1. ffmpeg
  2. Commit 87ce8b495f0e73625bc583e01c07bcd3e67b95ab
Browse code

Template dsputil_template.c with respect to pixel size, etc.

Preparatory patch for high bit depth h264 decoding support.

Signed-off-by: Ronald S. Bultje <rsbultje@gmail.com>

Oskar Arvidsson authored on 2011/03/30 00:48:55
Showing 1 changed files
libavcodec/dsputil_template.c
History View file @ 87ce8b4
... ... 
@@ -29,18 +29,46 @@
29 29
30 30 
 #include "dsputil.h"
31 31
32 
+#define BIT_DEPTH 8
33 
+
34 
+#define pixel  uint8_t
35 
+#define pixel2 uint16_t
36 
+#define pixel4 uint32_t
37 
+#define dctcoef int16_t
38 
+
39 
+#define FUNC(a)  a
40 
+#define FUNCC(a) a ## _c
41 
+#define INIT_CLIP uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;
42 
+#define CLIP(a) cm[a]
43 
+#define AV_RN2P AV_RN16
44 
+#define AV_RN4P AV_RN32
45 
+#define PIXEL_MAX ((1<<BIT_DEPTH)-1)
46 
+
47 
+#define no_rnd_avg_pixel4 no_rnd_avg32
48 
+#define    rnd_avg_pixel4    rnd_avg32
49 
+
32 50 
 /* draw the edges of width 'w' of an image of size width, height */
33 51 
 //FIXME check that this is ok for mpeg4 interlaced
34 
-static void draw_edges_c(uint8_t *buf, int wrap, int width, int height, int w, int sides)
52 
+static void FUNCC(draw_edges)(uint8_t *_buf, int _wrap, int width, int height, int w, int sides)
35 53 
 {
36 
-    uint8_t *ptr, *last_line;
54 
+    pixel *buf = (pixel*)_buf;
55 
+    int wrap = _wrap / sizeof(pixel);
56 
+    pixel *ptr, *last_line;
37 57 
     int i;
38 58
39 59 
     /* left and right */
40 60 
     ptr = buf;
41 61 
     for(i=0;i<height;i++) {
62 
+#if BIT_DEPTH > 8
63 
+        int j;
64 
+        for (j = 0; j < w; j++) {
65 
+            ptr[j-w] = ptr[0];
66 
+            ptr[j+width] = ptr[width-1];
67 
+        }
68 
+#else
42 69 
         memset(ptr - w, ptr[0], w);
43 70 
         memset(ptr + width, ptr[width-1], w);
71 
+#endif
44 72 
         ptr += wrap;
45 73 
     }
46 74
... ... 
@@ -49,10 +77,10 @@ static void draw_edges_c(uint8_t *buf, int wrap, int width, int height, int w, i
49 49 
     last_line = buf + (height - 1) * wrap;
50 50 
     if (sides & EDGE_TOP)
51 51 
         for(i = 0; i < w; i++)
52 
-            memcpy(buf - (i + 1) * wrap, buf, width + w + w); // top
52 
+            memcpy(buf - (i + 1) * wrap, buf, (width + w + w) * sizeof(pixel)); // top
53 53 
     if (sides & EDGE_BOTTOM)
54 54 
         for (i = 0; i < w; i++)
55 
-            memcpy(last_line + (i + 1) * wrap, last_line, width + w + w); // bottom
55 
+            memcpy(last_line + (i + 1) * wrap, last_line, (width + w + w) * sizeof(pixel)); // bottom
56 56 
 }
57 57
58 58 
 /**
... ... 
@@ -67,7 +95,7 @@ static void draw_edges_c(uint8_t *buf, int wrap, int width, int height, int w, i
67 67 
  * @param w width of the source buffer
68 68 
  * @param h height of the source buffer
69 69 
  */
70 
-void ff_emulated_edge_mc(uint8_t *buf, const uint8_t *src, int linesize, int block_w, int block_h,
70 
+void FUNC(ff_emulated_edge_mc)(uint8_t *buf, const uint8_t *src, int linesize, int block_w, int block_h,
71 71 
                                     int src_x, int src_y, int w, int h){
72 72 
     int x, y;
73 73 
     int start_y, start_x, end_y, end_x;
... ... 
@@ -80,10 +108,10 @@ void ff_emulated_edge_mc(uint8_t *buf, const uint8_t *src, int linesize, int blo
80 80 
         src_y=1-block_h;
81 81 
     }
82 82 
     if(src_x>= w){
83 
-        src+= (w-1-src_x);
83 
+        src+= (w-1-src_x)*sizeof(pixel);
84 84 
         src_x=w-1;
85 85 
     }else if(src_x<=-block_w){
86 
-        src+= (1-block_w-src_x);
86 
+        src+= (1-block_w-src_x)*sizeof(pixel);
87 87 
         src_x=1-block_w;
88 88 
     }
89 89
... ... 
@@ -95,18 +123,18 @@ void ff_emulated_edge_mc(uint8_t *buf, const uint8_t *src, int linesize, int blo
95 95 
     assert(start_x < end_x && block_w);
96 96
97 97 
     w    = end_x - start_x;
98 
-    src += start_y*linesize + start_x;
99 
-    buf += start_x;
98 
+    src += start_y*linesize + start_x*sizeof(pixel);
99 
+    buf += start_x*sizeof(pixel);
100 100
101 101 
     //top
102 102 
     for(y=0; y<start_y; y++){
103 
-        memcpy(buf, src, w);
103 
+        memcpy(buf, src, w*sizeof(pixel));
104 104 
         buf += linesize;
105 105 
     }
106 106
107 107 
     // copy existing part
108 108 
     for(; y<end_y; y++){
109 
-        memcpy(buf, src, w);
109 
+        memcpy(buf, src, w*sizeof(pixel));
110 110 
         src += linesize;
111 111 
         buf += linesize;
112 112 
     }
... ... 
@@ -114,28 +142,33 @@ void ff_emulated_edge_mc(uint8_t *buf, const uint8_t *src, int linesize, int blo
114 114 
     //bottom
115 115 
     src -= linesize;
116 116 
     for(; y<block_h; y++){
117 
-        memcpy(buf, src, w);
117 
+        memcpy(buf, src, w*sizeof(pixel));
118 118 
         buf += linesize;
119 119 
     }
120 120
121 
-    buf -= block_h * linesize + start_x;
121 
+    buf -= block_h * linesize + start_x*sizeof(pixel);
122 122 
     while (block_h--){
123 
+        pixel *bufp = (pixel*)buf;
123 124 
        //left
124 125 
         for(x=0; x<start_x; x++){
125 
-            buf[x] = buf[start_x];
126 
+            bufp[x] = bufp[start_x];
126 127 
         }
127 128
128 129 
        //right
129 130 
         for(x=end_x; x<block_w; x++){
130 
-            buf[x] = buf[end_x - 1];
131 
+            bufp[x] = bufp[end_x - 1];
131 132 
         }
132 133 
         buf += linesize;
133 134 
     }
134 135 
 }
135 136
136 
-static void add_pixels8_c(uint8_t *restrict pixels, DCTELEM *block, int line_size)
137 
+static void FUNCC(add_pixels8)(uint8_t *restrict _pixels, DCTELEM *_block, int line_size)
137 138 
 {
138 139 
     int i;
140 
+    pixel *restrict pixels = (pixel *restrict)_pixels;
141 
+    dctcoef *block = (dctcoef*)_block;
142 
+    line_size /= sizeof(pixel);
143 
+
139 144 
     for(i=0;i<8;i++) {
140 145 
         pixels[0] += block[0];
141 146 
         pixels[1] += block[1];
... ... 
@@ -150,9 +183,13 @@ static void add_pixels8_c(uint8_t *restrict pixels, DCTELEM *block, int line_siz
150 150 
     }
151 151 
 }
152 152
153 
-static void add_pixels4_c(uint8_t *restrict pixels, DCTELEM *block, int line_size)
153 
+static void FUNCC(add_pixels4)(uint8_t *restrict _pixels, DCTELEM *_block, int line_size)
154 154 
 {
155 155 
     int i;
156 
+    pixel *restrict pixels = (pixel *restrict)_pixels;
157 
+    dctcoef *block = (dctcoef*)_block;
158 
+    line_size /= sizeof(pixel);
159 
+
156 160 
     for(i=0;i<4;i++) {
157 161 
         pixels[0] += block[0];
158 162 
         pixels[1] += block[1];
... ... 
