X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Fdsputil.c;h=80efd07825bf026558abc0f16052592002ddde17;hb=6bf6a9301b97983367e19cefc8da9153a88a244e;hp=49bd1c4ca11bcfa38996a31236648e27460eba11;hpb=bb270c0896b39e1ae9277355e3c120ed3feb64a3;p=ffmpeg

diff --git a/libavcodec/dsputil.c b/libavcodec/dsputil.c
index 49bd1c4ca11..80efd07825b 100644
--- a/libavcodec/dsputil.c
+++ b/libavcodec/dsputil.c
@@ -3,21 +3,23 @@
  * Copyright (c) 2000, 2001 Fabrice Bellard.
  * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
  *
- * This library is free software; you can redistribute it and/or
+ * gmc & q-pel & 32/64 bit based MC by Michael Niedermayer <michaelni@gmx.at>
+ *
+ * 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 of the License, or (at your option) any later version.
+ * version 2.1 of the License, or (at your option) any later version.
  *
- * This library is distributed in the hope that it will be useful,
+ * 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 this library; if not, write to the Free Software
- * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
- *
- * gmc & q-pel & 32/64 bit based MC by Michael Niedermayer <michaelni@gmx.at>
+ * License along with FFmpeg; if not, write to the Free Software
+ * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
  */
 
 /**
@@ -27,15 +29,30 @@
 
 #include "avcodec.h"
 #include "dsputil.h"
-#include "mpegvideo.h"
 #include "simple_idct.h"
 #include "faandct.h"
+#include "faanidct.h"
+#include "h263.h"
+#include "snow.h"
 
 /* snow.c */
 void ff_spatial_dwt(int *buffer, int width, int height, int stride, int type, int decomposition_count);
 
-uint8_t cropTbl[256 + 2 * MAX_NEG_CROP] = {0, };
-uint32_t squareTbl[512] = {0, };
+/* vorbis.c */
+void vorbis_inverse_coupling(float *mag, float *ang, int blocksize);
+
+/* flacenc.c */
+void ff_flac_compute_autocorr(const int32_t *data, int len, int lag, double *autoc);
+
+/* pngdec.c */
+void ff_add_png_paeth_prediction(uint8_t *dst, uint8_t *src, uint8_t *top, int w, int bpp);
+
+uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP] = {0, };
+uint32_t ff_squareTbl[512] = {0, };
+
+// 0x7f7f7f7f or 0x7f7f7f7f7f7f7f7f or whatever, depending on the cpu's native arithmetic size
+#define pb_7f (~0UL/255 * 0x7f)
+#define pb_80 (~0UL/255 * 0x80)
 
 const uint8_t ff_zigzag_direct[64] = {
     0,   1,  8, 16,  9,  2,  3, 10,
@@ -62,7 +79,7 @@ const uint8_t ff_zigzag248_direct[64] = {
 };
 
 /* not permutated inverse zigzag_direct + 1 for MMX quantizer */
-uint16_t __align8 inv_zigzag_direct16[64] = {0, };
+DECLARE_ALIGNED_8(uint16_t, inv_zigzag_direct16[64]) = {0, };
 
 const uint8_t ff_alternate_horizontal_scan[64] = {
     0,  1,   2,  3,  8,  9, 16, 17,
@@ -87,7 +104,7 @@ const uint8_t ff_alternate_vertical_scan[64] = {
 };
 
 /* a*inverse[b]>>32 == a/b for all 0<=a<=65536 && 2<=b<=255 */
-const uint32_t inverse[256]={
+const uint32_t ff_inverse[256]={
          0, 4294967295U,2147483648U,1431655766, 1073741824,  858993460,  715827883,  613566757,
  536870912,  477218589,  429496730,  390451573,  357913942,  330382100,  306783379,  286331154,
  268435456,  252645136,  238609295,  226050911,  214748365,  204522253,  195225787,  186737709,
@@ -134,6 +151,32 @@ static const uint8_t simple_mmx_permutation[64]={
         0x32, 0x3A, 0x36, 0x3B, 0x33, 0x3E, 0x37, 0x3F,
 };
 
+static const uint8_t idct_sse2_row_perm[8] = {0, 4, 1, 5, 2, 6, 3, 7};
+
+void ff_init_scantable(uint8_t *permutation, ScanTable *st, const uint8_t *src_scantable){
+    int i;
+    int end;
+
+    st->scantable= src_scantable;
+
+    for(i=0; i<64; i++){
+        int j;
+        j = src_scantable[i];
+        st->permutated[i] = permutation[j];
+#ifdef ARCH_POWERPC
+        st->inverse[j] = i;
+#endif
+    }
+
+    end=-1;
+    for(i=0; i<64; i++){
+        int j;
+        j = st->permutated[i];
+        if(j>end) end=j;
+        st->raster_end[i]= end;
+    }
+}
+
 static int pix_sum_c(uint8_t * pix, int line_size)
 {
     int s, i, j;
@@ -159,7 +202,7 @@ static int pix_sum_c(uint8_t * pix, int line_size)
 static int pix_norm1_c(uint8_t * pix, int line_size)
 {
     int s, i, j;
-    uint32_t *sq = squareTbl + 256;
+    uint32_t *sq = ff_squareTbl + 256;
 
     s = 0;
     for (i = 0; i < 16; i++) {
@@ -204,7 +247,7 @@ static int pix_norm1_c(uint8_t * pix, int line_size)
     return s;
 }
 
-static void bswap_buf(uint32_t *dst, uint32_t *src, int w){
+static void bswap_buf(uint32_t *dst, const uint32_t *src, int w){
     int i;
 
     for(i=0; i+8<=w; i+=8){
@@ -225,7 +268,7 @@ static void bswap_buf(uint32_t *dst, uint32_t *src, int w){
 static int sse4_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h)
 {
     int s, i;
-    uint32_t *sq = squareTbl + 256;
+    uint32_t *sq = ff_squareTbl + 256;
 
     s = 0;
     for (i = 0; i < h; i++) {
@@ -242,7 +285,7 @@ static int sse4_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h)
 static int sse8_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h)
 {
     int s, i;
-    uint32_t *sq = squareTbl + 256;
+    uint32_t *sq = ff_squareTbl + 256;
 
     s = 0;
     for (i = 0; i < h; i++) {
@@ -263,7 +306,7 @@ static int sse8_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h)
 static int sse16_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
 {
     int s, i;
-    uint32_t *sq = squareTbl + 256;
+    uint32_t *sq = ff_squareTbl + 256;
 
     s = 0;
     for (i = 0; i < h; i++) {
@@ -291,35 +334,34 @@ static int sse16_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
 }
 
 
+#ifdef CONFIG_SNOW_ENCODER //dwt is in snow.c
 static inline int w_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int w, int h, int type){
-#ifdef CONFIG_SNOW_ENCODER //idwt is in snow.c
     int s, i, j;
     const int dec_count= w==8 ? 3 : 4;
-    int tmp[16*16];
-#if 0
+    int tmp[32*32];
     int level, ori;
     static const int scale[2][2][4][4]={
       {
         {
-            //8x8 dec=3
+            // 9/7 8x8 dec=3
             {268, 239, 239, 213},
             {  0, 224, 224, 152},
             {  0, 135, 135, 110},
         },{
-            //16x16 dec=4
+            // 9/7 16x16 or 32x32 dec=4
             {344, 310, 310, 280},
             {  0, 320, 320, 228},
             {  0, 175, 175, 136},
             {  0, 129, 129, 102},
         }
       },{
-        {//FIXME 5/3
-            //8x8 dec=3
+        {
+            // 5/3 8x8 dec=3
             {275, 245, 245, 218},
             {  0, 230, 230, 156},
             {  0, 138, 138, 113},
         },{
-            //16x16 dec=4
+            // 5/3 16x16 or 32x32 dec=4
             {352, 317, 317, 286},
             {  0, 328, 328, 233},
             {  0, 180, 180, 140},
@@ -327,51 +369,39 @@ static inline int w_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, in
         }
       }
     };
-#endif
 
     for (i = 0; i < h; i++) {
         for (j = 0; j < w; j+=4) {
-            tmp[16*i+j+0] = (pix1[j+0] - pix2[j+0])<<4;
-            tmp[16*i+j+1] = (pix1[j+1] - pix2[j+1])<<4;
-            tmp[16*i+j+2] = (pix1[j+2] - pix2[j+2])<<4;
-            tmp[16*i+j+3] = (pix1[j+3] - pix2[j+3])<<4;
+            tmp[32*i+j+0] = (pix1[j+0] - pix2[j+0])<<4;
+            tmp[32*i+j+1] = (pix1[j+1] - pix2[j+1])<<4;
+            tmp[32*i+j+2] = (pix1[j+2] - pix2[j+2])<<4;
+            tmp[32*i+j+3] = (pix1[j+3] - pix2[j+3])<<4;
         }
         pix1 += line_size;
         pix2 += line_size;
     }
 
-    ff_spatial_dwt(tmp, w, h, 16, type, dec_count);
+    ff_spatial_dwt(tmp, w, h, 32, type, dec_count);
 
     s=0;
-#if 0
+    assert(w==h);
     for(level=0; level<dec_count; level++){
         for(ori= level ? 1 : 0; ori<4; ori++){
-            int sx= (ori&1) ? 1<<level: 0;
-            int stride= 16<<(dec_count-level);
+            int size= w>>(dec_count-level);
+            int sx= (ori&1) ? size : 0;
+            int stride= 32<<(dec_count-level);
             int sy= (ori&2) ? stride>>1 : 0;
-            int size= 1<<level;
 
             for(i=0; i<size; i++){
                 for(j=0; j<size; j++){
                     int v= tmp[sx + sy + i*stride + j] * scale[type][dec_count-3][level][ori];
-                    s += ABS(v);
+                    s += FFABS(v);
                 }
             }
         }
     }
-#endif
-    for (i = 0; i < h; i++) {
-        for (j = 0; j < w; j+=4) {
-            s+= ABS(tmp[16*i+j+0]);
-            s+= ABS(tmp[16*i+j+1]);
-            s+= ABS(tmp[16*i+j+2]);
-            s+= ABS(tmp[16*i+j+3]);
-        }
-    }
     assert(s>=0);
-
-    return s>>2;
-#endif
+    return s>>9;
 }
 
 static int w53_8_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
@@ -390,6 +420,115 @@ static int w97_16_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int
     return w_c(v, pix1, pix2, line_size, 16, h, 0);
 }
 
