1 /*****************************************************************************
2 * predict.c: intra prediction
3 *****************************************************************************
4 * Copyright (C) 2003-2013 x264 project
6 * Authors: Laurent Aimar <fenrir@via.ecp.fr>
7 * Loren Merritt <lorenm@u.washington.edu>
8 * Fiona Glaser <fiona@x264.com>
9 * Henrik Gramner <hengar-6@student.ltu.se>
11 * This program is free software; you can redistribute it and/or modify
12 * it under the terms of the GNU General Public License as published by
13 * the Free Software Foundation; either version 2 of the License, or
14 * (at your option) any later version.
16 * This program is distributed in the hope that it will be useful,
17 * but WITHOUT ANY WARRANTY; without even the implied warranty of
18 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
19 * GNU General Public License for more details.
21 * You should have received a copy of the GNU General Public License
22 * along with this program; if not, write to the Free Software
23 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111, USA.
25 * This program is also available under a commercial proprietary license.
26 * For more information, contact us at licensing@x264.com.
27 *****************************************************************************/
29 /* predict4x4 are inspired from ffmpeg h264 decoder */
35 # include "x86/predict.h"
38 # include "ppc/predict.h"
41 # include "arm/predict.h"
44 /****************************************************************************
45 * 16x16 prediction for intra luma block
46 ****************************************************************************/
48 #define PREDICT_16x16_DC(v)\
49 for( int i = 0; i < 16; i++ )\
51 MPIXEL_X4( src+ 0 ) = v;\
52 MPIXEL_X4( src+ 4 ) = v;\
53 MPIXEL_X4( src+ 8 ) = v;\
54 MPIXEL_X4( src+12 ) = v;\
58 void x264_predict_16x16_dc_c( pixel *src )
62 for( int i = 0; i < 16; i++ )
64 dc += src[-1 + i * FDEC_STRIDE];
65 dc += src[i - FDEC_STRIDE];
67 pixel4 dcsplat = PIXEL_SPLAT_X4( ( dc + 16 ) >> 5 );
69 PREDICT_16x16_DC( dcsplat );
71 static void x264_predict_16x16_dc_left_c( pixel *src )
75 for( int i = 0; i < 16; i++ )
76 dc += src[-1 + i * FDEC_STRIDE];
77 pixel4 dcsplat = PIXEL_SPLAT_X4( ( dc + 8 ) >> 4 );
79 PREDICT_16x16_DC( dcsplat );
81 static void x264_predict_16x16_dc_top_c( pixel *src )
85 for( int i = 0; i < 16; i++ )
86 dc += src[i - FDEC_STRIDE];
87 pixel4 dcsplat = PIXEL_SPLAT_X4( ( dc + 8 ) >> 4 );
89 PREDICT_16x16_DC( dcsplat );