@@ -296,127 +333,128 @@ static void OPNAME ## _no_rnd_pixels_xy2_c(uint8_t *block, const uint8_t *pixels
296 296 
         }\
297 297 
 }\
298 298 
 \
299 
-CALL_2X_PIXELS(OPNAME ## _pixels16_c    , OPNAME ## _pixels_c    , 8)\
300 
-CALL_2X_PIXELS(OPNAME ## _pixels16_x2_c , OPNAME ## _pixels_x2_c , 8)\
301 
-CALL_2X_PIXELS(OPNAME ## _pixels16_y2_c , OPNAME ## _pixels_y2_c , 8)\
302 
-CALL_2X_PIXELS(OPNAME ## _pixels16_xy2_c, OPNAME ## _pixels_xy2_c, 8)\
303 
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_x2_c , OPNAME ## _no_rnd_pixels_x2_c , 8)\
304 
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_y2_c , OPNAME ## _no_rnd_pixels_y2_c , 8)\
305 
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_xy2_c, OPNAME ## _no_rnd_pixels_xy2_c, 8)
299 
+CALL_2X_PIXELS(OPNAME ## _pixels16_c    , OPNAME ## _pixels_c    , 8*sizeof(pixel))\
300 
+CALL_2X_PIXELS(OPNAME ## _pixels16_x2_c , OPNAME ## _pixels_x2_c , 8*sizeof(pixel))\
301 
+CALL_2X_PIXELS(OPNAME ## _pixels16_y2_c , OPNAME ## _pixels_y2_c , 8*sizeof(pixel))\
302 
+CALL_2X_PIXELS(OPNAME ## _pixels16_xy2_c, OPNAME ## _pixels_xy2_c, 8*sizeof(pixel))\
303 
+CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_x2_c , OPNAME ## _no_rnd_pixels_x2_c , 8*sizeof(pixel))\
304 
+CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_y2_c , OPNAME ## _no_rnd_pixels_y2_c , 8*sizeof(pixel))\
305 
+CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_xy2_c, OPNAME ## _no_rnd_pixels_xy2_c, 8*sizeof(pixel))
306 306
307 307 
 #define op_avg(a, b) a = ( ((a)|(b)) - ((((a)^(b))&0xFEFEFEFEFEFEFEFEULL)>>1) )
308 308 
 #else // 64 bit variant
309 309
310 310 
 #define PIXOP2(OPNAME, OP) \
311 
-static void OPNAME ## _pixels2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
311 
+static void FUNCC(OPNAME ## _pixels2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
312 312 
     int i;\
313 313 
     for(i=0; i<h; i++){\
314 
-        OP(*((uint16_t*)(block  )), AV_RN16(pixels  ));\
314 
+        OP(*((pixel2*)(block  )), AV_RN2P(pixels  ));\
315 315 
         pixels+=line_size;\
316 316 
         block +=line_size;\
317 317 
     }\
318 318 
 }\
319 
-static void OPNAME ## _pixels4_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
319 
+static void FUNCC(OPNAME ## _pixels4)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
320 320 
     int i;\
321 321 
     for(i=0; i<h; i++){\
322 
-        OP(*((uint32_t*)(block  )), AV_RN32(pixels  ));\
322 
+        OP(*((pixel4*)(block  )), AV_RN4P(pixels  ));\
323 323 
         pixels+=line_size;\
324 324 
         block +=line_size;\
325 325 
     }\
326 326 
 }\
327 
-static void OPNAME ## _pixels8_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
327 
+static void FUNCC(OPNAME ## _pixels8)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
328 328 
     int i;\
329 329 
     for(i=0; i<h; i++){\
330 
-        OP(*((uint32_t*)(block  )), AV_RN32(pixels  ));\
331 
-        OP(*((uint32_t*)(block+4)), AV_RN32(pixels+4));\
330 
+        OP(*((pixel4*)(block                )), AV_RN4P(pixels                ));\
331 
+        OP(*((pixel4*)(block+4*sizeof(pixel))), AV_RN4P(pixels+4*sizeof(pixel)));\
332 332 
         pixels+=line_size;\
333 333 
         block +=line_size;\
334 334 
     }\
335 335 
 }\
336 
-static inline void OPNAME ## _no_rnd_pixels8_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
337 
-    OPNAME ## _pixels8_c(block, pixels, line_size, h);\
336 
+static inline void FUNCC(OPNAME ## _no_rnd_pixels8)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
337 
+    FUNCC(OPNAME ## _pixels8)(block, pixels, line_size, h);\
338 338 
 }\
339 339 
 \
340 
-static inline void OPNAME ## _no_rnd_pixels8_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
340 
+static inline void FUNC(OPNAME ## _no_rnd_pixels8_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
341 341 
                                                 int src_stride1, int src_stride2, int h){\
342 342 
     int i;\
343 343 
     for(i=0; i<h; i++){\
344 
-        uint32_t a,b;\
345 
-        a= AV_RN32(&src1[i*src_stride1  ]);\
346 
-        b= AV_RN32(&src2[i*src_stride2  ]);\
347 
-        OP(*((uint32_t*)&dst[i*dst_stride  ]), no_rnd_avg32(a, b));\
348 
-        a= AV_RN32(&src1[i*src_stride1+4]);\
349 
-        b= AV_RN32(&src2[i*src_stride2+4]);\
350 
-        OP(*((uint32_t*)&dst[i*dst_stride+4]), no_rnd_avg32(a, b));\
344 
+        pixel4 a,b;\
345 
+        a= AV_RN4P(&src1[i*src_stride1  ]);\
346 
+        b= AV_RN4P(&src2[i*src_stride2  ]);\
347 
+        OP(*((pixel4*)&dst[i*dst_stride  ]), no_rnd_avg_pixel4(a, b));\
348 
+        a= AV_RN4P(&src1[i*src_stride1+4*sizeof(pixel)]);\
349 
+        b= AV_RN4P(&src2[i*src_stride2+4*sizeof(pixel)]);\
350 
+        OP(*((pixel4*)&dst[i*dst_stride+4*sizeof(pixel)]), no_rnd_avg_pixel4(a, b));\
351 351 
     }\
352 352 
 }\
353 353 
 \
354 
-static inline void OPNAME ## _pixels8_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
354 
+static inline void FUNC(OPNAME ## _pixels8_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
355 355 
                                                 int src_stride1, int src_stride2, int h){\
356 356 
     int i;\
357 357 
     for(i=0; i<h; i++){\
358 
-        uint32_t a,b;\
359 
-        a= AV_RN32(&src1[i*src_stride1  ]);\
360 
-        b= AV_RN32(&src2[i*src_stride2  ]);\
361 
-        OP(*((uint32_t*)&dst[i*dst_stride  ]), rnd_avg32(a, b));\
362 
-        a= AV_RN32(&src1[i*src_stride1+4]);\
363 
-        b= AV_RN32(&src2[i*src_stride2+4]);\
364 
-        OP(*((uint32_t*)&dst[i*dst_stride+4]), rnd_avg32(a, b));\
358 
+        pixel4 a,b;\
359 
+        a= AV_RN4P(&src1[i*src_stride1  ]);\
360 
+        b= AV_RN4P(&src2[i*src_stride2  ]);\
361 
+        OP(*((pixel4*)&dst[i*dst_stride  ]), rnd_avg_pixel4(a, b));\
362 
+        a= AV_RN4P(&src1[i*src_stride1+4*sizeof(pixel)]);\
363 
+        b= AV_RN4P(&src2[i*src_stride2+4*sizeof(pixel)]);\
364 
+        OP(*((pixel4*)&dst[i*dst_stride+4*sizeof(pixel)]), rnd_avg_pixel4(a, b));\
365 365 
     }\
366 366 
 }\
367 367 
 \
368 
-static inline void OPNAME ## _pixels4_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
368 
+static inline void FUNC(OPNAME ## _pixels4_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
369 369 
                                                 int src_stride1, int src_stride2, int h){\
370 370 
     int i;\
371 371 
     for(i=0; i<h; i++){\
372 
-        uint32_t a,b;\
373 
-        a= AV_RN32(&src1[i*src_stride1  ]);\
374 
-        b= AV_RN32(&src2[i*src_stride2  ]);\
375 
-        OP(*((uint32_t*)&dst[i*dst_stride  ]), rnd_avg32(a, b));\
372 
+        pixel4 a,b;\
373 
+        a= AV_RN4P(&src1[i*src_stride1  ]);\
374 
+        b= AV_RN4P(&src2[i*src_stride2  ]);\
375 
+        OP(*((pixel4*)&dst[i*dst_stride  ]), rnd_avg_pixel4(a, b));\
376 376 
     }\
377 377 
 }\
378 378 
 \
379 
-static inline void OPNAME ## _pixels2_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
379 
+static inline void FUNC(OPNAME ## _pixels2_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
380 380 
                                                 int src_stride1, int src_stride2, int h){\
381 381 
     int i;\
382 382 
     for(i=0; i<h; i++){\
383 
-        uint32_t a,b;\
384 
-        a= AV_RN16(&src1[i*src_stride1  ]);\
385 
-        b= AV_RN16(&src2[i*src_stride2  ]);\
386 
-        OP(*((uint16_t*)&dst[i*dst_stride  ]), rnd_avg32(a, b));\
383 
+        pixel4 a,b;\
384 
+        a= AV_RN2P(&src1[i*src_stride1  ]);\
385 
+        b= AV_RN2P(&src2[i*src_stride2  ]);\
386 
+        OP(*((pixel2*)&dst[i*dst_stride  ]), rnd_avg_pixel4(a, b));\
387 387 
     }\
388 388 
 }\
389 389 
 \
390 
-static inline void OPNAME ## _pixels16_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
390 
+static inline void FUNC(OPNAME ## _pixels16_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
391 391 
                                                 int src_stride1, int src_stride2, int h){\
392 
-    OPNAME ## _pixels8_l2(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);\
393 
-    OPNAME ## _pixels8_l2(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);\
392 
+    FUNC(OPNAME ## _pixels8_l2)(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);\
393 
+    FUNC(OPNAME ## _pixels8_l2)(dst+8*sizeof(pixel), src1+8*sizeof(pixel), src2+8*sizeof(pixel), dst_stride, src_stride1, src_stride2, h);\
394 394 
 }\
395 395 
 \
396 
-static inline void OPNAME ## _no_rnd_pixels16_l2(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
396 
+static inline void FUNC(OPNAME ## _no_rnd_pixels16_l2)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, int dst_stride, \
397 397 
                                                 int src_stride1, int src_stride2, int h){\
398 
-    OPNAME ## _no_rnd_pixels8_l2(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);\
399 