+int w53_32_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
+    return w_c(v, pix1, pix2, line_size, 32, h, 1);
+}
+
+int w97_32_c(void *v, uint8_t * pix1, uint8_t * pix2, int line_size, int h){
+    return w_c(v, pix1, pix2, line_size, 32, h, 0);
+}
+#endif
+
+/* draw the edges of width 'w' of an image of size width, height */
+//FIXME check that this is ok for mpeg4 interlaced
+static void draw_edges_c(uint8_t *buf, int wrap, int width, int height, int w)
+{
+    uint8_t *ptr, *last_line;
+    int i;
+
+    last_line = buf + (height - 1) * wrap;
+    for(i=0;i<w;i++) {
+        /* top and bottom */
+        memcpy(buf - (i + 1) * wrap, buf, width);
+        memcpy(last_line + (i + 1) * wrap, last_line, width);
+    }
+    /* left and right */
+    ptr = buf;
+    for(i=0;i<height;i++) {
+        memset(ptr - w, ptr[0], w);
+        memset(ptr + width, ptr[width-1], w);
+        ptr += wrap;
+    }
+    /* corners */
+    for(i=0;i<w;i++) {
+        memset(buf - (i + 1) * wrap - w, buf[0], w); /* top left */
+        memset(buf - (i + 1) * wrap + width, buf[width-1], w); /* top right */
+        memset(last_line + (i + 1) * wrap - w, last_line[0], w); /* top left */
+        memset(last_line + (i + 1) * wrap + width, last_line[width-1], w); /* top right */
+    }
+}
+
+/**
+ * Copies a rectangular area of samples to a temporary buffer and replicates the boarder samples.
+ * @param buf destination buffer
+ * @param src source buffer
+ * @param linesize number of bytes between 2 vertically adjacent samples in both the source and destination buffers
+ * @param block_w width of block
+ * @param block_h height of block
+ * @param src_x x coordinate of the top left sample of the block in the source buffer
+ * @param src_y y coordinate of the top left sample of the block in the source buffer
+ * @param w width of the source buffer
+ * @param h height of the source buffer
+ */
+void ff_emulated_edge_mc(uint8_t *buf, uint8_t *src, int linesize, int block_w, int block_h,
+                                    int src_x, int src_y, int w, int h){
+    int x, y;
+    int start_y, start_x, end_y, end_x;
+
+    if(src_y>= h){
+        src+= (h-1-src_y)*linesize;
+        src_y=h-1;
+    }else if(src_y<=-block_h){
+        src+= (1-block_h-src_y)*linesize;
+        src_y=1-block_h;
+    }
+    if(src_x>= w){
+        src+= (w-1-src_x);
+        src_x=w-1;
+    }else if(src_x<=-block_w){
+        src+= (1-block_w-src_x);
+        src_x=1-block_w;
+    }
+
+    start_y= FFMAX(0, -src_y);
+    start_x= FFMAX(0, -src_x);
+    end_y= FFMIN(block_h, h-src_y);
+    end_x= FFMIN(block_w, w-src_x);
+
+    // copy existing part
+    for(y=start_y; y<end_y; y++){
+        for(x=start_x; x<end_x; x++){
+            buf[x + y*linesize]= src[x + y*linesize];
+        }
+    }
+
+    //top
+    for(y=0; y<start_y; y++){
+        for(x=start_x; x<end_x; x++){
+            buf[x + y*linesize]= buf[x + start_y*linesize];
+        }
+    }
+
+    //bottom
+    for(y=end_y; y<block_h; y++){
+        for(x=start_x; x<end_x; x++){
+            buf[x + y*linesize]= buf[x + (end_y-1)*linesize];
+        }
+    }
+
+    for(y=0; y<block_h; y++){
+       //left
+        for(x=0; x<start_x; x++){
+            buf[x + y*linesize]= buf[start_x + y*linesize];
+        }
+
+       //right
+        for(x=end_x; x<block_w; x++){
+            buf[x + y*linesize]= buf[end_x - 1 + y*linesize];
+        }
+    }
+}
+
 static void get_pixels_c(DCTELEM *restrict block, const uint8_t *pixels, int line_size)
 {
     int i;
@@ -434,7 +573,7 @@ static void put_pixels_clamped_c(const DCTELEM *block, uint8_t *restrict pixels,
                                  int line_size)
 {
     int i;
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
     for(i=0;i<8;i++) {
@@ -456,7 +595,7 @@ static void put_pixels_clamped4_c(const DCTELEM *block, uint8_t *restrict pixels
                                  int line_size)
 {
     int i;
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
     for(i=0;i<4;i++) {
@@ -474,7 +613,7 @@ static void put_pixels_clamped2_c(const DCTELEM *block, uint8_t *restrict pixels
                                  int line_size)
 {
     int i;
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
     for(i=0;i<2;i++) {
@@ -511,7 +650,7 @@ static void add_pixels_clamped_c(const DCTELEM *block, uint8_t *restrict pixels,
                           int line_size)
 {
     int i;
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
     for(i=0;i<8;i++) {
@@ -532,7 +671,7 @@ static void add_pixels_clamped4_c(const DCTELEM *block, uint8_t *restrict pixels
                           int line_size)
 {
     int i;
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
     for(i=0;i<4;i++) {
@@ -549,7 +688,7 @@ static void add_pixels_clamped2_c(const DCTELEM *block, uint8_t *restrict pixels
                           int line_size)
 {
     int i;
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
     for(i=0;i<2;i++) {
@@ -590,6 +729,14 @@ static void add_pixels4_c(uint8_t *restrict pixels, DCTELEM *block, int line_siz
     }
 }
 
+static int sum_abs_dctelem_c(DCTELEM *block)
+{
+    int sum=0, i;
+    for(i=0; i<64; i++)
+        sum+= FFABS(block[i]);
+    return sum;
+}
+
 #if 0
 