91 static void x264_predict_16x16_dc_128_c( pixel *src )
93 PREDICT_16x16_DC( PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) ) );
95 void x264_predict_16x16_h_c( pixel *src )
97 for( int i = 0; i < 16; i++ )
99 const pixel4 v = PIXEL_SPLAT_X4( src[-1] );
100 MPIXEL_X4( src+ 0 ) = v;
101 MPIXEL_X4( src+ 4 ) = v;
102 MPIXEL_X4( src+ 8 ) = v;
103 MPIXEL_X4( src+12 ) = v;
107 void x264_predict_16x16_v_c( pixel *src )
109 pixel4 v0 = MPIXEL_X4( &src[ 0-FDEC_STRIDE] );
110 pixel4 v1 = MPIXEL_X4( &src[ 4-FDEC_STRIDE] );
111 pixel4 v2 = MPIXEL_X4( &src[ 8-FDEC_STRIDE] );
112 pixel4 v3 = MPIXEL_X4( &src[12-FDEC_STRIDE] );
114 for( int i = 0; i < 16; i++ )
116 MPIXEL_X4( src+ 0 ) = v0;
117 MPIXEL_X4( src+ 4 ) = v1;
118 MPIXEL_X4( src+ 8 ) = v2;
119 MPIXEL_X4( src+12 ) = v3;
123 void x264_predict_16x16_p_c( pixel *src )
127 /* calculate H and V */
128 for( int i = 0; i <= 7; i++ )
130 H += ( i + 1 ) * ( src[ 8 + i - FDEC_STRIDE ] - src[6 -i -FDEC_STRIDE] );
131 V += ( i + 1 ) * ( src[-1 + (8+i)*FDEC_STRIDE] - src[-1 + (6-i)*FDEC_STRIDE] );
134 int a = 16 * ( src[-1 + 15*FDEC_STRIDE] + src[15 - FDEC_STRIDE] );
135 int b = ( 5 * H + 32 ) >> 6;
136 int c = ( 5 * V + 32 ) >> 6;
138 int i00 = a - b * 7 - c * 7 + 16;
140 for( int y = 0; y < 16; y++ )
143 for( int x = 0; x < 16; x++ )
145 src[x] = x264_clip_pixel( pix>>5 );
154 /****************************************************************************
155 * 8x8 prediction for intra chroma block (4:2:0)
156 ****************************************************************************/
158 static void x264_predict_8x8c_dc_128_c( pixel *src )
160 for( int y = 0; y < 8; y++ )
162 MPIXEL_X4( src+0 ) = PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) );
163 MPIXEL_X4( src+4 ) = PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) );
167 static void x264_predict_8x8c_dc_left_c( pixel *src )
169 int dc0 = 0, dc1 = 0;
171 for( int y = 0; y < 4; y++ )
173 dc0 += src[y * FDEC_STRIDE - 1];
174 dc1 += src[(y+4) * FDEC_STRIDE - 1];
176 pixel4 dc0splat = PIXEL_SPLAT_X4( ( dc0 + 2 ) >> 2 );
177 pixel4 dc1splat = PIXEL_SPLAT_X4( ( dc1 + 2 ) >> 2 );
179 for( int y = 0; y < 4; y++ )
181 MPIXEL_X4( src+0 ) = dc0splat;
182 MPIXEL_X4( src+4 ) = dc0splat;
185 for( int y = 0; y < 4; y++ )
187 MPIXEL_X4( src+0 ) = dc1splat;
188 MPIXEL_X4( src+4 ) = dc1splat;
193 static void x264_predict_8x8c_dc_top_c( pixel *src )
195 int dc0 = 0, dc1 = 0;
197 for( int x = 0; x < 4; x++ )
199 dc0 += src[x - FDEC_STRIDE];
200 dc1 += src[x + 4 - FDEC_STRIDE];
202 pixel4 dc0splat = PIXEL_SPLAT_X4( ( dc0 + 2 ) >> 2 );