-    OPNAME ## _no_rnd_pixels8_l2(dst+8, src1+8, src2+8, dst_stride, src_stride1, src_stride2, h);\
398 
+    FUNC(OPNAME ## _no_rnd_pixels8_l2)(dst  , src1  , src2  , dst_stride, src_stride1, src_stride2, h);\
399 
+    FUNC(OPNAME ## _no_rnd_pixels8_l2)(dst+8*sizeof(pixel), src1+8*sizeof(pixel), src2+8*sizeof(pixel), dst_stride, src_stride1, src_stride2, h);\
400 400 
 }\
401 401 
 \
402 
-static inline void OPNAME ## _no_rnd_pixels8_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
403 
-    OPNAME ## _no_rnd_pixels8_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
402 
+static inline void FUNCC(OPNAME ## _no_rnd_pixels8_x2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
403 
+    FUNC(OPNAME ## _no_rnd_pixels8_l2)(block, pixels, pixels+sizeof(pixel), line_size, line_size, line_size, h);\
404 404 
 }\
405 405 
 \
406 
-static inline void OPNAME ## _pixels8_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
407 
-    OPNAME ## _pixels8_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
406 
+static inline void FUNCC(OPNAME ## _pixels8_x2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
407 
+    FUNC(OPNAME ## _pixels8_l2)(block, pixels, pixels+sizeof(pixel), line_size, line_size, line_size, h);\
408 408 
 }\
409 409 
 \
410 
-static inline void OPNAME ## _no_rnd_pixels8_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
411 
-    OPNAME ## _no_rnd_pixels8_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
410 
+static inline void FUNCC(OPNAME ## _no_rnd_pixels8_y2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
411 
+    FUNC(OPNAME ## _no_rnd_pixels8_l2)(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
412 412 
 }\
413 413 
 \
414 
-static inline void OPNAME ## _pixels8_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
415 
-    OPNAME ## _pixels8_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
414 
+static inline void FUNCC(OPNAME ## _pixels8_y2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
415 
+    FUNC(OPNAME ## _pixels8_l2)(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
416 416 
 }\
417 417 
 \
418 
-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,\
418 
+static inline void FUNC(OPNAME ## _pixels8_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
419 419 
                  int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
420 
+    /* FIXME HIGH BIT DEPTH */\
420 421 
     int i;\
421 422 
     for(i=0; i<h; i++){\
422 423 
         uint32_t a, b, c, d, l0, l1, h0, h1;\
... ... 
@@ -451,24 +489,25 @@ static inline void OPNAME ## _pixels8_l4(uint8_t *dst, const uint8_t *src1, cons
451 451 
     }\
452 452 
 }\
453 453 
 \
454 
-static inline void OPNAME ## _pixels4_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
455 
-    OPNAME ## _pixels4_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
454 
+static inline void FUNCC(OPNAME ## _pixels4_x2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
455 
+    FUNC(OPNAME ## _pixels4_l2)(block, pixels, pixels+sizeof(pixel), line_size, line_size, line_size, h);\
456 456 
 }\
457 457 
 \
458 
-static inline void OPNAME ## _pixels4_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
459 
-    OPNAME ## _pixels4_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
458 
+static inline void FUNCC(OPNAME ## _pixels4_y2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
459 
+    FUNC(OPNAME ## _pixels4_l2)(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
460 460 
 }\
461 461 
 \
462 
-static inline void OPNAME ## _pixels2_x2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
463 
-    OPNAME ## _pixels2_l2(block, pixels, pixels+1, line_size, line_size, line_size, h);\
462 
+static inline void FUNCC(OPNAME ## _pixels2_x2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
463 
+    FUNC(OPNAME ## _pixels2_l2)(block, pixels, pixels+sizeof(pixel), line_size, line_size, line_size, h);\
464 464 
 }\
465 465 
 \
466 
-static inline void OPNAME ## _pixels2_y2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
467 
-    OPNAME ## _pixels2_l2(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
466 
+static inline void FUNCC(OPNAME ## _pixels2_y2)(uint8_t *block, const uint8_t *pixels, int line_size, int h){\
467 
+    FUNC(OPNAME ## _pixels2_l2)(block, pixels, pixels+line_size, line_size, line_size, line_size, h);\
468 468 
 }\
469 469 
 \
470 
-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,\
470 
+static inline void FUNC(OPNAME ## _no_rnd_pixels8_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
471 471 
                  int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
472 
+    /* FIXME HIGH BIT DEPTH*/\
472 473 
     int i;\
473 474 
     for(i=0; i<h; i++){\
474 475 
         uint32_t a, b, c, d, l0, l1, h0, h1;\
... ... 
@@ -502,20 +541,23 @@ static inline void OPNAME ## _no_rnd_pixels8_l4(uint8_t *dst, const uint8_t *src
502 502 
         OP(*((uint32_t*)&dst[i*dst_stride+4]), h0+h1+(((l0+l1)>>2)&0x0F0F0F0FUL));\
503 503 
     }\
504 504 
 }\
505 
-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,\
505 
+static inline void FUNC(OPNAME ## _pixels16_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
506 506 
                  int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
507 
-    OPNAME ## _pixels8_l4(dst  , src1  , src2  , src3  , src4  , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
508 
-    OPNAME ## _pixels8_l4(dst+8, src1+8, src2+8, src3+8, src4+8, dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
507 
+    FUNC(OPNAME ## _pixels8_l4)(dst  , src1  , src2  , src3  , src4  , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
508 
+    FUNC(OPNAME ## _pixels8_l4)(dst+8*sizeof(pixel), src1+8*sizeof(pixel), src2+8*sizeof(pixel), src3+8*sizeof(pixel), src4+8*sizeof(pixel), dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
509 509 
 }\
510 
-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,\
510 
+static inline void FUNC(OPNAME ## _no_rnd_pixels16_l4)(uint8_t *dst, const uint8_t *src1, const uint8_t *src2, const uint8_t *src3, const uint8_t *src4,\
511 511 
                  int dst_stride, int src_stride1, int src_stride2,int src_stride3,int src_stride4, int h){\
512 
-    OPNAME ## _no_rnd_pixels8_l4(dst  , src1  , src2  , src3  , src4  , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
513 
-    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);\
512 
+    FUNC(OPNAME ## _no_rnd_pixels8_l4)(dst  , src1  , src2  , src3  , src4  , dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
513 
+    FUNC(OPNAME ## _no_rnd_pixels8_l4)(dst+8*sizeof(pixel), src1+8*sizeof(pixel), src2+8*sizeof(pixel), src3+8*sizeof(pixel), src4+8*sizeof(pixel), dst_stride, src_stride1, src_stride2, src_stride3, src_stride4, h);\
514 514 
 }\
515 515 
 \
516 
-static inline void OPNAME ## _pixels2_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
516 
+static inline void FUNCC(OPNAME ## _pixels2_xy2)(uint8_t *_block, const uint8_t *_pixels, int line_size, int h)\
517 517 
 {\
518 518 
         int i, a0, b0, a1, b1;\
519 
+        pixel *block = (pixel*)_block;\
520 
+        const pixel *pixels = (const pixel*)_pixels;\
521 
+        line_size /= sizeof(pixel);\
519 522 
         a0= pixels[0];\
520 523 
         b0= pixels[1] + 2;\
521 524 
         a0 += b0;\
... ... 
@@ -546,8 +588,9 @@ static inline void OPNAME ## _pixels2_xy2_c(uint8_t *block, const uint8_t *pixel
546 546 
         }\
547 547 
 }\
548 548 
 \
549 
-static inline void OPNAME ## _pixels4_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
549 
+static inline void FUNCC(OPNAME ## _pixels4_xy2)(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
550 550 
 {\
551 
+        /* FIXME HIGH BIT DEPTH */\
551 552 
         int i;\
552 553 
         const uint32_t a= AV_RN32(pixels  );\
553 554 
         const uint32_t b= AV_RN32(pixels+1);\
... ... 
@@ -582,8 +625,9 @@ static inline void OPNAME ## _pixels4_xy2_c(uint8_t *block, const uint8_t *pixel
582 582 
         }\
583 583 
 }\
584 584 
 \
585 
-static inline void OPNAME ## _pixels8_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
585 
+static inline void FUNCC(OPNAME ## _pixels8_xy2)(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
586 586 
 {\
587 
+    /* FIXME HIGH BIT DEPTH */\
587 588 
     int j;\
588 589 
     for(j=0; j<2; j++){\
589 590 
         int i;\
... ... 
@@ -623,8 +667,9 @@ static inline void OPNAME ## _pixels8_xy2_c(uint8_t *block, const uint8_t *pixel
623 623 
     }\
624 624 
 }\
625 625 
 \
626 
-static inline void OPNAME ## _no_rnd_pixels8_xy2_c(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
626 
+static inline void FUNCC(OPNAME ## _no_rnd_pixels8_xy2)(uint8_t *block, const uint8_t *pixels, int line_size, int h)\
627 627 
 {\
628 
+    /* FIXME HIGH BIT DEPTH */\
628 629 
     int j;\
629 630 
     for(j=0; j<2; j++){\
630 631 
         int i;\
... ... 