 #define PIXOP2(OPNAME, OP) \
@@ -597,7 +744,7 @@ static void OPNAME ## _pixels(uint8_t *block, const uint8_t *pixels, int line_si
 {\
     int i;\
     for(i=0; i<h; i++){\
-        OP(*((uint64_t*)block), LD64(pixels));\
+        OP(*((uint64_t*)block), AV_RN64(pixels));\
         pixels+=line_size;\
         block +=line_size;\
     }\
@@ -607,8 +754,8 @@ static void OPNAME ## _no_rnd_pixels_x2_c(uint8_t *block, const uint8_t *pixels,
 {\
     int i;\
     for(i=0; i<h; i++){\
-        const uint64_t a= LD64(pixels  );\
-        const uint64_t b= LD64(pixels+1);\
+        const uint64_t a= AV_RN64(pixels  );\
+        const uint64_t b= AV_RN64(pixels+1);\
         OP(*((uint64_t*)block), (a&b) + (((a^b)&0xFEFEFEFEFEFEFEFEULL)>>1));\
         pixels+=line_size;\
         block +=line_size;\
@@ -619,8 +766,8 @@ static void OPNAME ## _pixels_x2_c(uint8_t *block, const uint8_t *pixels, int li
 {\
     int i;\
     for(i=0; i<h; i++){\
-        const uint64_t a= LD64(pixels  );\
-        const uint64_t b= LD64(pixels+1);\
+        const uint64_t a= AV_RN64(pixels  );\
+        const uint64_t b= AV_RN64(pixels+1);\
         OP(*((uint64_t*)block), (a|b) - (((a^b)&0xFEFEFEFEFEFEFEFEULL)>>1));\
         pixels+=line_size;\
         block +=line_size;\
@@ -631,8 +778,8 @@ static void OPNAME ## _no_rnd_pixels_y2_c(uint8_t *block, const uint8_t *pixels,
 {\
     int i;\
     for(i=0; i<h; i++){\
-        const uint64_t a= LD64(pixels          );\
-        const uint64_t b= LD64(pixels+line_size);\
+        const uint64_t a= AV_RN64(pixels          );\
+        const uint64_t b= AV_RN64(pixels+line_size);\
         OP(*((uint64_t*)block), (a&b) + (((a^b)&0xFEFEFEFEFEFEFEFEULL)>>1));\
         pixels+=line_size;\
         block +=line_size;\
@@ -643,8 +790,8 @@ static void OPNAME ## _pixels_y2_c(uint8_t *block, const uint8_t *pixels, int li
 {\
     int i;\
     for(i=0; i<h; i++){\
-        const uint64_t a= LD64(pixels          );\
-        const uint64_t b= LD64(pixels+line_size);\
+        const uint64_t a= AV_RN64(pixels          );\
+        const uint64_t b= AV_RN64(pixels+line_size);\
         OP(*((uint64_t*)block), (a|b) - (((a^b)&0xFEFEFEFEFEFEFEFEULL)>>1));\
         pixels+=line_size;\
         block +=line_size;\
@@ -654,8 +801,8 @@ static void OPNAME ## _pixels_y2_c(uint8_t *block, const uint8_t *pixels, int li
 static void OPNAME ## _pixels_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
 {\
         int i;\
-        const uint64_t a= LD64(pixels  );\
-        const uint64_t b= LD64(pixels+1);\
+        const uint64_t a= AV_RN64(pixels  );\
+        const uint64_t b= AV_RN64(pixels+1);\
         uint64_t l0=  (a&0x0303030303030303ULL)\
                     + (b&0x0303030303030303ULL)\
                     + 0x0202020202020202ULL;\
@@ -665,8 +812,8 @@ static void OPNAME ## _pixels_xy2_c(uint8_t *block, const uint8_t *pixels, int l
 \
         pixels+=line_size;\
         for(i=0; i<h; i+=2){\
-            uint64_t a= LD64(pixels  );\
-            uint64_t b= LD64(pixels+1);\
+            uint64_t a= AV_RN64(pixels  );\
+            uint64_t b= AV_RN64(pixels+1);\
             l1=  (a&0x0303030303030303ULL)\
                + (b&0x0303030303030303ULL);\
             h1= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
@@ -674,8 +821,8 @@ static void OPNAME ## _pixels_xy2_c(uint8_t *block, const uint8_t *pixels, int l
             OP(*((uint64_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0F0F0F0F0FULL));\
             pixels+=line_size;\
             block +=line_size;\
-            a= LD64(pixels  );\
-            b= LD64(pixels+1);\
+            a= AV_RN64(pixels  );\
+            b= AV_RN64(pixels+1);\
             l0=  (a&0x0303030303030303ULL)\
                + (b&0x0303030303030303ULL)\
                + 0x0202020202020202ULL;\
@@ -690,8 +837,8 @@ static void OPNAME ## _pixels_xy2_c(uint8_t *block, const uint8_t *pixels, int l
 static void OPNAME ## _no_rnd_pixels_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
 {\
         int i;\
-        const uint64_t a= LD64(pixels  );\
-        const uint64_t b= LD64(pixels+1);\
+        const uint64_t a= AV_RN64(pixels  );\
+        const uint64_t b= AV_RN64(pixels+1);\
         uint64_t l0=  (a&0x0303030303030303ULL)\
                     + (b&0x0303030303030303ULL)\
                     + 0x0101010101010101ULL;\
@@ -701,8 +848,8 @@ static void OPNAME ## _no_rnd_pixels_xy2_c(uint8_t *block, const uint8_t *pixels
 \
         pixels+=line_size;\
         for(i=0; i<h; i+=2){\
-            uint64_t a= LD64(pixels  );\
-            uint64_t b= LD64(pixels+1);\
+            uint64_t a= AV_RN64(pixels  );\
+            uint64_t b= AV_RN64(pixels+1);\
             l1=  (a&0x0303030303030303ULL)\
                + (b&0x0303030303030303ULL);\
             h1= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
@@ -710,8 +857,8 @@ static void OPNAME ## _no_rnd_pixels_xy2_c(uint8_t *block, const uint8_t *pixels
             OP(*((uint64_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0F0F0F0F0FULL));\
             pixels+=line_size;\
             block +=line_size;\
-            a= LD64(pixels  );\
-            b= LD64(pixels+1);\
+            a= AV_RN64(pixels  );\
+            b= AV_RN64(pixels+1);\
             l0=  (a&0x0303030303030303ULL)\
                + (b&0x0303030303030303ULL)\
                + 0x0101010101010101ULL;\
@@ -738,7 +885,7 @@ CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_xy2_c, OPNAME ## _no_rnd_pixels_xy2_c,
 static void OPNAME ## _pixels2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
     int i;\
     for(i=0; i<h; i++){\
-        OP(*((uint16_t*)(block  )), LD16(pixels  ));\
+        OP(*((uint16_t*)(block  )), AV_RN16(pixels  ));\
         pixels+=line_size;\
         block +=line_size;\
     }\
@@ -746,7 +893,7 @@ static void OPNAME ## _pixels2_c(uint8_t *block, const uint8_t *pixels, int line
 static void OPNAME ## _pixels4_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
     int i;\
     for(i=0; i<h; i++){\
-        OP(*((uint32_t*)(block  )), LD32(pixels  ));\
+        OP(*((uint32_t*)(block  )), AV_RN32(pixels  ));\
         pixels+=line_size;\
         block +=line_size;\
     }\
@@ -754,8 +901,8 @@ static void OPNAME ## _pixels4_c(uint8_t *block, const uint8_t *pixels, int line
 static void OPNAME ## _pixels8_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
     int i;\
     for(i=0; i<h; i++){\
-        OP(*((uint32_t*)(block  )), LD32(pixels  ));\
-        OP(*((uint32_t*)(block+4)), LD32(pixels+4));\
+        OP(*((uint32_t*)(block  )), AV_RN32(pixels  ));\
+        OP(*((uint32_t*)(block+4)), AV_RN32(pixels+4));\
         pixels+=line_size;\
         block +=line_size;\
     }\
@@ -769,11 +916,11 @@ static inline void OPNAME ## _no_rnd_pixels8_l2(uint8_t *dst, const uint8_t *src
     int i;\
     for(i=0; i<h; i++){\
         uint32_t a,b;\
-        a= LD32(&src1[i*src_stride1  ]);\
-        b= LD32(&src2[i*src_stride2  ]);\
+        a= AV_RN32(&src1[i*src_stride1  ]);\
+        b= AV_RN32(&src2[i*src_stride2  ]);\
         OP(*((uint32_t*)&dst[i*dst_stride  ]), no_rnd_avg32(a, b));\
-        a= LD32(&src1[i*src_stride1+4]);\
-        b= LD32(&src2[i*src_stride2+4]);\
+        a= AV_RN32(&src1[i*src_stride1+4]);\
+        b= AV_RN32(&src2[i*src_stride2+4]);\
         OP(*((uint32_t*)&dst[i*dst_stride+4]), no_rnd_avg32(a, b));\
     }\
 }\
@@ -783,11 +930,11 @@ static inline void OPNAME ## _pixels8_l2(uint8_t *dst, const uint8_t *src1, cons
     int i;\
     for(i=0; i<h; i++){\
         uint32_t a,b;\
-        a= LD32(&src1[i*src_stride1  ]);\
-        b= LD32(&src2[i*src_stride2  ]);\
+        a= AV_RN32(&src1[i*src_stride1  ]);\
+        b= AV_RN32(&src2[i*src_stride2  ]);\
         OP(*((uint32_t*)&dst[i*dst_stride  ]), rnd_avg32(a, b));\
-        a= LD32(&src1[i*src_stride1+4]);\
-        b= LD32(&src2[i*src_stride2+4]);\
+        a= AV_RN32(&src1[i*src_stride1+4]);\
+        b= AV_RN32(&src2[i*src_stride2+4]);\
         OP(*((uint32_t*)&dst[i*dst_stride+4]), rnd_avg32(a, b));\
     }\
 }\
@@ -797,8 +944,8 @@ static inline void OPNAME ## _pixels4_l2(uint8_t *dst, const uint8_t *src1, cons
     int i;\
     for(i=0; i<h; i++){\
         uint32_t a,b;\
-        a= LD32(&src1[i*src_stride1  ]);\
-        b= LD32(&src2[i*src_stride2  ]);\
+        a= AV_RN32(&src1[i*src_stride1  ]);\
+        b= AV_RN32(&src2[i*src_stride2  ]);\
         OP(*((uint32_t*)&dst[i*dst_stride  ]), rnd_avg32(a, b));\
     }\
 }\
@@ -808,8 +955,8 @@ static inline void OPNAME ## _pixels2_l2(uint8_t *dst, const uint8_t *src1, cons
     int i;\
     for(i=0; i<h; i++){\
         uint32_t a,b;\
-        a= LD16(&src1[i*src_stride1  ]);\
-        b= LD16(&src2[i*src_stride2  ]);\
+        a= AV_RN16(&src1[i*src_stride1  ]);\
+        b= AV_RN16(&src2[i*src_stride2  ]);\
         OP(*((uint16_t*)&dst[i*dst_stride  ]), rnd_avg32(a, b));\
     }\
 }\
@@ -847,10 +994,10 @@ static inline void OPNAME ## _pixels8_l4(uint8_t *dst, const uint8_t *src1, uint
     int i;\
     for(i=0; i<h; i++){\
         uint32_t a, b, c, d, l0, l1, h0, h1;\
-        a= LD32(&src1[i*src_stride1]);\
-        b= LD32(&src2[i*src_stride2]);\
-        c= LD32(&src3[i*src_stride3]);\
-        d= LD32(&src4[i*src_stride4]);\
+        a= AV_RN32(&src1[i*src_stride1]);\
+        b= AV_RN32(&src2[i*src_stride2]);\
+        c= AV_RN32(&src3[i*src_stride3]);\
+        d= AV_RN32(&src4[i*src_stride4]);\
         l0=  (a&0x03030303UL)\
            + (b&0x03030303UL)\
            + 0x02020202UL;\
@@ -861,10 +1008,10 @@ static inline void OPNAME ## _pixels8_l4(uint8_t *dst, const uint8_t *src1, uint
         h1= ((c&0xFCFCFCFCUL)>>2)\
           + ((d&0xFCFCFCFCUL)>>2);\
         OP(*((uint32_t*)&dst[i*dst_stride]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-        a= LD32(&src1[i*src_stride1+4]);\
-        b= LD32(&src2[i*src_stride2+4]);\
-        c= LD32(&src3[i*src_stride3+4]);\
-        d= LD32(&src4[i*src_stride4+4]);\
+        a= AV_RN32(&src1[i*src_stride1+4]);\
+        b= AV_RN32(&src2[i*src_stride2+4]);\
+        c= AV_RN32(&src3[i*src_stride3+4]);\
+        d= AV_RN32(&src4[i*src_stride4+4]);\
         l0=  (a&0x03030303UL)\
            + (b&0x03030303UL)\
            + 0x02020202UL;\
@@ -899,10 +1046,10 @@ static inline void OPNAME ## _no_rnd_pixels8_l4(uint8_t *dst, const uint8_t *src
     int i;\
     for(i=0; i<h; i++){\
         uint32_t a, b, c, d, l0, l1, h0, h1;\
-        a= LD32(&src1[i*src_stride1]);\
-        b= LD32(&src2[i*src_stride2]);\
-        c= LD32(&src3[i*src_stride3]);\
-        d= LD32(&src4[i*src_stride4]);\
+        a= AV_RN32(&src1[i*src_stride1]);\
+        b= AV_RN32(&src2[i*src_stride2]);\
+        c= AV_RN32(&src3[i*src_stride3]);\
+        d= AV_RN32(&src4[i*src_stride4]);\
         l0=  (a&0x03030303UL)\
            + (b&0x03030303UL)\
            + 0x01010101UL;\
@@ -913,10 +1060,10 @@ static inline void OPNAME ## _no_rnd_pixels8_l4(uint8_t *dst, const uint8_t *src
         h1= ((c&0xFCFCFCFCUL)>>2)\
           + ((d&0xFCFCFCFCUL)>>2);\
         OP(*((uint32_t*)&dst[i*dst_stride]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-        a= LD32(&src1[i*src_stride1+4]);\
-        b= LD32(&src2[i*src_stride2+4]);\
-        c= LD32(&src3[i*src_stride3+4]);\
-        d= LD32(&src4[i*src_stride4+4]);\
+        a= AV_RN32(&src1[i*src_stride1+4]);\
+        b= AV_RN32(&src2[i*src_stride2+4]);\
+        c= AV_RN32(&src3[i*src_stride3+4]);\
+        d= AV_RN32(&src4[i*src_stride4+4]);\
         l0=  (a&0x03030303UL)\
            + (b&0x03030303UL)\
            + 0x01010101UL;\
@@ -976,8 +1123,8 @@ static inline void OPNAME ## _pixels2_xy2_c(uint8_t *block, const uint8_t *pixel
 static inline void OPNAME ## _pixels4_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
 {\
         int i;\
-        const uint32_t a= LD32(pixels  );\
-        const uint32_t b= LD32(pixels+1);\
+        const uint32_t a= AV_RN32(pixels  );\
+        const uint32_t b= AV_RN32(pixels+1);\
         uint32_t l0=  (a&0x03030303UL)\
                     + (b&0x03030303UL)\
                     + 0x02020202UL;\
@@ -987,8 +1134,8 @@ static inline void OPNAME ## _pixels4_xy2_c(uint8_t *block, const uint8_t *pixel
 \
         pixels+=line_size;\
         for(i=0; i<h; i+=2){\
-            uint32_t a= LD32(pixels  );\
-            uint32_t b= LD32(pixels+1);\
+            uint32_t a= AV_RN32(pixels  );\
+            uint32_t b= AV_RN32(pixels+1);\
             l1=  (a&0x03030303UL)\
                + (b&0x03030303UL);\
             h1= ((a&0xFCFCFCFCUL)>>2)\
@@ -996,8 +1143,8 @@ static inline void OPNAME ## _pixels4_xy2_c(uint8_t *block, const uint8_t *pixel
             OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
             pixels+=line_size;\
             block +=line_size;\
-            a= LD32(pixels  );\
-            b= LD32(pixels+1);\
+            a= AV_RN32(pixels  );\
+            b= AV_RN32(pixels+1);\
             l0=  (a&0x03030303UL)\
                + (b&0x03030303UL)\
                + 0x02020202UL;\
@@ -1014,8 +1161,8 @@ static inline void OPNAME ## _pixels8_xy2_c(uint8_t *block, const uint8_t *pixel
     int j;\
     for(j=0; j<2; j++){\
         int i;\
-        const uint32_t a= LD32(pixels  );\
-        const uint32_t b= LD32(pixels+1);\
+        const uint32_t a= AV_RN32(pixels  );\
+        const uint32_t b= AV_RN32(pixels+1);\
         uint32_t l0=  (a&0x03030303UL)\
                     + (b&0x03030303UL)\
                     + 0x02020202UL;\
@@ -1025,8 +1172,8 @@ static inline void OPNAME ## _pixels8_xy2_c(uint8_t *block, const uint8_t *pixel
 \
         pixels+=line_size;\
         for(i=0; i<h; i+=2){\
-            uint32_t a= LD32(pixels  );\
-            uint32_t b= LD32(pixels+1);\
+            uint32_t a= AV_RN32(pixels  );\
+            uint32_t b= AV_RN32(pixels+1);\
             l1=  (a&0x03030303UL)\
                + (b&0x03030303UL);\
             h1= ((a&0xFCFCFCFCUL)>>2)\
@@ -1034,8 +1181,8 @@ static inline void OPNAME ## _pixels8_xy2_c(uint8_t *block, const uint8_t *pixel
             OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
             pixels+=line_size;\
             block +=line_size;\
-            a= LD32(pixels  );\
-            b= LD32(pixels+1);\
+            a= AV_RN32(pixels  );\
+            b= AV_RN32(pixels+1);\
             l0=  (a&0x03030303UL)\
                + (b&0x03030303UL)\
                + 0x02020202UL;\
@@ -1055,8 +1202,8 @@ static inline void OPNAME ## _no_rnd_pixels8_xy2_c(uint8_t *block, const uint8_t
     int j;\
     for(j=0; j<2; j++){\
         int i;\
-        const uint32_t a= LD32(pixels  );\
-        const uint32_t b= LD32(pixels+1);\
+        const uint32_t a= AV_RN32(pixels  );\
+        const uint32_t b= AV_RN32(pixels+1);\
         uint32_t l0=  (a&0x03030303UL)\
                     + (b&0x03030303UL)\
                     + 0x01010101UL;\
@@ -1066,8 +1213,8 @@ static inline void OPNAME ## _no_rnd_pixels8_xy2_c(uint8_t *block, const uint8_t
 \
         pixels+=line_size;\
         for(i=0; i<h; i+=2){\
-            uint32_t a= LD32(pixels  );\
-            uint32_t b= LD32(pixels+1);\
+            uint32_t a= AV_RN32(pixels  );\
+            uint32_t b= AV_RN32(pixels+1);\
             l1=  (a&0x03030303UL)\
                + (b&0x03030303UL);\
             h1= ((a&0xFCFCFCFCUL)>>2)\
@@ -1075,8 +1222,8 @@ static inline void OPNAME ## _no_rnd_pixels8_xy2_c(uint8_t *block, const uint8_t
             OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
             pixels+=line_size;\
             block +=line_size;\
-            a= LD32(pixels  );\
-            b= LD32(pixels+1);\
+            a= AV_RN32(pixels  );\
+            b= AV_RN32(pixels+1);\
             l0=  (a&0x03030303UL)\
                + (b&0x03030303UL)\
                + 0x01010101UL;\
@@ -1143,7 +1290,7 @@ static void gmc1_c(uint8_t *dst, uint8_t *src, int stride, int h, int x16, int y
     }
 }
 