203 pixel4 dc1splat = PIXEL_SPLAT_X4( ( dc1 + 2 ) >> 2 );
205 for( int y = 0; y < 8; y++ )
207 MPIXEL_X4( src+0 ) = dc0splat;
208 MPIXEL_X4( src+4 ) = dc1splat;
212 void x264_predict_8x8c_dc_c( pixel *src )
214 int s0 = 0, s1 = 0, s2 = 0, s3 = 0;
221 for( int i = 0; i < 4; i++ )
223 s0 += src[i - FDEC_STRIDE];
224 s1 += src[i + 4 - FDEC_STRIDE];
225 s2 += src[-1 + i * FDEC_STRIDE];
226 s3 += src[-1 + (i+4)*FDEC_STRIDE];
232 pixel4 dc0 = PIXEL_SPLAT_X4( ( s0 + s2 + 4 ) >> 3 );
233 pixel4 dc1 = PIXEL_SPLAT_X4( ( s1 + 2 ) >> 2 );
234 pixel4 dc2 = PIXEL_SPLAT_X4( ( s3 + 2 ) >> 2 );
235 pixel4 dc3 = PIXEL_SPLAT_X4( ( s1 + s3 + 4 ) >> 3 );
237 for( int y = 0; y < 4; y++ )
239 MPIXEL_X4( src+0 ) = dc0;
240 MPIXEL_X4( src+4 ) = dc1;
244 for( int y = 0; y < 4; y++ )
246 MPIXEL_X4( src+0 ) = dc2;
247 MPIXEL_X4( src+4 ) = dc3;
251 void x264_predict_8x8c_h_c( pixel *src )
253 for( int i = 0; i < 8; i++ )
255 pixel4 v = PIXEL_SPLAT_X4( src[-1] );
256 MPIXEL_X4( src+0 ) = v;
257 MPIXEL_X4( src+4 ) = v;
261 void x264_predict_8x8c_v_c( pixel *src )
263 pixel4 v0 = MPIXEL_X4( src+0-FDEC_STRIDE );
264 pixel4 v1 = MPIXEL_X4( src+4-FDEC_STRIDE );
266 for( int i = 0; i < 8; i++ )
268 MPIXEL_X4( src+0 ) = v0;
269 MPIXEL_X4( src+4 ) = v1;
273 void x264_predict_8x8c_p_c( pixel *src )
277 for( int i = 0; i < 4; i++ )
279 H += ( i + 1 ) * ( src[4+i - FDEC_STRIDE] - src[2 - i -FDEC_STRIDE] );
280 V += ( i + 1 ) * ( src[-1 +(i+4)*FDEC_STRIDE] - src[-1+(2-i)*FDEC_STRIDE] );
283 int a = 16 * ( src[-1+7*FDEC_STRIDE] + src[7 - FDEC_STRIDE] );
284 int b = ( 17 * H + 16 ) >> 5;
285 int c = ( 17 * V + 16 ) >> 5;
286 int i00 = a -3*b -3*c + 16;
288 for( int y = 0; y < 8; y++ )
291 for( int x = 0; x < 8; x++ )
293 src[x] = x264_clip_pixel( pix>>5 );
301 /****************************************************************************
302 * 8x16 prediction for intra chroma block (4:2:2)
303 ****************************************************************************/
305 static void x264_predict_8x16c_dc_128_c( pixel *src )
307 for( int y = 0; y < 16; y++ )
309 MPIXEL_X4( src+0 ) = PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) );
310 MPIXEL_X4( src+4 ) = PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) );
314 static void x264_predict_8x16c_dc_left_c( pixel *src )
316 for( int i = 0; i < 4; i++ )
320 for( int y = 0; y < 4; y++ )
321 dc += src[y*FDEC_STRIDE - 1];
323 pixel4 dcsplat = PIXEL_SPLAT_X4( (dc + 2) >> 2 );
325 for( int y = 0; y < 4; y++ )
327 MPIXEL_X4( src+0 ) = dcsplat;
328 MPIXEL_X4( src+4 ) = dcsplat;