@@ -664,18 +709,17 @@ static inline void OPNAME ## _no_rnd_pixels8_xy2_c(uint8_t *block, const uint8_t
664 664 
     }\
665 665 
 }\
666 666 
 \
667 
-CALL_2X_PIXELS(OPNAME ## _pixels16_c  , OPNAME ## _pixels8_c  , 8)\
668 
-CALL_2X_PIXELS(OPNAME ## _pixels16_x2_c , OPNAME ## _pixels8_x2_c , 8)\
669 
-CALL_2X_PIXELS(OPNAME ## _pixels16_y2_c , OPNAME ## _pixels8_y2_c , 8)\
670 
-CALL_2X_PIXELS(OPNAME ## _pixels16_xy2_c, OPNAME ## _pixels8_xy2_c, 8)\
671 
-av_unused CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_c  , OPNAME ## _pixels8_c         , 8)\
672 
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_x2_c , OPNAME ## _no_rnd_pixels8_x2_c , 8)\
673 
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_y2_c , OPNAME ## _no_rnd_pixels8_y2_c , 8)\
674 
-CALL_2X_PIXELS(OPNAME ## _no_rnd_pixels16_xy2_c, OPNAME ## _no_rnd_pixels8_xy2_c, 8)\
667 
+CALL_2X_PIXELS(FUNCC(OPNAME ## _pixels16)    , FUNCC(OPNAME ## _pixels8)    , 8*sizeof(pixel))\
668 
+CALL_2X_PIXELS(FUNCC(OPNAME ## _pixels16_x2) , FUNCC(OPNAME ## _pixels8_x2) , 8*sizeof(pixel))\
669 
+CALL_2X_PIXELS(FUNCC(OPNAME ## _pixels16_y2) , FUNCC(OPNAME ## _pixels8_y2) , 8*sizeof(pixel))\
670 
+CALL_2X_PIXELS(FUNCC(OPNAME ## _pixels16_xy2), FUNCC(OPNAME ## _pixels8_xy2), 8*sizeof(pixel))\
671 
+av_unused CALL_2X_PIXELS(FUNCC(OPNAME ## _no_rnd_pixels16)    , FUNCC(OPNAME ## _pixels8) , 8*sizeof(pixel))\
672 
+CALL_2X_PIXELS(FUNCC(OPNAME ## _no_rnd_pixels16_x2) , FUNCC(OPNAME ## _no_rnd_pixels8_x2) , 8*sizeof(pixel))\
673 
+CALL_2X_PIXELS(FUNCC(OPNAME ## _no_rnd_pixels16_y2) , FUNCC(OPNAME ## _no_rnd_pixels8_y2) , 8*sizeof(pixel))\
674 
+CALL_2X_PIXELS(FUNCC(OPNAME ## _no_rnd_pixels16_xy2), FUNCC(OPNAME ## _no_rnd_pixels8_xy2), 8*sizeof(pixel))\
675 675
676 
-#define op_avg(a, b) a = rnd_avg32(a, b)
676 
+#define op_avg(a, b) a = rnd_avg_pixel4(a, b)
677 677 
 #endif
678 
-
679 678 
 #define op_put(a, b) a = b
680 679
681 680 
 PIXOP2(avg, op_avg)
... ... 
@@ -686,21 +730,24 @@ PIXOP2(put, op_put)
686 686 
 #define put_no_rnd_pixels8_c  put_pixels8_c
687 687 
 #define put_no_rnd_pixels16_c put_pixels16_c
688 688
689 
-static void put_no_rnd_pixels16_l2_c(uint8_t *dst, const uint8_t *a, const uint8_t *b, int stride, int h){
690 
-    put_no_rnd_pixels16_l2(dst, a, b, stride, stride, stride, h);
689 
+static void FUNCC(put_no_rnd_pixels16_l2)(uint8_t *dst, const uint8_t *a, const uint8_t *b, int stride, int h){
690 
+    FUNC(put_no_rnd_pixels16_l2)(dst, a, b, stride, stride, stride, h);
691 691 
 }
692 692
693 
-static void put_no_rnd_pixels8_l2_c(uint8_t *dst, const uint8_t *a, const uint8_t *b, int stride, int h){
694 
-    put_no_rnd_pixels8_l2(dst, a, b, stride, stride, stride, h);
693 
+static void FUNCC(put_no_rnd_pixels8_l2)(uint8_t *dst, const uint8_t *a, const uint8_t *b, int stride, int h){
694 
+    FUNC(put_no_rnd_pixels8_l2)(dst, a, b, stride, stride, stride, h);
695 695 
 }
696 696
697 697 
 #define H264_CHROMA_MC(OPNAME, OP)\
698 
-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){\
698 
+static void FUNCC(OPNAME ## h264_chroma_mc2)(uint8_t *_dst/*align 8*/, uint8_t *_src/*align 1*/, int stride, int h, int x, int y){\
699 
+    pixel *dst = (pixel*)_dst;\
700 
+    pixel *src = (pixel*)_src;\
699 701 
     const int A=(8-x)*(8-y);\
700 702 
     const int B=(  x)*(8-y);\
701 703 
     const int C=(8-x)*(  y);\
702 704 
     const int D=(  x)*(  y);\
703 705 
     int i;\
706 
+    stride /= sizeof(pixel);\
704 707 
     \
705 708 
     assert(x<8 && y<8 && x>=0 && y>=0);\
706 709 
 \
... ... 
@@ -723,12 +770,15 @@ static void OPNAME ## h264_chroma_mc2_c(uint8_t *dst/*align 8*/, uint8_t *src/*a
723 723 
     }\
724 724 
 }\
725 725 
 \
726 
-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){\
726 
+static void FUNCC(OPNAME ## h264_chroma_mc4)(uint8_t *_dst/*align 8*/, uint8_t *_src/*align 1*/, int stride, int h, int x, int y){\
727 
+    pixel *dst = (pixel*)_dst;\
728 
+    pixel *src = (pixel*)_src;\
727 729 
     const int A=(8-x)*(8-y);\
728 730 
     const int B=(  x)*(8-y);\
729 731 
     const int C=(8-x)*(  y);\
730 732 
     const int D=(  x)*(  y);\
731 733 
     int i;\
734 
+    stride /= sizeof(pixel);\
732 735 
     \
733 736 
     assert(x<8 && y<8 && x>=0 && y>=0);\
734 737 
 \
... ... 
@@ -755,12 +805,15 @@ static void OPNAME ## h264_chroma_mc4_c(uint8_t *dst/*align 8*/, uint8_t *src/*a
755 755 
     }\
756 756 
 }\
757 757 
 \
758 
-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){\
758 
+static void FUNCC(OPNAME ## h264_chroma_mc8)(uint8_t *_dst/*align 8*/, uint8_t *_src/*align 1*/, int stride, int h, int x, int y){\
759 
+    pixel *dst = (pixel*)_dst;\
760 
+    pixel *src = (pixel*)_src;\
759 761 
     const int A=(8-x)*(8-y);\
760 762 
     const int B=(  x)*(8-y);\
761 763 
     const int C=(8-x)*(  y);\
762 764 
     const int D=(  x)*(  y);\
763 765 
     int i;\
766 
+    stride /= sizeof(pixel);\
764 767 
     \
765 768 
     assert(x<8 && y<8 && x>=0 && y>=0);\
766 769 
 \
... ... 
@@ -804,10 +857,14 @@ H264_CHROMA_MC(avg_       , op_avg)
804 804 
 #undef op_put
805 805
806 806 
 #define H264_LOWPASS(OPNAME, OP, OP2) \
807 
-static av_unused void OPNAME ## h264_qpel2_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
807 
+static av_unused void FUNC(OPNAME ## h264_qpel2_h_lowpass)(uint8_t *_dst, uint8_t *_src, int dstStride, int srcStride){\
808 808 
     const int h=2;\
809 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
809 
+    INIT_CLIP\
810 810 
     int i;\
811 
+    pixel *dst = (pixel*)_dst;\
812 
+    pixel *src = (pixel*)_src;\
813 
+    dstStride /= sizeof(pixel);\
814 
+    srcStride /= sizeof(pixel);\
811 815 
     for(i=0; i<h; i++)\
812 816 
     {\
813 817 
         OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\
... ... 
@@ -817,10 +874,14 @@ static av_unused void OPNAME ## h264_qpel2_h_lowpass(uint8_t *dst, uint8_t *src,
817 817 
     }\
818 818 
 }\
819 819 
 \
820 
-static av_unused void OPNAME ## h264_qpel2_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
820 
+static av_unused void FUNC(OPNAME ## h264_qpel2_v_lowpass)(uint8_t *_dst, uint8_t *_src, int dstStride, int srcStride){\
821 821 
     const int w=2;\
822 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
822 
+    INIT_CLIP\
823 823 
     int i;\
824 
+    pixel *dst = (pixel*)_dst;\
825 
+    pixel *src = (pixel*)_src;\
826 
+    dstStride /= sizeof(pixel);\
827 
+    srcStride /= sizeof(pixel);\
824 828 
     for(i=0; i<w; i++)\
825 829 
     {\
826 830 
         const int srcB= src[-2*srcStride];\
... ... 
