1 /*****************************************************************************
2 * predict.c: intra prediction
3 *****************************************************************************
4 * Copyright (C) 2003-2011 x264 project
6 * Authors: Laurent Aimar <fenrir@via.ecp.fr>
7 * Loren Merritt <lorenm@u.washington.edu>
8 * Fiona Glaser <fiona@x264.com>
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
24 * This program is also available under a commercial proprietary license.
25 * For more information, contact us at licensing@x264.com.
26 *****************************************************************************/
28 /* predict4x4 are inspired from ffmpeg h264 decoder */
34 # include "x86/predict.h"
37 # include "ppc/predict.h"
40 # include "arm/predict.h"
43 /****************************************************************************
44 * 16x16 prediction for intra luma block
45 ****************************************************************************/
47 #define PREDICT_16x16_DC(v)\
48 for( int i = 0; i < 16; i++ )\
50 MPIXEL_X4( src+ 0 ) = v;\
51 MPIXEL_X4( src+ 4 ) = v;\
52 MPIXEL_X4( src+ 8 ) = v;\
53 MPIXEL_X4( src+12 ) = v;\
57 void x264_predict_16x16_dc_c( pixel *src )
61 for( int i = 0; i < 16; i++ )
63 dc += src[-1 + i * FDEC_STRIDE];
64 dc += src[i - FDEC_STRIDE];
66 pixel4 dcsplat = PIXEL_SPLAT_X4( ( dc + 16 ) >> 5 );
68 PREDICT_16x16_DC( dcsplat );
70 static void x264_predict_16x16_dc_left_c( pixel *src )
74 for( int i = 0; i < 16; i++ )
75 dc += src[-1 + i * FDEC_STRIDE];
76 pixel4 dcsplat = PIXEL_SPLAT_X4( ( dc + 8 ) >> 4 );
78 PREDICT_16x16_DC( dcsplat );
80 static void x264_predict_16x16_dc_top_c( pixel *src )
84 for( int i = 0; i < 16; i++ )
85 dc += src[i - FDEC_STRIDE];
86 pixel4 dcsplat = PIXEL_SPLAT_X4( ( dc + 8 ) >> 4 );
88 PREDICT_16x16_DC( dcsplat );
90 static void x264_predict_16x16_dc_128_c( pixel *src )
92 PREDICT_16x16_DC( PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) ) );
94 void x264_predict_16x16_h_c( pixel *src )
96 for( int i = 0; i < 16; i++ )
98 const pixel4 v = PIXEL_SPLAT_X4( src[-1] );
99 MPIXEL_X4( src+ 0 ) = v;
100 MPIXEL_X4( src+ 4 ) = v;
101 MPIXEL_X4( src+ 8 ) = v;
102 MPIXEL_X4( src+12 ) = v;
106 void x264_predict_16x16_v_c( pixel *src )
108 pixel4 v0 = MPIXEL_X4( &src[ 0-FDEC_STRIDE] );
109 pixel4 v1 = MPIXEL_X4( &src[ 4-FDEC_STRIDE] );
110 pixel4 v2 = MPIXEL_X4( &src[ 8-FDEC_STRIDE] );
111 pixel4 v3 = MPIXEL_X4( &src[12-FDEC_STRIDE] );
113 for( int i = 0; i < 16; i++ )
115 MPIXEL_X4( src+ 0 ) = v0;
116 MPIXEL_X4( src+ 4 ) = v1;
117 MPIXEL_X4( src+ 8 ) = v2;
118 MPIXEL_X4( src+12 ) = v3;
122 void x264_predict_16x16_p_c( pixel *src )
126 /* calculate H and V */
127 for( int i = 0; i <= 7; i++ )
129 H += ( i + 1 ) * ( src[ 8 + i - FDEC_STRIDE ] - src[6 -i -FDEC_STRIDE] );