333 static void x264_predict_8x16c_dc_top_c( pixel *src )
335 int dc0 = 0, dc1 = 0;
337 for(int x = 0; x < 4; x++ )
339 dc0 += src[x - FDEC_STRIDE];
340 dc1 += src[x + 4 - FDEC_STRIDE];
342 pixel4 dc0splat = PIXEL_SPLAT_X4( ( dc0 + 2 ) >> 2 );
343 pixel4 dc1splat = PIXEL_SPLAT_X4( ( dc1 + 2 ) >> 2 );
345 for( int y = 0; y < 16; y++ )
347 MPIXEL_X4( src+0 ) = dc0splat;
348 MPIXEL_X4( src+4 ) = dc1splat;
352 void x264_predict_8x16c_dc_c( pixel *src )
354 int s0 = 0, s1 = 0, s2 = 0, s3 = 0, s4 = 0, s5 = 0;
363 for( int i = 0; i < 4; i++ )
365 s0 += src[i+0 - FDEC_STRIDE];
366 s1 += src[i+4 - FDEC_STRIDE];
367 s2 += src[-1 + (i+0) * FDEC_STRIDE];
368 s3 += src[-1 + (i+4) * FDEC_STRIDE];
369 s4 += src[-1 + (i+8) * FDEC_STRIDE];
370 s5 += src[-1 + (i+12) * FDEC_STRIDE];
378 pixel4 dc0 = PIXEL_SPLAT_X4( ( s0 + s2 + 4 ) >> 3 );
379 pixel4 dc1 = PIXEL_SPLAT_X4( ( s1 + 2 ) >> 2 );
380 pixel4 dc2 = PIXEL_SPLAT_X4( ( s3 + 2 ) >> 2 );
381 pixel4 dc3 = PIXEL_SPLAT_X4( ( s1 + s3 + 4 ) >> 3 );
382 pixel4 dc4 = PIXEL_SPLAT_X4( ( s4 + 2 ) >> 2 );
383 pixel4 dc5 = PIXEL_SPLAT_X4( ( s1 + s4 + 4 ) >> 3 );
384 pixel4 dc6 = PIXEL_SPLAT_X4( ( s5 + 2 ) >> 2 );
385 pixel4 dc7 = PIXEL_SPLAT_X4( ( s1 + s5 + 4 ) >> 3 );
387 for( int y = 0; y < 4; y++ )
389 MPIXEL_X4( src+0 ) = dc0;
390 MPIXEL_X4( src+4 ) = dc1;
393 for( int y = 0; y < 4; y++ )
395 MPIXEL_X4( src+0 ) = dc2;
396 MPIXEL_X4( src+4 ) = dc3;
399 for( int y = 0; y < 4; y++ )
401 MPIXEL_X4( src+0 ) = dc4;
402 MPIXEL_X4( src+4 ) = dc5;
405 for( int y = 0; y < 4; y++ )
407 MPIXEL_X4( src+0 ) = dc6;
408 MPIXEL_X4( src+4 ) = dc7;
412 void x264_predict_8x16c_h_c( pixel *src )
414 for( int i = 0; i < 16; i++ )
416 pixel4 v = PIXEL_SPLAT_X4( src[-1] );
417 MPIXEL_X4( src+0 ) = v;
418 MPIXEL_X4( src+4 ) = v;
422 void x264_predict_8x16c_v_c( pixel *src )
424 pixel4 v0 = MPIXEL_X4( src+0-FDEC_STRIDE );
425 pixel4 v1 = MPIXEL_X4( src+4-FDEC_STRIDE );
427 for( int i = 0; i < 16; i++ )
429 MPIXEL_X4( src+0 ) = v0;
430 MPIXEL_X4( src+4 ) = v1;
434 void x264_predict_8x16c_p_c( pixel *src )
439 for( int i = 0; i < 4; i++ )
440 H += ( i + 1 ) * ( src[4 + i - FDEC_STRIDE] - src[2 - i - FDEC_STRIDE] );
441 for( int i = 0; i < 8; i++ )
442 V += ( i + 1 ) * ( src[-1 + (i+8)*FDEC_STRIDE] - src[-1 + (6-i)*FDEC_STRIDE] );
444 int a = 16 * ( src[-1 + 15*FDEC_STRIDE] + src[7 - FDEC_STRIDE] );
445 int b = ( 17 * H + 16 ) >> 5;
446 int c = ( 5 * V + 32 ) >> 6;
447 int i00 = a -3*b -7*c + 16;
449 for( int y = 0; y < 16; y++ )
452 for( int x = 0; x < 8; x++ )
454 src[x] = x264_clip_pixel( pix>>5 );