@@ -837,39 +898,48 @@ static av_unused void OPNAME ## h264_qpel2_v_lowpass(uint8_t *dst, uint8_t *src,
837 837 
     }\
838 838 
 }\
839 839 
 \
840 
-static av_unused void OPNAME ## h264_qpel2_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
840 
+static av_unused void FUNC(OPNAME ## h264_qpel2_hv_lowpass)(uint8_t *_dst, int16_t *tmp, uint8_t *_src, int dstStride, int tmpStride, int srcStride){\
841 841 
     const int h=2;\
842 842 
     const int w=2;\
843 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
843 
+    const int pad = (BIT_DEPTH > 9) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\
844 
+    INIT_CLIP\
844 845 
     int i;\
846 
+    pixel *dst = (pixel*)_dst;\
847 
+    pixel *src = (pixel*)_src;\
848 
+    dstStride /= sizeof(pixel);\
849 
+    srcStride /= sizeof(pixel);\
845 850 
     src -= 2*srcStride;\
846 851 
     for(i=0; i<h+5; i++)\
847 852 
     {\
848 
-        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]);\
849 
-        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]);\
853 
+        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]) + pad;\
854 
+        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]) + pad;\
850 855 
         tmp+=tmpStride;\
851 856 
         src+=srcStride;\
852 857 
     }\
853 858 
     tmp -= tmpStride*(h+5-2);\
854 859 
     for(i=0; i<w; i++)\
855 860 
     {\
856 
-        const int tmpB= tmp[-2*tmpStride];\
857 
-        const int tmpA= tmp[-1*tmpStride];\
858 
-        const int tmp0= tmp[0 *tmpStride];\
859 
-        const int tmp1= tmp[1 *tmpStride];\
860 
-        const int tmp2= tmp[2 *tmpStride];\
861 
-        const int tmp3= tmp[3 *tmpStride];\
862 
-        const int tmp4= tmp[4 *tmpStride];\
861 
+        const int tmpB= tmp[-2*tmpStride] - pad;\
862 
+        const int tmpA= tmp[-1*tmpStride] - pad;\
863 
+        const int tmp0= tmp[0 *tmpStride] - pad;\
864 
+        const int tmp1= tmp[1 *tmpStride] - pad;\
865 
+        const int tmp2= tmp[2 *tmpStride] - pad;\
866 
+        const int tmp3= tmp[3 *tmpStride] - pad;\
867 
+        const int tmp4= tmp[4 *tmpStride] - pad;\
863 868 
         OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
864 869 
         OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
865 870 
         dst++;\
866 871 
         tmp++;\
867 872 
     }\
868 873 
 }\
869 
-static void OPNAME ## h264_qpel4_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
874 
+static void FUNC(OPNAME ## h264_qpel4_h_lowpass)(uint8_t *_dst, uint8_t *_src, int dstStride, int srcStride){\
870 875 
     const int h=4;\
871 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
876 
+    INIT_CLIP\
872 877 
     int i;\
878 
+    pixel *dst = (pixel*)_dst;\
879 
+    pixel *src = (pixel*)_src;\
880 
+    dstStride /= sizeof(pixel);\
881 
+    srcStride /= sizeof(pixel);\
873 882 
     for(i=0; i<h; i++)\
874 883 
     {\
875 884 
         OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]));\
... ... 
@@ -881,10 +951,14 @@ static void OPNAME ## h264_qpel4_h_lowpass(uint8_t *dst, uint8_t *src, int dstSt
881 881 
     }\
882 882 
 }\
883 883 
 \
884 
-static void OPNAME ## h264_qpel4_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
884 
+static void FUNC(OPNAME ## h264_qpel4_v_lowpass)(uint8_t *_dst, uint8_t *_src, int dstStride, int srcStride){\
885 885 
     const int w=4;\
886 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
886 
+    INIT_CLIP\
887 887 
     int i;\
888 
+    pixel *dst = (pixel*)_dst;\
889 
+    pixel *src = (pixel*)_src;\
890 
+    dstStride /= sizeof(pixel);\
891 
+    srcStride /= sizeof(pixel);\
888 892 
     for(i=0; i<w; i++)\
889 893 
     {\
890 894 
         const int srcB= src[-2*srcStride];\
... ... 
@@ -905,33 +979,38 @@ static void OPNAME ## h264_qpel4_v_lowpass(uint8_t *dst, uint8_t *src, int dstSt
905 905 
     }\
906 906 
 }\
907 907 
 \
908 
-static void OPNAME ## h264_qpel4_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
908 
+static void FUNC(OPNAME ## h264_qpel4_hv_lowpass)(uint8_t *_dst, int16_t *tmp, uint8_t *_src, int dstStride, int tmpStride, int srcStride){\
909 909 
     const int h=4;\
910 910 
     const int w=4;\
911 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
911 
+    const int pad = (BIT_DEPTH > 9) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\
912 
+    INIT_CLIP\
912 913 
     int i;\
914 
+    pixel *dst = (pixel*)_dst;\
915 
+    pixel *src = (pixel*)_src;\
916 
+    dstStride /= sizeof(pixel);\
917 
+    srcStride /= sizeof(pixel);\
913 918 
     src -= 2*srcStride;\
914 919 
     for(i=0; i<h+5; i++)\
915 920 
     {\
916 
-        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]);\
917 
-        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]);\
918 
-        tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]);\
919 
-        tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]);\
921 
+        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3]) + pad;\
922 
+        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4]) + pad;\
923 
+        tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5]) + pad;\
924 
+        tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6]) + pad;\
920 925 
         tmp+=tmpStride;\
921 926 
         src+=srcStride;\
922 927 
     }\
923 928 
     tmp -= tmpStride*(h+5-2);\
924 929 
     for(i=0; i<w; i++)\
925 930 
     {\
926 
-        const int tmpB= tmp[-2*tmpStride];\
927 
-        const int tmpA= tmp[-1*tmpStride];\
928 
-        const int tmp0= tmp[0 *tmpStride];\
929 
-        const int tmp1= tmp[1 *tmpStride];\
930 
-        const int tmp2= tmp[2 *tmpStride];\
931 
-        const int tmp3= tmp[3 *tmpStride];\
932 
-        const int tmp4= tmp[4 *tmpStride];\
933 
-        const int tmp5= tmp[5 *tmpStride];\
934 
-        const int tmp6= tmp[6 *tmpStride];\
931 
+        const int tmpB= tmp[-2*tmpStride] - pad;\
932 
+        const int tmpA= tmp[-1*tmpStride] - pad;\
933 
+        const int tmp0= tmp[0 *tmpStride] - pad;\
934 
+        const int tmp1= tmp[1 *tmpStride] - pad;\
935 
+        const int tmp2= tmp[2 *tmpStride] - pad;\
936 
+        const int tmp3= tmp[3 *tmpStride] - pad;\
937 
+        const int tmp4= tmp[4 *tmpStride] - pad;\
938 
+        const int tmp5= tmp[5 *tmpStride] - pad;\
939 
+        const int tmp6= tmp[6 *tmpStride] - pad;\
935 940 
         OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
936 941 
         OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
937 942 
         OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\
... ... 
@@ -941,10 +1020,14 @@ static void OPNAME ## h264_qpel4_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t
941 941 
     }\
942 942 
 }\
943 943 
 \
944 
-static void OPNAME ## h264_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
944 
+static void FUNC(OPNAME ## h264_qpel8_h_lowpass)(uint8_t *_dst, uint8_t *_src, int dstStride, int srcStride){\
945 945 
     const int h=8;\
946 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
946 
+    INIT_CLIP\
947 947 
     int i;\
948 
+    pixel *dst = (pixel*)_dst;\
949 
+    pixel *src = (pixel*)_src;\
950 
+    dstStride /= sizeof(pixel);\
951 
+    srcStride /= sizeof(pixel);\
948 952 
     for(i=0; i<h; i++)\
949 953 
     {\
950 954 
         OP(dst[0], (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]));\
... ... 
@@ -960,10 +1043,14 @@ static void OPNAME ## h264_qpel8_h_lowpass(uint8_t *dst, uint8_t *src, int dstSt
960 960 
     }\
961 961 
 }\
962 962 
 \
963 
-static void OPNAME ## h264_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
963 
+static void FUNC(OPNAME ## h264_qpel8_v_lowpass)(uint8_t *_dst, uint8_t *_src, int dstStride, int srcStride){\
964 964 
     const int w=8;\
965 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
965 
+    INIT_CLIP\
966 966 
     int i;\
967 
+    pixel *dst = (pixel*)_dst;\
968 
+    pixel *src = (pixel*)_src;\
969 
+    dstStride /= sizeof(pixel);\
970 
+    srcStride /= sizeof(pixel);\
967 971 
     for(i=0; i<w; i++)\
968 972 
     {\
969 973 
         const int srcB= src[-2*srcStride];\
... ... 
