X-Git-Url: https://git.sesse.net/?a=blobdiff_plain;f=libavcodec%2Fdsputil.c;h=5c705ea169af3d1dcd7dd909efb1c2818e6852f1;hb=cc5e9e5ff052fe31aa757de79f2d11fb21df3fba;hp=1627262c223390cf1ef26cb8597bab8c853c7264;hpb=2912e87a6c9264d556734e2bf94a99c64cf9b102;p=ffmpeg

diff --git a/libavcodec/dsputil.c b/libavcodec/dsputil.c
index 1627262c223..5c705ea169a 100644
--- a/libavcodec/dsputil.c
+++ b/libavcodec/dsputil.c
@@ -36,13 +36,22 @@
 #include "mathops.h"
 #include "mpegvideo.h"
 #include "config.h"
-#include "ac3dec.h"
 #include "vorbis.h"
-#include "png.h"
 
 uint8_t ff_cropTbl[256 + 2 * MAX_NEG_CROP] = {0, };
 uint32_t ff_squareTbl[512] = {0, };
 
+#define BIT_DEPTH 9
+#include "dsputil_template.c"
+#undef BIT_DEPTH
+
+#define BIT_DEPTH 10
+#include "dsputil_template.c"
+#undef BIT_DEPTH
+
+#define BIT_DEPTH 8
+#include "dsputil_template.c"
+
 // 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)
@@ -72,7 +81,7 @@ const uint8_t ff_zigzag248_direct[64] = {
 };
 
 /* not permutated inverse zigzag_direct + 1 for MMX quantizer */
-DECLARE_ALIGNED(16, uint16_t, inv_zigzag_direct16)[64];
+DECLARE_ALIGNED(16, uint16_t, ff_inv_zigzag_direct16)[64];
 
 const uint8_t ff_alternate_horizontal_scan[64] = {
     0,  1,   2,  3,  8,  9, 16, 17,
@@ -120,9 +129,6 @@ void ff_init_scantable(uint8_t *permutation, ScanTable *st, const uint8_t *src_s
         int j;
         j = src_scantable[i];
         st->permutated[i] = permutation[j];
-#if ARCH_PPC
-        st->inverse[j] = i;
-#endif
     }
 
     end=-1;
@@ -134,6 +140,41 @@ void ff_init_scantable(uint8_t *permutation, ScanTable *st, const uint8_t *src_s
     }
 }
 
+void ff_init_scantable_permutation(uint8_t *idct_permutation,
+                                   int idct_permutation_type)
+{
+    int i;
+
+    switch(idct_permutation_type){
+    case FF_NO_IDCT_PERM:
+        for(i=0; i<64; i++)
+            idct_permutation[i]= i;
+        break;
+    case FF_LIBMPEG2_IDCT_PERM:
+        for(i=0; i<64; i++)
+            idct_permutation[i]= (i & 0x38) | ((i & 6) >> 1) | ((i & 1) << 2);
+        break;
+    case FF_SIMPLE_IDCT_PERM:
+        for(i=0; i<64; i++)
+            idct_permutation[i]= simple_mmx_permutation[i];
+        break;
+    case FF_TRANSPOSE_IDCT_PERM:
+        for(i=0; i<64; i++)
+            idct_permutation[i]= ((i&7)<<3) | (i>>3);
+        break;
+    case FF_PARTTRANS_IDCT_PERM:
+        for(i=0; i<64; i++)
+            idct_permutation[i]= (i&0x24) | ((i&3)<<3) | ((i>>3)&3);
+        break;
+    case FF_SSE2_IDCT_PERM:
+        for(i=0; i<64; i++)
+            idct_permutation[i]= (i&0x38) | idct_sse2_row_perm[i&7];
+        break;
+    default:
+        av_log(NULL, AV_LOG_ERROR, "Internal error, IDCT permutation not set\n");
+    }
+}
+
 static int pix_sum_c(uint8_t * pix, int line_size)
 {
     int s, i, j;
@@ -174,7 +215,7 @@ static int pix_norm1_c(uint8_t * pix, int line_size)
             s += sq[pix[6]];
             s += sq[pix[7]];
 #else
-#if LONG_MAX > 2147483647
+#if HAVE_FAST_64BIT
             register uint64_t x=*(uint64_t*)pix;
             s += sq[x&0xff];
             s += sq[(x>>8)&0xff];
@@ -296,132 +337,6 @@ static int sse16_c(void *v, uint8_t *pix1, uint8_t *pix2, int line_size, int h)
     return s;
 }
 
-/* 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 */
-    }
-}
-
-/**
- * Copy a rectangular area of samples to a temporary buffer and replicate the border 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, const 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);
-    assert(start_y < end_y && block_h);
-    assert(start_x < end_x && block_w);
-
-    w    = end_x - start_x;
-    src += start_y*linesize + start_x;
-    buf += start_x;
-
-    //top
-    for(y=0; y<start_y; y++){
-        memcpy(buf, src, w);
-        buf += linesize;
-    }
-
-    // copy existing part
-    for(; y<end_y; y++){
-        memcpy(buf, src, w);
-        src += linesize;
-        buf += linesize;
-    }
-
-    //bottom
-    src -= linesize;
-    for(; y<block_h; y++){
-        memcpy(buf, src, w);
-        buf += linesize;
-    }
-
-    buf -= block_h * linesize + start_x;
-    while (block_h--){
-       //left
-        for(x=0; x<start_x; x++){
-            buf[x] = buf[start_x];
-        }
-
-       //right
-        for(x=end_x; x<block_w; x++){
-            buf[x] = buf[end_x - 1];
-        }
-        buf += linesize;
-    }
-}
-
-static void get_pixels_c(DCTELEM *restrict block, const uint8_t *pixels, int line_size)
-{
-    int i;
-
-    /* read the pixels */
-    for(i=0;i<8;i++) {
-        block[0] = pixels[0];
-        block[1] = pixels[1];
-        block[2] = pixels[2];
-        block[3] = pixels[3];
-        block[4] = pixels[4];
-        block[5] = pixels[5];
-        block[6] = pixels[6];
-        block[7] = pixels[7];
-        pixels += line_size;
-        block += 8;
-    }
-}
-
 static void diff_pixels_c(DCTELEM *restrict block, const uint8_t *s1,
                           const uint8_t *s2, int stride){
     int i;
@@ -443,65 +358,30 @@ static void diff_pixels_c(DCTELEM *restrict block, const uint8_t *s1,
 }
 
 
-void ff_put_pixels_clamped_c(const DCTELEM *block, uint8_t *restrict pixels,
-                             int line_size)
-{
-    int i;
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
-
-    /* read the pixels */
-    for(i=0;i<8;i++) {
-        pixels[0] = cm[block[0]];
-        pixels[1] = cm[block[1]];
-        pixels[2] = cm[block[2]];
-        pixels[3] = cm[block[3]];
-        pixels[4] = cm[block[4]];
-        pixels[5] = cm[block[5]];
-        pixels[6] = cm[block[6]];
-        pixels[7] = cm[block[7]];
-
-        pixels += line_size;
-        block += 8;
-    }
-}
-
-static void put_pixels_clamped4_c(const DCTELEM *block, uint8_t *restrict pixels,
-                                 int line_size)
-{
-    int i;
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
-
-    /* read the pixels */
-    for(i=0;i<4;i++) {
-        pixels[0] = cm[block[0]];
-        pixels[1] = cm[block[1]];
-        pixels[2] = cm[block[2]];
-        pixels[3] = cm[block[3]];
-
-        pixels += line_size;
-        block += 8;
-    }
-}
-
-static void put_pixels_clamped2_c(const DCTELEM *block, uint8_t *restrict pixels,
+static void put_pixels_clamped_c(const DCTELEM *block, uint8_t *restrict pixels,
                                  int line_size)
 {
     int i;
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
-    for(i=0;i<2;i++) {
-        pixels[0] = cm[block[0]];
-        pixels[1] = cm[block[1]];
+    for(i=0;i<8;i++) {
+        pixels[0] = av_clip_uint8(block[0]);
+        pixels[1] = av_clip_uint8(block[1]);
+        pixels[2] = av_clip_uint8(block[2]);
+        pixels[3] = av_clip_uint8(block[3]);
+        pixels[4] = av_clip_uint8(block[4]);
+        pixels[5] = av_clip_uint8(block[5]);
+        pixels[6] = av_clip_uint8(block[6]);
+        pixels[7] = av_clip_uint8(block[7]);
 
         pixels += line_size;
         block += 8;
     }
 }
 
-void ff_put_signed_pixels_clamped_c(const DCTELEM *block,
-                                    uint8_t *restrict pixels,
-                                    int line_size)
+static void put_signed_pixels_clamped_c(const DCTELEM *block,
+                                        uint8_t *restrict pixels,
+                                        int line_size)
 {
     int i, j;
 
@@ -520,110 +400,26 @@ void ff_put_signed_pixels_clamped_c(const DCTELEM *block,
     }
 }
 
-static void put_pixels_nonclamped_c(const DCTELEM *block, uint8_t *restrict pixels,
-                                    int line_size)
-{
-    int i;
-
-    /* read the pixels */
-    for(i=0;i<8;i++) {
-        pixels[0] = block[0];
-        pixels[1] = block[1];
-        pixels[2] = block[2];
-        pixels[3] = block[3];
-        pixels[4] = block[4];
-        pixels[5] = block[5];
-        pixels[6] = block[6];
-        pixels[7] = block[7];
-
-        pixels += line_size;
-        block += 8;
-    }
-}
-
-void ff_add_pixels_clamped_c(const DCTELEM *block, uint8_t *restrict pixels,
-                             int line_size)
-{
-    int i;
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
-
-    /* read the pixels */
-    for(i=0;i<8;i++) {
-        pixels[0] = cm[pixels[0] + block[0]];
-        pixels[1] = cm[pixels[1] + block[1]];
-        pixels[2] = cm[pixels[2] + block[2]];
-        pixels[3] = cm[pixels[3] + block[3]];
-        pixels[4] = cm[pixels[4] + block[4]];
-        pixels[5] = cm[pixels[5] + block[5]];
-        pixels[6] = cm[pixels[6] + block[6]];
-        pixels[7] = cm[pixels[7] + block[7]];
-        pixels += line_size;
-        block += 8;
-    }
-}
-
-static void add_pixels_clamped4_c(const DCTELEM *block, uint8_t *restrict pixels,
-                          int line_size)
-{
-    int i;
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
-
-    /* read the pixels */
-    for(i=0;i<4;i++) {
-        pixels[0] = cm[pixels[0] + block[0]];
-        pixels[1] = cm[pixels[1] + block[1]];
-        pixels[2] = cm[pixels[2] + block[2]];
-        pixels[3] = cm[pixels[3] + block[3]];
-        pixels += line_size;
-        block += 8;
-    }
-}
-
-static void add_pixels_clamped2_c(const DCTELEM *block, uint8_t *restrict pixels,
-                          int line_size)
+static void add_pixels_clamped_c(const DCTELEM *block, uint8_t *restrict pixels,
+                                 int line_size)
 {
     int i;
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
 
     /* read the pixels */
-    for(i=0;i<2;i++) {
-        pixels[0] = cm[pixels[0] + block[0]];
-        pixels[1] = cm[pixels[1] + block[1]];
-        pixels += line_size;
-        block += 8;
-    }
-}
-
-static void add_pixels8_c(uint8_t *restrict pixels, DCTELEM *block, int line_size)
-{
-    int i;
     for(i=0;i<8;i++) {
-        pixels[0] += block[0];
-        pixels[1] += block[1];
-        pixels[2] += block[2];
-        pixels[3] += block[3];
-        pixels[4] += block[4];
-        pixels[5] += block[5];
-        pixels[6] += block[6];
-        pixels[7] += block[7];
+        pixels[0] = av_clip_uint8(pixels[0] + block[0]);
+        pixels[1] = av_clip_uint8(pixels[1] + block[1]);
+        pixels[2] = av_clip_uint8(pixels[2] + block[2]);
+        pixels[3] = av_clip_uint8(pixels[3] + block[3]);
+        pixels[4] = av_clip_uint8(pixels[4] + block[4]);
+        pixels[5] = av_clip_uint8(pixels[5] + block[5]);
+        pixels[6] = av_clip_uint8(pixels[6] + block[6]);
+        pixels[7] = av_clip_uint8(pixels[7] + block[7]);
         pixels += line_size;
         block += 8;
     }
 }
 