130 V += ( i + 1 ) * ( src[-1 + (8+i)*FDEC_STRIDE] - src[-1 + (6-i)*FDEC_STRIDE] );
133 int a = 16 * ( src[-1 + 15*FDEC_STRIDE] + src[15 - FDEC_STRIDE] );
134 int b = ( 5 * H + 32 ) >> 6;
135 int c = ( 5 * V + 32 ) >> 6;
137 int i00 = a - b * 7 - c * 7 + 16;
139 for( int y = 0; y < 16; y++ )
142 for( int x = 0; x < 16; x++ )
144 src[x] = x264_clip_pixel( pix>>5 );
153 /****************************************************************************
154 * 8x8 prediction for intra chroma block
155 ****************************************************************************/
157 static void x264_predict_8x8c_dc_128_c( pixel *src )
159 for( int y = 0; y < 8; y++ )
161 MPIXEL_X4( src+0 ) = PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) );
162 MPIXEL_X4( src+4 ) = PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) );
166 static void x264_predict_8x8c_dc_left_c( pixel *src )
168 int dc0 = 0, dc1 = 0;
170 for( int y = 0; y < 4; y++ )
172 dc0 += src[y * FDEC_STRIDE - 1];
173 dc1 += src[(y+4) * FDEC_STRIDE - 1];
175 pixel4 dc0splat = PIXEL_SPLAT_X4( ( dc0 + 2 ) >> 2 );
176 pixel4 dc1splat = PIXEL_SPLAT_X4( ( dc1 + 2 ) >> 2 );
178 for( int y = 0; y < 4; y++ )
180 MPIXEL_X4( src+0 ) = dc0splat;
181 MPIXEL_X4( src+4 ) = dc0splat;
184 for( int y = 0; y < 4; y++ )
186 MPIXEL_X4( src+0 ) = dc1splat;
187 MPIXEL_X4( src+4 ) = dc1splat;
192 static void x264_predict_8x8c_dc_top_c( pixel *src )
194 int dc0 = 0, dc1 = 0;
196 for( int x = 0; x < 4; x++ )
198 dc0 += src[x - FDEC_STRIDE];
199 dc1 += src[x + 4 - FDEC_STRIDE];
201 pixel4 dc0splat = PIXEL_SPLAT_X4( ( dc0 + 2 ) >> 2 );
202 pixel4 dc1splat = PIXEL_SPLAT_X4( ( dc1 + 2 ) >> 2 );
204 for( int y = 0; y < 8; y++ )
206 MPIXEL_X4( src+0 ) = dc0splat;
207 MPIXEL_X4( src+4 ) = dc1splat;
211 void x264_predict_8x8c_dc_c( pixel *src )
213 int s0 = 0, s1 = 0, s2 = 0, s3 = 0;
220 for( int i = 0; i < 4; i++ )
222 s0 += src[i - FDEC_STRIDE];
223 s1 += src[i + 4 - FDEC_STRIDE];
224 s2 += src[-1 + i * FDEC_STRIDE];
225 s3 += src[-1 + (i+4)*FDEC_STRIDE];
231 pixel4 dc0 = PIXEL_SPLAT_X4( ( s0 + s2 + 4 ) >> 3 );
232 pixel4 dc1 = PIXEL_SPLAT_X4( ( s1 + 2 ) >> 2 );
233 pixel4 dc2 = PIXEL_SPLAT_X4( ( s3 + 2 ) >> 2 );
234 pixel4 dc3 = PIXEL_SPLAT_X4( ( s1 + s3 + 4 ) >> 3 );
236 for( int y = 0; y < 4; y++ )
238 MPIXEL_X4( src+0 ) = dc0;
239 MPIXEL_X4( src+4 ) = dc1;
243 for( int y = 0; y < 4; y++ )
245 MPIXEL_X4( src+0 ) = dc2;
246 MPIXEL_X4( src+4 ) = dc3;
250 void x264_predict_8x8c_h_c( pixel *src )
252 for( int i = 0; i < 8; i++ )
254 pixel4 v = PIXEL_SPLAT_X4( src[-1] );
255 MPIXEL_X4( src+0 ) = v;
256 MPIXEL_X4( src+4 ) = v;
260 void x264_predict_8x8c_v_c( pixel *src )
262 pixel4 v0 = MPIXEL_X4( src+0-FDEC_STRIDE );
263 pixel4 v1 = MPIXEL_X4( src+4-FDEC_STRIDE );
265 for( int i = 0; i < 8; i++ )
267 MPIXEL_X4( src+0 ) = v0;