@@ -992,41 +1079,46 @@ static void OPNAME ## h264_qpel8_v_lowpass(uint8_t *dst, uint8_t *src, int dstSt
992 992 
     }\
993 993 
 }\
994 994 
 \
995 
-static void OPNAME ## h264_qpel8_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
995 
+static void FUNC(OPNAME ## h264_qpel8_hv_lowpass)(uint8_t *_dst, int16_t *tmp, uint8_t *_src, int dstStride, int tmpStride, int srcStride){\
996 996 
     const int h=8;\
997 997 
     const int w=8;\
998 
-    uint8_t *cm = ff_cropTbl + MAX_NEG_CROP;\
998 
+    const int pad = (BIT_DEPTH > 9) ? (-10 * ((1<<BIT_DEPTH)-1)) : 0;\
999 
+    INIT_CLIP\
999 1000 
     int i;\
1001 
+    pixel *dst = (pixel*)_dst;\
1002 
+    pixel *src = (pixel*)_src;\
1003 
+    dstStride /= sizeof(pixel);\
1004 
+    srcStride /= sizeof(pixel);\
1000 1005 
     src -= 2*srcStride;\
1001 1006 
     for(i=0; i<h+5; i++)\
1002 1007 
     {\
1003 
-        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]);\
1004 
-        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]);\
1005 
-        tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]);\
1006 
-        tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]);\
1007 
-        tmp[4]= (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]);\
1008 
-        tmp[5]= (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]);\
1009 
-        tmp[6]= (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]);\
1010 
-        tmp[7]= (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]);\
1008 
+        tmp[0]= (src[0]+src[1])*20 - (src[-1]+src[2])*5 + (src[-2]+src[3 ]) + pad;\
1009 
+        tmp[1]= (src[1]+src[2])*20 - (src[0 ]+src[3])*5 + (src[-1]+src[4 ]) + pad;\
1010 
+        tmp[2]= (src[2]+src[3])*20 - (src[1 ]+src[4])*5 + (src[0 ]+src[5 ]) + pad;\
1011 
+        tmp[3]= (src[3]+src[4])*20 - (src[2 ]+src[5])*5 + (src[1 ]+src[6 ]) + pad;\
1012 
+        tmp[4]= (src[4]+src[5])*20 - (src[3 ]+src[6])*5 + (src[2 ]+src[7 ]) + pad;\
1013 
+        tmp[5]= (src[5]+src[6])*20 - (src[4 ]+src[7])*5 + (src[3 ]+src[8 ]) + pad;\
1014 
+        tmp[6]= (src[6]+src[7])*20 - (src[5 ]+src[8])*5 + (src[4 ]+src[9 ]) + pad;\
1015 
+        tmp[7]= (src[7]+src[8])*20 - (src[6 ]+src[9])*5 + (src[5 ]+src[10]) + pad;\
1011 1016 
         tmp+=tmpStride;\
1012 1017 
         src+=srcStride;\
1013 1018 
     }\
1014 1019 
     tmp -= tmpStride*(h+5-2);\
1015 1020 
     for(i=0; i<w; i++)\
1016 1021 
     {\
1017 
-        const int tmpB= tmp[-2*tmpStride];\
1018 
-        const int tmpA= tmp[-1*tmpStride];\
1019 
-        const int tmp0= tmp[0 *tmpStride];\
1020 
-        const int tmp1= tmp[1 *tmpStride];\
1021 
-        const int tmp2= tmp[2 *tmpStride];\
1022 
-        const int tmp3= tmp[3 *tmpStride];\
1023 
-        const int tmp4= tmp[4 *tmpStride];\
1024 
-        const int tmp5= tmp[5 *tmpStride];\
1025 
-        const int tmp6= tmp[6 *tmpStride];\
1026 
-        const int tmp7= tmp[7 *tmpStride];\
1027 
-        const int tmp8= tmp[8 *tmpStride];\
1028 
-        const int tmp9= tmp[9 *tmpStride];\
1029 
-        const int tmp10=tmp[10*tmpStride];\
1022 
+        const int tmpB= tmp[-2*tmpStride] - pad;\
1023 
+        const int tmpA= tmp[-1*tmpStride] - pad;\
1024 
+        const int tmp0= tmp[0 *tmpStride] - pad;\
1025 
+        const int tmp1= tmp[1 *tmpStride] - pad;\
1026 
+        const int tmp2= tmp[2 *tmpStride] - pad;\
1027 
+        const int tmp3= tmp[3 *tmpStride] - pad;\
1028 
+        const int tmp4= tmp[4 *tmpStride] - pad;\
1029 
+        const int tmp5= tmp[5 *tmpStride] - pad;\
1030 
+        const int tmp6= tmp[6 *tmpStride] - pad;\
1031 
+        const int tmp7= tmp[7 *tmpStride] - pad;\
1032 
+        const int tmp8= tmp[8 *tmpStride] - pad;\
1033 
+        const int tmp9= tmp[9 *tmpStride] - pad;\
1034 
+        const int tmp10=tmp[10*tmpStride] - pad;\
1030 1035 
         OP2(dst[0*dstStride], (tmp0+tmp1)*20 - (tmpA+tmp2)*5 + (tmpB+tmp3));\
1031 1036 
         OP2(dst[1*dstStride], (tmp1+tmp2)*20 - (tmp0+tmp3)*5 + (tmpA+tmp4));\
1032 1037 
         OP2(dst[2*dstStride], (tmp2+tmp3)*20 - (tmp1+tmp4)*5 + (tmp0+tmp5));\
... ... 
@@ -1040,175 +1132,175 @@ static void OPNAME ## h264_qpel8_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t
1040 1040 
     }\
1041 1041 
 }\
1042 1042 
 \
1043 
-static void OPNAME ## h264_qpel16_v_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
1044 
-    OPNAME ## h264_qpel8_v_lowpass(dst  , src  , dstStride, srcStride);\
1045 
-    OPNAME ## h264_qpel8_v_lowpass(dst+8, src+8, dstStride, srcStride);\
1043 
+static void FUNC(OPNAME ## h264_qpel16_v_lowpass)(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
1044 
+    FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst                , src                , dstStride, srcStride);\
1045 
+    FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
1046 1046 
     src += 8*srcStride;\
1047 1047 
     dst += 8*dstStride;\
1048 
-    OPNAME ## h264_qpel8_v_lowpass(dst  , src  , dstStride, srcStride);\
1049 
-    OPNAME ## h264_qpel8_v_lowpass(dst+8, src+8, dstStride, srcStride);\
1048 
+    FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst                , src                , dstStride, srcStride);\
1049 
+    FUNC(OPNAME ## h264_qpel8_v_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
1050 1050 
 }\
1051 1051 
 \
1052 
-static void OPNAME ## h264_qpel16_h_lowpass(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
1053 
-    OPNAME ## h264_qpel8_h_lowpass(dst  , src  , dstStride, srcStride);\
1054 
-    OPNAME ## h264_qpel8_h_lowpass(dst+8, src+8, dstStride, srcStride);\
1052 
+static void FUNC(OPNAME ## h264_qpel16_h_lowpass)(uint8_t *dst, uint8_t *src, int dstStride, int srcStride){\
1053 
+    FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst                , src                , dstStride, srcStride);\
1054 
+    FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
1055 1055 
     src += 8*srcStride;\
1056 1056 
     dst += 8*dstStride;\
1057 
-    OPNAME ## h264_qpel8_h_lowpass(dst  , src  , dstStride, srcStride);\
1058 
-    OPNAME ## h264_qpel8_h_lowpass(dst+8, src+8, dstStride, srcStride);\
1057 
+    FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst                , src                , dstStride, srcStride);\
1058 
+    FUNC(OPNAME ## h264_qpel8_h_lowpass)(dst+8*sizeof(pixel), src+8*sizeof(pixel), dstStride, srcStride);\
1059 1059 
 }\
1060 1060 
 \
1061 
-static void OPNAME ## h264_qpel16_hv_lowpass(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
1062 
-    OPNAME ## h264_qpel8_hv_lowpass(dst  , tmp  , src  , dstStride, tmpStride, srcStride);\
1063 
-    OPNAME ## h264_qpel8_hv_lowpass(dst+8, tmp+8, src+8, dstStride, tmpStride, srcStride);\
1061 
+static void FUNC(OPNAME ## h264_qpel16_hv_lowpass)(uint8_t *dst, int16_t *tmp, uint8_t *src, int dstStride, int tmpStride, int srcStride){\
1062 
+    FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst                , tmp  , src                , dstStride, tmpStride, srcStride);\
1063 
+    FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst+8*sizeof(pixel), tmp+8, src+8*sizeof(pixel), dstStride, tmpStride, srcStride);\
1064 1064 
     src += 8*srcStride;\
1065 1065 
     dst += 8*dstStride;\
1066 
-    OPNAME ## h264_qpel8_hv_lowpass(dst  , tmp  , src  , dstStride, tmpStride, srcStride);\
1067 
-    OPNAME ## h264_qpel8_hv_lowpass(dst+8, tmp+8, src+8, dstStride, tmpStride, srcStride);\
1066 
+    FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst                , tmp  , src                , dstStride, tmpStride, srcStride);\
1067 
+    FUNC(OPNAME ## h264_qpel8_hv_lowpass)(dst+8*sizeof(pixel), tmp+8, src+8*sizeof(pixel), dstStride, tmpStride, srcStride);\
1068 1068 
 }\
1069 1069
1070 1070 
 #define H264_MC(OPNAME, SIZE) \
1071 
-static av_unused void OPNAME ## h264_qpel ## SIZE ## _mc00_c (uint8_t *dst, uint8_t *src, int stride){\
1072 
-    OPNAME ## pixels ## SIZE ## _c(dst, src, stride, SIZE);\
1071 
+static av_unused void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc00)(uint8_t *dst, uint8_t *src, int stride){\
1072 
+    FUNCC(OPNAME ## pixels ## SIZE)(dst, src, stride, SIZE);\
1073 1073 
 }\
1074 1074 
 \
1075 
-static void OPNAME ## h264_qpel ## SIZE ## _mc10_c(uint8_t *dst, uint8_t *src, int stride){\
1076 
-    uint8_t half[SIZE*SIZE];\
1077 
-    put_h264_qpel ## SIZE ## _h_lowpass(half, src, SIZE, stride);\
1078 
-    OPNAME ## pixels ## SIZE ## _l2(dst, src, half, stride, stride, SIZE, SIZE);\
1075 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc10)(uint8_t *dst, uint8_t *src, int stride){\
1076 
+    uint8_t half[SIZE*SIZE*sizeof(pixel)];\