-static void gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
+void ff_gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
                   int dxx, int dxy, int dyx, int dyy, int shift, int r, int width, int height)
 {
     int y, vx, vy;
@@ -1176,19 +1323,19 @@ static void gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
                                            + src[index+stride+1]*   frac_x )*   frac_y
                                         + r)>>(shift*2);
                 }else{
-                    index= src_x + clip(src_y, 0, height)*stride;
+                    index= src_x + av_clip(src_y, 0, height)*stride;
                     dst[y*stride + x]= ( (  src[index         ]*(s-frac_x)
                                           + src[index       +1]*   frac_x )*s
                                         + r)>>(shift*2);
                 }
             }else{
                 if((unsigned)src_y < height){
-                    index= clip(src_x, 0, width) + src_y*stride;
+                    index= av_clip(src_x, 0, width) + src_y*stride;
                     dst[y*stride + x]= (  (  src[index         ]*(s-frac_y)
                                            + src[index+stride  ]*   frac_y )*s
                                         + r)>>(shift*2);
                 }else{
-                    index= clip(src_x, 0, width) + clip(src_y, 0, height)*stride;
+                    index= av_clip(src_x, 0, width) + av_clip(src_y, 0, height)*stride;
                     dst[y*stride + x]=    src[index         ];
                 }
             }
@@ -1426,12 +1573,22 @@ static void OPNAME ## h264_chroma_mc2_c(uint8_t *dst/*align 8*/, uint8_t *src/*a
     \
     assert(x<8 && y<8 && x>=0 && y>=0);\
 \
-    for(i=0; i<h; i++)\
-    {\
-        OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
-        OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
-        dst+= stride;\
-        src+= stride;\
+    if(D){\
+        for(i=0; i<h; i++){\
+            OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
+            OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
+            dst+= stride;\
+            src+= stride;\
+        }\
+    }else{\
+        const int E= B+C;\
+        const int step= C ? stride : 1;\
+        for(i=0; i<h; i++){\
+            OP(dst[0], (A*src[0] + E*src[step+0]));\
+            OP(dst[1], (A*src[1] + E*src[step+1]));\
+            dst+= stride;\
+            src+= stride;\
+        }\
     }\
 }\
 \
@@ -1444,14 +1601,26 @@ static void OPNAME ## h264_chroma_mc4_c(uint8_t *dst/*align 8*/, uint8_t *src/*a
     \
     assert(x<8 && y<8 && x>=0 && y>=0);\
 \
-    for(i=0; i<h; i++)\
-    {\
-        OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
-        OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
-        OP(dst[2], (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3]));\
-        OP(dst[3], (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4]));\
-        dst+= stride;\
-        src+= stride;\
+    if(D){\
+        for(i=0; i<h; i++){\
+            OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
+            OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
+            OP(dst[2], (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3]));\
+            OP(dst[3], (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4]));\
+            dst+= stride;\
+            src+= stride;\
+        }\
+    }else{\
+        const int E= B+C;\
+        const int step= C ? stride : 1;\
+        for(i=0; i<h; i++){\
+            OP(dst[0], (A*src[0] + E*src[step+0]));\
+            OP(dst[1], (A*src[1] + E*src[step+1]));\
+            OP(dst[2], (A*src[2] + E*src[step+2]));\
+            OP(dst[3], (A*src[3] + E*src[step+3]));\
+            dst+= stride;\
+            src+= stride;\
+        }\
     }\
 }\
 \
@@ -1464,18 +1633,34 @@ static void OPNAME ## h264_chroma_mc8_c(uint8_t *dst/*align 8*/, uint8_t *src/*a
     \
     assert(x<8 && y<8 && x>=0 && y>=0);\
 \
-    for(i=0; i<h; i++)\
-    {\
-        OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
-        OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
-        OP(dst[2], (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3]));\
-        OP(dst[3], (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4]));\
-        OP(dst[4], (A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5]));\
-        OP(dst[5], (A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6]));\
-        OP(dst[6], (A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7]));\
-        OP(dst[7], (A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8]));\
-        dst+= stride;\
-        src+= stride;\
+    if(D){\
+        for(i=0; i<h; i++){\
+            OP(dst[0], (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1]));\
+            OP(dst[1], (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2]));\
+            OP(dst[2], (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3]));\
+            OP(dst[3], (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4]));\
+            OP(dst[4], (A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5]));\
+            OP(dst[5], (A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6]));\
+            OP(dst[6], (A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7]));\
+            OP(dst[7], (A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8]));\
+            dst+= stride;\
+            src+= stride;\
+        }\
+    }else{\
+        const int E= B+C;\
+        const int step= C ? stride : 1;\
+        for(i=0; i<h; i++){\
+            OP(dst[0], (A*src[0] + E*src[step+0]));\
+            OP(dst[1], (A*src[1] + E*src[step+1]));\
+            OP(dst[2], (A*src[2] + E*src[step+2]));\
+            OP(dst[3], (A*src[3] + E*src[step+3]));\
+            OP(dst[4], (A*src[4] + E*src[step+4]));\
+            OP(dst[5], (A*src[5] + E*src[step+5]));\
+            OP(dst[6], (A*src[6] + E*src[step+6]));\
+            OP(dst[7], (A*src[7] + E*src[step+7]));\
+            dst+= stride;\
+            src+= stride;\
+        }\
     }\
 }
 
@@ -1487,75 +1672,33 @@ H264_CHROMA_MC(avg_       , op_avg)
 #undef op_avg
 #undef op_put
 
-static inline void copy_block4(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
-{
+static void put_no_rnd_h264_chroma_mc8_c(uint8_t *dst/*align 8*/, uint8_t *src/*align 1*/, int stride, int h, int x, int y){
+    const int A=(8-x)*(8-y);
+    const int B=(  x)*(8-y);
+    const int C=(8-x)*(  y);
+    const int D=(  x)*(  y);
     int i;
-    for(i=0; i<h; i++)
-    {
-        ST32(dst   , LD32(src   ));
-        dst+=dstStride;
-        src+=srcStride;
-    }
-}
-
-static inline void copy_block8(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
-{
-    int i;
-    for(i=0; i<h; i++)
-    {
-        ST32(dst   , LD32(src   ));
-        ST32(dst+4 , LD32(src+4 ));
-        dst+=dstStride;
-        src+=srcStride;
-    }
-}
-
-static inline void copy_block16(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
-{
-    int i;
-    for(i=0; i<h; i++)
-    {
-        ST32(dst   , LD32(src   ));
-        ST32(dst+4 , LD32(src+4 ));
-        ST32(dst+8 , LD32(src+8 ));
-        ST32(dst+12, LD32(src+12));
-        dst+=dstStride;
-        src+=srcStride;
-    }
-}
 
-static inline void copy_block17(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
-{
-    int i;
-    for(i=0; i<h; i++)
-    {
-        ST32(dst   , LD32(src   ));
-        ST32(dst+4 , LD32(src+4 ));
-        ST32(dst+8 , LD32(src+8 ));
-        ST32(dst+12, LD32(src+12));
-        dst[16]= src[16];
-        dst+=dstStride;
-        src+=srcStride;
-    }
-}
+    assert(x<8 && y<8 && x>=0 && y>=0);
 
-static inline void copy_block9(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h)
-{
-    int i;
     for(i=0; i<h; i++)
     {
-        ST32(dst   , LD32(src   ));
-        ST32(dst+4 , LD32(src+4 ));
-        dst[8]= src[8];
-        dst+=dstStride;
-        src+=srcStride;
+        dst[0] = (A*src[0] + B*src[1] + C*src[stride+0] + D*src[stride+1] + 32 - 4) >> 6;
+        dst[1] = (A*src[1] + B*src[2] + C*src[stride+1] + D*src[stride+2] + 32 - 4) >> 6;
+        dst[2] = (A*src[2] + B*src[3] + C*src[stride+2] + D*src[stride+3] + 32 - 4) >> 6;
+        dst[3] = (A*src[3] + B*src[4] + C*src[stride+3] + D*src[stride+4] + 32 - 4) >> 6;
+        dst[4] = (A*src[4] + B*src[5] + C*src[stride+4] + D*src[stride+5] + 32 - 4) >> 6;
+        dst[5] = (A*src[5] + B*src[6] + C*src[stride+5] + D*src[stride+6] + 32 - 4) >> 6;
+        dst[6] = (A*src[6] + B*src[7] + C*src[stride+6] + D*src[stride+7] + 32 - 4) >> 6;
+        dst[7] = (A*src[7] + B*src[8] + C*src[stride+7] + D*src[stride+8] + 32 - 4) >> 6;
+        dst+= stride;
+        src+= stride;
     }
 }
 
-
 #define QPEL_MC(r, OPNAME, RND, OP) \
 static void OPNAME ## mpeg4_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<h; i++)\
     {\
@@ -1574,7 +1717,7 @@ static void OPNAME ## mpeg4_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstS
 \
 static void OPNAME ## mpeg4_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
     const int w=8;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<w; i++)\
     {\
@@ -1601,7 +1744,7 @@ static void OPNAME ## mpeg4_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstS
 }\
 \
 static void OPNAME ## mpeg4_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     \
     for(i=0; i<h; i++)\
@@ -1628,7 +1771,7 @@ static void OPNAME ## mpeg4_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dst
 }\
 \
 static void OPNAME ## mpeg4_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     const int w=16;\
     for(i=0; i<w; i++)\
@@ -2052,9 +2195,71 @@ QPEL_MC(0, avg_       , _       , op_avg)
 
 #if 1
 #define H264_LOWPASS(OPNAME, OP, OP2) \
+static av_unused void OPNAME ## h264_qpel2_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int h=2;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<h; i++)\
+    {\
+        OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\
+        OP(dst[1], (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]));\
+        dst+=dstStride;\
+        src+=srcStride;\
+    }\
+}\
+\
+static av_unused void OPNAME ## h264_qpel2_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
+    const int w=2;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+    int i;\
+    for(i=0; i<w; i++)\
+    {\
+        const int srcB= src[-2*srcStride];\
+        const int srcA= src[-1*srcStride];\
+        const int src0= src[0 *srcStride];\
+        const int src1= src[1 *srcStride];\
+        const int src2= src[2 *srcStride];\
+        const int src3= src[3 *srcStride];\
+        const int src4= src[4 *srcStride];\
+        OP(dst[0*dstStride], (src0+src1)*20 - (srcA+src2)*5 + (srcB+src3));\
+        OP(dst[1*dstStride], (src1+src2)*20 - (src0+src3)*5 + (srcA+src4));\
+        dst++;\
+        src++;\
+    }\
+}\
+\
+static av_unused void OPNAME ## h264_qpel2_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
+    const int h=2;\
+    const int w=2;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
+    int i;\
+    src -= 2*srcStride;\
+    for(i=0; i<h+5; i++)\
+    {\
+        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]);\
+        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]);\
+        tmp+=tmpStride;\
+        src+=srcStride;\
+    }\
+    tmp -= tmpStride*(h+5-2);\
+    for(i=0; i<w; i++)\
+    {\
+        const int tmpB= tmp[-2*tmpStride];\
+        const int tmpA= tmp[-1*tmpStride];\
+        const int tmp0= tmp[0 *tmpStride];\
+        const int tmp1= tmp[1 *tmpStride];\
+        const int tmp2= tmp[2 *tmpStride];\
+        const int tmp3= tmp[3 *tmpStride];\
+        const int tmp4= tmp[4 *tmpStride];\
+        OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
+        OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
+        dst++;\
+        tmp++;\
+    }\
+}\
 static void OPNAME ## h264_qpel4_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
     const int h=4;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<h; i++)\
     {\
@@ -2069,7 +2274,7 @@ static void OPNAME ## h264_qpel4_h_lowpass(uint8_t *dst, uint8_t *src, int dstSt
 \
 static void OPNAME ## h264_qpel4_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
     const int w=4;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<w; i++)\
     {\
@@ -2094,7 +2299,7 @@ static void OPNAME ## h264_qpel4_v_lowpass(uint8_t *dst, uint8_t *src, int dstSt
 static void OPNAME ## h264_qpel4_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
     const int h=4;\
     const int w=4;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     src -= 2*srcStride;\
     for(i=0; i<h+5; i++)\
@@ -2129,7 +2334,7 @@ static void OPNAME ## h264_qpel4_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t
 \
 static void OPNAME ## h264_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
     const int h=8;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<h; i++)\
     {\
@@ -2148,7 +2353,7 @@ static void OPNAME ## h264_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstSt
 \
 static void OPNAME ## h264_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
     const int w=8;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     for(i=0; i<w; i++)\
     {\
@@ -2181,7 +2386,7 @@ static void OPNAME ## h264_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstSt
 static void OPNAME ## h264_qpel8_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
     const int h=8;\
     const int w=8;\
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;\
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
     int i;\
     src -= 2*srcStride;\
     for(i=0; i<h+5; i++)\
@@ -2398,6 +2603,7 @@ static void OPNAME ## h264_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, i
 