268 MPIXEL_X4( src+4 ) = v1;
272 void x264_predict_8x8c_p_c( pixel *src )
276 for( int i = 0; i < 4; i++ )
278 H += ( i + 1 ) * ( src[4+i - FDEC_STRIDE] - src[2 - i -FDEC_STRIDE] );
279 V += ( i + 1 ) * ( src[-1 +(i+4)*FDEC_STRIDE] - src[-1+(2-i)*FDEC_STRIDE] );
282 int a = 16 * ( src[-1+7*FDEC_STRIDE] + src[7 - FDEC_STRIDE] );
283 int b = ( 17 * H + 16 ) >> 5;
284 int c = ( 17 * V + 16 ) >> 5;
285 int i00 = a -3*b -3*c + 16;
287 for( int y = 0; y < 8; y++ )
290 for( int x = 0; x < 8; x++ )
292 src[x] = x264_clip_pixel( pix>>5 );
300 /****************************************************************************
301 * 4x4 prediction for intra luma block
302 ****************************************************************************/
304 #define SRC(x,y) src[(x)+(y)*FDEC_STRIDE]
305 #define SRC_X4(x,y) MPIXEL_X4( &SRC(x,y) )
307 #define PREDICT_4x4_DC(v)\
308 SRC_X4(0,0) = SRC_X4(0,1) = SRC_X4(0,2) = SRC_X4(0,3) = v;
310 static void x264_predict_4x4_dc_128_c( pixel *src )
312 PREDICT_4x4_DC( PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) ) );
314 static void x264_predict_4x4_dc_left_c( pixel *src )
316 pixel4 dc = PIXEL_SPLAT_X4( (SRC(-1,0) + SRC(-1,1) + SRC(-1,2) + SRC(-1,3) + 2) >> 2 );
317 PREDICT_4x4_DC( dc );
319 static void x264_predict_4x4_dc_top_c( pixel *src )
321 pixel4 dc = PIXEL_SPLAT_X4( (SRC(0,-1) + SRC(1,-1) + SRC(2,-1) + SRC(3,-1) + 2) >> 2 );
322 PREDICT_4x4_DC( dc );
324 void x264_predict_4x4_dc_c( pixel *src )
326 pixel4 dc = PIXEL_SPLAT_X4( (SRC(-1,0) + SRC(-1,1) + SRC(-1,2) + SRC(-1,3) +
327 SRC(0,-1) + SRC(1,-1) + SRC(2,-1) + SRC(3,-1) + 4) >> 3 );
328 PREDICT_4x4_DC( dc );
330 void x264_predict_4x4_h_c( pixel *src )
332 SRC_X4(0,0) = PIXEL_SPLAT_X4( SRC(-1,0) );
333 SRC_X4(0,1) = PIXEL_SPLAT_X4( SRC(-1,1) );
334 SRC_X4(0,2) = PIXEL_SPLAT_X4( SRC(-1,2) );
335 SRC_X4(0,3) = PIXEL_SPLAT_X4( SRC(-1,3) );
337 void x264_predict_4x4_v_c( pixel *src )
339 PREDICT_4x4_DC(SRC_X4(0,-1));
342 #define PREDICT_4x4_LOAD_LEFT\
346 UNUSED int l3 = SRC(-1,3);
348 #define PREDICT_4x4_LOAD_TOP\
352 UNUSED int t3 = SRC(3,-1);
354 #define PREDICT_4x4_LOAD_TOP_RIGHT\
358 UNUSED int t7 = SRC(7,-1);
360 #define F1(a,b) (((a)+(b)+1)>>1)
361 #define F2(a,b,c) (((a)+2*(b)+(c)+2)>>2)
363 static void x264_predict_4x4_ddl_c( pixel *src )
366 PREDICT_4x4_LOAD_TOP_RIGHT
367 SRC(0,0)= F2(t0,t1,t2);
368 SRC(1,0)=SRC(0,1)= F2(t1,t2,t3);
369 SRC(2,0)=SRC(1,1)=SRC(0,2)= F2(t2,t3,t4);
370 SRC(3,0)=SRC(2,1)=SRC(1,2)=SRC(0,3)= F2(t3,t4,t5);
371 SRC(3,1)=SRC(2,2)=SRC(1,3)= F2(t4,t5,t6);
372 SRC(3,2)=SRC(2,3)= F2(t5,t6,t7);
373 SRC(3,3)= F2(t6,t7,t7);
375 static void x264_predict_4x4_ddr_c( pixel *src )
378 PREDICT_4x4_LOAD_LEFT
380 SRC(3,0)= F2(t3,t2,t1);
381 SRC(2,0)=SRC(3,1)= F2(t2,t1,t0);
382 SRC(1,0)=SRC(2,1)=SRC(3,2)= F2(t1,t0,lt);