1077 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(half, src, SIZE*sizeof(pixel), stride);\
1078 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, src, half, stride, stride, SIZE*sizeof(pixel), SIZE);\
1079 1079 
 }\
1080 1080 
 \
1081 
-static void OPNAME ## h264_qpel ## SIZE ## _mc20_c(uint8_t *dst, uint8_t *src, int stride){\
1082 
-    OPNAME ## h264_qpel ## SIZE ## _h_lowpass(dst, src, stride, stride);\
1081 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc20)(uint8_t *dst, uint8_t *src, int stride){\
1082 
+    FUNC(OPNAME ## h264_qpel ## SIZE ## _h_lowpass)(dst, src, stride, stride);\
1083 1083 
 }\
1084 1084 
 \
1085 
-static void OPNAME ## h264_qpel ## SIZE ## _mc30_c(uint8_t *dst, uint8_t *src, int stride){\
1086 
-    uint8_t half[SIZE*SIZE];\
1087 
-    put_h264_qpel ## SIZE ## _h_lowpass(half, src, SIZE, stride);\
1088 
-    OPNAME ## pixels ## SIZE ## _l2(dst, src+1, half, stride, stride, SIZE, SIZE);\
1085 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc30)(uint8_t *dst, uint8_t *src, int stride){\
1086 
+    uint8_t half[SIZE*SIZE*sizeof(pixel)];\
1087 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(half, src, SIZE*sizeof(pixel), stride);\
1088 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, src+sizeof(pixel), half, stride, stride, SIZE*sizeof(pixel), SIZE);\
1089 1089 
 }\
1090 1090 
 \
1091 
-static void OPNAME ## h264_qpel ## SIZE ## _mc01_c(uint8_t *dst, uint8_t *src, int stride){\
1092 
-    uint8_t full[SIZE*(SIZE+5)];\
1093 
-    uint8_t * const full_mid= full + SIZE*2;\
1094 
-    uint8_t half[SIZE*SIZE];\
1095 
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
1096 
-    put_h264_qpel ## SIZE ## _v_lowpass(half, full_mid, SIZE, SIZE);\
1097 
-    OPNAME ## pixels ## SIZE ## _l2(dst, full_mid, half, stride, SIZE, SIZE, SIZE);\
1091 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc01)(uint8_t *dst, uint8_t *src, int stride){\
1092 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1093 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1094 
+    uint8_t half[SIZE*SIZE*sizeof(pixel)];\
1095 
+    FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel),  stride, SIZE + 5);\
1096 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(half, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1097 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid, half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1098 1098 
 }\
1099 1099 
 \
1100 
-static void OPNAME ## h264_qpel ## SIZE ## _mc02_c(uint8_t *dst, uint8_t *src, int stride){\
1101 
-    uint8_t full[SIZE*(SIZE+5)];\
1102 
-    uint8_t * const full_mid= full + SIZE*2;\
1103 
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
1104 
-    OPNAME ## h264_qpel ## SIZE ## _v_lowpass(dst, full_mid, stride, SIZE);\
1100 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc02)(uint8_t *dst, uint8_t *src, int stride){\
1101 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1102 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1103 
+    FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel),  stride, SIZE + 5);\
1104 
+    FUNC(OPNAME ## h264_qpel ## SIZE ## _v_lowpass)(dst, full_mid, stride, SIZE*sizeof(pixel));\
1105 1105 
 }\
1106 1106 
 \
1107 
-static void OPNAME ## h264_qpel ## SIZE ## _mc03_c(uint8_t *dst, uint8_t *src, int stride){\
1108 
-    uint8_t full[SIZE*(SIZE+5)];\
1109 
-    uint8_t * const full_mid= full + SIZE*2;\
1110 
-    uint8_t half[SIZE*SIZE];\
1111 
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
1112 
-    put_h264_qpel ## SIZE ## _v_lowpass(half, full_mid, SIZE, SIZE);\
1113 
-    OPNAME ## pixels ## SIZE ## _l2(dst, full_mid+SIZE, half, stride, SIZE, SIZE, SIZE);\
1107 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc03)(uint8_t *dst, uint8_t *src, int stride){\
1108 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1109 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1110 
+    uint8_t half[SIZE*SIZE*sizeof(pixel)];\
1111 
+    FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel),  stride, SIZE + 5);\
1112 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(half, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1113 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, full_mid+SIZE*sizeof(pixel), half, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1114 1114 
 }\
1115 1115 
 \
1116 
-static void OPNAME ## h264_qpel ## SIZE ## _mc11_c(uint8_t *dst, uint8_t *src, int stride){\
1117 
-    uint8_t full[SIZE*(SIZE+5)];\
1118 
-    uint8_t * const full_mid= full + SIZE*2;\
1119 
-    uint8_t halfH[SIZE*SIZE];\
1120 
-    uint8_t halfV[SIZE*SIZE];\
1121 
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);\
1122 
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
1123 
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
1124 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
1116 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc11)(uint8_t *dst, uint8_t *src, int stride){\
1117 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1118 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1119 
+    uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
1120 
+    uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
1121 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src, SIZE*sizeof(pixel), stride);\
1122 
+    FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel),  stride, SIZE + 5);\
1123 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1124 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1125 1125 
 }\
1126 1126 
 \
1127 
-static void OPNAME ## h264_qpel ## SIZE ## _mc31_c(uint8_t *dst, uint8_t *src, int stride){\
1128 
-    uint8_t full[SIZE*(SIZE+5)];\
1129 
-    uint8_t * const full_mid= full + SIZE*2;\
1130 
-    uint8_t halfH[SIZE*SIZE];\
1131 
-    uint8_t halfV[SIZE*SIZE];\
1132 
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);\
1133 
-    copy_block ## SIZE (full, src - stride*2 + 1, SIZE,  stride, SIZE + 5);\
1134 
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
1135 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
1127 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc31)(uint8_t *dst, uint8_t *src, int stride){\
1128 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1129 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1130 
+    uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
1131 
+    uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
1132 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src, SIZE*sizeof(pixel), stride);\
1133 
+    FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel),  stride, SIZE + 5);\
1134 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1135 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1136 1136 
 }\
1137 1137 
 \
1138 
-static void OPNAME ## h264_qpel ## SIZE ## _mc13_c(uint8_t *dst, uint8_t *src, int stride){\
1139 
-    uint8_t full[SIZE*(SIZE+5)];\
1140 
-    uint8_t * const full_mid= full + SIZE*2;\
1141 
-    uint8_t halfH[SIZE*SIZE];\
1142 
-    uint8_t halfV[SIZE*SIZE];\
1143 
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);\
1144 
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
1145 
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
1146 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
1138 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc13)(uint8_t *dst, uint8_t *src, int stride){\
1139 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1140 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1141 
+    uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
1142 
+    uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
1143 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\
1144 
+    FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel),  stride, SIZE + 5);\
1145 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1146 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1147 1147 
 }\
1148 1148 
 \
1149 
-static void OPNAME ## h264_qpel ## SIZE ## _mc33_c(uint8_t *dst, uint8_t *src, int stride){\
1150 
-    uint8_t full[SIZE*(SIZE+5)];\
1151 
-    uint8_t * const full_mid= full + SIZE*2;\
1152 
-    uint8_t halfH[SIZE*SIZE];\
1153 
-    uint8_t halfV[SIZE*SIZE];\
1154 
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);\
1155 
-    copy_block ## SIZE (full, src - stride*2 + 1, SIZE,  stride, SIZE + 5);\
1156 
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
1157 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfV, stride, SIZE, SIZE, SIZE);\
1149 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc33)(uint8_t *dst, uint8_t *src, int stride){\