462 /****************************************************************************
463 * 4x4 prediction for intra luma block
464 ****************************************************************************/
466 #define SRC(x,y) src[(x)+(y)*FDEC_STRIDE]
467 #define SRC_X4(x,y) MPIXEL_X4( &SRC(x,y) )
469 #define PREDICT_4x4_DC(v)\
470 SRC_X4(0,0) = SRC_X4(0,1) = SRC_X4(0,2) = SRC_X4(0,3) = v;
472 static void x264_predict_4x4_dc_128_c( pixel *src )
474 PREDICT_4x4_DC( PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) ) );
476 static void x264_predict_4x4_dc_left_c( pixel *src )
478 pixel4 dc = PIXEL_SPLAT_X4( (SRC(-1,0) + SRC(-1,1) + SRC(-1,2) + SRC(-1,3) + 2) >> 2 );
479 PREDICT_4x4_DC( dc );
481 static void x264_predict_4x4_dc_top_c( pixel *src )
483 pixel4 dc = PIXEL_SPLAT_X4( (SRC(0,-1) + SRC(1,-1) + SRC(2,-1) + SRC(3,-1) + 2) >> 2 );
484 PREDICT_4x4_DC( dc );
486 void x264_predict_4x4_dc_c( pixel *src )
488 pixel4 dc = PIXEL_SPLAT_X4( (SRC(-1,0) + SRC(-1,1) + SRC(-1,2) + SRC(-1,3) +
489 SRC(0,-1) + SRC(1,-1) + SRC(2,-1) + SRC(3,-1) + 4) >> 3 );
490 PREDICT_4x4_DC( dc );
492 void x264_predict_4x4_h_c( pixel *src )
494 SRC_X4(0,0) = PIXEL_SPLAT_X4( SRC(-1,0) );
495 SRC_X4(0,1) = PIXEL_SPLAT_X4( SRC(-1,1) );
496 SRC_X4(0,2) = PIXEL_SPLAT_X4( SRC(-1,2) );
497 SRC_X4(0,3) = PIXEL_SPLAT_X4( SRC(-1,3) );
499 void x264_predict_4x4_v_c( pixel *src )
501 PREDICT_4x4_DC(SRC_X4(0,-1));
504 #define PREDICT_4x4_LOAD_LEFT\
508 UNUSED int l3 = SRC(-1,3);
510 #define PREDICT_4x4_LOAD_TOP\
514 UNUSED int t3 = SRC(3,-1);
516 #define PREDICT_4x4_LOAD_TOP_RIGHT\
520 UNUSED int t7 = SRC(7,-1);
522 #define F1(a,b) (((a)+(b)+1)>>1)
523 #define F2(a,b,c) (((a)+2*(b)+(c)+2)>>2)
525 static void x264_predict_4x4_ddl_c( pixel *src )
528 PREDICT_4x4_LOAD_TOP_RIGHT
529 SRC(0,0)= F2(t0,t1,t2);
530 SRC(1,0)=SRC(0,1)= F2(t1,t2,t3);
531 SRC(2,0)=SRC(1,1)=SRC(0,2)= F2(t2,t3,t4);
532 SRC(3,0)=SRC(2,1)=SRC(1,2)=SRC(0,3)= F2(t3,t4,t5);
533 SRC(3,1)=SRC(2,2)=SRC(1,3)= F2(t4,t5,t6);
534 SRC(3,2)=SRC(2,3)= F2(t5,t6,t7);
535 SRC(3,3)= F2(t6,t7,t7);
537 static void x264_predict_4x4_ddr_c( pixel *src )
540 PREDICT_4x4_LOAD_LEFT
542 SRC(3,0)= F2(t3,t2,t1);
543 SRC(2,0)=SRC(3,1)= F2(t2,t1,t0);
544 SRC(1,0)=SRC(2,1)=SRC(3,2)= F2(t1,t0,lt);
545 SRC(0,0)=SRC(1,1)=SRC(2,2)=SRC(3,3)= F2(t0,lt,l0);
546 SRC(0,1)=SRC(1,2)=SRC(2,3)= F2(lt,l0,l1);
547 SRC(0,2)=SRC(1,3)= F2(l0,l1,l2);
548 SRC(0,3)= F2(l1,l2,l3);
551 static void x264_predict_4x4_vr_c( pixel *src )
554 PREDICT_4x4_LOAD_LEFT
556 SRC(0,3)= F2(l2,l1,l0);