383 SRC(0,0)=SRC(1,1)=SRC(2,2)=SRC(3,3)= F2(t0,lt,l0);
384 SRC(0,1)=SRC(1,2)=SRC(2,3)= F2(lt,l0,l1);
385 SRC(0,2)=SRC(1,3)= F2(l0,l1,l2);
386 SRC(0,3)= F2(l1,l2,l3);
389 static void x264_predict_4x4_vr_c( pixel *src )
392 PREDICT_4x4_LOAD_LEFT
394 SRC(0,3)= F2(l2,l1,l0);
395 SRC(0,2)= F2(l1,l0,lt);
396 SRC(0,1)=SRC(1,3)= F2(l0,lt,t0);
397 SRC(0,0)=SRC(1,2)= F1(lt,t0);
398 SRC(1,1)=SRC(2,3)= F2(lt,t0,t1);
399 SRC(1,0)=SRC(2,2)= F1(t0,t1);
400 SRC(2,1)=SRC(3,3)= F2(t0,t1,t2);
401 SRC(2,0)=SRC(3,2)= F1(t1,t2);
402 SRC(3,1)= F2(t1,t2,t3);
406 static void x264_predict_4x4_hd_c( pixel *src )
409 PREDICT_4x4_LOAD_LEFT
412 SRC(1,3)= F2(l1,l2,l3);
413 SRC(0,2)=SRC(2,3)= F1(l1,l2);
414 SRC(1,2)=SRC(3,3)= F2(l0,l1,l2);
415 SRC(0,1)=SRC(2,2)= F1(l0,l1);
416 SRC(1,1)=SRC(3,2)= F2(lt,l0,l1);
417 SRC(0,0)=SRC(2,1)= F1(lt,l0);
418 SRC(1,0)=SRC(3,1)= F2(t0,lt,l0);
419 SRC(2,0)= F2(t1,t0,lt);
420 SRC(3,0)= F2(t2,t1,t0);
423 static void x264_predict_4x4_vl_c( pixel *src )
426 PREDICT_4x4_LOAD_TOP_RIGHT
428 SRC(0,1)= F2(t0,t1,t2);
429 SRC(1,0)=SRC(0,2)= F1(t1,t2);
430 SRC(1,1)=SRC(0,3)= F2(t1,t2,t3);
431 SRC(2,0)=SRC(1,2)= F1(t2,t3);
432 SRC(2,1)=SRC(1,3)= F2(t2,t3,t4);
433 SRC(3,0)=SRC(2,2)= F1(t3,t4);
434 SRC(3,1)=SRC(2,3)= F2(t3,t4,t5);
436 SRC(3,3)= F2(t4,t5,t6);
439 static void x264_predict_4x4_hu_c( pixel *src )
441 PREDICT_4x4_LOAD_LEFT
443 SRC(1,0)= F2(l0,l1,l2);
444 SRC(2,0)=SRC(0,1)= F1(l1,l2);
445 SRC(3,0)=SRC(1,1)= F2(l1,l2,l3);
446 SRC(2,1)=SRC(0,2)= F1(l2,l3);
447 SRC(3,1)=SRC(1,2)= F2(l2,l3,l3);
448 SRC(3,2)=SRC(1,3)=SRC(0,3)=
449 SRC(2,2)=SRC(2,3)=SRC(3,3)= l3;
452 /****************************************************************************
453 * 8x8 prediction for intra luma block
454 ****************************************************************************/
457 edge[14-y] = F2(SRC(-1,y-1), SRC(-1,y), SRC(-1,y+1));
459 edge[16+x] = F2(SRC(x-1,-1), SRC(x,-1), SRC(x+1,-1));
461 static void x264_predict_8x8_filter_c( pixel *src, pixel edge[33], int i_neighbor, int i_filters )
463 /* edge[7..14] = l7..l0
465 * edge[16..31] = t0 .. t15
468 int have_lt = i_neighbor & MB_TOPLEFT;
469 if( i_filters & MB_LEFT )
471 edge[15] = (SRC(0,-1) + 2*SRC(-1,-1) + SRC(-1,0) + 2) >> 2;
472 edge[14] = ((have_lt ? SRC(-1,-1) : SRC(-1,0))
473 + 2*SRC(-1,0) + SRC(-1,1) + 2) >> 2;
474 PL(1) PL(2) PL(3) PL(4) PL(5) PL(6)
475 edge[7] = (SRC(-1,6) + 3*SRC(-1,7) + 2) >> 2;
478 if( i_filters & MB_TOP )
480 int have_tr = i_neighbor & MB_TOPRIGHT;
481 edge[16] = ((have_lt ? SRC(-1,-1) : SRC(0,-1))
482 + 2*SRC(0,-1) + SRC(1,-1) + 2) >> 2;
483 PT(1) PT(2) PT(3) PT(4) PT(5) PT(6)
484 edge[23] = (SRC(6,-1) + 2*SRC(7,-1)
485 + (have_tr ? SRC(8,-1) : SRC(7,-1)) + 2) >> 2;
487 if( i_filters & MB_TOPRIGHT )
491 PT(8) PT(9) PT(10) PT(11) PT(12) PT(13) PT(14)
493 edge[32] = (SRC(14,-1) + 3*SRC(15,-1) + 2) >> 2;