-static void add_pixels4_c(uint8_t *restrict pixels, DCTELEM *block, int line_size)
-{
-    int i;
-    for(i=0;i<4;i++) {
-        pixels[0] += block[0];
-        pixels[1] += block[1];
-        pixels[2] += block[2];
-        pixels[3] += block[3];
-        pixels += line_size;
-        block += 4;
-    }
-}
-
 static int sum_abs_dctelem_c(DCTELEM *block)
 {
     int sum=0, i;
@@ -644,563 +440,17 @@ static void fill_block16_c(uint8_t *block, uint8_t value, int line_size, int h)
 
 static void fill_block8_c(uint8_t *block, uint8_t value, int line_size, int h)
 {
-    int i;
-
-    for (i = 0; i < h; i++) {
-        memset(block, value, 8);
-        block += line_size;
-    }
-}
-
-static void scale_block_c(const uint8_t src[64]/*align 8*/, uint8_t *dst/*align 8*/, int linesize)
-{
-    int i, j;
-    uint16_t *dst1 = (uint16_t *) dst;
-    uint16_t *dst2 = (uint16_t *)(dst + linesize);
-
-    for (j = 0; j < 8; j++) {
-        for (i = 0; i < 8; i++) {
-            dst1[i] = dst2[i] = src[i] * 0x0101;
-        }
-        src  += 8;
-        dst1 += linesize;
-        dst2 += linesize;
-    }
-}
-
-#if 0
-
-#define PIXOP2(OPNAME, OP) \
-static void OPNAME ## _pixels(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-    int i;\
-    for(i=0; i<h; i++){\
-        OP(*((uint64_t*)block), AV_RN64(pixels));\
-        pixels+=line_size;\
-        block +=line_size;\
-    }\
-}\
-\
-static void OPNAME ## _no_rnd_pixels_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-    int i;\
-    for(i=0; i<h; i++){\
-        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;\
-    }\
-}\
-\
-static void OPNAME ## _pixels_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-    int i;\
-    for(i=0; i<h; i++){\
-        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;\
-    }\
-}\
-\
-static void OPNAME ## _no_rnd_pixels_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-    int i;\
-    for(i=0; i<h; i++){\
-        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;\
-    }\
-}\
-\
-static void OPNAME ## _pixels_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-    int i;\
-    for(i=0; i<h; i++){\
-        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;\
-    }\
-}\
-\
-static void OPNAME ## _pixels_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-        int i;\
-        const uint64_t a= AV_RN64(pixels  );\
-        const uint64_t b= AV_RN64(pixels+1);\
-        uint64_t l0=  (a&0x0303030303030303ULL)\
-                    + (b&0x0303030303030303ULL)\
-                    + 0x0202020202020202ULL;\
-        uint64_t h0= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
-                   + ((b&0xFCFCFCFCFCFCFCFCULL)>>2);\
-        uint64_t l1,h1;\
-\
-        pixels+=line_size;\
-        for(i=0; i<h; i+=2){\
-            uint64_t a= AV_RN64(pixels  );\
-            uint64_t b= AV_RN64(pixels+1);\
-            l1=  (a&0x0303030303030303ULL)\
-               + (b&0x0303030303030303ULL);\
-            h1= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
-              + ((b&0xFCFCFCFCFCFCFCFCULL)>>2);\
-            OP(*((uint64_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0F0F0F0F0FULL));\
-            pixels+=line_size;\
-            block +=line_size;\
-            a= AV_RN64(pixels  );\
-            b= AV_RN64(pixels+1);\
-            l0=  (a&0x0303030303030303ULL)\
-               + (b&0x0303030303030303ULL)\
-               + 0x0202020202020202ULL;\
-            h0= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
-              + ((b&0xFCFCFCFCFCFCFCFCULL)>>2);\
-            OP(*((uint64_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0F0F0F0F0FULL));\
-            pixels+=line_size;\
-            block +=line_size;\
-        }\
-}\
-\
-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= AV_RN64(pixels  );\
-        const uint64_t b= AV_RN64(pixels+1);\
-        uint64_t l0=  (a&0x0303030303030303ULL)\
-                    + (b&0x0303030303030303ULL)\
-                    + 0x0101010101010101ULL;\
-        uint64_t h0= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
-                   + ((b&0xFCFCFCFCFCFCFCFCULL)>>2);\
-        uint64_t l1,h1;\
-\
-        pixels+=line_size;\
-        for(i=0; i<h; i+=2){\
-            uint64_t a= AV_RN64(pixels  );\
-            uint64_t b= AV_RN64(pixels+1);\
-            l1=  (a&0x0303030303030303ULL)\
-               + (b&0x0303030303030303ULL);\
-            h1= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
-              + ((b&0xFCFCFCFCFCFCFCFCULL)>>2);\
-            OP(*((uint64_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0F0F0F0F0FULL));\
-            pixels+=line_size;\
-            block +=line_size;\
-            a= AV_RN64(pixels  );\
-            b= AV_RN64(pixels+1);\
-            l0=  (a&0x0303030303030303ULL)\
-               + (b&0x0303030303030303ULL)\
-               + 0x0101010101010101ULL;\
-            h0= ((a&0xFCFCFCFCFCFCFCFCULL)>>2)\
-              + ((b&0xFCFCFCFCFCFCFCFCULL)>>2);\
-            OP(*((uint64_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0F0F0F0F0FULL));\
-            pixels+=line_size;\
-            block +=line_size;\
-        }\
-}\
-\
-CALL_2X_PIXELS(OPNAME ## _pixels16_c    , OPNAME ## _pixels_c    , 8)\
-CALL_2X_PIXELS(OPNAME ## _pixels16_x2_c , OPNAME ## _pixels_x2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _pixels16_y2_c , OPNAME ## _pixels_y2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _pixels16_xy2_c, OPNAME ## _pixels_xy2_c, 8)\
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_x2_c , OPNAME ## _no_rnd_pixels_x2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_y2_c , OPNAME ## _no_rnd_pixels_y2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_xy2_c, OPNAME ## _no_rnd_pixels_xy2_c, 8)
-
-#define op_avg(a, b) a = ( ((a)|(b)) - ((((a)^(b))&0xFEFEFEFEFEFEFEFEULL)>>1) )
-#else // 64 bit variant
-
-#define PIXOP2(OPNAME, OP) \
-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  )), AV_RN16(pixels  ));\
-        pixels+=line_size;\
-        block +=line_size;\
-    }\
-}\
-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  )), AV_RN32(pixels  ));\
-        pixels+=line_size;\
-        block +=line_size;\
-    }\
-}\
-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  )), AV_RN32(pixels  ));\
-        OP(*((uint32_t*)(block+4)), AV_RN32(pixels+4));\
-        pixels+=line_size;\
-        block +=line_size;\
-    }\
-}\
-static inline void OPNAME ## _no_rnd_pixels8_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _pixels8_c(block, pixels, line_size, h);\
-}\
-\
-static inline void OPNAME ## _no_rnd_pixels8_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
-                                                int src_stride1, int src_stride2, int h){\
-    int i;\
-    for(i=0; i<h; i++){\
-        uint32_t a,b;\
-        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= 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));\
-    }\
-}\
-\
-static inline void OPNAME ## _pixels8_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
-                                                int src_stride1, int src_stride2, int h){\
-    int i;\
-    for(i=0; i<h; i++){\
-        uint32_t a,b;\
-        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= 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));\
-    }\
-}\
-\
-static inline void OPNAME ## _pixels4_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
-                                                int src_stride1, int src_stride2, int h){\
-    int i;\
-    for(i=0; i<h; i++){\
-        uint32_t a,b;\
-        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));\
-    }\
-}\
-\
-static inline void OPNAME ## _pixels2_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
-                                                int src_stride1, int src_stride2, int h){\
-    int i;\
-    for(i=0; i<h; i++){\
-        uint32_t a,b;\
-        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));\
-    }\
-}\
-\
-static inline void OPNAME ## _pixels16_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
-                                                int src_stride1, int src_stride2, int h){\
-    OPNAME ## _pixels8_l2(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);\
-    OPNAME ## _pixels8_l2(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);\
-}\
-\
-static inline void OPNAME ## _no_rnd_pixels16_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
-                                                int src_stride1, int src_stride2, int h){\
-    OPNAME ## _no_rnd_pixels8_l2(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);\
-    OPNAME ## _no_rnd_pixels8_l2(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);\
-}\
-\
-static inline void OPNAME ## _no_rnd_pixels8_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _no_rnd_pixels8_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _pixels8_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _pixels8_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _no_rnd_pixels8_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _no_rnd_pixels8_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _pixels8_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _pixels8_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _pixels8_l4(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
-                 int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
-    int i;\
-    for(i=0; i<h; i++){\
-        uint32_t a, b, c, d, l0, l1, h0, h1;\
-        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;\
-        h0= ((a&0xFCFCFCFCUL)>>2)\
-          + ((b&0xFCFCFCFCUL)>>2);\
-        l1=  (c&0x03030303UL)\
-           + (d&0x03030303UL);\
-        h1= ((c&0xFCFCFCFCUL)>>2)\
-          + ((d&0xFCFCFCFCUL)>>2);\
-        OP(*((uint32_t*)&dst[i*dst_stride]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-        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;\
-        h0= ((a&0xFCFCFCFCUL)>>2)\
-          + ((b&0xFCFCFCFCUL)>>2);\
-        l1=  (c&0x03030303UL)\
-           + (d&0x03030303UL);\
-        h1= ((c&0xFCFCFCFCUL)>>2)\
-          + ((d&0xFCFCFCFCUL)>>2);\
-        OP(*((uint32_t*)&dst[i*dst_stride+4]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-    }\
-}\
-\
-static inline void OPNAME ## _pixels4_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _pixels4_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _pixels4_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _pixels4_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _pixels2_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _pixels2_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _pixels2_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
-    OPNAME ## _pixels2_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
-}\
-\
-static inline void OPNAME ## _no_rnd_pixels8_l4(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
-                 int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
-    int i;\
-    for(i=0; i<h; i++){\
-        uint32_t a, b, c, d, l0, l1, h0, h1;\
-        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;\
-        h0= ((a&0xFCFCFCFCUL)>>2)\
-          + ((b&0xFCFCFCFCUL)>>2);\
-        l1=  (c&0x03030303UL)\
-           + (d&0x03030303UL);\
-        h1= ((c&0xFCFCFCFCUL)>>2)\
-          + ((d&0xFCFCFCFCUL)>>2);\
-        OP(*((uint32_t*)&dst[i*dst_stride]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-        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;\
-        h0= ((a&0xFCFCFCFCUL)>>2)\
-          + ((b&0xFCFCFCFCUL)>>2);\
-        l1=  (c&0x03030303UL)\
-           + (d&0x03030303UL);\
-        h1= ((c&0xFCFCFCFCUL)>>2)\
-          + ((d&0xFCFCFCFCUL)>>2);\
-        OP(*((uint32_t*)&dst[i*dst_stride+4]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-    }\
-}\
-static inline void OPNAME ## _pixels16_l4(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
-                 int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
-    OPNAME ## _pixels8_l4(dst  , src1  , src2  , src3  , src4  , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
-    OPNAME ## _pixels8_l4(dst+8, src1+8, src2+8, src3+8, src4+8, dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
-}\
-static inline void OPNAME ## _no_rnd_pixels16_l4(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
-                 int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
-    OPNAME ## _no_rnd_pixels8_l4(dst  , src1  , src2  , src3  , src4  , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
-    OPNAME ## _no_rnd_pixels8_l4(dst+8, src1+8, src2+8, src3+8, src4+8, dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
-}\
-\
-static inline void OPNAME ## _pixels2_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-        int i, a0, b0, a1, b1;\
-        a0= pixels[0];\
-        b0= pixels[1] + 2;\
-        a0 += b0;\
-        b0 += pixels[2];\
-\
-        pixels+=line_size;\
-        for(i=0; i<h; i+=2){\
-            a1= pixels[0];\
-            b1= pixels[1];\
-            a1 += b1;\
-            b1 += pixels[2];\
-\
-            block[0]= (a1+a0)>>2; /* FIXME non put */\
-            block[1]= (b1+b0)>>2;\
-\
-            pixels+=line_size;\
-            block +=line_size;\
-\
-            a0= pixels[0];\
-            b0= pixels[1] + 2;\
-            a0 += b0;\
-            b0 += pixels[2];\
-\
-            block[0]= (a1+a0)>>2;\
-            block[1]= (b1+b0)>>2;\
-            pixels+=line_size;\
-            block +=line_size;\
-        }\
-}\
-\
-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= AV_RN32(pixels  );\
-        const uint32_t b= AV_RN32(pixels+1);\
-        uint32_t l0=  (a&0x03030303UL)\
-                    + (b&0x03030303UL)\
-                    + 0x02020202UL;\
-        uint32_t h0= ((a&0xFCFCFCFCUL)>>2)\
-                   + ((b&0xFCFCFCFCUL)>>2);\
-        uint32_t l1,h1;\
-\
-        pixels+=line_size;\
-        for(i=0; i<h; i+=2){\
-            uint32_t a= AV_RN32(pixels  );\
-            uint32_t b= AV_RN32(pixels+1);\
-            l1=  (a&0x03030303UL)\
-               + (b&0x03030303UL);\
-            h1= ((a&0xFCFCFCFCUL)>>2)\
-              + ((b&0xFCFCFCFCUL)>>2);\
-            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-            pixels+=line_size;\
-            block +=line_size;\
-            a= AV_RN32(pixels  );\
-            b= AV_RN32(pixels+1);\
-            l0=  (a&0x03030303UL)\
-               + (b&0x03030303UL)\
-               + 0x02020202UL;\
-            h0= ((a&0xFCFCFCFCUL)>>2)\
-              + ((b&0xFCFCFCFCUL)>>2);\
-            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-            pixels+=line_size;\
-            block +=line_size;\
-        }\
-}\
-\
-static inline void OPNAME ## _pixels8_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-    int j;\
-    for(j=0; j<2; j++){\
-        int i;\
-        const uint32_t a= AV_RN32(pixels  );\
-        const uint32_t b= AV_RN32(pixels+1);\
-        uint32_t l0=  (a&0x03030303UL)\
-                    + (b&0x03030303UL)\
-                    + 0x02020202UL;\
-        uint32_t h0= ((a&0xFCFCFCFCUL)>>2)\
-                   + ((b&0xFCFCFCFCUL)>>2);\
-        uint32_t l1,h1;\
-\
-        pixels+=line_size;\
-        for(i=0; i<h; i+=2){\
-            uint32_t a= AV_RN32(pixels  );\
-            uint32_t b= AV_RN32(pixels+1);\
-            l1=  (a&0x03030303UL)\
-               + (b&0x03030303UL);\
-            h1= ((a&0xFCFCFCFCUL)>>2)\
-              + ((b&0xFCFCFCFCUL)>>2);\
-            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-            pixels+=line_size;\
-            block +=line_size;\
-            a= AV_RN32(pixels  );\
-            b= AV_RN32(pixels+1);\
-            l0=  (a&0x03030303UL)\
-               + (b&0x03030303UL)\
-               + 0x02020202UL;\
-            h0= ((a&0xFCFCFCFCUL)>>2)\
-              + ((b&0xFCFCFCFCUL)>>2);\
-            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-            pixels+=line_size;\
-            block +=line_size;\
-        }\
-        pixels+=4-line_size*(h+1);\
-        block +=4-line_size*h;\
-    }\
-}\
-\
-static inline void OPNAME ## _no_rnd_pixels8_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
-{\
-    int j;\
-    for(j=0; j<2; j++){\
-        int i;\
-        const uint32_t a= AV_RN32(pixels  );\
-        const uint32_t b= AV_RN32(pixels+1);\
-        uint32_t l0=  (a&0x03030303UL)\
-                    + (b&0x03030303UL)\
-                    + 0x01010101UL;\
-        uint32_t h0= ((a&0xFCFCFCFCUL)>>2)\
-                   + ((b&0xFCFCFCFCUL)>>2);\
-        uint32_t l1,h1;\
-\
-        pixels+=line_size;\
-        for(i=0; i<h; i+=2){\
-            uint32_t a= AV_RN32(pixels  );\
-            uint32_t b= AV_RN32(pixels+1);\
-            l1=  (a&0x03030303UL)\
-               + (b&0x03030303UL);\
-            h1= ((a&0xFCFCFCFCUL)>>2)\
-              + ((b&0xFCFCFCFCUL)>>2);\
-            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-            pixels+=line_size;\
-            block +=line_size;\
-            a= AV_RN32(pixels  );\
-            b= AV_RN32(pixels+1);\
-            l0=  (a&0x03030303UL)\
-               + (b&0x03030303UL)\
-               + 0x01010101UL;\
-            h0= ((a&0xFCFCFCFCUL)>>2)\
-              + ((b&0xFCFCFCFCUL)>>2);\
-            OP(*((uint32_t*)block), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
-            pixels+=line_size;\
-            block +=line_size;\
-        }\
-        pixels+=4-line_size*(h+1);\
-        block +=4-line_size*h;\
-    }\
-}\
-\
-CALL_2X_PIXELS(OPNAME ## _pixels16_c  , OPNAME ## _pixels8_c  , 8)\
-CALL_2X_PIXELS(OPNAME ## _pixels16_x2_c , OPNAME ## _pixels8_x2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _pixels16_y2_c , OPNAME ## _pixels8_y2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _pixels16_xy2_c, OPNAME ## _pixels8_xy2_c, 8)\
-av_unused CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_c  , OPNAME ## _pixels8_c         , 8)\
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_x2_c , OPNAME ## _no_rnd_pixels8_x2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_y2_c , OPNAME ## _no_rnd_pixels8_y2_c , 8)\
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_xy2_c, OPNAME ## _no_rnd_pixels8_xy2_c, 8)\
-
-#define op_avg(a, b) a = rnd_avg32(a, b)
-#endif
-#define op_put(a, b) a = b
-
-PIXOP2(avg, op_avg)
-PIXOP2(put, op_put)
-#undef op_avg
-#undef op_put
+    int i;
 