 H264_LOWPASS(put_       , op_put, op2_put)
 H264_LOWPASS(avg_       , op_avg, op2_avg)
+H264_MC(put_, 2)
 H264_MC(put_, 4)
 H264_MC(put_, 8)
 H264_MC(put_, 16)
@@ -2411,11 +2617,11 @@ H264_MC(avg_, 16)
 #undef op2_put
 #endif
 
-#define op_scale1(x)  block[x] = clip_uint8( (block[x]*weight + offset) >> log2_denom )
-#define op_scale2(x)  dst[x] = clip_uint8( (src[x]*weights + dst[x]*weightd + offset) >> (log2_denom+1))
+#define op_scale1(x)  block[x] = av_clip_uint8( (block[x]*weight + offset) >> log2_denom )
+#define op_scale2(x)  dst[x] = av_clip_uint8( (src[x]*weights + dst[x]*weightd + offset) >> (log2_denom+1))
 #define H264_WEIGHT(W,H) \
 static void weight_h264_pixels ## W ## x ## H ## _c(uint8_t *block, int stride, int log2_denom, int weight, int offset){ \
-    int attribute_unused x, y; \
+    int y; \
     offset <<= log2_denom; \
     if(log2_denom) offset += 1<<(log2_denom-1); \
     for(y=0; y<H; y++, block += stride){ \
@@ -2440,10 +2646,9 @@ static void weight_h264_pixels ## W ## x ## H ## _c(uint8_t *block, int stride,
         op_scale1(15); \
     } \
 } \
-static void biweight_h264_pixels ## W ## x ## H ## _c(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offsetd, int offsets){ \
-    int attribute_unused x, y; \
-    int offset = (offsets + offsetd + 1) >> 1; \
-    offset = ((offset << 1) + 1) << log2_denom; \
+static void biweight_h264_pixels ## W ## x ## H ## _c(uint8_t *dst, uint8_t *src, int stride, int log2_denom, int weightd, int weights, int offset){ \
+    int y; \
+    offset = ((offset + 1) | 1) << log2_denom; \
     for(y=0; y<H; y++, dst += stride, src += stride){ \
         op_scale2(0); \
         op_scale2(1); \
@@ -2483,7 +2688,7 @@ H264_WEIGHT(2,2)
 #undef H264_WEIGHT
 
 static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
     int i;
 
     for(i=0; i<h; i++){
@@ -2500,8 +2705,40 @@ static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int
     }
 }
 
+#ifdef CONFIG_CAVS_DECODER
+/* AVS specific */
+void ff_cavsdsp_init(DSPContext* c, AVCodecContext *avctx);
+
+void ff_put_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
+    put_pixels8_c(dst, src, stride, 8);
+}
+void ff_avg_cavs_qpel8_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
+    avg_pixels8_c(dst, src, stride, 8);
+}
+void ff_put_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
+    put_pixels16_c(dst, src, stride, 16);
+}
+void ff_avg_cavs_qpel16_mc00_c(uint8_t *dst, uint8_t *src, int stride) {
+    avg_pixels16_c(dst, src, stride, 16);
+}
+#endif /* CONFIG_CAVS_DECODER */
+
+#if defined(CONFIG_VC1_DECODER) || defined(CONFIG_WMV3_DECODER)
+/* VC-1 specific */
+void ff_vc1dsp_init(DSPContext* c, AVCodecContext *avctx);
+
+void ff_put_vc1_mspel_mc00_c(uint8_t *dst, uint8_t *src, int stride, int rnd) {
+    put_pixels8_c(dst, src, stride, 8);
+}
+#endif /* CONFIG_VC1_DECODER||CONFIG_WMV3_DECODER */
+
+void ff_intrax8dsp_init(DSPContext* c, AVCodecContext *avctx);
+
+/* H264 specific */
+void ff_h264dspenc_init(DSPContext* c, AVCodecContext *avctx);
+
 static void wmv2_mspel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int w){
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
     int i;
 
     for(i=0; i<w; i++){
@@ -2578,6 +2815,7 @@ static void put_mspel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){
 }
 
 static void h263_v_loop_filter_c(uint8_t *src, int stride, int qscale){
+    if(ENABLE_ANY_H263) {
     int x;
     const int strength= ff_h263_loop_filter_strength[qscale];
 
@@ -2603,16 +2841,18 @@ static void h263_v_loop_filter_c(uint8_t *src, int stride, int qscale){
         src[x-1*stride] = p1;
         src[x+0*stride] = p2;
 
-        ad1= ABS(d1)>>1;
+        ad1= FFABS(d1)>>1;
 
-        d2= clip((p0-p3)/4, -ad1, ad1);
+        d2= av_clip((p0-p3)/4, -ad1, ad1);
 
         src[x-2*stride] = p0 - d2;
         src[x+  stride] = p3 + d2;
     }
+    }
 }
 
 static void h263_h_loop_filter_c(uint8_t *src, int stride, int qscale){
+    if(ENABLE_ANY_H263) {
     int y;
     const int strength= ff_h263_loop_filter_strength[qscale];
 
@@ -2638,13 +2878,14 @@ static void h263_h_loop_filter_c(uint8_t *src, int stride, int qscale){
         src[y*stride-1] = p1;
         src[y*stride+0] = p2;
 
-        ad1= ABS(d1)>>1;
+        ad1= FFABS(d1)>>1;
 
-        d2= clip((p0-p3)/4, -ad1, ad1);
+        d2= av_clip((p0-p3)/4, -ad1, ad1);
 
         src[y*stride-2] = p0 - d2;
         src[y*stride+1] = p3 + d2;
     }
+    }
 }
 
 static void h261_loop_filter_c(uint8_t *src, int stride){
@@ -2690,25 +2931,25 @@ static inline void h264_loop_filter_luma_c(uint8_t *pix, int xstride, int ystrid
             const int q1 = pix[1*xstride];
             const int q2 = pix[2*xstride];
 
-            if( ABS( p0 - q0 ) < alpha &&
-                ABS( p1 - p0 ) < beta &&
-                ABS( q1 - q0 ) < beta ) {
+            if( FFABS( p0 - q0 ) < alpha &&
+                FFABS( p1 - p0 ) < beta &&
+                FFABS( q1 - q0 ) < beta ) {
 
                 int tc = tc0[i];
                 int i_delta;
 
-                if( ABS( p2 - p0 ) < beta ) {
-                    pix[-2*xstride] = p1 + clip( (( p2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - p1, -tc0[i], tc0[i] );
+                if( FFABS( p2 - p0 ) < beta ) {
+                    pix[-2*xstride] = p1 + av_clip( (( p2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - p1, -tc0[i], tc0[i] );
                     tc++;
                 }
-                if( ABS( q2 - q0 ) < beta ) {
-                    pix[   xstride] = q1 + clip( (( q2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - q1, -tc0[i], tc0[i] );
+                if( FFABS( q2 - q0 ) < beta ) {
+                    pix[   xstride] = q1 + av_clip( (( q2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - q1, -tc0[i], tc0[i] );
                     tc++;
                 }
 
-                i_delta = clip( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -tc, tc );
-                pix[-xstride] = clip_uint8( p0 + i_delta );    /* p0' */
-                pix[0]        = clip_uint8( q0 - i_delta );    /* q0' */
+                i_delta = av_clip( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -tc, tc );
+                pix[-xstride] = av_clip_uint8( p0 + i_delta );    /* p0' */
+                pix[0]        = av_clip_uint8( q0 - i_delta );    /* q0' */
             }
             pix += ystride;
         }
@@ -2738,14 +2979,14 @@ static inline void h264_loop_filter_chroma_c(uint8_t *pix, int xstride, int ystr
             const int q0 = pix[0];
             const int q1 = pix[1*xstride];
 
-            if( ABS( p0 - q0 ) < alpha &&
-                ABS( p1 - p0 ) < beta &&
-                ABS( q1 - q0 ) < beta ) {
+            if( FFABS( p0 - q0 ) < alpha &&
+                FFABS( p1 - p0 ) < beta &&
+                FFABS( q1 - q0 ) < beta ) {
 
-                int delta = clip( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -tc, tc );
+                int delta = av_clip( (((q0 - p0 ) << 2) + (p1 - q1) + 4) >> 3, -tc, tc );
 
-                pix[-xstride] = clip_uint8( p0 + delta );    /* p0' */
-                pix[0]        = clip_uint8( q0 - delta );    /* q0' */
+                pix[-xstride] = av_clip_uint8( p0 + delta );    /* p0' */
+                pix[0]        = av_clip_uint8( q0 - delta );    /* q0' */
             }
             pix += ystride;
         }
@@ -2769,9 +3010,9 @@ static inline void h264_loop_filter_chroma_intra_c(uint8_t *pix, int xstride, in
         const int q0 = pix[0];
         const int q1 = pix[1*xstride];
 
-        if( ABS( p0 - q0 ) < alpha &&
-            ABS( p1 - p0 ) < beta &&
-            ABS( q1 - q0 ) < beta ) {
+        if( FFABS( p0 - q0 ) < alpha &&
+            FFABS( p1 - p0 ) < beta &&
+            FFABS( q1 - q0 ) < beta ) {
 
             pix[-xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;   /* p0' */
             pix[0]        = ( 2*q1 + q0 + p1 + 2 ) >> 2;   /* q0' */
@@ -3000,9 +3241,9 @@ static int nsse16_c(void *v, uint8_t *s1, uint8_t *s2, int stride, int h){
         }
         if(y+1<h){
             for(x=0; x<15; x++){
-                score2+= ABS(  s1[x  ] - s1[x  +stride]
+                score2+= FFABS(  s1[x  ] - s1[x  +stride]
                              - s1[x+1] + s1[x+1+stride])
-                        -ABS(  s2[x  ] - s2[x  +stride]
+                        -FFABS(  s2[x  ] - s2[x  +stride]
                              - s2[x+1] + s2[x+1+stride]);
             }
         }
@@ -3010,8 +3251,8 @@ static int nsse16_c(void *v, uint8_t *s1, uint8_t *s2, int stride, int h){
         s2+= stride;
     }
 
-    if(c) return score1 + ABS(score2)*c->avctx->nsse_weight;
-    else  return score1 + ABS(score2)*8;
+    if(c) return score1 + FFABS(score2)*c->avctx->nsse_weight;
+    else  return score1 + FFABS(score2)*8;
 }
 
 static int nsse8_c(void *v, uint8_t *s1, uint8_t *s2, int stride, int h){
@@ -3026,9 +3267,9 @@ static int nsse8_c(void *v, uint8_t *s1, uint8_t *s2, int stride, int h){
         }
         if(y+1<h){
             for(x=0; x<7; x++){
-                score2+= ABS(  s1[x  ] - s1[x  +stride]
+                score2+= FFABS(  s1[x  ] - s1[x  +stride]
                              - s1[x+1] + s1[x+1+stride])
-                        -ABS(  s2[x  ] - s2[x  +stride]
+                        -FFABS(  s2[x  ] - s2[x  +stride]
                              - s2[x+1] + s2[x+1+stride]);
             }
         }
@@ -3036,8 +3277,8 @@ static int nsse8_c(void *v, uint8_t *s1, uint8_t *s2, int stride, int h){
         s2+= stride;
     }
 