557 SRC(0,2)= F2(l1,l0,lt);
558 SRC(0,1)=SRC(1,3)= F2(l0,lt,t0);
559 SRC(0,0)=SRC(1,2)= F1(lt,t0);
560 SRC(1,1)=SRC(2,3)= F2(lt,t0,t1);
561 SRC(1,0)=SRC(2,2)= F1(t0,t1);
562 SRC(2,1)=SRC(3,3)= F2(t0,t1,t2);
563 SRC(2,0)=SRC(3,2)= F1(t1,t2);
564 SRC(3,1)= F2(t1,t2,t3);
568 static void x264_predict_4x4_hd_c( pixel *src )
571 PREDICT_4x4_LOAD_LEFT
574 SRC(1,3)= F2(l1,l2,l3);
575 SRC(0,2)=SRC(2,3)= F1(l1,l2);
576 SRC(1,2)=SRC(3,3)= F2(l0,l1,l2);
577 SRC(0,1)=SRC(2,2)= F1(l0,l1);
578 SRC(1,1)=SRC(3,2)= F2(lt,l0,l1);
579 SRC(0,0)=SRC(2,1)= F1(lt,l0);
580 SRC(1,0)=SRC(3,1)= F2(t0,lt,l0);
581 SRC(2,0)= F2(t1,t0,lt);
582 SRC(3,0)= F2(t2,t1,t0);
585 static void x264_predict_4x4_vl_c( pixel *src )
588 PREDICT_4x4_LOAD_TOP_RIGHT
590 SRC(0,1)= F2(t0,t1,t2);
591 SRC(1,0)=SRC(0,2)= F1(t1,t2);
592 SRC(1,1)=SRC(0,3)= F2(t1,t2,t3);
593 SRC(2,0)=SRC(1,2)= F1(t2,t3);
594 SRC(2,1)=SRC(1,3)= F2(t2,t3,t4);
595 SRC(3,0)=SRC(2,2)= F1(t3,t4);
596 SRC(3,1)=SRC(2,3)= F2(t3,t4,t5);
598 SRC(3,3)= F2(t4,t5,t6);
601 static void x264_predict_4x4_hu_c( pixel *src )
603 PREDICT_4x4_LOAD_LEFT
605 SRC(1,0)= F2(l0,l1,l2);
606 SRC(2,0)=SRC(0,1)= F1(l1,l2);
607 SRC(3,0)=SRC(1,1)= F2(l1,l2,l3);
608 SRC(2,1)=SRC(0,2)= F1(l2,l3);
609 SRC(3,1)=SRC(1,2)= F2(l2,l3,l3);
610 SRC(3,2)=SRC(1,3)=SRC(0,3)=
611 SRC(2,2)=SRC(2,3)=SRC(3,3)= l3;
614 /****************************************************************************
615 * 8x8 prediction for intra luma block
616 ****************************************************************************/
619 edge[14-y] = F2(SRC(-1,y-1), SRC(-1,y), SRC(-1,y+1));
621 edge[16+x] = F2(SRC(x-1,-1), SRC(x,-1), SRC(x+1,-1));
623 static void x264_predict_8x8_filter_c( pixel *src, pixel edge[36], int i_neighbor, int i_filters )
625 /* edge[7..14] = l7..l0
627 * edge[16..31] = t0 .. t15
630 int have_lt = i_neighbor & MB_TOPLEFT;
631 if( i_filters & MB_LEFT )
633 edge[15] = (SRC(0,-1) + 2*SRC(-1,-1) + SRC(-1,0) + 2) >> 2;
634 edge[14] = ((have_lt ? SRC(-1,-1) : SRC(-1,0))
635 + 2*SRC(-1,0) + SRC(-1,1) + 2) >> 2;
636 PL(1) PL(2) PL(3) PL(4) PL(5) PL(6)
638 edge[7] = (SRC(-1,6) + 3*SRC(-1,7) + 2) >> 2;
641 if( i_filters & MB_TOP )
643 int have_tr = i_neighbor & MB_TOPRIGHT;
644 edge[16] = ((have_lt ? SRC(-1,-1) : SRC(0,-1))
645 + 2*SRC(0,-1) + SRC(1,-1) + 2) >> 2;
646 PT(1) PT(2) PT(3) PT(4) PT(5) PT(6)
647 edge[23] = (SRC(6,-1) + 2*SRC(7,-1)
648 + (have_tr ? SRC(8,-1) : SRC(7,-1)) + 2) >> 2;
650 if( i_filters & MB_TOPRIGHT )
654 PT(8) PT(9) PT(10) PT(11) PT(12) PT(13) PT(14)
656 edge[32] = (SRC(14,-1) + 3*SRC(15,-1) + 2) >> 2;