-#define put_no_rnd_pixels8_c  put_pixels8_c
-#define put_no_rnd_pixels16_c put_pixels16_c
+    for (i = 0; i < h; i++) {
+        memset(block, value, 8);
+        block += line_size;
+    }
+}
 
 #define avg2(a,b) ((a+b+1)>>1)
 #define avg4(a,b,c,d) ((a+b+c+d+2)>>2)
 
-static void put_no_rnd_pixels16_l2_c(uint8_t *dst, const uint8_t *a, const uint8_t *b, int stride, int h){
-    put_no_rnd_pixels16_l2(dst, a, b, stride, stride, stride, h);
-}
-
-static void put_no_rnd_pixels8_l2_c(uint8_t *dst, const uint8_t *a, const uint8_t *b, int stride, int h){
-    put_no_rnd_pixels8_l2(dst, a, b, stride, stride, stride, h);
-}
-
 static void gmc1_c(uint8_t *dst, uint8_t *src, int stride, int h, int x16, int y16, int rounder)
 {
     const int A=(16-x16)*(16-y16);
@@ -1284,10 +534,10 @@ void ff_gmc_c(uint8_t *dst, uint8_t *src, int stride, int h, int ox, int oy,
 
 static inline void put_tpel_pixels_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
     switch(width){
-    case 2: put_pixels2_c (dst, src, stride, height); break;
-    case 4: put_pixels4_c (dst, src, stride, height); break;
-    case 8: put_pixels8_c (dst, src, stride, height); break;
-    case 16:put_pixels16_c(dst, src, stride, height); break;
+    case 2: put_pixels2_8_c (dst, src, stride, height); break;
+    case 4: put_pixels4_8_c (dst, src, stride, height); break;
+    case 8: put_pixels8_8_c (dst, src, stride, height); break;
+    case 16:put_pixels16_8_c(dst, src, stride, height); break;
     }
 }
 
@@ -1381,10 +631,10 @@ static inline void put_tpel_pixels_mc22_c(uint8_t *dst, const uint8_t *src, int
 
 static inline void avg_tpel_pixels_mc00_c(uint8_t *dst, const uint8_t *src, int stride, int width, int height){
     switch(width){
-    case 2: avg_pixels2_c (dst, src, stride, height); break;
-    case 4: avg_pixels4_c (dst, src, stride, height); break;
-    case 8: avg_pixels8_c (dst, src, stride, height); break;
-    case 16:avg_pixels16_c(dst, src, stride, height); break;
+    case 2: avg_pixels2_8_c (dst, src, stride, height); break;
+    case 4: avg_pixels4_8_c (dst, src, stride, height); break;
+    case 8: avg_pixels8_8_c (dst, src, stride, height); break;
+    case 16:avg_pixels16_8_c(dst, src, stride, height); break;
     }
 }
 
@@ -1475,136 +725,6 @@ static inline void avg_tpel_pixels_mc22_c(uint8_t *dst, const uint8_t *src, int
       dst += stride;
     }
 }
-#if 0
-#define TPEL_WIDTH(width)\
-static void put_tpel_pixels ## width ## _mc00_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc00_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc10_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc10_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc20_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc20_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc01_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc01_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc11_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc11_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc21_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc21_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc02_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc02_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc12_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc12_c(dst, src, stride, width, height);}\
-static void put_tpel_pixels ## width ## _mc22_c(uint8_t *dst, const uint8_t *src, int stride, int height){\
-    void put_tpel_pixels_mc22_c(dst, src, stride, width, height);}
-#endif
-
-#define H264_CHROMA_MC(OPNAME, OP)\
-static void OPNAME ## h264_chroma_mc2_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;\
-    \
-    assert(x<8 && y<8 && x>=0 && y>=0);\
-\
-    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;\
-        }\
-    }\
-}\
-\
-static void OPNAME ## h264_chroma_mc4_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;\
-    \
-    assert(x<8 && y<8 && x>=0 && y>=0);\
-\
-    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;\
-        }\
-    }\
-}\
-\
-static void OPNAME ## 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;\
-    \
-    assert(x<8 && y<8 && x>=0 && y>=0);\
-\
-    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;\
-        }\
-    }\
-}
-
-#define op_avg(a, b) a = (((a)+(((b) + 32)>>6)+1)>>1)
-#define op_put(a, b) a = (((b) + 32)>>6)
-
-H264_CHROMA_MC(put_       , op_put)
-H264_CHROMA_MC(avg_       , op_avg)
-#undef op_avg
-#undef op_put
 