-    if(c) return score1 + ABS(score2)*c->avctx->nsse_weight;
-    else  return score1 + ABS(score2)*8;
+    if(c) return score1 + FFABS(score2)*c->avctx->nsse_weight;
+    else  return score1 + FFABS(score2)*8;
 }
 
 static int try_8x8basis_c(int16_t rem[64], int16_t weight[64], int16_t basis[64], int scale){
@@ -3077,7 +3318,7 @@ void ff_block_permute(DCTELEM *block, uint8_t *permutation, const uint8_t *scant
     DCTELEM temp[64];
 
     if(last<=0) return;
-    //if(permutation[1]==1) return; //FIXME its ok but not clean and might fail for some perms
+    //if(permutation[1]==1) return; //FIXME it is ok but not clean and might fail for some permutations
 
     for(i=0; i<=last; i++){
         const int j= scantable[i];
@@ -3115,6 +3356,9 @@ void ff_set_cmp(DSPContext* c, me_cmp_func *cmp, int type){
         case FF_CMP_DCT:
             cmp[i]= c->dct_sad[i];
             break;
+        case FF_CMP_DCT264:
+            cmp[i]= c->dct264_sad[i];
+            break;
         case FF_CMP_DCTMAX:
             cmp[i]= c->dct_max[i];
             break;
@@ -3139,12 +3383,14 @@ void ff_set_cmp(DSPContext* c, me_cmp_func *cmp, int type){
         case FF_CMP_NSSE:
             cmp[i]= c->nsse[i];
             break;
+#ifdef CONFIG_SNOW_ENCODER
         case FF_CMP_W53:
             cmp[i]= c->w53[i];
             break;
         case FF_CMP_W97:
             cmp[i]= c->w97[i];
             break;
+#endif
         default:
             av_log(NULL, AV_LOG_ERROR,"internal error in cmp function selection\n");
         }
@@ -3160,32 +3406,47 @@ static void clear_blocks_c(DCTELEM *blocks)
 }
 
 static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
-    int i;
-    for(i=0; i+7<w; i+=8){
-        dst[i+0] += src[i+0];
-        dst[i+1] += src[i+1];
-        dst[i+2] += src[i+2];
-        dst[i+3] += src[i+3];
-        dst[i+4] += src[i+4];
-        dst[i+5] += src[i+5];
-        dst[i+6] += src[i+6];
-        dst[i+7] += src[i+7];
+    long i;
+    for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
+        long a = *(long*)(src+i);
+        long b = *(long*)(dst+i);
+        *(long*)(dst+i) = ((a&pb_7f) + (b&pb_7f)) ^ ((a^b)&pb_80);
     }
     for(; i<w; i++)
         dst[i+0] += src[i+0];
 }
 
+static void add_bytes_l2_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
+    long i;
+    for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
+        long a = *(long*)(src1+i);
+        long b = *(long*)(src2+i);
+        *(long*)(dst+i) = ((a&pb_7f) + (b&pb_7f)) ^ ((a^b)&pb_80);
+    }
+    for(; i<w; i++)
+        dst[i] = src1[i]+src2[i];
+}
+
 static void diff_bytes_c(uint8_t *dst, uint8_t *src1, uint8_t *src2, int w){
-    int i;
-    for(i=0; i+7<w; i+=8){
-        dst[i+0] = src1[i+0]-src2[i+0];
-        dst[i+1] = src1[i+1]-src2[i+1];
-        dst[i+2] = src1[i+2]-src2[i+2];
-        dst[i+3] = src1[i+3]-src2[i+3];
-        dst[i+4] = src1[i+4]-src2[i+4];
-        dst[i+5] = src1[i+5]-src2[i+5];
-        dst[i+6] = src1[i+6]-src2[i+6];
-        dst[i+7] = src1[i+7]-src2[i+7];
+    long i;
+#ifndef HAVE_FAST_UNALIGNED
+    if((long)src2 & (sizeof(long)-1)){
+        for(i=0; i+7<w; i+=8){
+            dst[i+0] = src1[i+0]-src2[i+0];
+            dst[i+1] = src1[i+1]-src2[i+1];
+            dst[i+2] = src1[i+2]-src2[i+2];
+            dst[i+3] = src1[i+3]-src2[i+3];
+            dst[i+4] = src1[i+4]-src2[i+4];
+            dst[i+5] = src1[i+5]-src2[i+5];
+            dst[i+6] = src1[i+6]-src2[i+6];
+            dst[i+7] = src1[i+7]-src2[i+7];
+        }
+    }else
+#endif
+    for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
+        long a = *(long*)(src1+i);
+        long b = *(long*)(src2+i);
+        *(long*)(dst+i) = ((a|pb_80) - (b&pb_7f)) ^ ((a^b^pb_80)&pb_80);
     }
     for(; i<w; i++)
         dst[i+0] = src1[i+0]-src2[i+0];
@@ -3222,7 +3483,7 @@ o2= (i1)-(i2);
     y= a-b;\
 }
 
-#define BUTTERFLYA(x,y) (ABS((x)+(y)) + ABS((x)-(y)))
+#define BUTTERFLYA(x,y) (FFABS((x)+(y)) + FFABS((x)-(y)))
 
 static int hadamard8_diff8x8_c(/*MpegEncContext*/ void *s, uint8_t *dst, uint8_t *src, int stride, int h){
     int i;
@@ -3319,31 +3580,79 @@ static int hadamard8_intra8x8_c(/*MpegEncContext*/ void *s, uint8_t *src, uint8_
             +BUTTERFLYA(temp[8*3+i], temp[8*7+i]);
     }
 
-    sum -= ABS(temp[8*0] + temp[8*4]); // -mean
+    sum -= FFABS(temp[8*0] + temp[8*4]); // -mean
 
     return sum;
 }
 
 static int dct_sad8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
     MpegEncContext * const s= (MpegEncContext *)c;
-    uint64_t __align8 aligned_temp[sizeof(DCTELEM)*64/8];
+    DECLARE_ALIGNED_16(uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
     DCTELEM * const temp= (DCTELEM*)aligned_temp;
-    int sum=0, i;
 
     assert(h==8);
 
     s->dsp.diff_pixels(temp, src1, src2, stride);
     s->dsp.fdct(temp);
+    return s->dsp.sum_abs_dctelem(temp);
+}
 
-    for(i=0; i<64; i++)
-        sum+= ABS(temp[i]);
+#ifdef CONFIG_GPL
+#define DCT8_1D {\
+    const int s07 = SRC(0) + SRC(7);\
+    const int s16 = SRC(1) + SRC(6);\
+    const int s25 = SRC(2) + SRC(5);\
+    const int s34 = SRC(3) + SRC(4);\
+    const int a0 = s07 + s34;\
+    const int a1 = s16 + s25;\
+    const int a2 = s07 - s34;\
+    const int a3 = s16 - s25;\
+    const int d07 = SRC(0) - SRC(7);\
+    const int d16 = SRC(1) - SRC(6);\
+    const int d25 = SRC(2) - SRC(5);\
+    const int d34 = SRC(3) - SRC(4);\
+    const int a4 = d16 + d25 + (d07 + (d07>>1));\
+    const int a5 = d07 - d34 - (d25 + (d25>>1));\
+    const int a6 = d07 + d34 - (d16 + (d16>>1));\
+    const int a7 = d16 - d25 + (d34 + (d34>>1));\
+    DST(0,  a0 + a1     ) ;\
+    DST(1,  a4 + (a7>>2)) ;\
+    DST(2,  a2 + (a3>>1)) ;\
+    DST(3,  a5 + (a6>>2)) ;\
+    DST(4,  a0 - a1     ) ;\
+    DST(5,  a6 - (a5>>2)) ;\
+    DST(6, (a2>>1) - a3 ) ;\
+    DST(7, (a4>>2) - a7 ) ;\
+}
+
+static int dct264_sad8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
+    MpegEncContext * const s= (MpegEncContext *)c;
+    DCTELEM dct[8][8];
+    int i;
+    int sum=0;
 
+    s->dsp.diff_pixels(dct[0], src1, src2, stride);
+
+#define SRC(x) dct[i][x]
+#define DST(x,v) dct[i][x]= v
+    for( i = 0; i < 8; i++ )
+        DCT8_1D
+#undef SRC
+#undef DST
+
+#define SRC(x) dct[x][i]
+#define DST(x,v) sum += FFABS(v)
+    for( i = 0; i < 8; i++ )
+        DCT8_1D
+#undef SRC
+#undef DST
     return sum;
 }
+#endif
 
 static int dct_max8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
     MpegEncContext * const s= (MpegEncContext *)c;
-    uint64_t __align8 aligned_temp[sizeof(DCTELEM)*64/8];
+    DECLARE_ALIGNED_8(uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
     DCTELEM * const temp= (DCTELEM*)aligned_temp;
     int sum=0, i;
 
@@ -3353,16 +3662,14 @@ static int dct_max8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2
     s->dsp.fdct(temp);
 
     for(i=0; i<64; i++)
-        sum= FFMAX(sum, ABS(temp[i]));
+        sum= FFMAX(sum, FFABS(temp[i]));
 
     return sum;
 }
 
-void simple_idct(DCTELEM *block); //FIXME
-
 static int quant_psnr8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
     MpegEncContext * const s= (MpegEncContext *)c;
-    uint64_t __align8 aligned_temp[sizeof(DCTELEM)*64*2/8];
+    DECLARE_ALIGNED_8 (uint64_t, aligned_temp[sizeof(DCTELEM)*64*2/8]);
     DCTELEM * const temp= (DCTELEM*)aligned_temp;
     DCTELEM * const bak = ((DCTELEM*)aligned_temp)+64;
     int sum=0, i;
@@ -3376,7 +3683,7 @@ static int quant_psnr8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *s
 
     s->block_last_index[0/*FIXME*/]= s->fast_dct_quantize(s, temp, 0/*FIXME*/, s->qscale, &i);
     s->dct_unquantize_inter(s, temp, 0, s->qscale);
-    simple_idct(temp); //FIXME
+    ff_simple_idct(temp); //FIXME
 
     for(i=0; i<64; i++)
         sum+= (temp[i]-bak[i])*(temp[i]-bak[i]);
@@ -3387,11 +3694,11 @@ static int quant_psnr8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *s
 static int rd8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
     MpegEncContext * const s= (MpegEncContext *)c;
     const uint8_t *scantable= s->intra_scantable.permutated;
-    uint64_t __align8 aligned_temp[sizeof(DCTELEM)*64/8];
-    uint64_t __align8 aligned_bak[stride];
+    DECLARE_ALIGNED_8 (uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
+    DECLARE_ALIGNED_8 (uint64_t, aligned_bak[stride]);
     DCTELEM * const temp= (DCTELEM*)aligned_temp;
     uint8_t * const bak= (uint8_t*)aligned_bak;
-    int i, last, run, bits, level, distoration, start_i;
+    int i, last, run, bits, level, distortion, start_i;
     const int esc_length= s->ac_esc_length;
     uint8_t * length;
     uint8_t * last_length;
@@ -3458,15 +3765,15 @@ static int rd8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int
 
     s->dsp.idct_add(bak, stride, temp);
 
-    distoration= s->dsp.sse[1](NULL, bak, src1, stride, 8);
+    distortion= s->dsp.sse[1](NULL, bak, src1, stride, 8);
 
