0, [r0+r2*8-1]  ; advance 8 rows = lines
 %endif
     dec            r3
     jg .next8px
-%ifidn %2, v
     REP_RET
-%else ; h
-    mov           rsp, r5           ; restore stack pointer
-    RET
-%endif
 %else ; sse2
-%ifidn %2, h
-    mov           rsp, r5           ; restore stack pointer
-%endif
     RET
 %endif
 %endmacro
 
 INIT_MMX
+%define SPLATB_REG SPLATB_REG_MMX
 SIMPLE_LOOPFILTER mmx,    v, 4
-SIMPLE_LOOPFILTER mmx,    h, 6
+SIMPLE_LOOPFILTER mmx,    h, 5
+%define SPLATB_REG SPLATB_REG_MMXEXT
 SIMPLE_LOOPFILTER mmxext, v, 4
-SIMPLE_LOOPFILTER mmxext, h, 6
+SIMPLE_LOOPFILTER mmxext, h, 5
 INIT_XMM
+%define SPLATB_REG SPLATB_REG_SSE2
+%define WRITE_8W   WRITE_8W_SSE2
 SIMPLE_LOOPFILTER sse2,   v, 3
-SIMPLE_LOOPFILTER sse2,   h, 6
+SIMPLE_LOOPFILTER sse2,   h, 5
+%define SPLATB_REG SPLATB_REG_SSSE3
 SIMPLE_LOOPFILTER ssse3,  v, 3
-SIMPLE_LOOPFILTER ssse3,  h, 6
+SIMPLE_LOOPFILTER ssse3,  h, 5
+%define WRITE_8W   WRITE_8W_SSE4
+SIMPLE_LOOPFILTER sse4,   h, 5
 
 ;-----------------------------------------------------------------------------
 ; void vp8_h/v_loop_filter<size>_inner_<opt>(uint8_t *dst, [uint8_t *v,] int stride,
@@ -1573,15 +1646,17 @@ cglobal vp8_%2_loop_filter16y_inner_%1, 5, %3, %5
 %define stack_reg   hev_thr_reg
 %endif
 
-%ifidn %1, ssse3
+%ifnidn %1, sse2
+%if mmsize == 16
     pxor             m7, m7
 %endif
+%endif
 
 %ifndef m8 ; mmx/mmxext or sse2 on x86-32
     ; splat function arguments
-    SPLATB_REG       m0, E_reg, %1, m7 ; E
-    SPLATB_REG       m1, I_reg, %1, m7 ; I
-    SPLATB_REG       m2, hev_thr_reg, %1, m7 ; hev_thresh
+    SPLATB_REG       m0, E_reg, m7   ; E
+    SPLATB_REG       m1, I_reg, m7   ; I
+    SPLATB_REG       m2, hev_thr_reg, m7 ; hev_thresh
 
     ; align stack
     mov       stack_reg, rsp         ; backup stack pointer
@@ -1614,9 +1689,9 @@ cglobal vp8_%2_loop_filter16y_inner_%1, 5, %3, %5
 %define q0backup m8
 
     ; splat function arguments
-    SPLATB_REG   flim_E, E_reg, %1, m7 ; E
-    SPLATB_REG   flim_I, I_reg, %1, m7 ; I
-    SPLATB_REG  hev_thr, hev_thr_reg, %1, m7 ; hev_thresh
+    SPLATB_REG   flim_E, E_reg, m7   ; E
+    SPLATB_REG   flim_I, I_reg, m7   ; I
+    SPLATB_REG  hev_thr, hev_thr_reg, m7 ; hev_thresh
 %endif
 
 %if mmsize == 8 && %4 == 16 ; mmx/mmxext
@@ -2028,17 +2103,20 @@ cglobal vp8_%2_loop_filter16y_inner_%1, 5, %3, %5
 %endmacro
 
 INIT_MMX
+%define SPLATB_REG SPLATB_REG_MMX
 INNER_LOOPFILTER mmx,    v, 6, 16, 0
 INNER_LOOPFILTER mmx,    h, 6, 16, 0
-INNER_LOOPFILTER mmxext, v, 6, 16, 0
-INNER_LOOPFILTER mmxext, h, 6, 16, 0
-
 INNER_LOOPFILTER mmx,    v, 6,  8, 0
 INNER_LOOPFILTER mmx,    h, 6,  8, 0
+
+%define SPLATB_REG SPLATB_REG_MMXEXT
+INNER_LOOPFILTER mmxext, v, 6, 16, 0
+INNER_LOOPFILTER mmxext, h, 6, 16, 0
 INNER_LOOPFILTER mmxext, v, 6,  8, 0
 INNER_LOOPFILTER mmxext, h, 6,  8, 0
 