1150 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1151 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1152 
+    uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
1153 
+    uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
1154 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\
1155 
+    FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel),  stride, SIZE + 5);\
1156 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1157 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1158 1158 
 }\
1159 1159 
 \
1160 
-static void OPNAME ## h264_qpel ## SIZE ## _mc22_c(uint8_t *dst, uint8_t *src, int stride){\
1161 
-    int16_t tmp[SIZE*(SIZE+5)];\
1162 
-    OPNAME ## h264_qpel ## SIZE ## _hv_lowpass(dst, tmp, src, stride, SIZE, stride);\
1160 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc22)(uint8_t *dst, uint8_t *src, int stride){\
1161 
+    int16_t tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
1162 
+    FUNC(OPNAME ## h264_qpel ## SIZE ## _hv_lowpass)(dst, tmp, src, stride, SIZE*sizeof(pixel), stride);\
1163 1163 
 }\
1164 1164 
 \
1165 
-static void OPNAME ## h264_qpel ## SIZE ## _mc21_c(uint8_t *dst, uint8_t *src, int stride){\
1166 
-    int16_t tmp[SIZE*(SIZE+5)];\
1167 
-    uint8_t halfH[SIZE*SIZE];\
1168 
-    uint8_t halfHV[SIZE*SIZE];\
1169 
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src, SIZE, stride);\
1170 
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
1171 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfHV, stride, SIZE, SIZE, SIZE);\
1165 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc21)(uint8_t *dst, uint8_t *src, int stride){\
1166 
+    int16_t tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
1167 
+    uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
1168 
+    uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
1169 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src, SIZE*sizeof(pixel), stride);\
1170 
+    FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
1171 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1172 1172 
 }\
1173 1173 
 \
1174 
-static void OPNAME ## h264_qpel ## SIZE ## _mc23_c(uint8_t *dst, uint8_t *src, int stride){\
1175 
-    int16_t tmp[SIZE*(SIZE+5)];\
1176 
-    uint8_t halfH[SIZE*SIZE];\
1177 
-    uint8_t halfHV[SIZE*SIZE];\
1178 
-    put_h264_qpel ## SIZE ## _h_lowpass(halfH, src + stride, SIZE, stride);\
1179 
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
1180 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfH, halfHV, stride, SIZE, SIZE, SIZE);\
1174 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc23)(uint8_t *dst, uint8_t *src, int stride){\
1175 
+    int16_t tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
1176 
+    uint8_t halfH[SIZE*SIZE*sizeof(pixel)];\
1177 
+    uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
1178 
+    FUNC(put_h264_qpel ## SIZE ## _h_lowpass)(halfH, src + stride, SIZE*sizeof(pixel), stride);\
1179 
+    FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
1180 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfH, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1181 1181 
 }\
1182 1182 
 \
1183 
-static void OPNAME ## h264_qpel ## SIZE ## _mc12_c(uint8_t *dst, uint8_t *src, int stride){\
1184 
-    uint8_t full[SIZE*(SIZE+5)];\
1185 
-    uint8_t * const full_mid= full + SIZE*2;\
1186 
-    int16_t tmp[SIZE*(SIZE+5)];\
1187 
-    uint8_t halfV[SIZE*SIZE];\
1188 
-    uint8_t halfHV[SIZE*SIZE];\
1189 
-    copy_block ## SIZE (full, src - stride*2, SIZE,  stride, SIZE + 5);\
1190 
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
1191 
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
1192 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfV, halfHV, stride, SIZE, SIZE, SIZE);\
1183 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc12)(uint8_t *dst, uint8_t *src, int stride){\
1184 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1185 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1186 
+    int16_t tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
1187 
+    uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
1188 
+    uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
1189 
+    FUNC(copy_block ## SIZE )(full, src - stride*2, SIZE*sizeof(pixel),  stride, SIZE + 5);\
1190 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1191 
+    FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
1192 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfV, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1193 1193 
 }\
1194 1194 
 \
1195 
-static void OPNAME ## h264_qpel ## SIZE ## _mc32_c(uint8_t *dst, uint8_t *src, int stride){\
1196 
-    uint8_t full[SIZE*(SIZE+5)];\
1197 
-    uint8_t * const full_mid= full + SIZE*2;\
1198 
-    int16_t tmp[SIZE*(SIZE+5)];\
1199 
-    uint8_t halfV[SIZE*SIZE];\
1200 
-    uint8_t halfHV[SIZE*SIZE];\
1201 
-    copy_block ## SIZE (full, src - stride*2 + 1, SIZE,  stride, SIZE + 5);\
1202 
-    put_h264_qpel ## SIZE ## _v_lowpass(halfV, full_mid, SIZE, SIZE);\
1203 
-    put_h264_qpel ## SIZE ## _hv_lowpass(halfHV, tmp, src, SIZE, SIZE, stride);\
1204 
-    OPNAME ## pixels ## SIZE ## _l2(dst, halfV, halfHV, stride, SIZE, SIZE, SIZE);\
1195 
+static void FUNCC(OPNAME ## h264_qpel ## SIZE ## _mc32)(uint8_t *dst, uint8_t *src, int stride){\
1196 
+    uint8_t full[SIZE*(SIZE+5)*sizeof(pixel)];\
1197 
+    uint8_t * const full_mid= full + SIZE*2*sizeof(pixel);\
1198 
+    int16_t tmp[SIZE*(SIZE+5)*sizeof(pixel)];\
1199 
+    uint8_t halfV[SIZE*SIZE*sizeof(pixel)];\
1200 
+    uint8_t halfHV[SIZE*SIZE*sizeof(pixel)];\
1201 
+    FUNC(copy_block ## SIZE )(full, src - stride*2 + sizeof(pixel), SIZE*sizeof(pixel),  stride, SIZE + 5);\
1202 
+    FUNC(put_h264_qpel ## SIZE ## _v_lowpass)(halfV, full_mid, SIZE*sizeof(pixel), SIZE*sizeof(pixel));\
1203 
+    FUNC(put_h264_qpel ## SIZE ## _hv_lowpass)(halfHV, tmp, src, SIZE*sizeof(pixel), SIZE*sizeof(pixel), stride);\
1204 
+    FUNC(OPNAME ## pixels ## SIZE ## _l2)(dst, halfV, halfHV, stride, SIZE*sizeof(pixel), SIZE*sizeof(pixel), SIZE);\
1205 1205 
 }\
1206 1206
1207 
-#define op_avg(a, b)  a = (((a)+cm[((b) + 16)>>5]+1)>>1)
1207 
+#define op_avg(a, b)  a = (((a)+CLIP(((b) + 16)>>5)+1)>>1)
1208 1208 
 //#define op_avg2(a, b) a = (((a)*w1+cm[((b) + 16)>>5]*w2 + o + 64)>>7)
1209 
-#define op_put(a, b)  a = cm[((b) + 16)>>5]
1210 
-#define op2_avg(a, b)  a = (((a)+cm[((b) + 512)>>10]+1)>>1)
1211 
-#define op2_put(a, b)  a = cm[((b) + 512)>>10]
1209 
+#define op_put(a, b)  a = CLIP(((b) + 16)>>5)
1210 
+#define op2_avg(a, b)  a = (((a)+CLIP(((b) + 512)>>10)+1)>>1)
1211 
+#define op2_put(a, b)  a = CLIP(((b) + 512)>>10)
1212 1212
1213 1213 
 H264_LOWPASS(put_       , op_put, op2_put)
1214 1214 
 H264_LOWPASS(avg_       , op_avg, op2_avg)
... ... 
@@ -1230,28 +1322,28 @@ H264_MC(avg_, 16)
1230 1230 
 #define put_h264_qpel16_mc00_c ff_put_pixels16x16_c
1231 1231 
 #define avg_h264_qpel16_mc00_c ff_avg_pixels16x16_c
1232 1232
1233 
-void ff_put_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride) {
1234 
-    put_pixels8_c(dst, src, stride, 8);
1233 
+void FUNCC(ff_put_pixels8x8)(uint8_t *dst, uint8_t *src, int stride) {
1234 
+    FUNCC(put_pixels8)(dst, src, stride, 8);
1235 1235 
 }
1236 
-void ff_avg_pixels8x8_c(uint8_t *dst, uint8_t *src, int stride) {
1237 
-    avg_pixels8_c(dst, src, stride, 8);
1236 
+void FUNCC(ff_avg_pixels8x8)(uint8_t *dst, uint8_t *src, int stride) {
1237 
+    FUNCC(avg_pixels8)(dst, src, stride, 8);
1238 1238 
 }
1239 
-void ff_put_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride) {
1240 
-    put_pixels16_c(dst, src, stride, 16);
1239 
+void FUNCC(ff_put_pixels16x16)(uint8_t *dst, uint8_t *src, int stride) {
1240 
+    FUNCC(put_pixels16)(dst, src, stride, 16);
1241 1241 
 }
1242 
-void ff_avg_pixels16x16_c(uint8_t *dst, uint8_t *src, int stride) {
1243 
-    avg_pixels16_c(dst, src, stride, 16);
1242 
+void FUNCC(ff_avg_pixels16x16)(uint8_t *dst, uint8_t *src, int stride) {
1243 
+    FUNCC(avg_pixels16)(dst, src, stride, 16);
1244 1244 
 }
1245 1245
1246 
-static void clear_block_c(DCTELEM *block)
1246 
+static void FUNCC(clear_block)(DCTELEM *block)
1247 1247 
 {
1248 
-    memset(block, 0, sizeof(DCTELEM)*64);
1248 
+    memset(block, 0, sizeof(dctcoef)*64);
1249 1249 
 }
1250 1250
1251 1251 
 /**
1252 1252 
  * memset(blocks, 0, sizeof(DCTELEM)*6*64)
1253 1253 
  */
1254 
-static void clear_blocks_c(DCTELEM *blocks)
1254 
+static void FUNCC(clear_blocks)(DCTELEM *blocks)
1255 1255 
 {
1256 
-    memset(blocks, 0, sizeof(DCTELEM)*6*64);
1256 
+    memset(blocks, 0, sizeof(dctcoef)*6*64);
1257 1257 
 }