-    return distoration + ((bits*s->qscale*s->qscale*109 + 64)>>7);
+    return distortion + ((bits*s->qscale*s->qscale*109 + 64)>>7);
 }
 
 static int bit8x8_c(/*MpegEncContext*/ void *c, uint8_t *src1, uint8_t *src2, int stride, int h){
     MpegEncContext * const s= (MpegEncContext *)c;
     const uint8_t *scantable= s->intra_scantable.permutated;
-    uint64_t __align8 aligned_temp[sizeof(DCTELEM)*64/8];
+    DECLARE_ALIGNED_8 (uint64_t, aligned_temp[sizeof(DCTELEM)*64/8]);
     DCTELEM * const temp= (DCTELEM*)aligned_temp;
     int i, last, run, bits, level, start_i;
     const int esc_length= s->ac_esc_length;
@@ -3529,8 +3836,8 @@ static int vsad_intra16_c(/*MpegEncContext*/ void *c, uint8_t *s, uint8_t *dummy
 
     for(y=1; y<h; y++){
         for(x=0; x<16; x+=4){
-            score+= ABS(s[x  ] - s[x  +stride]) + ABS(s[x+1] - s[x+1+stride])
-                   +ABS(s[x+2] - s[x+2+stride]) + ABS(s[x+3] - s[x+3+stride]);
+            score+= FFABS(s[x  ] - s[x  +stride]) + FFABS(s[x+1] - s[x+1+stride])
+                   +FFABS(s[x+2] - s[x+2+stride]) + FFABS(s[x+3] - s[x+3+stride]);
         }
         s+= stride;
     }
@@ -3544,7 +3851,7 @@ static int vsad16_c(/*MpegEncContext*/ void *c, uint8_t *s1, uint8_t *s2, int st
 
     for(y=1; y<h; y++){
         for(x=0; x<16; x++){
-            score+= ABS(s1[x  ] - s2[x ] - s1[x  +stride] + s2[x +stride]);
+            score+= FFABS(s1[x  ] - s2[x ] - s1[x  +stride] + s2[x +stride]);
         }
         s1+= stride;
         s2+= stride;
@@ -3584,16 +3891,143 @@ static int vsse16_c(/*MpegEncContext*/ void *c, uint8_t *s1, uint8_t *s2, int st
     return score;
 }
 
-WARPER8_16_SQ(hadamard8_diff8x8_c, hadamard8_diff16_c)
-WARPER8_16_SQ(hadamard8_intra8x8_c, hadamard8_intra16_c)
-WARPER8_16_SQ(dct_sad8x8_c, dct_sad16_c)
-WARPER8_16_SQ(dct_max8x8_c, dct_max16_c)
-WARPER8_16_SQ(quant_psnr8x8_c, quant_psnr16_c)
-WARPER8_16_SQ(rd8x8_c, rd16_c)
-WARPER8_16_SQ(bit8x8_c, bit16_c)
+static int ssd_int8_vs_int16_c(const int8_t *pix1, const int16_t *pix2,
+                               int size){
+    int score=0;
+    int i;
+    for(i=0; i<size; i++)
+        score += (pix1[i]-pix2[i])*(pix1[i]-pix2[i]);
+    return score;
+}
+
+WRAPPER8_16_SQ(hadamard8_diff8x8_c, hadamard8_diff16_c)
+WRAPPER8_16_SQ(hadamard8_intra8x8_c, hadamard8_intra16_c)
+WRAPPER8_16_SQ(dct_sad8x8_c, dct_sad16_c)
+#ifdef CONFIG_GPL
+WRAPPER8_16_SQ(dct264_sad8x8_c, dct264_sad16_c)
+#endif
+WRAPPER8_16_SQ(dct_max8x8_c, dct_max16_c)
+WRAPPER8_16_SQ(quant_psnr8x8_c, quant_psnr16_c)
+WRAPPER8_16_SQ(rd8x8_c, rd16_c)
+WRAPPER8_16_SQ(bit8x8_c, bit16_c)
+
+static void vector_fmul_c(float *dst, const float *src, int len){
+    int i;
+    for(i=0; i<len; i++)
+        dst[i] *= src[i];
+}
+
+static void vector_fmul_reverse_c(float *dst, const float *src0, const float *src1, int len){
+    int i;
+    src1 += len-1;
+    for(i=0; i<len; i++)
+        dst[i] = src0[i] * src1[-i];
+}
 
+void ff_vector_fmul_add_add_c(float *dst, const float *src0, const float *src1, const float *src2, int src3, int len, int step){
+    int i;
+    for(i=0; i<len; i++)
+        dst[i*step] = src0[i] * src1[i] + src2[i] + src3;
+}
+
+void ff_float_to_int16_c(int16_t *dst, const float *src, int len){
+    int i;
+    for(i=0; i<len; i++) {
+        int_fast32_t tmp = ((const int32_t*)src)[i];
+        if(tmp & 0xf0000){
+            tmp = (0x43c0ffff - tmp)>>31;
+            // is this faster on some gcc/cpu combinations?
+//          if(tmp > 0x43c0ffff) tmp = 0xFFFF;
+//          else                 tmp = 0;
+        }
+        dst[i] = tmp - 0x8000;
+    }
+}
+
+#define W0 2048
+#define W1 2841 /* 2048*sqrt (2)*cos (1*pi/16) */
+#define W2 2676 /* 2048*sqrt (2)*cos (2*pi/16) */
+#define W3 2408 /* 2048*sqrt (2)*cos (3*pi/16) */
+#define W4 2048 /* 2048*sqrt (2)*cos (4*pi/16) */
+#define W5 1609 /* 2048*sqrt (2)*cos (5*pi/16) */
+#define W6 1108 /* 2048*sqrt (2)*cos (6*pi/16) */
+#define W7 565  /* 2048*sqrt (2)*cos (7*pi/16) */
+
+static void wmv2_idct_row(short * b)
+{
+    int s1,s2;
+    int a0,a1,a2,a3,a4,a5,a6,a7;
+    /*step 1*/
+    a1 = W1*b[1]+W7*b[7];
+    a7 = W7*b[1]-W1*b[7];
+    a5 = W5*b[5]+W3*b[3];
+    a3 = W3*b[5]-W5*b[3];
+    a2 = W2*b[2]+W6*b[6];
+    a6 = W6*b[2]-W2*b[6];
+    a0 = W0*b[0]+W0*b[4];
+    a4 = W0*b[0]-W0*b[4];
+    /*step 2*/
+    s1 = (181*(a1-a5+a7-a3)+128)>>8;//1,3,5,7,
+    s2 = (181*(a1-a5-a7+a3)+128)>>8;
+    /*step 3*/
+    b[0] = (a0+a2+a1+a5 + (1<<7))>>8;
+    b[1] = (a4+a6 +s1   + (1<<7))>>8;
+    b[2] = (a4-a6 +s2   + (1<<7))>>8;
+    b[3] = (a0-a2+a7+a3 + (1<<7))>>8;
+    b[4] = (a0-a2-a7-a3 + (1<<7))>>8;
+    b[5] = (a4-a6 -s2   + (1<<7))>>8;
+    b[6] = (a4+a6 -s1   + (1<<7))>>8;
+    b[7] = (a0+a2-a1-a5 + (1<<7))>>8;
+}
+static void wmv2_idct_col(short * b)
+{
+    int s1,s2;
+    int a0,a1,a2,a3,a4,a5,a6,a7;
+    /*step 1, with extended precision*/
+    a1 = (W1*b[8*1]+W7*b[8*7] + 4)>>3;
+    a7 = (W7*b[8*1]-W1*b[8*7] + 4)>>3;
+    a5 = (W5*b[8*5]+W3*b[8*3] + 4)>>3;
+    a3 = (W3*b[8*5]-W5*b[8*3] + 4)>>3;
+    a2 = (W2*b[8*2]+W6*b[8*6] + 4)>>3;
+    a6 = (W6*b[8*2]-W2*b[8*6] + 4)>>3;
+    a0 = (W0*b[8*0]+W0*b[8*4]    )>>3;
+    a4 = (W0*b[8*0]-W0*b[8*4]    )>>3;
+    /*step 2*/
+    s1 = (181*(a1-a5+a7-a3)+128)>>8;
+    s2 = (181*(a1-a5-a7+a3)+128)>>8;
+    /*step 3*/
+    b[8*0] = (a0+a2+a1+a5 + (1<<13))>>14;
+    b[8*1] = (a4+a6 +s1   + (1<<13))>>14;
+    b[8*2] = (a4-a6 +s2   + (1<<13))>>14;
+    b[8*3] = (a0-a2+a7+a3 + (1<<13))>>14;
+
+    b[8*4] = (a0-a2-a7-a3 + (1<<13))>>14;
+    b[8*5] = (a4-a6 -s2   + (1<<13))>>14;
+    b[8*6] = (a4+a6 -s1   + (1<<13))>>14;
+    b[8*7] = (a0+a2-a1-a5 + (1<<13))>>14;
+}
+void ff_wmv2_idct_c(short * block){
+    int i;
+
+    for(i=0;i<64;i+=8){
+        wmv2_idct_row(block+i);
+    }
+    for(i=0;i<8;i++){
+        wmv2_idct_col(block+i);
+    }
+}
 /* XXX: those functions should be suppressed ASAP when all IDCTs are
  converted */
+static void ff_wmv2_idct_put_c(uint8_t *dest, int line_size, DCTELEM *block)
+{
+    ff_wmv2_idct_c(block);
+    put_pixels_clamped_c(block, dest, line_size);
+}
+static void ff_wmv2_idct_add_c(uint8_t *dest, int line_size, DCTELEM *block)
+{
+    ff_wmv2_idct_c(block);
+    add_pixels_clamped_c(block, dest, line_size);
+}
 static void ff_jref_idct_put(uint8_t *dest, int line_size, DCTELEM *block)
 {
     j_rev_dct (block);
@@ -3629,40 +4063,63 @@ static void ff_jref_idct2_add(uint8_t *dest, int line_size, DCTELEM *block)
 
 static void ff_jref_idct1_put(uint8_t *dest, int line_size, DCTELEM *block)
 {
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     dest[0] = cm[(block[0] + 4)>>3];
 }
 static void ff_jref_idct1_add(uint8_t *dest, int line_size, DCTELEM *block)
 {
-    uint8_t *cm = cropTbl + MAX_NEG_CROP;
+    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     dest[0] = cm[dest[0] + ((block[0] + 4)>>3)];
 }
 
+static void just_return(void *mem av_unused, int stride av_unused, int h av_unused) { return; }
+
 /* init static data */
 void dsputil_static_init(void)
 {
     int i;
 
-    for(i=0;i<256;i++) cropTbl[i + MAX_NEG_CROP] = i;
+    for(i=0;i<256;i++) ff_cropTbl[i + MAX_NEG_CROP] = i;
     for(i=0;i<MAX_NEG_CROP;i++) {
-        cropTbl[i] = 0;
-        cropTbl[i + MAX_NEG_CROP + 256] = 255;
+        ff_cropTbl[i] = 0;
+        ff_cropTbl[i + MAX_NEG_CROP + 256] = 255;
     }
 
     for(i=0;i<512;i++) {
-        squareTbl[i] = (i - 256) * (i - 256);
+        ff_squareTbl[i] = (i - 256) * (i - 256);
     }
 
     for(i=0; i<64; i++) inv_zigzag_direct16[ff_zigzag_direct[i]]= i+1;
 }
 
+int ff_check_alignment(void){
+    static int did_fail=0;
+    DECLARE_ALIGNED_16(int, aligned);
+
+    if((long)&aligned & 15){
+        if(!did_fail){
+#if defined(HAVE_MMX) || defined(HAVE_ALTIVEC)
+            av_log(NULL, AV_LOG_ERROR,
+                "Compiler did not align stack variables. Libavcodec has been miscompiled\n"
+                "and may be very slow or crash. This is not a bug in libavcodec,\n"
+                "but in the compiler. You may try recompiling using gcc >= 4.2.\n"
+                "Do not report crashes to FFmpeg developers.\n");
+#endif
+            did_fail=1;
+        }
+        return -1;
+    }
+    return 0;
+}
 
 void dsputil_init(DSPContext* c, AVCodecContext *avctx)
 {
     int i;
 