 INIT_XMM
+%define SPLATB_REG SPLATB_REG_SSE2
 INNER_LOOPFILTER sse2,   v, 5, 16, 13
 %ifdef m8
 INNER_LOOPFILTER sse2,   h, 5, 16, 13
@@ -2048,6 +2126,7 @@ INNER_LOOPFILTER sse2,   h, 6, 16, 13
 INNER_LOOPFILTER sse2,   v, 6,  8, 13
 INNER_LOOPFILTER sse2,   h, 6,  8, 13
 
+%define SPLATB_REG SPLATB_REG_SSSE3
 INNER_LOOPFILTER ssse3,  v, 5, 16, 13
 %ifdef m8
 INNER_LOOPFILTER ssse3,  h, 5, 16, 13
@@ -2062,67 +2141,6 @@ INNER_LOOPFILTER ssse3,  h, 6,  8, 13
 ;                                            int flimE, int flimI, int hev_thr);
 ;-----------------------------------------------------------------------------
 
-; write 4 or 8 words in the mmx/xmm registers as 8 lines
-; 1 and 2 are the registers to write, this can be the same (for SSE2)
-; for pre-SSE4:
-; 3 is a general-purpose register that we will clobber
-; for SSE4:
-; 3 is a pointer to the destination's 5th line
-; 4 is a pointer to the destination's 4th line
-; 5/6 is -stride and +stride
-; 7 is optimization string
-%macro WRITE_8W 7
-%ifidn %7, sse4
-    pextrw    [%4+%5*4], %1, 0
-    pextrw    [%3+%5*4], %1, 1
-    pextrw    [%4+%5*2], %1, 2
-    pextrw    [%4+%5  ], %1, 3
-    pextrw    [%4     ], %1, 4
-    pextrw    [%3     ], %1, 5
-    pextrw    [%3+%6  ], %1, 6
-    pextrw    [%3+%6*2], %1, 7
-%else
-    movd             %3, %1
-%if mmsize == 8
-    punpckhdq        %1, %1
-%else
-    psrldq           %1, 4
-%endif
-    mov       [%4+%5*4], %3w
-    shr              %3, 16
-    add              %4, %6
-    mov       [%4+%5*4], %3w
-
-    movd             %3, %1
-%if mmsize == 16
-    psrldq           %1, 4
-%endif
-    add              %4, %5
-    mov       [%4+%5*2], %3w
-    shr              %3, 16
-    mov       [%4+%5  ], %3w
-
-    movd             %3, %2
-%if mmsize == 8
-    punpckhdq        %2, %2
-%else
-    psrldq           %2, 4
-%endif
-    mov       [%4     ], %3w
-    shr              %3, 16
-    mov       [%4+%6  ], %3w
-
-    movd             %3, %2
-    add              %4, %6
-    mov       [%4+%6  ], %3w
-    shr              %3, 16
-    mov       [%4+%6*2], %3w
-%if mmsize == 8
-    add              %4, %5
-%endif
-%endif
-%endmacro
-
 %macro MBEDGE_LOOPFILTER 5
 %if %4 == 8 ; chroma
 cglobal vp8_%2_loop_filter8uv_mbedge_%1, 6, %3, %5
@@ -2152,24 +2170,35 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
 %define stack_reg   hev_thr_reg
 %endif
 
-%ifidn %1, ssse3
+%define ssse3_or_higher 0
+%ifnidn %1, sse2
+%if mmsize == 16
+%define ssse3_or_higher 1
+%endif
+%endif
+
+%if ssse3_or_higher
     pxor             m7, m7
 %endif
 
 %ifndef m8 ; mmx/mmxext or sse2 on x86-32
     ; splat function arguments
-    SPLATB_REG       m0, E_reg, %1, m7 ; E
-    SPLATB_REG       m1, I_reg, %1, m7 ; I
-    SPLATB_REG       m2, hev_thr_reg, %1, m7 ; hev_thresh
+    SPLATB_REG       m0, E_reg, m7   ; E
+    SPLATB_REG       m1, I_reg, m7   ; I
+    SPLATB_REG       m2, hev_thr_reg, m7 ; hev_thresh
 