 #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){\
@@ -1727,7 +847,7 @@ static void OPNAME ## mpeg4_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dst
 static void OPNAME ## qpel8_mc10_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[64];\
     put ## RND ## mpeg4_qpel8_h_lowpass(half, src, 8, stride, 8);\
-    OPNAME ## pixels8_l2(dst, src, half, stride, stride, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, src, half, stride, stride, 8, 8);\
 }\
 \
 static void OPNAME ## qpel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1737,7 +857,7 @@ static void OPNAME ## qpel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
 static void OPNAME ## qpel8_mc30_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[64];\
     put ## RND ## mpeg4_qpel8_h_lowpass(half, src, 8, stride, 8);\
-    OPNAME ## pixels8_l2(dst, src+1, half, stride, stride, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, src+1, half, stride, stride, 8, 8);\
 }\
 \
 static void OPNAME ## qpel8_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1745,7 +865,7 @@ static void OPNAME ## qpel8_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[64];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(half, full, 8, 16);\
-    OPNAME ## pixels8_l2(dst, full, half, stride, 16, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, full, half, stride, 16, 8, 8);\
 }\
 \
 static void OPNAME ## qpel8_mc02_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1759,7 +879,7 @@ static void OPNAME ## qpel8_mc03_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[64];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(half, full, 8, 16);\
-    OPNAME ## pixels8_l2(dst, full+16, half, stride, 16, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, full+16, half, stride, 16, 8, 8);\
 }\
 void ff_ ## OPNAME ## qpel8_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1770,7 +890,7 @@ void ff_ ## OPNAME ## qpel8_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfV, full, 8, 16);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l4(dst, full, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
+    OPNAME ## pixels8_l4_8(dst, full, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1778,9 +898,9 @@ static void OPNAME ## qpel8_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t halfHV[64];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
-    put ## RND ## pixels8_l2(halfH, halfH, full, 8, 8, 16, 9);\
+    put ## RND ## pixels8_l2_8(halfH, halfH, full, 8, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfH, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
 }\
 void ff_ ## OPNAME ## qpel8_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1791,7 +911,7 @@ void ff_ ## OPNAME ## qpel8_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfV, full+1, 8, 16);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l4(dst, full+1, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
+    OPNAME ## pixels8_l4_8(dst, full+1, halfH, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1799,9 +919,9 @@ static void OPNAME ## qpel8_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t halfHV[64];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
-    put ## RND ## pixels8_l2(halfH, halfH, full+1, 8, 8, 16, 9);\
+    put ## RND ## pixels8_l2_8(halfH, halfH, full+1, 8, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfH, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
 }\
 void ff_ ## OPNAME ## qpel8_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1812,7 +932,7 @@ void ff_ ## OPNAME ## qpel8_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfV, full, 8, 16);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l4(dst, full+16, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
+    OPNAME ## pixels8_l4_8(dst, full+16, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1820,9 +940,9 @@ static void OPNAME ## qpel8_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t halfHV[64];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
-    put ## RND ## pixels8_l2(halfH, halfH, full, 8, 8, 16, 9);\
+    put ## RND ## pixels8_l2_8(halfH, halfH, full, 8, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfH+8, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
 }\
 void ff_ ## OPNAME ## qpel8_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1833,7 +953,7 @@ void ff_ ## OPNAME ## qpel8_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full  , 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfV, full+1, 8, 16);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l4(dst, full+17, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
+    OPNAME ## pixels8_l4_8(dst, full+17, halfH+8, halfV, halfHV, stride, 16, 8, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1841,23 +961,23 @@ static void OPNAME ## qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t halfHV[64];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
-    put ## RND ## pixels8_l2(halfH, halfH, full+1, 8, 8, 16, 9);\
+    put ## RND ## pixels8_l2_8(halfH, halfH, full+1, 8, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfH+8, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc21_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t halfH[72];\
     uint8_t halfHV[64];\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfH, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfH, halfHV, stride, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc23_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t halfH[72];\
     uint8_t halfHV[64];\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, src, 8, stride, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfH+8, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfH+8, halfHV, stride, 8, 8, 8);\
 }\
 void ff_ ## OPNAME ## qpel8_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
@@ -1868,14 +988,14 @@ void ff_ ## OPNAME ## qpel8_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfV, full, 8, 16);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfV, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
     uint8_t halfH[72];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
-    put ## RND ## pixels8_l2(halfH, halfH, full, 8, 8, 16, 9);\
+    put ## RND ## pixels8_l2_8(halfH, halfH, full, 8, 8, 16, 9);\
     OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
 }\
 void ff_ ## OPNAME ## qpel8_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1887,14 +1007,14 @@ void ff_ ## OPNAME ## qpel8_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfV, full+1, 8, 16);\
     put ## RND ## mpeg4_qpel8_v_lowpass(halfHV, halfH, 8, 8);\
-    OPNAME ## pixels8_l2(dst, halfV, halfHV, stride, 8, 8, 8);\
+    OPNAME ## pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);\
 }\
 static void OPNAME ## qpel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[16*9];\
     uint8_t halfH[72];\
     copy_block9(full, src, 16, stride, 9);\
     put ## RND ## mpeg4_qpel8_h_lowpass(halfH, full, 8, 16, 9);\
-    put ## RND ## pixels8_l2(halfH, halfH, full+1, 8, 8, 16, 9);\
+    put ## RND ## pixels8_l2_8(halfH, halfH, full+1, 8, 8, 16, 9);\
     OPNAME ## mpeg4_qpel8_v_lowpass(dst, halfH, stride, 8);\
 }\
 static void OPNAME ## qpel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1906,7 +1026,7 @@ static void OPNAME ## qpel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
 static void OPNAME ## qpel16_mc10_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[256];\
     put ## RND ## mpeg4_qpel16_h_lowpass(half, src, 16, stride, 16);\
-    OPNAME ## pixels16_l2(dst, src, half, stride, stride, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, src, half, stride, stride, 16, 16);\
 }\
 \
 static void OPNAME ## qpel16_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1916,7 +1036,7 @@ static void OPNAME ## qpel16_mc20_c(uint8_t *dst, uint8_t *src, int stride){\
 static void OPNAME ## qpel16_mc30_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[256];\
     put ## RND ## mpeg4_qpel16_h_lowpass(half, src, 16, stride, 16);\
-    OPNAME ## pixels16_l2(dst, src+1, half, stride, stride, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, src+1, half, stride, stride, 16, 16);\
 }\
 \
 static void OPNAME ## qpel16_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1924,7 +1044,7 @@ static void OPNAME ## qpel16_mc01_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[256];\
     copy_block17(full, src, 24, stride, 17);\
     put ## RND ## mpeg4_qpel16_v_lowpass(half, full, 16, 24);\
-    OPNAME ## pixels16_l2(dst, full, half, stride, 24, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, full, half, stride, 24, 16, 16);\
 }\
 \
 static void OPNAME ## qpel16_mc02_c(uint8_t *dst, uint8_t *src, int stride){\
@@ -1938,7 +1058,7 @@ static void OPNAME ## qpel16_mc03_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t half[256];\
     copy_block17(full, src, 24, stride, 17);\
     put ## RND ## mpeg4_qpel16_v_lowpass(half, full, 16, 24);\
-    OPNAME ## pixels16_l2(dst, full+24, half, stride, 24, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, full+24, half, stride, 24, 16, 16);\
 }\
 void ff_ ## OPNAME ## qpel16_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[24*17];\
@@ -1949,7 +1069,7 @@ void ff_ ## OPNAME ## qpel16_mc11_old_c(uint8_t *dst, uint8_t *src, int stride){
     put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
     put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full, 16, 24);\
     put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l4(dst, full, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
+    OPNAME ## pixels16_l4_8(dst, full, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
 }\
 static void OPNAME ## qpel16_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[24*17];\
@@ -1957,9 +1077,9 @@ static void OPNAME ## qpel16_mc11_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t halfHV[256];\
     copy_block17(full, src, 24, stride, 17);\
     put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## pixels16_l2(halfH, halfH, full, 16, 16, 24, 17);\
+    put ## RND ## pixels16_l2_8(halfH, halfH, full, 16, 16, 24, 17);\
     put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfH, halfHV, stride, 16, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
 }\
 void ff_ ## OPNAME ## qpel16_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){\
     uint8_t full[24*17];\
@@ -1967,570 +1087,141 @@ void ff_ ## OPNAME ## qpel16_mc31_old_c(uint8_t *dst, uint8_t *src, int stride){
     uint8_t halfV[256];\
     uint8_t halfHV[256];\
     copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full+1, 16, 24);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l4(dst, full+1, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
-}\
-static void OPNAME ## qpel16_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    uint8_t halfHV[256];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## pixels16_l2(halfH, halfH, full+1, 16, 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfH, halfHV, stride, 16, 16, 16);\
-}\
-void ff_ ## OPNAME ## qpel16_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    uint8_t halfV[256];\
-    uint8_t halfHV[256];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full, 16, 24);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l4(dst, full+24, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
-}\
-static void OPNAME ## qpel16_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    uint8_t halfHV[256];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## pixels16_l2(halfH, halfH, full, 16, 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfH+16, halfHV, stride, 16, 16, 16);\
-}\
-void ff_ ## OPNAME ## qpel16_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    uint8_t halfV[256];\
-    uint8_t halfHV[256];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full  , 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full+1, 16, 24);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l4(dst, full+25, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
-}\
-static void OPNAME ## qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    uint8_t halfHV[256];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## pixels16_l2(halfH, halfH, full+1, 16, 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfH+16, halfHV, stride, 16, 16, 16);\
-}\
-static void OPNAME ## qpel16_mc21_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t halfH[272];\
-    uint8_t halfHV[256];\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfH, halfHV, stride, 16, 16, 16);\
-}\
-static void OPNAME ## qpel16_mc23_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t halfH[272];\
-    uint8_t halfHV[256];\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfH+16, halfHV, stride, 16, 16, 16);\
-}\
-void ff_ ## OPNAME ## qpel16_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    uint8_t halfV[256];\
-    uint8_t halfHV[256];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full, 16, 24);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfV, halfHV, stride, 16, 16, 16);\
-}\
-static void OPNAME ## qpel16_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## pixels16_l2(halfH, halfH, full, 16, 16, 24, 17);\
-    OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
-}\
-void ff_ ## OPNAME ## qpel16_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    uint8_t halfV[256];\
-    uint8_t halfHV[256];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full+1, 16, 24);\
-    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
-    OPNAME ## pixels16_l2(dst, halfV, halfHV, stride, 16, 16, 16);\
-}\
-static void OPNAME ## qpel16_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[24*17];\
-    uint8_t halfH[272];\
-    copy_block17(full, src, 24, stride, 17);\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
-    put ## RND ## pixels16_l2(halfH, halfH, full+1, 16, 16, 24, 17);\
-    OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
-}\
-static void OPNAME ## qpel16_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t halfH[272];\
-    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
-    OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
-}
-
-#define op_avg(a, b) a = (((a)+cm[((b) + 16)>>5]+1)>>1)
-#define op_avg_no_rnd(a, b) a = (((a)+cm[((b) + 15)>>5])>>1)
-#define op_put(a, b) a = cm[((b) + 16)>>5]
-#define op_put_no_rnd(a, b) a = cm[((b) + 15)>>5]
-
-QPEL_MC(0, put_       , _       , op_put)
-QPEL_MC(1, put_no_rnd_, _no_rnd_, op_put_no_rnd)
-QPEL_MC(0, avg_       , _       , op_avg)
-//QPEL_MC(1, avg_no_rnd , _       , op_avg)
-#undef op_avg
-#undef op_avg_no_rnd
-#undef op_put
-#undef op_put_no_rnd
-
-#define put_qpel8_mc00_c  ff_put_pixels8x8_c
-#define avg_qpel8_mc00_c  ff_avg_pixels8x8_c
-#define put_qpel16_mc00_c ff_put_pixels16x16_c
-#define avg_qpel16_mc00_c ff_avg_pixels16x16_c
-#define put_no_rnd_qpel8_mc00_c  ff_put_pixels8x8_c
-#define put_no_rnd_qpel16_mc00_c ff_put_pixels16x16_c
-
-#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 = 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]));\
-        OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]));\
-        OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]));\
-        dst+=dstStride;\
-        src+=srcStride;\
-    }\
-}\
-\
-static void OPNAME ## h264_qpel4_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
-    const int w=4;\
-    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];\
-        const int src5= src[5 *srcStride];\
-        const int src6= src[6 *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));\
-        OP(dst[2*dstStride], (src2+src3)*20 - (src1+src4)*5 + (src0+src5));\
-        OP(dst[3*dstStride], (src3+src4)*20 - (src2+src5)*5 + (src1+src6));\
-        dst++;\
-        src++;\
-    }\
-}\
-\
-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 = 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[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]);\
-        tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]);\
-        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];\
-        const int tmp5= tmp[5 *tmpStride];\
-        const int tmp6= tmp[6 *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));\
-        OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\
-        OP2(dst[3*dstStride], (tmp3+tmp4)*20 - (tmp2+tmp5)*5 + (tmp1+tmp6));\
-        dst++;\
-        tmp++;\
-    }\
-}\
-\
-static void OPNAME ## h264_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
-    const int h=8;\
-    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 ]));\
-        OP(dst[2], (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]));\
-        OP(dst[3], (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]));\
-        OP(dst[4], (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]));\
-        OP(dst[5], (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]));\
-        OP(dst[6], (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]));\
-        OP(dst[7], (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]));\
-        dst+=dstStride;\
-        src+=srcStride;\
-    }\
-}\
-\
-static void OPNAME ## h264_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
-    const int w=8;\
-    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];\
-        const int src5= src[5 *srcStride];\
-        const int src6= src[6 *srcStride];\
-        const int src7= src[7 *srcStride];\
-        const int src8= src[8 *srcStride];\
-        const int src9= src[9 *srcStride];\
-        const int src10=src[10*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));\
-        OP(dst[2*dstStride], (src2+src3)*20 - (src1+src4)*5 + (src0+src5));\
-        OP(dst[3*dstStride], (src3+src4)*20 - (src2+src5)*5 + (src1+src6));\
-        OP(dst[4*dstStride], (src4+src5)*20 - (src3+src6)*5 + (src2+src7));\
-        OP(dst[5*dstStride], (src5+src6)*20 - (src4+src7)*5 + (src3+src8));\
-        OP(dst[6*dstStride], (src6+src7)*20 - (src5+src8)*5 + (src4+src9));\
-        OP(dst[7*dstStride], (src7+src8)*20 - (src6+src9)*5 + (src5+src10));\
-        dst++;\
-        src++;\
-    }\
-}\
-\
-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 = 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[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]);\
-        tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]);\
-        tmp[4]= (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]);\
-        tmp[5]= (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]);\
-        tmp[6]= (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]);\
-        tmp[7]= (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]);\
-        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];\
-        const int tmp5= tmp[5 *tmpStride];\
-        const int tmp6= tmp[6 *tmpStride];\
-        const int tmp7= tmp[7 *tmpStride];\
-        const int tmp8= tmp[8 *tmpStride];\
-        const int tmp9= tmp[9 *tmpStride];\
-        const int tmp10=tmp[10*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));\
-        OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\
-        OP2(dst[3*dstStride], (tmp3+tmp4)*20 - (tmp2+tmp5)*5 + (tmp1+tmp6));\
-        OP2(dst[4*dstStride], (tmp4+tmp5)*20 - (tmp3+tmp6)*5 + (tmp2+tmp7));\
-        OP2(dst[5*dstStride], (tmp5+tmp6)*20 - (tmp4+tmp7)*5 + (tmp3+tmp8));\
-        OP2(dst[6*dstStride], (tmp6+tmp7)*20 - (tmp5+tmp8)*5 + (tmp4+tmp9));\
-        OP2(dst[7*dstStride], (tmp7+tmp8)*20 - (tmp6+tmp9)*5 + (tmp5+tmp10));\
-        dst++;\
-        tmp++;\
-    }\
-}\
-\
-static void OPNAME ## h264_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
-    OPNAME ## h264_qpel8_v_lowpass(dst  , src  , dstStride, srcStride);\
-    OPNAME ## h264_qpel8_v_lowpass(dst+8, src+8, dstStride, srcStride);\
-    src += 8*srcStride;\
-    dst += 8*dstStride;\
-    OPNAME ## h264_qpel8_v_lowpass(dst  , src  , dstStride, srcStride);\
-    OPNAME ## h264_qpel8_v_lowpass(dst+8, src+8, dstStride, srcStride);\
-}\
-\
-static void OPNAME ## h264_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
-    OPNAME ## h264_qpel8_h_lowpass(dst  , src  , dstStride, srcStride);\
-    OPNAME ## h264_qpel8_h_lowpass(dst+8, src+8, dstStride, srcStride);\
-    src += 8*srcStride;\
-    dst += 8*dstStride;\
-    OPNAME ## h264_qpel8_h_lowpass(dst  , src  , dstStride, srcStride);\
-    OPNAME ## h264_qpel8_h_lowpass(dst+8, src+8, dstStride, srcStride);\
-}\
-\
-static void OPNAME ## h264_qpel16_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
-    OPNAME ## h264_qpel8_hv_lowpass(dst  , tmp  , src  , dstStride, tmpStride, srcStride);\
-    OPNAME ## h264_qpel8_hv_lowpass(dst+8, tmp+8, src+8, dstStride, tmpStride, srcStride);\
-    src += 8*srcStride;\
-    dst += 8*dstStride;\
-    OPNAME ## h264_qpel8_hv_lowpass(dst  , tmp  , src  , dstStride, tmpStride, srcStride);\
-    OPNAME ## h264_qpel8_hv_lowpass(dst+8, tmp+8, src+8, dstStride, tmpStride, srcStride);\
-}\
-
-#define H264_MC(OPNAME, SIZE) \
-static av_unused void OPNAME ## h264_qpel ## SIZE ## _mc00_c (uint8_t *dst, uint8_t *src, int stride){\
-    OPNAME ## pixels ## SIZE ## _c(dst, src, stride, SIZE);\
-}\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t half[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(half, src, SIZE, stride);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, src, half, stride, stride, SIZE, SIZE);\
-}\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
-    OPNAME ## h264_qpel ## SIZE ## _h_lowpass(dst, src, stride, stride);\
-}\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t half[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(half, src, SIZE, stride);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, src+1, half, stride, stride, SIZE, SIZE);\
-}\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    uint8_t half[SIZE*SIZE];\
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(half, full_mid, SIZE, SIZE);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, full_mid, half, stride, SIZE, SIZE, SIZE);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full+1, 16, 24);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l4_8(dst, full+1, halfH, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
-    OPNAME ## h264_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE);\
+static void OPNAME ## qpel16_mc31_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    uint8_t halfHV[256];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## pixels16_l2_8(halfH, halfH, full+1, 16, 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    uint8_t half[SIZE*SIZE];\
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(half, full_mid, SIZE, SIZE);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, full_mid+SIZE, half, stride, SIZE, SIZE, SIZE);\
+void ff_ ## OPNAME ## qpel16_mc13_old_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    uint8_t halfV[256];\
+    uint8_t halfHV[256];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full, 16, 24);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l4_8(dst, full+24, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    uint8_t halfH[SIZE*SIZE];\
-    uint8_t halfV[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);\
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
+static void OPNAME ## qpel16_mc13_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    uint8_t halfHV[256];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## pixels16_l2_8(halfH, halfH, full, 16, 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    uint8_t halfH[SIZE*SIZE];\
-    uint8_t halfV[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);\
-    copy_block ## SIZE (full, src - stride*2 + 1, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
+void ff_ ## OPNAME ## qpel16_mc33_old_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    uint8_t halfV[256];\
+    uint8_t halfHV[256];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full  , 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full+1, 16, 24);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l4_8(dst, full+25, halfH+16, halfV, halfHV, stride, 24, 16, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    uint8_t halfH[SIZE*SIZE];\
-    uint8_t halfV[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);\
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
+static void OPNAME ## qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    uint8_t halfHV[256];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## pixels16_l2_8(halfH, halfH, full+1, 16, 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc33_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    uint8_t halfH[SIZE*SIZE];\
-    uint8_t halfV[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);\
-    copy_block ## SIZE (full, src - stride*2 + 1, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
+static void OPNAME ## qpel16_mc21_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t halfH[272];\
+    uint8_t halfHV[256];\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfH, halfHV, stride, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
-    int16_t tmp[SIZE*(SIZE+5)];\
-    OPNAME ## h264_qpel ## SIZE ## _hv_lowpass(dst, tmp, src, stride, SIZE, stride);\
+static void OPNAME ## qpel16_mc23_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t halfH[272];\
+    uint8_t halfHV[256];\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfH+16, halfHV, stride, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
-    int16_t tmp[SIZE*(SIZE+5)];\
-    uint8_t halfH[SIZE*SIZE];\
-    uint8_t halfHV[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);\
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfHV, stride, SIZE, SIZE, SIZE);\
+void ff_ ## OPNAME ## qpel16_mc12_old_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    uint8_t halfV[256];\
+    uint8_t halfHV[256];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full, 16, 24);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfV, halfHV, stride, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, int stride){\
-    int16_t tmp[SIZE*(SIZE+5)];\
-    uint8_t halfH[SIZE*SIZE];\
-    uint8_t halfHV[SIZE*SIZE];\
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);\
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfHV, stride, SIZE, SIZE, SIZE);\
+static void OPNAME ## qpel16_mc12_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## pixels16_l2_8(halfH, halfH, full, 16, 16, 24, 17);\
+    OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    int16_t tmp[SIZE*(SIZE+5)];\
-    uint8_t halfV[SIZE*SIZE];\
-    uint8_t halfHV[SIZE*SIZE];\
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfV, halfHV, stride, SIZE, SIZE, SIZE);\
+void ff_ ## OPNAME ## qpel16_mc32_old_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    uint8_t halfV[256];\
+    uint8_t halfHV[256];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfV, full+1, 16, 24);\
+    put ## RND ## mpeg4_qpel16_v_lowpass(halfHV, halfH, 16, 16);\
+    OPNAME ## pixels16_l2_8(dst, halfV, halfHV, stride, 16, 16, 16);\
 }\
-\
-static void OPNAME ## h264_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, int stride){\
-    uint8_t full[SIZE*(SIZE+5)];\
-    uint8_t * const full_mid= full + SIZE*2;\
-    int16_t tmp[SIZE*(SIZE+5)];\
-    uint8_t halfV[SIZE*SIZE];\
-    uint8_t halfHV[SIZE*SIZE];\
-    copy_block ## SIZE (full, src - stride*2 + 1, SIZE,  stride, SIZE + 5);\
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfV, halfHV, stride, SIZE, SIZE, SIZE);\
+static void OPNAME ## qpel16_mc32_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t full[24*17];\
+    uint8_t halfH[272];\
+    copy_block17(full, src, 24, stride, 17);\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, full, 16, 24, 17);\
+    put ## RND ## pixels16_l2_8(halfH, halfH, full+1, 16, 16, 24, 17);\
+    OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
 }\
+static void OPNAME ## qpel16_mc22_c(uint8_t *dst, uint8_t *src, int stride){\
+    uint8_t halfH[272];\
+    put ## RND ## mpeg4_qpel16_h_lowpass(halfH, src, 16, stride, 17);\
+    OPNAME ## mpeg4_qpel16_v_lowpass(dst, halfH, stride, 16);\
+}
 