+    ff_check_alignment();
+
 #ifdef CONFIG_ENCODERS
     if(avctx->dct_algo==FF_DCT_FASTINT) {
         c->fdct = fdct_ifast;
@@ -3679,7 +4136,7 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
 #endif //CONFIG_ENCODERS
 
     if(avctx->lowres==1){
-        if(avctx->idct_algo==FF_IDCT_INT || avctx->idct_algo==FF_IDCT_AUTO){
+        if(avctx->idct_algo==FF_IDCT_INT || avctx->idct_algo==FF_IDCT_AUTO || !ENABLE_H264_DECODER){
             c->idct_put= ff_jref_idct4_put;
             c->idct_add= ff_jref_idct4_add;
         }else{
@@ -3704,21 +4161,36 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
             c->idct_add= ff_jref_idct_add;
             c->idct    = j_rev_dct;
             c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM;
-        }else if(avctx->idct_algo==FF_IDCT_VP3){
+        }else if((ENABLE_VP3_DECODER || ENABLE_VP5_DECODER || ENABLE_VP6_DECODER || ENABLE_THEORA_DECODER ) &&
+                avctx->idct_algo==FF_IDCT_VP3){
             c->idct_put= ff_vp3_idct_put_c;
             c->idct_add= ff_vp3_idct_add_c;
             c->idct    = ff_vp3_idct_c;
             c->idct_permutation_type= FF_NO_IDCT_PERM;
+        }else if(avctx->idct_algo==FF_IDCT_WMV2){
+            c->idct_put= ff_wmv2_idct_put_c;
+            c->idct_add= ff_wmv2_idct_add_c;
+            c->idct    = ff_wmv2_idct_c;
+            c->idct_permutation_type= FF_NO_IDCT_PERM;
+        }else if(avctx->idct_algo==FF_IDCT_FAAN){
+            c->idct_put= ff_faanidct_put;
+            c->idct_add= ff_faanidct_add;
+            c->idct    = ff_faanidct;
+            c->idct_permutation_type= FF_NO_IDCT_PERM;
         }else{ //accurate/default
-            c->idct_put= simple_idct_put;
-            c->idct_add= simple_idct_add;
-            c->idct    = simple_idct;
+            c->idct_put= ff_simple_idct_put;
+            c->idct_add= ff_simple_idct_add;
+            c->idct    = ff_simple_idct;
             c->idct_permutation_type= FF_NO_IDCT_PERM;
         }
     }
 
-    c->h264_idct_add= ff_h264_idct_add_c;
-    c->h264_idct8_add= ff_h264_idct8_add_c;
+    if (ENABLE_H264_DECODER) {
+        c->h264_idct_add= ff_h264_idct_add_c;
+        c->h264_idct8_add= ff_h264_idct8_add_c;
+        c->h264_idct_dc_add= ff_h264_idct_dc_add_c;
+        c->h264_idct8_dc_add= ff_h264_idct8_dc_add_c;
+    }
 
     c->get_pixels = get_pixels_c;
     c->diff_pixels = diff_pixels_c;
@@ -3727,8 +4199,9 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->add_pixels_clamped = add_pixels_clamped_c;
     c->add_pixels8 = add_pixels8_c;
     c->add_pixels4 = add_pixels4_c;
+    c->sum_abs_dctelem = sum_abs_dctelem_c;
     c->gmc1 = gmc1_c;
-    c->gmc = gmc_c;
+    c->gmc = ff_gmc_c;
     c->clear_blocks = clear_blocks_c;
     c->pix_sum = pix_sum_c;
     c->pix_norm1 = pix_norm1_c;
@@ -3820,6 +4293,7 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     dspfunc(put_h264_qpel, 0, 16);
     dspfunc(put_h264_qpel, 1, 8);
     dspfunc(put_h264_qpel, 2, 4);
+    dspfunc(put_h264_qpel, 3, 2);
     dspfunc(avg_h264_qpel, 0, 16);
     dspfunc(avg_h264_qpel, 1, 8);
     dspfunc(avg_h264_qpel, 2, 4);
@@ -3831,6 +4305,7 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->avg_h264_chroma_pixels_tab[0]= avg_h264_chroma_mc8_c;
     c->avg_h264_chroma_pixels_tab[1]= avg_h264_chroma_mc4_c;
     c->avg_h264_chroma_pixels_tab[2]= avg_h264_chroma_mc2_c;
+    c->put_no_rnd_h264_chroma_pixels_tab[0]= put_no_rnd_h264_chroma_mc8_c;
 
     c->weight_h264_pixels_tab[0]= weight_h264_pixels16x16_c;
     c->weight_h264_pixels_tab[1]= weight_h264_pixels16x8_c;
@@ -3853,6 +4328,21 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->biweight_h264_pixels_tab[8]= biweight_h264_pixels2x4_c;
     c->biweight_h264_pixels_tab[9]= biweight_h264_pixels2x2_c;
 
+    c->draw_edges = draw_edges_c;
+
+#ifdef CONFIG_CAVS_DECODER
+    ff_cavsdsp_init(c,avctx);
+#endif
+#if defined(CONFIG_VC1_DECODER) || defined(CONFIG_WMV3_DECODER)
+    ff_vc1dsp_init(c,avctx);
+#endif
+#if defined(CONFIG_WMV2_DECODER) || defined(CONFIG_VC1_DECODER) || defined(CONFIG_WMV3_DECODER)
+    ff_intrax8dsp_init(c,avctx);
+#endif
+#if defined(CONFIG_H264_ENCODER)
+    ff_h264dspenc_init(c,avctx);
+#endif
+
     c->put_mspel_pixels_tab[0]= put_mspel8_mc00_c;
     c->put_mspel_pixels_tab[1]= put_mspel8_mc10_c;
     c->put_mspel_pixels_tab[2]= put_mspel8_mc20_c;
@@ -3870,6 +4360,9 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->hadamard8_diff[4]= hadamard8_intra16_c;
     SET_CMP_FUNC(dct_sad)
     SET_CMP_FUNC(dct_max)
+#ifdef CONFIG_GPL
+    SET_CMP_FUNC(dct264_sad)
+#endif
     c->sad[0]= pix_abs16_c;
     c->sad[1]= pix_abs8_c;
     c->sse[0]= sse16_c;
@@ -3884,15 +4377,23 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->vsse[4]= vsse_intra16_c;
     c->nsse[0]= nsse16_c;
     c->nsse[1]= nsse8_c;
+#ifdef CONFIG_SNOW_ENCODER
     c->w53[0]= w53_16_c;
     c->w53[1]= w53_8_c;
     c->w97[0]= w97_16_c;
     c->w97[1]= w97_8_c;
+#endif
+
+    c->ssd_int8_vs_int16 = ssd_int8_vs_int16_c;
 
     c->add_bytes= add_bytes_c;
+    c->add_bytes_l2= add_bytes_l2_c;
     c->diff_bytes= diff_bytes_c;
     c->sub_hfyu_median_prediction= sub_hfyu_median_prediction_c;
     c->bswap_buf= bswap_buf;
+#ifdef CONFIG_PNG_DECODER
+    c->add_png_paeth_prediction= ff_add_png_paeth_prediction;
+#endif
 
     c->h264_v_loop_filter_luma= h264_v_loop_filter_luma_c;
     c->h264_h_loop_filter_luma= h264_h_loop_filter_luma_c;
@@ -3900,39 +4401,61 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->h264_h_loop_filter_chroma= h264_h_loop_filter_chroma_c;
     c->h264_v_loop_filter_chroma_intra= h264_v_loop_filter_chroma_intra_c;
     c->h264_h_loop_filter_chroma_intra= h264_h_loop_filter_chroma_intra_c;
+    c->h264_loop_filter_strength= NULL;
 
-    c->h263_h_loop_filter= h263_h_loop_filter_c;
-    c->h263_v_loop_filter= h263_v_loop_filter_c;
+    if (ENABLE_ANY_H263) {
+        c->h263_h_loop_filter= h263_h_loop_filter_c;
+        c->h263_v_loop_filter= h263_v_loop_filter_c;
+    }
 
     c->h261_loop_filter= h261_loop_filter_c;
 
     c->try_8x8basis= try_8x8basis_c;
     c->add_8x8basis= add_8x8basis_c;
 
-#ifdef HAVE_MMX
-    dsputil_init_mmx(c, avctx);
-#endif
-#ifdef ARCH_ARMV4L
-    dsputil_init_armv4l(c, avctx);
-#endif
-#ifdef HAVE_MLIB
-    dsputil_init_mlib(c, avctx);
+#ifdef CONFIG_SNOW_DECODER
+    c->vertical_compose97i = ff_snow_vertical_compose97i;
+    c->horizontal_compose97i = ff_snow_horizontal_compose97i;
+    c->inner_add_yblock = ff_snow_inner_add_yblock;
 #endif
-#ifdef ARCH_SPARC
-   dsputil_init_vis(c,avctx);
-#endif
-#ifdef ARCH_ALPHA
-    dsputil_init_alpha(c, avctx);
-#endif
-#ifdef ARCH_POWERPC
-    dsputil_init_ppc(c, avctx);
-#endif
-#ifdef HAVE_MMI
-    dsputil_init_mmi(c, avctx);
+
+#ifdef CONFIG_VORBIS_DECODER
+    c->vorbis_inverse_coupling = vorbis_inverse_coupling;
 #endif
-#ifdef ARCH_SH4
-    dsputil_init_sh4(c,avctx);
+#ifdef CONFIG_FLAC_ENCODER
+    c->flac_compute_autocorr = ff_flac_compute_autocorr;
 #endif
+    c->vector_fmul = vector_fmul_c;
+    c->vector_fmul_reverse = vector_fmul_reverse_c;
+    c->vector_fmul_add_add = ff_vector_fmul_add_add_c;
+    c->float_to_int16 = ff_float_to_int16_c;
+
+    c->shrink[0]= ff_img_copy_plane;
+    c->shrink[1]= ff_shrink22;
+    c->shrink[2]= ff_shrink44;
+    c->shrink[3]= ff_shrink88;
+
+    c->prefetch= just_return;
+
+    memset(c->put_2tap_qpel_pixels_tab, 0, sizeof(c->put_2tap_qpel_pixels_tab));
+    memset(c->avg_2tap_qpel_pixels_tab, 0, sizeof(c->avg_2tap_qpel_pixels_tab));
+
+    if (ENABLE_MMX)      dsputil_init_mmx   (c, avctx);
+    if (ENABLE_ARMV4L)   dsputil_init_armv4l(c, avctx);
+    if (ENABLE_MLIB)     dsputil_init_mlib  (c, avctx);
+    if (ENABLE_VIS)      dsputil_init_vis   (c, avctx);
+    if (ENABLE_ALPHA)    dsputil_init_alpha (c, avctx);
+    if (ENABLE_POWERPC)  dsputil_init_ppc   (c, avctx);
+    if (ENABLE_MMI)      dsputil_init_mmi   (c, avctx);
+    if (ENABLE_SH4)      dsputil_init_sh4   (c, avctx);
+    if (ENABLE_BFIN)     dsputil_init_bfin  (c, avctx);
+
+    for(i=0; i<64; i++){
+        if(!c->put_2tap_qpel_pixels_tab[0][i])
+            c->put_2tap_qpel_pixels_tab[0][i]= c->put_h264_qpel_pixels_tab[0][i];
+        if(!c->avg_2tap_qpel_pixels_tab[0][i])
+            c->avg_2tap_qpel_pixels_tab[0][i]= c->avg_h264_qpel_pixels_tab[0][i];
+    }
 
     switch(c->idct_permutation_type){
     case FF_NO_IDCT_PERM:
@@ -3955,6 +4478,10 @@ void dsputil_init(DSPContext* c, AVCodecContext *avctx)
         for(i=0; i<64; i++)
             c->idct_permutation[i]= (i&0x24) | ((i&3)<<3) | ((i>>3)&3);
         break;
+    case FF_SSE2_IDCT_PERM:
+        for(i=0; i<64; i++)
+            c->idct_permutation[i]= (i&0x38) | idct_sse2_row_perm[i&7];
+        break;
     default:
         av_log(avctx, AV_LOG_ERROR, "Internal error, IDCT permutation not set\n");
     }