497 MPIXEL_X4( edge+24 ) = PIXEL_SPLAT_X4( SRC(7,-1) );
498 MPIXEL_X4( edge+28 ) = PIXEL_SPLAT_X4( SRC(7,-1) );
499 edge[32] = SRC(7,-1);
509 UNUSED int l##y = edge[14-y];
511 UNUSED int t##x = edge[16+x];
512 #define PREDICT_8x8_LOAD_TOPLEFT \
514 #define PREDICT_8x8_LOAD_LEFT \
515 PL(0) PL(1) PL(2) PL(3) PL(4) PL(5) PL(6) PL(7)
516 #define PREDICT_8x8_LOAD_TOP \
517 PT(0) PT(1) PT(2) PT(3) PT(4) PT(5) PT(6) PT(7)
518 #define PREDICT_8x8_LOAD_TOPRIGHT \
519 PT(8) PT(9) PT(10) PT(11) PT(12) PT(13) PT(14) PT(15)
521 #define PREDICT_8x8_DC(v) \
522 for( int y = 0; y < 8; y++ ) { \
523 MPIXEL_X4( src+0 ) = v; \
524 MPIXEL_X4( src+4 ) = v; \
525 src += FDEC_STRIDE; \
528 static void x264_predict_8x8_dc_128_c( pixel *src, pixel edge[33] )
530 PREDICT_8x8_DC( PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) ) );
532 static void x264_predict_8x8_dc_left_c( pixel *src, pixel edge[33] )
534 PREDICT_8x8_LOAD_LEFT
535 pixel4 dc = PIXEL_SPLAT_X4( (l0+l1+l2+l3+l4+l5+l6+l7+4) >> 3 );
536 PREDICT_8x8_DC( dc );
538 static void x264_predict_8x8_dc_top_c( pixel *src, pixel edge[33] )
541 pixel4 dc = PIXEL_SPLAT_X4( (t0+t1+t2+t3+t4+t5+t6+t7+4) >> 3 );
542 PREDICT_8x8_DC( dc );
544 void x264_predict_8x8_dc_c( pixel *src, pixel edge[33] )
546 PREDICT_8x8_LOAD_LEFT
548 pixel4 dc = PIXEL_SPLAT_X4( (l0+l1+l2+l3+l4+l5+l6+l7+t0+t1+t2+t3+t4+t5+t6+t7+8) >> 4 );
549 PREDICT_8x8_DC( dc );
551 void x264_predict_8x8_h_c( pixel *src, pixel edge[33] )
553 PREDICT_8x8_LOAD_LEFT
554 #define ROW(y) MPIXEL_X4( src+y*FDEC_STRIDE+0 ) =\
555 MPIXEL_X4( src+y*FDEC_STRIDE+4 ) = PIXEL_SPLAT_X4( l##y );
556 ROW(0); ROW(1); ROW(2); ROW(3); ROW(4); ROW(5); ROW(6); ROW(7);
559 void x264_predict_8x8_v_c( pixel *src, pixel edge[33] )
561 pixel4 top[2] = { MPIXEL_X4( edge+16 ),
562 MPIXEL_X4( edge+20 ) };
563 for( int y = 0; y < 8; y++ )
565 MPIXEL_X4( src+y*FDEC_STRIDE+0 ) = top[0];
566 MPIXEL_X4( src+y*FDEC_STRIDE+4 ) = top[1];
569 static void x264_predict_8x8_ddl_c( pixel *src, pixel edge[33] )
572 PREDICT_8x8_LOAD_TOPRIGHT
573 SRC(0,0)= F2(t0,t1,t2);
574 SRC(0,1)=SRC(1,0)= F2(t1,t2,t3);
575 SRC(0,2)=SRC(1,1)=SRC(2,0)= F2(t2,t3,t4);
576 SRC(0,3)=SRC(1,2)=SRC(2,1)=SRC(3,0)= F2(t3,t4,t5);
577 SRC(0,4)=SRC(1,3)=SRC(2,2)=SRC(3,1)=SRC(4,0)= F2(t4,t5,t6);
578 SRC(0,5)=SRC(1,4)=SRC(2,3)=SRC(3,2)=SRC(4,1)=SRC(5,0)= F2(t5,t6,t7);
579 SRC(0,6)=SRC(1,5)=SRC(2,4)=SRC(3,3)=SRC(4,2)=SRC(5,1)=SRC(6,0)= F2(t6,t7,t8);
580 SRC(0,7)=SRC(1,6)=SRC(2,5)=SRC(3,4)=SRC(4,3)=SRC(5,2)=SRC(6,1)=SRC(7,0)= F2(t7,t8,t9);
581 SRC(1,7)=SRC(2,6)=SRC(3,5)=SRC(4,4)=SRC(5,3)=SRC(6,2)=SRC(7,1)= F2(t8,t9,t10);
582 SRC(2,7)=SRC(3,6)=SRC(4,5)=SRC(5,4)=SRC(6,3)=SRC(7,2)= F2(t9,t10,t11);
583 SRC(3,7)=SRC(4,6)=SRC(5,5)=SRC(6,4)=SRC(7,3)= F2(t10,t11,t12);
584 SRC(4,7)=SRC(5,6)=SRC(6,5)=SRC(7,4)= F2(t11,t12,t13);