-#define op_avg(a, b)  a = (((a)+cm[((b) + 16)>>5]+1)>>1)
-//#define op_avg2(a, b) a = (((a)*w1+cm[((b) + 16)>>5]*w2 + o + 64)>>7)
-#define op_put(a, b)  a = cm[((b) + 16)>>5]
-#define op2_avg(a, b)  a = (((a)+cm[((b) + 512)>>10]+1)>>1)
-#define op2_put(a, b)  a = cm[((b) + 512)>>10]
-
-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)
-H264_MC(avg_, 4)
-H264_MC(avg_, 8)
-H264_MC(avg_, 16)
+#define op_avg(a, b) a = (((a)+cm[((b) + 16)>>5]+1)>>1)
+#define op_avg_no_rnd(a, b) a = (((a)+cm[((b) + 15)>>5])>>1)
+#define op_put(a, b) a = cm[((b) + 16)>>5]
+#define op_put_no_rnd(a, b) a = cm[((b) + 15)>>5]
 
+QPEL_MC(0, put_       , _       , op_put)
+QPEL_MC(1, put_no_rnd_, _no_rnd_, op_put_no_rnd)
+QPEL_MC(0, avg_       , _       , op_avg)
+//QPEL_MC(1, avg_no_rnd , _       , op_avg)
 #undef op_avg
+#undef op_avg_no_rnd
 #undef op_put
-#undef op2_avg
-#undef op2_put
-#endif
+#undef op_put_no_rnd
 
-#define put_h264_qpel8_mc00_c  ff_put_pixels8x8_c
-#define avg_h264_qpel8_mc00_c  ff_avg_pixels8x8_c
-#define put_h264_qpel16_mc00_c ff_put_pixels16x16_c
-#define avg_h264_qpel16_mc00_c ff_avg_pixels16x16_c
+#define put_qpel8_mc00_c  ff_put_pixels8x8_c
+#define avg_qpel8_mc00_c  ff_avg_pixels8x8_c
+#define put_qpel16_mc00_c ff_put_pixels16x16_c
+#define avg_qpel16_mc00_c ff_avg_pixels16x16_c
+#define put_no_rnd_qpel8_mc00_c  ff_put_pixels8x8_c
+#define put_no_rnd_qpel16_mc00_c ff_put_pixels16x16_8_c
 
 static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride, int h){
     uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
@@ -2550,31 +1241,18 @@ static void wmv2_mspel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int
     }
 }
 
-void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride) {
-    put_pixels8_c(dst, src, stride, 8);
-}
-void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride) {
-    avg_pixels8_c(dst, src, stride, 8);
-}
-void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride) {
-    put_pixels16_c(dst, src, stride, 16);
-}
-void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride) {
-    avg_pixels16_c(dst, src, stride, 16);
-}
-
 #if CONFIG_RV40_DECODER
-static void put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
-    put_pixels16_xy2_c(dst, src, stride, 16);
+void ff_put_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+    put_pixels16_xy2_8_c(dst, src, stride, 16);
 }
-static void avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
-    avg_pixels16_xy2_c(dst, src, stride, 16);
+void ff_avg_rv40_qpel16_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+    avg_pixels16_xy2_8_c(dst, src, stride, 16);
 }
-static void put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
-    put_pixels8_xy2_c(dst, src, stride, 8);
+void ff_put_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+    put_pixels8_xy2_8_c(dst, src, stride, 8);
 }
-static void avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
-    avg_pixels8_xy2_c(dst, src, stride, 8);
+void ff_avg_rv40_qpel8_mc33_c(uint8_t *dst, uint8_t *src, int stride){
+    avg_pixels8_xy2_8_c(dst, src, stride, 8);
 }
 #endif /* CONFIG_RV40_DECODER */
 