660 MPIXEL_X4( edge+24 ) = PIXEL_SPLAT_X4( SRC(7,-1) );
661 MPIXEL_X4( edge+28 ) = PIXEL_SPLAT_X4( SRC(7,-1) );
662 edge[32] = SRC(7,-1);
672 UNUSED int l##y = edge[14-y];
674 UNUSED int t##x = edge[16+x];
675 #define PREDICT_8x8_LOAD_TOPLEFT \
677 #define PREDICT_8x8_LOAD_LEFT \
678 PL(0) PL(1) PL(2) PL(3) PL(4) PL(5) PL(6) PL(7)
679 #define PREDICT_8x8_LOAD_TOP \
680 PT(0) PT(1) PT(2) PT(3) PT(4) PT(5) PT(6) PT(7)
681 #define PREDICT_8x8_LOAD_TOPRIGHT \
682 PT(8) PT(9) PT(10) PT(11) PT(12) PT(13) PT(14) PT(15)
684 #define PREDICT_8x8_DC(v) \
685 for( int y = 0; y < 8; y++ ) { \
686 MPIXEL_X4( src+0 ) = v; \
687 MPIXEL_X4( src+4 ) = v; \
688 src += FDEC_STRIDE; \
691 static void x264_predict_8x8_dc_128_c( pixel *src, pixel edge[36] )
693 PREDICT_8x8_DC( PIXEL_SPLAT_X4( 1 << (BIT_DEPTH-1) ) );
695 static void x264_predict_8x8_dc_left_c( pixel *src, pixel edge[36] )
697 PREDICT_8x8_LOAD_LEFT
698 pixel4 dc = PIXEL_SPLAT_X4( (l0+l1+l2+l3+l4+l5+l6+l7+4) >> 3 );
699 PREDICT_8x8_DC( dc );
701 static void x264_predict_8x8_dc_top_c( pixel *src, pixel edge[36] )
704 pixel4 dc = PIXEL_SPLAT_X4( (t0+t1+t2+t3+t4+t5+t6+t7+4) >> 3 );
705 PREDICT_8x8_DC( dc );
707 void x264_predict_8x8_dc_c( pixel *src, pixel edge[36] )
709 PREDICT_8x8_LOAD_LEFT
711 pixel4 dc = PIXEL_SPLAT_X4( (l0+l1+l2+l3+l4+l5+l6+l7+t0+t1+t2+t3+t4+t5+t6+t7+8) >> 4 );
712 PREDICT_8x8_DC( dc );
714 void x264_predict_8x8_h_c( pixel *src, pixel edge[36] )
716 PREDICT_8x8_LOAD_LEFT
717 #define ROW(y) MPIXEL_X4( src+y*FDEC_STRIDE+0 ) =\
718 MPIXEL_X4( src+y*FDEC_STRIDE+4 ) = PIXEL_SPLAT_X4( l##y );
719 ROW(0); ROW(1); ROW(2); ROW(3); ROW(4); ROW(5); ROW(6); ROW(7);
722 void x264_predict_8x8_v_c( pixel *src, pixel edge[36] )
724 pixel4 top[2] = { MPIXEL_X4( edge+16 ),
725 MPIXEL_X4( edge+20 ) };
726 for( int y = 0; y < 8; y++ )
728 MPIXEL_X4( src+y*FDEC_STRIDE+0 ) = top[0];
729 MPIXEL_X4( src+y*FDEC_STRIDE+4 ) = top[1];
732 static void x264_predict_8x8_ddl_c( pixel *src, pixel edge[36] )
735 PREDICT_8x8_LOAD_TOPRIGHT
736 SRC(0,0)= F2(t0,t1,t2);
737 SRC(0,1)=SRC(1,0)= F2(t1,t2,t3);
738 SRC(0,2)=SRC(1,1)=SRC(2,0)= F2(t2,t3,t4);
739 SRC(0,3)=SRC(1,2)=SRC(2,1)=SRC(3,0)= F2(t3,t4,t5);
740 SRC(0,4)=SRC(1,3)=SRC(2,2)=SRC(3,1)=SRC(4,0)= F2(t4,t5,t6);
741 SRC(0,5)=SRC(1,4)=SRC(2,3)=SRC(3,2)=SRC(4,1)=SRC(5,0)= F2(t5,t6,t7);
742 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);
743 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);
744 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);
745 SRC(2,7)=SRC(3,6)=SRC(4,5)=SRC(5,4)=SRC(6,3)=SRC(7,2)= F2(t9,t10,t11);