585 SRC(5,7)=SRC(6,6)=SRC(7,5)= F2(t12,t13,t14);
586 SRC(6,7)=SRC(7,6)= F2(t13,t14,t15);
587 SRC(7,7)= F2(t14,t15,t15);
589 static void x264_predict_8x8_ddr_c( pixel *src, pixel edge[33] )
592 PREDICT_8x8_LOAD_LEFT
593 PREDICT_8x8_LOAD_TOPLEFT
594 SRC(0,7)= F2(l7,l6,l5);
595 SRC(0,6)=SRC(1,7)= F2(l6,l5,l4);
596 SRC(0,5)=SRC(1,6)=SRC(2,7)= F2(l5,l4,l3);
597 SRC(0,4)=SRC(1,5)=SRC(2,6)=SRC(3,7)= F2(l4,l3,l2);
598 SRC(0,3)=SRC(1,4)=SRC(2,5)=SRC(3,6)=SRC(4,7)= F2(l3,l2,l1);
599 SRC(0,2)=SRC(1,3)=SRC(2,4)=SRC(3,5)=SRC(4,6)=SRC(5,7)= F2(l2,l1,l0);
600 SRC(0,1)=SRC(1,2)=SRC(2,3)=SRC(3,4)=SRC(4,5)=SRC(5,6)=SRC(6,7)= F2(l1,l0,lt);
601 SRC(0,0)=SRC(1,1)=SRC(2,2)=SRC(3,3)=SRC(4,4)=SRC(5,5)=SRC(6,6)=SRC(7,7)= F2(l0,lt,t0);
602 SRC(1,0)=SRC(2,1)=SRC(3,2)=SRC(4,3)=SRC(5,4)=SRC(6,5)=SRC(7,6)= F2(lt,t0,t1);
603 SRC(2,0)=SRC(3,1)=SRC(4,2)=SRC(5,3)=SRC(6,4)=SRC(7,5)= F2(t0,t1,t2);
604 SRC(3,0)=SRC(4,1)=SRC(5,2)=SRC(6,3)=SRC(7,4)= F2(t1,t2,t3);
605 SRC(4,0)=SRC(5,1)=SRC(6,2)=SRC(7,3)= F2(t2,t3,t4);
606 SRC(5,0)=SRC(6,1)=SRC(7,2)= F2(t3,t4,t5);
607 SRC(6,0)=SRC(7,1)= F2(t4,t5,t6);
608 SRC(7,0)= F2(t5,t6,t7);
611 static void x264_predict_8x8_vr_c( pixel *src, pixel edge[33] )
614 PREDICT_8x8_LOAD_LEFT
615 PREDICT_8x8_LOAD_TOPLEFT
616 SRC(0,6)= F2(l5,l4,l3);
617 SRC(0,7)= F2(l6,l5,l4);
618 SRC(0,4)=SRC(1,6)= F2(l3,l2,l1);
619 SRC(0,5)=SRC(1,7)= F2(l4,l3,l2);
620 SRC(0,2)=SRC(1,4)=SRC(2,6)= F2(l1,l0,lt);
621 SRC(0,3)=SRC(1,5)=SRC(2,7)= F2(l2,l1,l0);
622 SRC(0,1)=SRC(1,3)=SRC(2,5)=SRC(3,7)= F2(l0,lt,t0);
623 SRC(0,0)=SRC(1,2)=SRC(2,4)=SRC(3,6)= F1(lt,t0);
624 SRC(1,1)=SRC(2,3)=SRC(3,5)=SRC(4,7)= F2(lt,t0,t1);
625 SRC(1,0)=SRC(2,2)=SRC(3,4)=SRC(4,6)= F1(t0,t1);
626 SRC(2,1)=SRC(3,3)=SRC(4,5)=SRC(5,7)= F2(t0,t1,t2);
627 SRC(2,0)=SRC(3,2)=SRC(4,4)=SRC(5,6)= F1(t1,t2);
628 SRC(3,1)=SRC(4,3)=SRC(5,5)=SRC(6,7)= F2(t1,t2,t3);
629 SRC(3,0)=SRC(4,2)=SRC(5,4)=SRC(6,6)= F1(t2,t3);
630 SRC(4,1)=SRC(5,3)=SRC(6,5)=SRC(7,7)= F2(t2,t3,t4);
631 SRC(4,0)=SRC(5,2)=SRC(6,4)=SRC(7,6)= F1(t3,t4);
632 SRC(5,1)=SRC(6,3)=SRC(7,5)= F2(t3,t4,t5);
633 SRC(5,0)=SRC(6,2)=SRC(7,4)= F1(t4,t5);
634 SRC(6,1)=SRC(7,3)= F2(t4,t5,t6);
635 SRC(6,0)=SRC(7,2)= F1(t5,t6);
636 SRC(7,1)= F2(t5,t6,t7);
639 static void x264_predict_8x8_hd_c( pixel *src, pixel edge[33] )
642 PREDICT_8x8_LOAD_LEFT
643 PREDICT_8x8_LOAD_TOPLEFT
644 int p1 = pack_pixel_1to2(F1(l6,l7), F2(l5,l6,l7));
645 int p2 = pack_pixel_1to2(F1(l5,l6), F2(l4,l5,l6));
646 int p3 = pack_pixel_1to2(F1(l4,l5), F2(l3,l4,l5));
647 int p4 = pack_pixel_1to2(F1(l3,l4), F2(l2,l3,l4));
648 int p5 = pack_pixel_1to2(F1(l2,l3), F2(l1,l2,l3));
649 int p6 = pack_pixel_1to2(F1(l1,l2), F2(l0,l1,l2));
650 int p7 = pack_pixel_1to2(F1(l0,l1), F2(lt,l0,l1));
651 int p8 = pack_pixel_1to2(F1(lt,l0), F2(l0,lt,t0));
652 int p9 = pack_pixel_1to2(F2(t1,t0,lt), F2(t2,t1,t0));