     ; align stack
     mov       stack_reg, rsp         ; backup stack pointer
     and             rsp, ~(mmsize-1) ; align stack
+%if mmsize == 16
+    sub             rsp, mmsize * 7
+%else
     sub             rsp, mmsize * 8  ; stack layout: [0]=E, [1]=I, [2]=hev_thr
                                      ;               [3]=hev() result
                                      ;               [4]=filter tmp result
                                      ;               [5]/[6] = p2/q2 backup
                                      ;               [7]=lim_res sign result
+%endif
 
 %define flim_E   [rsp]
 %define flim_I   [rsp+mmsize]
@@ -2180,7 +2209,11 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
 %define q0backup [rsp+mmsize*4]
 %define p2backup [rsp+mmsize*5]
 %define q2backup [rsp+mmsize*6]
+%if mmsize == 16
+%define lim_sign [rsp]
+%else
 %define lim_sign [rsp+mmsize*7]
+%endif
 
     mova         flim_E, m0
     mova         flim_I, m1
@@ -2197,12 +2230,12 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
 %define q0backup m8
 %define p2backup m13
 %define q2backup m14
-%define lim_sign m15
+%define lim_sign m9
 
     ; splat function arguments
-    SPLATB_REG   flim_E, E_reg, %1, m7 ; E
-    SPLATB_REG   flim_I, I_reg, %1, m7 ; I
-    SPLATB_REG  hev_thr, hev_thr_reg, %1, m7 ; hev_thresh
+    SPLATB_REG   flim_E, E_reg, m7   ; E
+    SPLATB_REG   flim_I, I_reg, m7   ; I
+    SPLATB_REG  hev_thr, hev_thr_reg, m7 ; hev_thresh
 %endif
 
 %if mmsize == 8 && %4 == 16 ; mmx/mmxext
@@ -2543,7 +2576,11 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
     paddusb          m4, m1          ; q0-f1
 
     ; filter_mbedge (m2-m5 = p1-q1; lim_res carries w)
+%if ssse3_or_higher
+    mova             m7, [pb_1]
+%else
     mova             m7, [pw_63]
+%endif
 %ifdef m8
     SWAP              1, 8
 %else
@@ -2552,15 +2589,40 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
     pxor             m0, m0
     mova             m6, m1
     pcmpgtb          m0, m1         ; which are negative
+%if ssse3_or_higher
+    punpcklbw        m6, m7         ; interleave with "1" for rounding
+    punpckhbw        m1, m7
+%else
     punpcklbw        m6, m0         ; signed byte->word
     punpckhbw        m1, m0
+%endif
     mova       lim_sign, m0
+%if ssse3_or_higher
+    mova             m7, [pb_27_63]
+%ifndef m8
+    mova        lim_res, m1
+%endif
+%ifdef m10
+    SWAP              0, 10         ; don't lose lim_sign copy
+%endif
+    mova             m0, m7
+    pmaddubsw        m7, m6
+    SWAP              6, 7
+    pmaddubsw        m0, m1
+    SWAP              1, 0
+%ifdef m10
+    SWAP              0, 10
+%else
+    mova             m0, lim_sign
+%endif
+%else
     mova       mask_res, m6         ; backup for later in filter
     mova        lim_res, m1
     pmullw          m6, [pw_27]
     pmullw          m1, [pw_27]
     paddw           m6, m7
     paddw           m1, m7
+%endif
     psraw           m6, 7
     psraw           m1, 7
     packsswb        m6, m1          ; a0
@@ -2568,18 +2630,39 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
     psubb           m1, m6
     pand            m1, m0          ; -a0
     pandn           m0, m6          ; +a0
+%if ssse3_or_higher
+    mova            m6, [pb_18_63]  ; pipelining
+%endif
     psubusb         m3, m1
     paddusb         m4, m1
     paddusb         m3, m0          ; p0+a0
     psubusb         m4, m0          ; q0-a0
 