@@ -2610,7 +1288,7 @@ static void wmv2_mspel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int
 static void put_mspel8_mc10_c(uint8_t *dst, uint8_t *src, int stride){
     uint8_t half[64];
     wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
-    put_pixels8_l2(dst, src, half, stride, stride, 8, 8);
+    put_pixels8_l2_8(dst, src, half, stride, stride, 8, 8);
 }
 
 static void put_mspel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){
@@ -2620,7 +1298,7 @@ static void put_mspel8_mc20_c(uint8_t *dst, uint8_t *src, int stride){
 static void put_mspel8_mc30_c(uint8_t *dst, uint8_t *src, int stride){
     uint8_t half[64];
     wmv2_mspel8_h_lowpass(half, src, 8, stride, 8);
-    put_pixels8_l2(dst, src+1, half, stride, stride, 8, 8);
+    put_pixels8_l2_8(dst, src+1, half, stride, stride, 8, 8);
 }
 
 static void put_mspel8_mc02_c(uint8_t *dst, uint8_t *src, int stride){
@@ -2634,7 +1312,7 @@ static void put_mspel8_mc12_c(uint8_t *dst, uint8_t *src, int stride){
     wmv2_mspel8_h_lowpass(halfH, src-stride, 8, stride, 11);
     wmv2_mspel8_v_lowpass(halfV, src, 8, stride, 8);
     wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
-    put_pixels8_l2(dst, halfV, halfHV, stride, 8, 8, 8);
+    put_pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);
 }
 static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){
     uint8_t halfH[88];
@@ -2643,7 +1321,7 @@ static void put_mspel8_mc32_c(uint8_t *dst, uint8_t *src, int stride){
     wmv2_mspel8_h_lowpass(halfH, src-stride, 8, stride, 11);
     wmv2_mspel8_v_lowpass(halfV, src+1, 8, stride, 8);
     wmv2_mspel8_v_lowpass(halfHV, halfH+8, 8, 8, 8);
-    put_pixels8_l2(dst, halfV, halfHV, stride, 8, 8, 8);
+    put_pixels8_l2_8(dst, halfV, halfHV, stride, 8, 8, 8);
 }
 static void put_mspel8_mc22_c(uint8_t *dst, uint8_t *src, int stride){
     uint8_t halfH[88];
@@ -3028,7 +1706,7 @@ static void add_8x8basis_c(int16_t rem[64], int16_t basis[64], int scale){
 }
 
 /**
- * permutes an 8x8 block.
+ * Permute an 8x8 block.
  * @param block the block which will be permuted according to the given permutation vector
  * @param permutation the permutation vector
  * @param last the last non zero coefficient in scantable order, used to speed the permutation up
@@ -3120,19 +1798,6 @@ void ff_set_cmp(DSPContext* c, me_cmp_func *cmp, int type){
     }
 }
 
-static void clear_block_c(DCTELEM *block)
-{
-    memset(block, 0, sizeof(DCTELEM)*64);
-}
-
-/**
- * memset(blocks, 0, sizeof(DCTELEM)*6*64)
- */
-static void clear_blocks_c(DCTELEM *blocks)
-{
-    memset(blocks, 0, sizeof(DCTELEM)*6*64);
-}
-
 static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
     long i;
     for(i=0; i<=w-sizeof(long); i+=sizeof(long)){
@@ -3144,17 +1809,6 @@ static void add_bytes_c(uint8_t *dst, uint8_t *src, int w){
         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){
     long i;
 #if !HAVE_FAST_UNALIGNED
@@ -3332,13 +1986,6 @@ static int hadamard8_diff8x8_c(/*MpegEncContext*/ void *s, uint8_t *dst, uint8_t
             +BUTTERFLYA(temp[8*2+i], temp[8*6+i])
             +BUTTERFLYA(temp[8*3+i], temp[8*7+i]);
     }
-#if 0
-static int maxi=0;
-if(sum>maxi){
-    maxi=sum;
-    printf("MAX:%d\n", maxi);
-}
-#endif
     return sum;
 }
 
@@ -3485,7 +2132,7 @@ 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);
-    ff_simple_idct(temp); //FIXME
+    ff_simple_idct_8(temp); //FIXME
 
     for(i=0; i<64; i++)
         sum+= (temp[i]-bak[i])*(temp[i]-bak[i]);
@@ -3715,12 +2362,6 @@ 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 *src0, const float *src1, int len){
-    int i;
-    for(i=0; i<len; i++)
-        dst[i] = src0[i] * src1[i];
-}
-
 static void vector_fmul_reverse_c(float *dst, const float *src0, const float *src1, int len){
     int i;
     src1 += len-1;
@@ -3759,50 +2400,6 @@ static void vector_fmul_scalar_c(float *dst, const float *src, float mul,
         dst[i] = src[i] * mul;
 }
 
-static void vector_fmul_sv_scalar_2_c(float *dst, const float *src,
-                                      const float **sv, float mul, int len)
-{
-    int i;
-    for (i = 0; i < len; i += 2, sv++) {
-        dst[i  ] = src[i  ] * sv[0][0] * mul;
-        dst[i+1] = src[i+1] * sv[0][1] * mul;
-    }
-}
-
-static void vector_fmul_sv_scalar_4_c(float *dst, const float *src,
-                                      const float **sv, float mul, int len)
-{
-    int i;
-    for (i = 0; i < len; i += 4, sv++) {
-        dst[i  ] = src[i  ] * sv[0][0] * mul;
-        dst[i+1] = src[i+1] * sv[0][1] * mul;
-        dst[i+2] = src[i+2] * sv[0][2] * mul;
-        dst[i+3] = src[i+3] * sv[0][3] * mul;
-    }
-}
-
-static void sv_fmul_scalar_2_c(float *dst, const float **sv, float mul,
-                               int len)
-{
-    int i;
-    for (i = 0; i < len; i += 2, sv++) {
-        dst[i  ] = sv[0][0] * mul;
-        dst[i+1] = sv[0][1] * mul;
-    }
-}
-
-static void sv_fmul_scalar_4_c(float *dst, const float **sv, float mul,
-                               int len)
-{
-    int i;
-    for (i = 0; i < len; i += 4, sv++) {
-        dst[i  ] = sv[0][0] * mul;
-        dst[i+1] = sv[0][1] * mul;
-        dst[i+2] = sv[0][2] * mul;
-        dst[i+3] = sv[0][3] * mul;
-    }
-}
-
 static void butterflies_float_c(float *restrict v1, float *restrict v2,
                                 int len)
 {
@@ -3814,7 +2411,19 @@ static void butterflies_float_c(float *restrict v1, float *restrict v2,
     }
 }
 
-static float scalarproduct_float_c(const float *v1, const float *v2, int len)
+static void butterflies_float_interleave_c(float *dst, const float *src0,
+                                           const float *src1, int len)
+{
+    int i;
+    for (i = 0; i < len; i++) {
+        float f1 = src0[i];
+        float f2 = src1[i];
+        dst[2*i    ] = f1 + f2;
+        dst[2*i + 1] = f1 - f2;
+    }
+}
+
+float ff_scalarproduct_float_c(const float *v1, const float *v2, int len)
 {
     float p = 0.0;
     int i;
@@ -3830,7 +2439,7 @@ static inline uint32_t clipf_c_one(uint32_t a, uint32_t mini,
 {
 
     if(a > mini) return mini;
-    else if((a^(1<<31)) > maxisign) return maxi;
+    else if((a^(1U<<31)) > maxisign) return maxi;
     else return a;
 }
 
@@ -3838,7 +2447,7 @@ static void vector_clipf_c_opposite_sign(float *dst, const float *src, float *mi
     int i;
     uint32_t mini = *(uint32_t*)min;
     uint32_t maxi = *(uint32_t*)max;
-    uint32_t maxisign = maxi ^ (1<<31);
+    uint32_t maxisign = maxi ^ (1U<<31);
     uint32_t *dsti = (uint32_t*)dst;
     const uint32_t *srci = (const uint32_t*)src;
     for(i=0; i<len; i+=8) {
@@ -3870,12 +2479,12 @@ static void vector_clipf_c(float *dst, const float *src, float min, float max, i
     }
 }
 
-static int32_t scalarproduct_int16_c(const int16_t * v1, const int16_t * v2, int order, int shift)
+static int32_t scalarproduct_int16_c(const int16_t * v1, const int16_t * v2, int order)
 {
     int res = 0;
 
     while (order--)
-        res += (*v1++ * *v2++) >> shift;
+        res += *v1++ * *v2++;
 
     return res;
 }
@@ -3890,6 +2499,35 @@ static int32_t scalarproduct_and_madd_int16_c(int16_t *v1, const int16_t *v2, co
     return res;
 }
 
+static void apply_window_int16_c(int16_t *output, const int16_t *input,
+                                 const int16_t *window, unsigned int len)
+{
+    int i;
+    int len2 = len >> 1;
+
+    for (i = 0; i < len2; i++) {
+        int16_t w       = window[i];
+        output[i]       = (MUL16(input[i],       w) + (1 << 14)) >> 15;
+        output[len-i-1] = (MUL16(input[len-i-1], w) + (1 << 14)) >> 15;
+    }
+}
+
+static void vector_clip_int32_c(int32_t *dst, const int32_t *src, int32_t min,
+                                int32_t max, unsigned int len)
+{
+    do {
+        *dst++ = av_clip(*src++, min, max);
+        *dst++ = av_clip(*src++, min, max);
+        *dst++ = av_clip(*src++, min, max);
+        *dst++ = av_clip(*src++, min, max);
+        *dst++ = av_clip(*src++, min, max);
+        *dst++ = av_clip(*src++, min, max);
+        *dst++ = av_clip(*src++, min, max);
+        *dst++ = av_clip(*src++, min, max);
+        len -= 8;
+    } while (len > 0);
+}
+
 #define W0 2048
 #define W1 2841 /* 2048*sqrt (2)*cos (1*pi/16) */
 #define W2 2676 /* 2048*sqrt (2)*cos (2*pi/16) */
@@ -3967,63 +2605,28 @@ void ff_wmv2_idct_c(short * block){
 static void ff_wmv2_idct_put_c(uint8_t *dest, int line_size, DCTELEM *block)
 {
     ff_wmv2_idct_c(block);
-    ff_put_pixels_clamped_c(block, dest, line_size);
+    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);
-    ff_add_pixels_clamped_c(block, dest, line_size);
+    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);
-    ff_put_pixels_clamped_c(block, dest, line_size);
+    ff_j_rev_dct (block);
+    put_pixels_clamped_c(block, dest, line_size);
 }
 static void ff_jref_idct_add(uint8_t *dest, int line_size, DCTELEM *block)
 {
-    j_rev_dct (block);
-    ff_add_pixels_clamped_c(block, dest, line_size);
-}
-
-static void ff_jref_idct4_put(uint8_t *dest, int line_size, DCTELEM *block)
-{
-    j_rev_dct4 (block);
-    put_pixels_clamped4_c(block, dest, line_size);
-}
-static void ff_jref_idct4_add(uint8_t *dest, int line_size, DCTELEM *block)
-{
-    j_rev_dct4 (block);
-    add_pixels_clamped4_c(block, dest, line_size);
-}
-
-static void ff_jref_idct2_put(uint8_t *dest, int line_size, DCTELEM *block)
-{
-    j_rev_dct2 (block);
-    put_pixels_clamped2_c(block, dest, line_size);
-}
-static void ff_jref_idct2_add(uint8_t *dest, int line_size, DCTELEM *block)
-{
-    j_rev_dct2 (block);
-    add_pixels_clamped2_c(block, dest, line_size);
-}
-
-static void ff_jref_idct1_put(uint8_t *dest, int line_size, DCTELEM *block)
-{
-    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 = ff_cropTbl + MAX_NEG_CROP;
-
-    dest[0] = cm[dest[0] + ((block[0] + 4)>>3)];
+    ff_j_rev_dct (block);
+    add_pixels_clamped_c(block, dest, line_size);
 }
 
 static void just_return(void *mem av_unused, int stride av_unused, int h av_unused) { return; }
 