653 int p10 = pack_pixel_1to2(F2(t3,t2,t1), F2(t4,t3,t2));
654 int p11 = pack_pixel_1to2(F2(t5,t4,t3), F2(t6,t5,t4));
655 SRC_X4(0,7)= pack_pixel_2to4(p1,p2);
656 SRC_X4(0,6)= pack_pixel_2to4(p2,p3);
657 SRC_X4(4,7)=SRC_X4(0,5)= pack_pixel_2to4(p3,p4);
658 SRC_X4(4,6)=SRC_X4(0,4)= pack_pixel_2to4(p4,p5);
659 SRC_X4(4,5)=SRC_X4(0,3)= pack_pixel_2to4(p5,p6);
660 SRC_X4(4,4)=SRC_X4(0,2)= pack_pixel_2to4(p6,p7);
661 SRC_X4(4,3)=SRC_X4(0,1)= pack_pixel_2to4(p7,p8);
662 SRC_X4(4,2)=SRC_X4(0,0)= pack_pixel_2to4(p8,p9);
663 SRC_X4(4,1)= pack_pixel_2to4(p9,p10);
664 SRC_X4(4,0)= pack_pixel_2to4(p10,p11);
666 static void x264_predict_8x8_vl_c( pixel *src, pixel edge[33] )
669 PREDICT_8x8_LOAD_TOPRIGHT
671 SRC(0,1)= F2(t0,t1,t2);
672 SRC(0,2)=SRC(1,0)= F1(t1,t2);
673 SRC(0,3)=SRC(1,1)= F2(t1,t2,t3);
674 SRC(0,4)=SRC(1,2)=SRC(2,0)= F1(t2,t3);
675 SRC(0,5)=SRC(1,3)=SRC(2,1)= F2(t2,t3,t4);
676 SRC(0,6)=SRC(1,4)=SRC(2,2)=SRC(3,0)= F1(t3,t4);
677 SRC(0,7)=SRC(1,5)=SRC(2,3)=SRC(3,1)= F2(t3,t4,t5);
678 SRC(1,6)=SRC(2,4)=SRC(3,2)=SRC(4,0)= F1(t4,t5);
679 SRC(1,7)=SRC(2,5)=SRC(3,3)=SRC(4,1)= F2(t4,t5,t6);
680 SRC(2,6)=SRC(3,4)=SRC(4,2)=SRC(5,0)= F1(t5,t6);
681 SRC(2,7)=SRC(3,5)=SRC(4,3)=SRC(5,1)= F2(t5,t6,t7);
682 SRC(3,6)=SRC(4,4)=SRC(5,2)=SRC(6,0)= F1(t6,t7);
683 SRC(3,7)=SRC(4,5)=SRC(5,3)=SRC(6,1)= F2(t6,t7,t8);
684 SRC(4,6)=SRC(5,4)=SRC(6,2)=SRC(7,0)= F1(t7,t8);
685 SRC(4,7)=SRC(5,5)=SRC(6,3)=SRC(7,1)= F2(t7,t8,t9);
686 SRC(5,6)=SRC(6,4)=SRC(7,2)= F1(t8,t9);
687 SRC(5,7)=SRC(6,5)=SRC(7,3)= F2(t8,t9,t10);
688 SRC(6,6)=SRC(7,4)= F1(t9,t10);
689 SRC(6,7)=SRC(7,5)= F2(t9,t10,t11);
690 SRC(7,6)= F1(t10,t11);
691 SRC(7,7)= F2(t10,t11,t12);
693 static void x264_predict_8x8_hu_c( pixel *src, pixel edge[33] )
695 PREDICT_8x8_LOAD_LEFT
696 int p1 = pack_pixel_1to2(F1(l0,l1), F2(l0,l1,l2));
697 int p2 = pack_pixel_1to2(F1(l1,l2), F2(l1,l2,l3));
698 int p3 = pack_pixel_1to2(F1(l2,l3), F2(l2,l3,l4));
699 int p4 = pack_pixel_1to2(F1(l3,l4), F2(l3,l4,l5));
700 int p5 = pack_pixel_1to2(F1(l4,l5), F2(l4,l5,l6));
701 int p6 = pack_pixel_1to2(F1(l5,l6), F2(l5,l6,l7));
702 int p7 = pack_pixel_1to2(F1(l6,l7), F2(l6,l7,l7));
703 int p8 = pack_pixel_1to2(l7,l7);
704 SRC_X4(0,0)= pack_pixel_2to4(p1,p2);
705 SRC_X4(0,1)= pack_pixel_2to4(p2,p3);
706 SRC_X4(4,0)=SRC_X4(0,2)= pack_pixel_2to4(p3,p4);
707 SRC_X4(4,1)=SRC_X4(0,3)= pack_pixel_2to4(p4,p5);
708 SRC_X4(4,2)=SRC_X4(0,4)= pack_pixel_2to4(p5,p6);
709 SRC_X4(4,3)=SRC_X4(0,5)= pack_pixel_2to4(p6,p7);
710 SRC_X4(4,4)=SRC_X4(0,6)= pack_pixel_2to4(p7,p8);
711 SRC_X4(4,5)=SRC_X4(4,6)= SRC_X4(0,7) = SRC_X4(4,7) = pack_pixel_2to4(p8,p8);
714 /****************************************************************************
715 * Exported functions:
716 ****************************************************************************/