-    mova            m6, mask_res
+%if ssse3_or_higher
+    SWAP             6, 7
+%ifdef m10
+    SWAP             1, 10
+%else
     mova            m1, lim_res
+%endif
+    mova            m0, m7
+    pmaddubsw       m7, m6
+    SWAP             6, 7
+    pmaddubsw       m0, m1
+    SWAP             1, 0
+%ifdef m10
+    SWAP             0, 10
+%endif
     mova            m0, lim_sign
+%else
+    mova            m6, mask_res
+    mova            m1, lim_res
     pmullw          m6, [pw_18]
     pmullw          m1, [pw_18]
     paddw           m6, m7
     paddw           m1, m7
+%endif
+    mova            m0, lim_sign
     psraw           m6, 7
     psraw           m1, 7
     packsswb        m6, m1          ; a1
@@ -2587,11 +2670,27 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
     psubb           m1, m6
     pand            m1, m0          ; -a1
     pandn           m0, m6          ; +a1
+%if ssse3_or_higher
+    mova            m6, [pb_9_63]
+%endif
     psubusb         m2, m1
     paddusb         m5, m1
     paddusb         m2, m0          ; p1+a1
     psubusb         m5, m0          ; q1-a1
 
+%if ssse3_or_higher
+    SWAP             6, 7
+%ifdef m10
+    SWAP             1, 10
+%else
+    mova            m1, lim_res
+%endif
+    mova            m0, m7
+    pmaddubsw       m7, m6
+    SWAP             6, 7
+    pmaddubsw       m0, m1
+    SWAP             1, 0
+%else
 %ifdef m8
     SWAP             6, 12
     SWAP             1, 8
@@ -2603,8 +2702,9 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
     pmullw          m1, [pw_9]
     paddw           m6, m7
     paddw           m1, m7
-%ifdef m15
-    SWAP             7, 15
+%endif
+%ifdef m9
+    SWAP             7, 9
 %else
     mova            m7, lim_sign
 %endif
@@ -2656,17 +2756,20 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
 %if mmsize == 8 ; mmx/mmxext (h)
     WRITE_4x2D        1, 2, 3, 4, dst_reg, dst2_reg, mstride_reg, stride_reg
     add         dst_reg, 4
-    WRITE_8W         m5, m6, dst2_reg, dst_reg, mstride_reg, stride_reg, %4
+    WRITE_2x4W       m5, m6, dst2_reg, dst_reg, mstride_reg, stride_reg
 %else ; sse2 (h)
     lea        dst8_reg, [dst8_reg+mstride_reg+1]
     WRITE_4x4D        1, 2, 3, 4, dst_reg, dst2_reg, dst8_reg, mstride_reg, stride_reg, %4
     lea         dst_reg, [dst2_reg+mstride_reg+4]
     lea        dst8_reg, [dst8_reg+mstride_reg+4]
-    WRITE_8W         m5, m5, dst2_reg, dst_reg,  mstride_reg, stride_reg, %2
-%ifidn %2, sse4
-    lea         dst_reg, [dst8_reg+ stride_reg]
+%ifidn %1, sse4
+    add        dst2_reg, 4
+%endif
+    WRITE_8W         m5, dst2_reg, dst_reg,  mstride_reg, stride_reg
+%ifidn %1, sse4
+    lea        dst2_reg, [dst8_reg+ stride_reg]
 %endif
-    WRITE_8W         m6, m6, dst2_reg, dst8_reg, mstride_reg, stride_reg, %2
+    WRITE_8W         m6, dst2_reg, dst8_reg, mstride_reg, stride_reg
 %endif
 %endif
 
@@ -2696,38 +2799,44 @@ cglobal vp8_%2_loop_filter16y_mbedge_%1, 5, %3, %5
 %endmacro
 
 INIT_MMX
+%define SPLATB_REG SPLATB_REG_MMX
 MBEDGE_LOOPFILTER mmx,    v, 6, 16, 0
 MBEDGE_LOOPFILTER mmx,    h, 6, 16, 0
-MBEDGE_LOOPFILTER mmxext, v, 6, 16, 0
-MBEDGE_LOOPFILTER mmxext, h, 6, 16, 0
-
 MBEDGE_LOOPFILTER mmx,    v, 6,  8, 0
 MBEDGE_LOOPFILTER mmx,    h, 6,  8, 0
+
+%define SPLATB_REG SPLATB_REG_MMXEXT
+MBEDGE_LOOPFILTER mmxext, v, 6, 16, 0
+MBEDGE_LOOPFILTER mmxext, h, 6, 16, 0
 MBEDGE_LOOPFILTER mmxext, v, 6,  8, 0
 MBEDGE_LOOPFILTER mmxext, h, 6,  8, 0
 