 /* init static data */
-av_cold void dsputil_static_init(void)
+av_cold void ff_dsputil_static_init(void)
 {
     int i;
 
@@ -4037,21 +2640,21 @@ av_cold void dsputil_static_init(void)
         ff_squareTbl[i] = (i - 256) * (i - 256);
     }
 
-    for(i=0; i<64; i++) inv_zigzag_direct16[ff_zigzag_direct[i]]= i+1;
+    for(i=0; i<64; i++) ff_inv_zigzag_direct16[ff_zigzag_direct[i]]= i+1;
 }
 
 int ff_check_alignment(void){
     static int did_fail=0;
-    DECLARE_ALIGNED(16, int, aligned);
+    LOCAL_ALIGNED_16(int, aligned, [4]);
 
-    if((intptr_t)&aligned & 15){
+    if((intptr_t)aligned & 15){
         if(!did_fail){
 #if HAVE_MMX || 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");
+                "Do not report crashes to Libav developers.\n");
 #endif
             did_fail=1;
         }
@@ -4060,59 +2663,43 @@ int ff_check_alignment(void){
     return 0;
 }
 
-av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
+av_cold void ff_dsputil_init(DSPContext* c, AVCodecContext *avctx)
 {
-    int i;
+    int i, j;
 
     ff_check_alignment();
 
 #if CONFIG_ENCODERS
-    if(avctx->dct_algo==FF_DCT_FASTINT) {
-        c->fdct = fdct_ifast;
-        c->fdct248 = fdct_ifast248;
-    }
-    else if(avctx->dct_algo==FF_DCT_FAAN) {
-        c->fdct = ff_faandct;
-        c->fdct248 = ff_faandct248;
-    }
-    else {
-        c->fdct = ff_jpeg_fdct_islow; //slow/accurate/default
-        c->fdct248 = ff_fdct248_islow;
+    if (avctx->bits_per_raw_sample == 10) {
+        c->fdct    = ff_jpeg_fdct_islow_10;
+        c->fdct248 = ff_fdct248_islow_10;
+    } else {
+        if(avctx->dct_algo==FF_DCT_FASTINT) {
+            c->fdct    = ff_fdct_ifast;
+            c->fdct248 = ff_fdct_ifast248;
+        }
+        else if(avctx->dct_algo==FF_DCT_FAAN) {
+            c->fdct    = ff_faandct;
+            c->fdct248 = ff_faandct248;
+        }
+        else {
+            c->fdct    = ff_jpeg_fdct_islow_8; //slow/accurate/default
+            c->fdct248 = ff_fdct248_islow_8;
+        }
     }
 #endif //CONFIG_ENCODERS
 
-    if(avctx->lowres==1){
-        if(avctx->idct_algo==FF_IDCT_INT || avctx->idct_algo==FF_IDCT_AUTO || !CONFIG_H264_DECODER){
-            c->idct_put= ff_jref_idct4_put;
-            c->idct_add= ff_jref_idct4_add;
-        }else{
-            c->idct_put= ff_h264_lowres_idct_put_c;
-            c->idct_add= ff_h264_lowres_idct_add_c;
-        }
-        c->idct    = j_rev_dct4;
-        c->idct_permutation_type= FF_NO_IDCT_PERM;
-    }else if(avctx->lowres==2){
-        c->idct_put= ff_jref_idct2_put;
-        c->idct_add= ff_jref_idct2_add;
-        c->idct    = j_rev_dct2;
-        c->idct_permutation_type= FF_NO_IDCT_PERM;
-    }else if(avctx->lowres==3){
-        c->idct_put= ff_jref_idct1_put;
-        c->idct_add= ff_jref_idct1_add;
-        c->idct    = j_rev_dct1;
-        c->idct_permutation_type= FF_NO_IDCT_PERM;
-    }else{
+    if (avctx->bits_per_raw_sample == 10) {
+        c->idct_put              = ff_simple_idct_put_10;
+        c->idct_add              = ff_simple_idct_add_10;
+        c->idct                  = ff_simple_idct_10;
+        c->idct_permutation_type = FF_NO_IDCT_PERM;
+    } else {
         if(avctx->idct_algo==FF_IDCT_INT){
             c->idct_put= ff_jref_idct_put;
             c->idct_add= ff_jref_idct_add;
-            c->idct    = j_rev_dct;
+            c->idct    = ff_j_rev_dct;
             c->idct_permutation_type= FF_LIBMPEG2_IDCT_PERM;
-        }else if((CONFIG_VP3_DECODER || CONFIG_VP5_DECODER || CONFIG_VP6_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;
@@ -4123,42 +2710,26 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
             c->idct_add= ff_faanidct_add;
             c->idct    = ff_faanidct;
             c->idct_permutation_type= FF_NO_IDCT_PERM;
-        }else if(CONFIG_EATGQ_DECODER && avctx->idct_algo==FF_IDCT_EA) {
-            c->idct_put= ff_ea_idct_put_c;
-            c->idct_permutation_type= FF_NO_IDCT_PERM;
-        }else if(CONFIG_BINK_DECODER && avctx->idct_algo==FF_IDCT_BINK) {
-            c->idct     = ff_bink_idct_c;
-            c->idct_add = ff_bink_idct_add_c;
-            c->idct_put = ff_bink_idct_put_c;
-            c->idct_permutation_type = FF_NO_IDCT_PERM;
         }else{ //accurate/default
-            c->idct_put= ff_simple_idct_put;
-            c->idct_add= ff_simple_idct_add;
-            c->idct    = ff_simple_idct;
+            c->idct_put = ff_simple_idct_put_8;
+            c->idct_add = ff_simple_idct_add_8;
+            c->idct     = ff_simple_idct_8;
             c->idct_permutation_type= FF_NO_IDCT_PERM;
         }
     }
 
-    c->get_pixels = get_pixels_c;
     c->diff_pixels = diff_pixels_c;
-    c->put_pixels_clamped = ff_put_pixels_clamped_c;
-    c->put_signed_pixels_clamped = ff_put_signed_pixels_clamped_c;
-    c->put_pixels_nonclamped = put_pixels_nonclamped_c;
-    c->add_pixels_clamped = ff_add_pixels_clamped_c;
-    c->add_pixels8 = add_pixels8_c;
-    c->add_pixels4 = add_pixels4_c;
+    c->put_pixels_clamped = put_pixels_clamped_c;
+    c->put_signed_pixels_clamped = put_signed_pixels_clamped_c;
+    c->add_pixels_clamped = add_pixels_clamped_c;
     c->sum_abs_dctelem = sum_abs_dctelem_c;
-    c->emulated_edge_mc = ff_emulated_edge_mc;
     c->gmc1 = gmc1_c;
     c->gmc = ff_gmc_c;
-    c->clear_block = clear_block_c;
-    c->clear_blocks = clear_blocks_c;
     c->pix_sum = pix_sum_c;
     c->pix_norm1 = pix_norm1_c;
 
     c->fill_block_tab[0] = fill_block16_c;
     c->fill_block_tab[1] = fill_block8_c;
-    c->scale_block = scale_block_c;
 
     /* TODO [0] 16  [1] 8 */
     c->pix_abs[0][0] = pix_abs16_c;
@@ -4170,30 +2741,6 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->pix_abs[1][2] = pix_abs8_y2_c;
     c->pix_abs[1][3] = pix_abs8_xy2_c;
 
-#define dspfunc(PFX, IDX, NUM) \
-    c->PFX ## _pixels_tab[IDX][0] = PFX ## _pixels ## NUM ## _c;     \
-    c->PFX ## _pixels_tab[IDX][1] = PFX ## _pixels ## NUM ## _x2_c;  \
-    c->PFX ## _pixels_tab[IDX][2] = PFX ## _pixels ## NUM ## _y2_c;  \
-    c->PFX ## _pixels_tab[IDX][3] = PFX ## _pixels ## NUM ## _xy2_c
-
-    dspfunc(put, 0, 16);
-    dspfunc(put_no_rnd, 0, 16);
-    dspfunc(put, 1, 8);
-    dspfunc(put_no_rnd, 1, 8);
-    dspfunc(put, 2, 4);
-    dspfunc(put, 3, 2);
-
-    dspfunc(avg, 0, 16);
-    dspfunc(avg_no_rnd, 0, 16);
-    dspfunc(avg, 1, 8);
-    dspfunc(avg_no_rnd, 1, 8);
-    dspfunc(avg, 2, 4);
-    dspfunc(avg, 3, 2);
-#undef dspfunc
-
-    c->put_no_rnd_pixels_l2[0]= put_no_rnd_pixels16_l2_c;
-    c->put_no_rnd_pixels_l2[1]= put_no_rnd_pixels8_l2_c;
-
     c->put_tpel_pixels_tab[ 0] = put_tpel_pixels_mc00_c;
     c->put_tpel_pixels_tab[ 1] = put_tpel_pixels_mc10_c;
     c->put_tpel_pixels_tab[ 2] = put_tpel_pixels_mc20_c;
@@ -4244,40 +2791,11 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     dspfunc(avg_qpel, 1, 8);
     /* dspfunc(avg_no_rnd_qpel, 1, 8); */
 
-    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);
-
 #undef dspfunc
-    c->put_h264_chroma_pixels_tab[0]= put_h264_chroma_mc8_c;
-    c->put_h264_chroma_pixels_tab[1]= put_h264_chroma_mc4_c;
-    c->put_h264_chroma_pixels_tab[2]= put_h264_chroma_mc2_c;
-    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->draw_edges = draw_edges_c;
 
 #if CONFIG_MLP_DECODER || CONFIG_TRUEHD_DECODER
     ff_mlp_init(c, avctx);
 #endif
-#if CONFIG_WMV2_DECODER || CONFIG_VC1_DECODER
-    ff_intrax8dsp_init(c,avctx);
-#endif
-#if CONFIG_RV30_DECODER
-    ff_rv30dsp_init(c,avctx);
-#endif
-#if CONFIG_RV40_DECODER
-    ff_rv40dsp_init(c,avctx);
-    c->put_rv40_qpel_pixels_tab[0][15] = put_rv40_qpel16_mc33_c;
-    c->avg_rv40_qpel_pixels_tab[0][15] = avg_rv40_qpel16_mc33_c;
-    c->put_rv40_qpel_pixels_tab[1][15] = put_rv40_qpel8_mc33_c;
-    c->avg_rv40_qpel_pixels_tab[1][15] = avg_rv40_qpel8_mc33_c;
-#endif
 
     c->put_mspel_pixels_tab[0]= ff_put_pixels8x8_c;
     c->put_mspel_pixels_tab[1]= put_mspel8_mc10_c;
@@ -4323,7 +2841,6 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     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->add_hfyu_median_prediction= add_hfyu_median_prediction_c;
     c->sub_hfyu_median_prediction= sub_hfyu_median_prediction_c;
@@ -4331,49 +2848,33 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     c->add_hfyu_left_prediction_bgr32 = add_hfyu_left_prediction_bgr32_c;
     c->bswap_buf= bswap_buf;
     c->bswap16_buf = bswap16_buf;
-#if CONFIG_PNG_DECODER
-    c->add_png_paeth_prediction= ff_add_png_paeth_prediction;
-#endif
 
     if (CONFIG_H263_DECODER || CONFIG_H263_ENCODER) {
         c->h263_h_loop_filter= h263_h_loop_filter_c;
         c->h263_v_loop_filter= h263_v_loop_filter_c;
     }
 
-    if (CONFIG_VP3_DECODER) {
-        c->vp3_h_loop_filter= ff_vp3_h_loop_filter_c;
-        c->vp3_v_loop_filter= ff_vp3_v_loop_filter_c;
-        c->vp3_idct_dc_add= ff_vp3_idct_dc_add_c;
-    }
-
     c->h261_loop_filter= h261_loop_filter_c;
 
     c->try_8x8basis= try_8x8basis_c;
     c->add_8x8basis= add_8x8basis_c;
 