717 void x264_predict_16x16_init( int cpu, x264_predict_t pf[7] )
719 pf[I_PRED_16x16_V ] = x264_predict_16x16_v_c;
720 pf[I_PRED_16x16_H ] = x264_predict_16x16_h_c;
721 pf[I_PRED_16x16_DC] = x264_predict_16x16_dc_c;
722 pf[I_PRED_16x16_P ] = x264_predict_16x16_p_c;
723 pf[I_PRED_16x16_DC_LEFT]= x264_predict_16x16_dc_left_c;
724 pf[I_PRED_16x16_DC_TOP ]= x264_predict_16x16_dc_top_c;
725 pf[I_PRED_16x16_DC_128 ]= x264_predict_16x16_dc_128_c;
728 x264_predict_16x16_init_mmx( cpu, pf );
732 if( cpu&X264_CPU_ALTIVEC )
733 x264_predict_16x16_init_altivec( pf );
737 x264_predict_16x16_init_arm( cpu, pf );
741 void x264_predict_8x8c_init( int cpu, x264_predict_t pf[7] )
743 pf[I_PRED_CHROMA_V ] = x264_predict_8x8c_v_c;
744 pf[I_PRED_CHROMA_H ] = x264_predict_8x8c_h_c;
745 pf[I_PRED_CHROMA_DC] = x264_predict_8x8c_dc_c;
746 pf[I_PRED_CHROMA_P ] = x264_predict_8x8c_p_c;
747 pf[I_PRED_CHROMA_DC_LEFT]= x264_predict_8x8c_dc_left_c;
748 pf[I_PRED_CHROMA_DC_TOP ]= x264_predict_8x8c_dc_top_c;
749 pf[I_PRED_CHROMA_DC_128 ]= x264_predict_8x8c_dc_128_c;
752 x264_predict_8x8c_init_mmx( cpu, pf );
756 if( cpu&X264_CPU_ALTIVEC )
757 x264_predict_8x8c_init_altivec( pf );
761 x264_predict_8x8c_init_arm( cpu, pf );
765 void x264_predict_8x8_init( int cpu, x264_predict8x8_t pf[12], x264_predict_8x8_filter_t *predict_filter )
767 pf[I_PRED_8x8_V] = x264_predict_8x8_v_c;
768 pf[I_PRED_8x8_H] = x264_predict_8x8_h_c;
769 pf[I_PRED_8x8_DC] = x264_predict_8x8_dc_c;
770 pf[I_PRED_8x8_DDL] = x264_predict_8x8_ddl_c;
771 pf[I_PRED_8x8_DDR] = x264_predict_8x8_ddr_c;
772 pf[I_PRED_8x8_VR] = x264_predict_8x8_vr_c;
773 pf[I_PRED_8x8_HD] = x264_predict_8x8_hd_c;
774 pf[I_PRED_8x8_VL] = x264_predict_8x8_vl_c;
775 pf[I_PRED_8x8_HU] = x264_predict_8x8_hu_c;
776 pf[I_PRED_8x8_DC_LEFT]= x264_predict_8x8_dc_left_c;
777 pf[I_PRED_8x8_DC_TOP] = x264_predict_8x8_dc_top_c;
778 pf[I_PRED_8x8_DC_128] = x264_predict_8x8_dc_128_c;
779 *predict_filter = x264_predict_8x8_filter_c;
782 x264_predict_8x8_init_mmx( cpu, pf, predict_filter );
786 x264_predict_8x8_init_arm( cpu, pf, predict_filter );
790 void x264_predict_4x4_init( int cpu, x264_predict_t pf[12] )
792 pf[I_PRED_4x4_V] = x264_predict_4x4_v_c;
793 pf[I_PRED_4x4_H] = x264_predict_4x4_h_c;
794 pf[I_PRED_4x4_DC] = x264_predict_4x4_dc_c;
795 pf[I_PRED_4x4_DDL] = x264_predict_4x4_ddl_c;
796 pf[I_PRED_4x4_DDR] = x264_predict_4x4_ddr_c;
797 pf[I_PRED_4x4_VR] = x264_predict_4x4_vr_c;
798 pf[I_PRED_4x4_HD] = x264_predict_4x4_hd_c;
799 pf[I_PRED_4x4_VL] = x264_predict_4x4_vl_c;
800 pf[I_PRED_4x4_HU] = x264_predict_4x4_hu_c;
801 pf[I_PRED_4x4_DC_LEFT]= x264_predict_4x4_dc_left_c;
802 pf[I_PRED_4x4_DC_TOP] = x264_predict_4x4_dc_top_c;
803 pf[I_PRED_4x4_DC_128] = x264_predict_4x4_dc_128_c;
806 x264_predict_4x4_init_mmx( cpu, pf );
810 x264_predict_4x4_init_arm( cpu, pf );