 INIT_XMM
-MBEDGE_LOOPFILTER sse2,   v, 5, 16, 16
+%define SPLATB_REG SPLATB_REG_SSE2
+%define WRITE_8W   WRITE_8W_SSE2
+MBEDGE_LOOPFILTER sse2,   v, 5, 16, 15
 %ifdef m8
-MBEDGE_LOOPFILTER sse2,   h, 5, 16, 16
+MBEDGE_LOOPFILTER sse2,   h, 5, 16, 15
 %else
-MBEDGE_LOOPFILTER sse2,   h, 6, 16, 16
+MBEDGE_LOOPFILTER sse2,   h, 6, 16, 15
 %endif
-MBEDGE_LOOPFILTER sse2,   v, 6,  8, 16
-MBEDGE_LOOPFILTER sse2,   h, 6,  8, 16
+MBEDGE_LOOPFILTER sse2,   v, 6,  8, 15
+MBEDGE_LOOPFILTER sse2,   h, 6,  8, 15
 
-MBEDGE_LOOPFILTER ssse3,  v, 5, 16, 16
+%define SPLATB_REG SPLATB_REG_SSSE3
+MBEDGE_LOOPFILTER ssse3,  v, 5, 16, 15
 %ifdef m8
-MBEDGE_LOOPFILTER ssse3,  h, 5, 16, 16
+MBEDGE_LOOPFILTER ssse3,  h, 5, 16, 15
 %else
-MBEDGE_LOOPFILTER ssse3,  h, 6, 16, 16
+MBEDGE_LOOPFILTER ssse3,  h, 6, 16, 15
 %endif
-MBEDGE_LOOPFILTER ssse3,  v, 6,  8, 16
-MBEDGE_LOOPFILTER ssse3,  h, 6,  8, 16
+MBEDGE_LOOPFILTER ssse3,  v, 6,  8, 15
+MBEDGE_LOOPFILTER ssse3,  h, 6,  8, 15
 
+%define WRITE_8W   WRITE_8W_SSE4
 %ifdef m8
-MBEDGE_LOOPFILTER sse4,   h, 5, 16, 16
+MBEDGE_LOOPFILTER sse4,   h, 5, 16, 15
 %else
-MBEDGE_LOOPFILTER sse4,   h, 6, 16, 16
+MBEDGE_LOOPFILTER sse4,   h, 6, 16, 15
 %endif
-MBEDGE_LOOPFILTER sse4,   h, 6,  8, 16
+MBEDGE_LOOPFILTER sse4,   h, 6,  8, 15
diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/avcore.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/avcore.h
new file mode 100644
index 000000000..5a2a7c73f
--- /dev/null
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/avcore.h
@@ -0,0 +1,58 @@
+/*

+ * This file is part of FFmpeg.

+ *

+ * FFmpeg is free software; you can redistribute it and/or

+ * modify it under the terms of the GNU Lesser General Public

+ * License as published by the Free Software Foundation; either

+ * version 2.1 of the License, or (at your option) any later version.

+ *

+ * FFmpeg is distributed in the hope that it will be useful,

+ * but WITHOUT ANY WARRANTY; without even the implied warranty of

+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

+ * Lesser General Public License for more details.

+ *

+ * You should have received a copy of the GNU Lesser General Public

+ * License along with FFmpeg; if not, write to the Free Software

+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

+ */

+

+#ifndef AVCORE_AVCORE_H

+#define AVCORE_AVCORE_H

+

+/**

+ * @file

+ * shared media utilities for the libav* libraries

+ */

+

+#include <libavutil/avutil.h>

+

+#define LIBAVCORE_VERSION_MAJOR  0

+#define LIBAVCORE_VERSION_MINOR  2

+#define LIBAVCORE_VERSION_MICRO  0

+

+#define LIBAVCORE_VERSION_INT   AV_VERSION_INT(LIBAVCORE_VERSION_MAJOR, \

+                                               LIBAVCORE_VERSION_MINOR, \

+                                               LIBAVCORE_VERSION_MICRO)

+#define LIBAVCORE_VERSION       AV_VERSION(LIBAVCORE_VERSION_MAJOR,     \

+                                           LIBAVCORE_VERSION_MINOR,     \

+                                           LIBAVCORE_VERSION_MICRO)

+#define LIBAVCORE_BUILD         LIBAVCORE_VERSION_INT

+

+#define LIBAVCORE_IDENT         "Lavcore" AV_STRINGIFY(LIBAVCORE_VERSION)

+

+/**

+ * Return the LIBAVCORE_VERSION_INT constant.