 #if CONFIG_VORBIS_DECODER
-    c->vorbis_inverse_coupling = vorbis_inverse_coupling;
+    c->vorbis_inverse_coupling = ff_vorbis_inverse_coupling;
 #endif
-#if CONFIG_AC3_DECODER
-    c->ac3_downmix = ff_ac3_downmix_c;
-#endif
-    c->vector_fmul = vector_fmul_c;
     c->vector_fmul_reverse = vector_fmul_reverse_c;
     c->vector_fmul_add = vector_fmul_add_c;
     c->vector_fmul_window = vector_fmul_window_c;
     c->vector_clipf = vector_clipf_c;
     c->scalarproduct_int16 = scalarproduct_int16_c;
     c->scalarproduct_and_madd_int16 = scalarproduct_and_madd_int16_c;
-    c->scalarproduct_float = scalarproduct_float_c;
+    c->apply_window_int16 = apply_window_int16_c;
+    c->vector_clip_int32 = vector_clip_int32_c;
+    c->scalarproduct_float = ff_scalarproduct_float_c;
     c->butterflies_float = butterflies_float_c;
+    c->butterflies_float_interleave = butterflies_float_interleave_c;
     c->vector_fmul_scalar = vector_fmul_scalar_c;
 
-    c->vector_fmul_sv_scalar[0] = vector_fmul_sv_scalar_2_c;
-    c->vector_fmul_sv_scalar[1] = vector_fmul_sv_scalar_4_c;
-
-    c->sv_fmul_scalar[0] = sv_fmul_scalar_2_c;
-    c->sv_fmul_scalar[1] = sv_fmul_scalar_4_c;
-
     c->shrink[0]= av_image_copy_plane;
     c->shrink[1]= ff_shrink22;
     c->shrink[2]= ff_shrink44;
@@ -4384,60 +2885,116 @@ av_cold void dsputil_init(DSPContext* c, AVCodecContext *avctx)
     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 (HAVE_MMX)        dsputil_init_mmx   (c, avctx);
-    if (ARCH_ARM)        dsputil_init_arm   (c, avctx);
-    if (CONFIG_MLIB)     dsputil_init_mlib  (c, avctx);
-    if (HAVE_VIS)        dsputil_init_vis   (c, avctx);
-    if (ARCH_ALPHA)      dsputil_init_alpha (c, avctx);
-    if (ARCH_PPC)        dsputil_init_ppc   (c, avctx);
-    if (HAVE_MMI)        dsputil_init_mmi   (c, avctx);
-    if (ARCH_SH4)        dsputil_init_sh4   (c, avctx);
-    if (ARCH_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];
-    }
-
-    c->put_rv30_tpel_pixels_tab[0][0] = c->put_h264_qpel_pixels_tab[0][0];
-    c->put_rv30_tpel_pixels_tab[1][0] = c->put_h264_qpel_pixels_tab[1][0];
-    c->avg_rv30_tpel_pixels_tab[0][0] = c->avg_h264_qpel_pixels_tab[0][0];
-    c->avg_rv30_tpel_pixels_tab[1][0] = c->avg_h264_qpel_pixels_tab[1][0];
-
-    c->put_rv40_qpel_pixels_tab[0][0] = c->put_h264_qpel_pixels_tab[0][0];
-    c->put_rv40_qpel_pixels_tab[1][0] = c->put_h264_qpel_pixels_tab[1][0];
-    c->avg_rv40_qpel_pixels_tab[0][0] = c->avg_h264_qpel_pixels_tab[0][0];
-    c->avg_rv40_qpel_pixels_tab[1][0] = c->avg_h264_qpel_pixels_tab[1][0];
-
-    switch(c->idct_permutation_type){
-    case FF_NO_IDCT_PERM:
-        for(i=0; i<64; i++)
-            c->idct_permutation[i]= i;
-        break;
-    case FF_LIBMPEG2_IDCT_PERM:
-        for(i=0; i<64; i++)
-            c->idct_permutation[i]= (i & 0x38) | ((i & 6) >> 1) | ((i & 1) << 2);
-        break;
-    case FF_SIMPLE_IDCT_PERM:
-        for(i=0; i<64; i++)
-            c->idct_permutation[i]= simple_mmx_permutation[i];
-        break;
-    case FF_TRANSPOSE_IDCT_PERM:
-        for(i=0; i<64; i++)
-            c->idct_permutation[i]= ((i&7)<<3) | (i>>3);
-        break;
-    case FF_PARTTRANS_IDCT_PERM:
-        for(i=0; i<64; i++)
-            c->idct_permutation[i]= (i&0x24) | ((i&3)<<3) | ((i>>3)&3);
+#undef FUNC
+#undef FUNCC
+#define FUNC(f, depth) f ## _ ## depth
+#define FUNCC(f, depth) f ## _ ## depth ## _c
+
+#define dspfunc1(PFX, IDX, NUM, depth)\
+    c->PFX ## _pixels_tab[IDX][0] = FUNCC(PFX ## _pixels ## NUM        , depth);\
+    c->PFX ## _pixels_tab[IDX][1] = FUNCC(PFX ## _pixels ## NUM ## _x2 , depth);\
+    c->PFX ## _pixels_tab[IDX][2] = FUNCC(PFX ## _pixels ## NUM ## _y2 , depth);\
+    c->PFX ## _pixels_tab[IDX][3] = FUNCC(PFX ## _pixels ## NUM ## _xy2, depth)
+
+#define dspfunc2(PFX, IDX, NUM, depth)\
+    c->PFX ## _pixels_tab[IDX][ 0] = FUNCC(PFX ## NUM ## _mc00, depth);\
+    c->PFX ## _pixels_tab[IDX][ 1] = FUNCC(PFX ## NUM ## _mc10, depth);\
+    c->PFX ## _pixels_tab[IDX][ 2] = FUNCC(PFX ## NUM ## _mc20, depth);\
+    c->PFX ## _pixels_tab[IDX][ 3] = FUNCC(PFX ## NUM ## _mc30, depth);\
+    c->PFX ## _pixels_tab[IDX][ 4] = FUNCC(PFX ## NUM ## _mc01, depth);\
+    c->PFX ## _pixels_tab[IDX][ 5] = FUNCC(PFX ## NUM ## _mc11, depth);\
+    c->PFX ## _pixels_tab[IDX][ 6] = FUNCC(PFX ## NUM ## _mc21, depth);\
+    c->PFX ## _pixels_tab[IDX][ 7] = FUNCC(PFX ## NUM ## _mc31, depth);\
+    c->PFX ## _pixels_tab[IDX][ 8] = FUNCC(PFX ## NUM ## _mc02, depth);\
+    c->PFX ## _pixels_tab[IDX][ 9] = FUNCC(PFX ## NUM ## _mc12, depth);\
+    c->PFX ## _pixels_tab[IDX][10] = FUNCC(PFX ## NUM ## _mc22, depth);\
+    c->PFX ## _pixels_tab[IDX][11] = FUNCC(PFX ## NUM ## _mc32, depth);\
+    c->PFX ## _pixels_tab[IDX][12] = FUNCC(PFX ## NUM ## _mc03, depth);\
+    c->PFX ## _pixels_tab[IDX][13] = FUNCC(PFX ## NUM ## _mc13, depth);\
+    c->PFX ## _pixels_tab[IDX][14] = FUNCC(PFX ## NUM ## _mc23, depth);\
+    c->PFX ## _pixels_tab[IDX][15] = FUNCC(PFX ## NUM ## _mc33, depth)
+
+
+#define BIT_DEPTH_FUNCS(depth, dct)\
+    c->get_pixels                    = FUNCC(get_pixels   ## dct   , depth);\
+    c->draw_edges                    = FUNCC(draw_edges            , depth);\
+    c->emulated_edge_mc              = FUNC (ff_emulated_edge_mc   , depth);\
+    c->clear_block                   = FUNCC(clear_block  ## dct   , depth);\
+    c->clear_blocks                  = FUNCC(clear_blocks ## dct   , depth);\
+    c->add_pixels8                   = FUNCC(add_pixels8  ## dct   , depth);\
+    c->add_pixels4                   = FUNCC(add_pixels4  ## dct   , depth);\
+    c->put_no_rnd_pixels_l2[0]       = FUNCC(put_no_rnd_pixels16_l2, depth);\
+    c->put_no_rnd_pixels_l2[1]       = FUNCC(put_no_rnd_pixels8_l2 , depth);\
+\
+    c->put_h264_chroma_pixels_tab[0] = FUNCC(put_h264_chroma_mc8   , depth);\
+    c->put_h264_chroma_pixels_tab[1] = FUNCC(put_h264_chroma_mc4   , depth);\
+    c->put_h264_chroma_pixels_tab[2] = FUNCC(put_h264_chroma_mc2   , depth);\
+    c->avg_h264_chroma_pixels_tab[0] = FUNCC(avg_h264_chroma_mc8   , depth);\
+    c->avg_h264_chroma_pixels_tab[1] = FUNCC(avg_h264_chroma_mc4   , depth);\
+    c->avg_h264_chroma_pixels_tab[2] = FUNCC(avg_h264_chroma_mc2   , depth);\
+\
+    dspfunc1(put       , 0, 16, depth);\
+    dspfunc1(put       , 1,  8, depth);\
+    dspfunc1(put       , 2,  4, depth);\
+    dspfunc1(put       , 3,  2, depth);\
+    dspfunc1(put_no_rnd, 0, 16, depth);\
+    dspfunc1(put_no_rnd, 1,  8, depth);\
+    dspfunc1(avg       , 0, 16, depth);\
+    dspfunc1(avg       , 1,  8, depth);\
+    dspfunc1(avg       , 2,  4, depth);\
+    dspfunc1(avg       , 3,  2, depth);\
+    dspfunc1(avg_no_rnd, 0, 16, depth);\
+    dspfunc1(avg_no_rnd, 1,  8, depth);\
+\
+    dspfunc2(put_h264_qpel, 0, 16, depth);\
+    dspfunc2(put_h264_qpel, 1,  8, depth);\
+    dspfunc2(put_h264_qpel, 2,  4, depth);\
+    dspfunc2(put_h264_qpel, 3,  2, depth);\
+    dspfunc2(avg_h264_qpel, 0, 16, depth);\
+    dspfunc2(avg_h264_qpel, 1,  8, depth);\
+    dspfunc2(avg_h264_qpel, 2,  4, depth);
+
+    switch (avctx->bits_per_raw_sample) {
+    case 9:
+        if (c->dct_bits == 32) {
+            BIT_DEPTH_FUNCS(9, _32);
+        } else {
+            BIT_DEPTH_FUNCS(9, _16);
+        }
         break;
-    case FF_SSE2_IDCT_PERM:
-        for(i=0; i<64; i++)
-            c->idct_permutation[i]= (i&0x38) | idct_sse2_row_perm[i&7];
+    case 10:
+        if (c->dct_bits == 32) {
+            BIT_DEPTH_FUNCS(10, _32);
+        } else {
+            BIT_DEPTH_FUNCS(10, _16);
+        }
         break;
     default:
-        av_log(avctx, AV_LOG_ERROR, "Internal error, IDCT permutation not set\n");
+        BIT_DEPTH_FUNCS(8, _16);
+        break;
     }
-}
 
+
+    if (HAVE_MMX)        ff_dsputil_init_mmx   (c, avctx);
+    if (ARCH_ARM)        ff_dsputil_init_arm   (c, avctx);
+    if (HAVE_VIS)        ff_dsputil_init_vis   (c, avctx);
+    if (ARCH_ALPHA)      ff_dsputil_init_alpha (c, avctx);
+    if (ARCH_PPC)        ff_dsputil_init_ppc   (c, avctx);
+    if (HAVE_MMI)        ff_dsputil_init_mmi   (c, avctx);
+    if (ARCH_SH4)        ff_dsputil_init_sh4   (c, avctx);
+    if (ARCH_BFIN)       ff_dsputil_init_bfin  (c, avctx);
+
+    for (i = 0; i < 4; i++) {
+        for (j = 0; j < 16; j++) {
+            if(!c->put_2tap_qpel_pixels_tab[i][j])
+                c->put_2tap_qpel_pixels_tab[i][j] =
+                    c->put_h264_qpel_pixels_tab[i][j];
+            if(!c->avg_2tap_qpel_pixels_tab[i][j])
+                c->avg_2tap_qpel_pixels_tab[i][j] =
+                    c->avg_h264_qpel_pixels_tab[i][j];
+        }
+    }
+
+    ff_init_scantable_permutation(c->idct_permutation,
+                                  c->idct_permutation_type);
+}