+ */

+unsigned avcore_version(void);

+

+/**

+ * Return the libavcore build-time configuration.

+ */

+const char *avcore_configuration(void);

+

+/**

+ * Return the libavcore license.

+ */

+const char *avcore_license(void);

+

+#endif /* AVCORE_AVCORE_H */

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/avcore_utils.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/avcore_utils.c
new file mode 100644
index 000000000..badb9ee36
--- /dev/null
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/avcore_utils.c
@@ -0,0 +1,43 @@
+/*

+ * This file is part of FFmpeg.

+ *

+ * FFmpeg is free software; you can redistribute it and/or

+ * modify it under the terms of the GNU Lesser General Public

+ * License as published by the Free Software Foundation; either

+ * version 2.1 of the License, or (at your option) any later version.

+ *

+ * FFmpeg is distributed in the hope that it will be useful,

+ * but WITHOUT ANY WARRANTY; without even the implied warranty of

+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

+ * Lesser General Public License for more details.

+ *

+ * You should have received a copy of the GNU Lesser General Public

+ * License along with FFmpeg; if not, write to the Free Software

+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

+ */

+

+#include "config.h"

+#include "avcore.h"

+

+/**

+ * @file

+ * various utility functions

+ */

+

+unsigned avcore_version(void)

+{

+    return LIBAVCORE_VERSION_INT;

+}

+

+#if 0

+const char *avcore_configuration(void)

+{

+    return FFMPEG_CONFIGURATION;

+}

+

+const char *avcore_license(void)

+{

+#define LICENSE_PREFIX "libavcore license: "

+    return LICENSE_PREFIX FFMPEG_LICENSE + sizeof(LICENSE_PREFIX) - 1;

+}

+#endif

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/imgutils.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/imgutils.c
new file mode 100644
index 000000000..b14dd32a8
--- /dev/null
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/imgutils.c
@@ -0,0 +1,97 @@
+/*

+ * This file is part of FFmpeg.

+ *

+ * FFmpeg is free software; you can redistribute it and/or

+ * modify it under the terms of the GNU Lesser General Public

+ * License as published by the Free Software Foundation; either

+ * version 2.1 of the License, or (at your option) any later version.

+ *

+ * FFmpeg is distributed in the hope that it will be useful,

+ * but WITHOUT ANY WARRANTY; without even the implied warranty of

+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

+ * Lesser General Public License for more details.

+ *

+ * You should have received a copy of the GNU Lesser General Public

+ * License along with FFmpeg; if not, write to the Free Software

+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

+ */

+

+/**

+ * @file

+ * misc image utilities

+ */

+

+#include "imgutils.h"

+#include "libavutil/pixdesc.h"

+

+int av_fill_image_linesizes(int linesizes[4], enum PixelFormat pix_fmt, int width)

+{

+    int i;

+    const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[pix_fmt];

+    int max_step     [4];       /* max pixel step for each plane */

+    int max_step_comp[4];       /* the component for each plane which has the max pixel step */

+

+    memset(linesizes, 0, 4*sizeof(linesizes[0]));

+

+    if (desc->flags & PIX_FMT_HWACCEL)

+        return AVERROR(EINVAL);

+

+    if (desc->flags & PIX_FMT_BITSTREAM) {

+        linesizes[0] = (width * (desc->comp[0].step_minus1+1) + 7) >> 3;

+        return 0;

+    }

+

+    memset(max_step     , 0, sizeof(max_step     ));

+    memset(max_step_comp, 0, sizeof(max_step_comp));

+    for (i = 0; i < 4; i++) {

+        const AVComponentDescriptor *comp = &(desc->comp[i]);

+        if ((comp->step_minus1+1) > max_step[comp->plane]) {

+            max_step     [comp->plane] = comp->step_minus1+1;

+            max_step_comp[comp->plane] = i;

+        }

+    }

+

+    for (i = 0; i < 4; i++) {

+        int s = (max_step_comp[i] == 1 || max_step_comp[i] == 2) ? desc->log2_chroma_w : 0;

+        linesizes[i] = max_step[i] * (((width + (1 << s) - 1)) >> s);

+    }

+

+    return 0;

+}

+

+int av_fill_image_pointers(uint8_t *data[4], enum PixelFormat pix_fmt, int height,

+                           uint8_t *ptr, const int linesizes[4])

+{

+    int i, total_size, size[4], has_plane[4];

+

+    const AVPixFmtDescriptor *desc = &av_pix_fmt_descriptors[pix_fmt];

+    memset(data     , 0, sizeof(data[0])*4);

+    memset(size     , 0, sizeof(size));

+    memset(has_plane, 0, sizeof(has_plane));

+

+    if (desc->flags & PIX_FMT_HWACCEL)

+        return AVERROR(EINVAL);

+

+    data[0] = ptr;

+    size[0] = linesizes[0] * height;

+

+    if (desc->flags & PIX_FMT_PAL) {

+        size[0] = (size[0] + 3) & ~3;

+        data[1] = ptr + size[0]; /* palette is stored here as 256 32 bits words */

+        return size[0] + 256 * 4;

+    }

+

+    for (i = 0; i < 4; i++)

+        has_plane[desc->comp[i].plane] = 1;

+

+    total_size = size[0];

+    for (i = 1; has_plane[i] && i < 4; i++) {

+        int h, s = (i == 1 || i == 2) ? desc->log2_chroma_h : 0;

+        data[i] = data[i-1] + size[i-1];

+        h = (height + (1 << s) - 1) >> s;

+        size[i] = h * linesizes[i];

+        total_size += size[i];

+    }

+

+    return total_size;

+}

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/imgutils.h b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/imgutils.h
new file mode 100644
index 000000000..b8024de8e
--- /dev/null
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavcore/imgutils.h
@@ -0,0 +1,53 @@
+/*

+ * This file is part of FFmpeg.

+ *

+ * FFmpeg is free software; you can redistribute it and/or

+ * modify it under the terms of the GNU Lesser General Public

+ * License as published by the Free Software Foundation; either

+ * version 2.1 of the License, or (at your option) any later version.

+ *

+ * FFmpeg is distributed in the hope that it will be useful,

+ * but WITHOUT ANY WARRANTY; without even the implied warranty of

+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU

+ * Lesser General Public License for more details.

+ *

+ * You should have received a copy of the GNU Lesser General Public

+ * License along with FFmpeg; if not, write to the Free Software

+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA

+ */

+

+#ifndef AVCORE_IMGUTILS_H

+#define AVCORE_IMGUTILS_H

+

+/**

+ * @file

+ * misc image utilities

+ */

+

+#include "libavutil/pixfmt.h"

+#include "avcore.h"

+

+/**

+ * Fill plane linesizes for an image with pixel format pix_fmt and

+ * width width.

+ *

+ * @param linesizes array to be filled with the linesize for each plane

+ * @return >= 0 in case of success, a negative error code otherwise

+ */

+int av_fill_image_linesizes(int linesizes[4], enum PixelFormat pix_fmt, int width);

+

+/**

+ * Fill plane data pointers for an image with pixel format pix_fmt and

+ * height height.

+ *

+ * @param data pointers array to be filled with the pointer for each image plane

+ * @param ptr the pointer to a buffer which will contain the image

+ * @param linesizes[4] the array containing the linesize for each

+ * plane, should be filled by av_fill_image_linesizes()

+ * @return the size in bytes required for the image buffer, a negative

+ * error code in case of failure

+ */

+int av_fill_image_pointers(uint8_t *data[4], enum PixelFormat pix_fmt, int height,

+                           uint8_t *ptr, const int linesizes[4]);

+

+#endif /* AVCORE_IMGUTILS_H */

diff --git a/src/filters/transform/MPCVideoDec/ffmpeg/libavutil/log.c b/src/filters/transform/MPCVideoDec/ffmpeg/libavutil/log.c
index b40d0e591..b9f4e902a 100644
--- a/src/filters/transform/MPCVideoDec/ffmpeg/libavutil/log.c
+++ b/src/filters/transform/MPCVideoDec/ffmpeg/libavutil/log.c
@@ -63,6 +63,7 @@ void av_log_default_callback(void* ptr, int level, const char* fmt, va_list vl)
     print_prefix= line[strlen(line)-1] == '\n';

     if(print_prefix && !strcmp(line, prev)){

         count++;

+        fprintf(stderr, "    Last message repeated %d times\r", count);

         return;

     }

     if(count>0){