746 SRC(3,7)=SRC(4,6)=SRC(5,5)=SRC(6,4)=SRC(7,3)= F2(t10,t11,t12);
747 SRC(4,7)=SRC(5,6)=SRC(6,5)=SRC(7,4)= F2(t11,t12,t13);
748 SRC(5,7)=SRC(6,6)=SRC(7,5)= F2(t12,t13,t14);
749 SRC(6,7)=SRC(7,6)= F2(t13,t14,t15);
750 SRC(7,7)= F2(t14,t15,t15);
752 static void x264_predict_8x8_ddr_c( pixel *src, pixel edge[36] )
755 PREDICT_8x8_LOAD_LEFT
756 PREDICT_8x8_LOAD_TOPLEFT
757 SRC(0,7)= F2(l7,l6,l5);
758 SRC(0,6)=SRC(1,7)= F2(l6,l5,l4);
759 SRC(0,5)=SRC(1,6)=SRC(2,7)= F2(l5,l4,l3);
760 SRC(0,4)=SRC(1,5)=SRC(2,6)=SRC(3,7)= F2(l4,l3,l2);
761 SRC(0,3)=SRC(1,4)=SRC(2,5)=SRC(3,6)=SRC(4,7)= F2(l3,l2,l1);
762 SRC(0,2)=SRC(1,3)=SRC(2,4)=SRC(3,5)=SRC(4,6)=SRC(5,7)= F2(l2,l1,l0);
763 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);
764 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);
765 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);
766 SRC(2,0)=SRC(3,1)=SRC(4,2)=SRC(5,3)=SRC(6,4)=SRC(7,5)= F2(t0,t1,t2);
767 SRC(3,0)=SRC(4,1)=SRC(5,2)=SRC(6,3)=SRC(7,4)= F2(t1,t2,t3);
768 SRC(4,0)=SRC(5,1)=SRC(6,2)=SRC(7,3)= F2(t2,t3,t4);
769 SRC(5,0)=SRC(6,1)=SRC(7,2)= F2(t3,t4,t5);
770 SRC(6,0)=SRC(7,1)= F2(t4,t5,t6);
771 SRC(7,0)= F2(t5,t6,t7);
774 static void x264_predict_8x8_vr_c( pixel *src, pixel edge[36] )
777 PREDICT_8x8_LOAD_LEFT
778 PREDICT_8x8_LOAD_TOPLEFT
779 SRC(0,6)= F2(l5,l4,l3);
780 SRC(0,7)= F2(l6,l5,l4);
781 SRC(0,4)=SRC(1,6)= F2(l3,l2,l1);
782 SRC(0,5)=SRC(1,7)= F2(l4,l3,l2);
783 SRC(0,2)=SRC(1,4)=SRC(2,6)= F2(l1,l0,lt);
784 SRC(0,3)=SRC(1,5)=SRC(2,7)= F2(l2,l1,l0);
785 SRC(0,1)=SRC(1,3)=SRC(2,5)=SRC(3,7)= F2(l0,lt,t0);
786 SRC(0,0)=SRC(1,2)=SRC(2,4)=SRC(3,6)= F1(lt,t0);
787 SRC(1,1)=SRC(2,3)=SRC(3,5)=SRC(4,7)= F2(lt,t0,t1);
788 SRC(1,0)=SRC(2,2)=SRC(3,4)=SRC(4,6)= F1(t0,t1);
789 SRC(2,1)=SRC(3,3)=SRC(4,5)=SRC(5,7)= F2(t0,t1,t2);
790 SRC(2,0)=SRC(3,2)=SRC(4,4)=SRC(5,6)= F1(t1,t2);
791 SRC(3,1)=SRC(4,3)=SRC(5,5)=SRC(6,7)= F2(t1,t2,t3);
792 SRC(3,0)=SRC(4,2)=SRC(5,4)=SRC(6,6)= F1(t2,t3);
793 SRC(4,1)=SRC(5,3)=SRC(6,5)=SRC(7,7)= F2(t2,t3,t4);
794 SRC(4,0)=SRC(5,2)=SRC(6,4)=SRC(7,6)= F1(t3,t4);
795 SRC(5,1)=SRC(6,3)=SRC(7,5)= F2(t3,t4,t5);
796 SRC(5,0)=SRC(6,2)=SRC(7,4)= F1(t4,t5);
797 SRC(6,1)=SRC(7,3)= F2(t4,t5,t6);
798 SRC(6,0)=SRC(7,2)= F1(t5,t6);
799 SRC(7,1)= F2(t5,t6,t7);
802 static void x264_predict_8x8_hd_c( pixel *src, pixel edge[36] )
805 PREDICT_8x8_LOAD_LEFT
806 PREDICT_8x8_LOAD_TOPLEFT
807 int p1 = pack_pixel_1to2(F1(l6,l7), F2(l5,l6,l7));
808 int p2 = pack_pixel_1to2(F1(l5,l6), F2(l4,l5,l6));
809 int p3 = pack_pixel_1to2(F1(l4,l5), F2(l3,l4,l5));
810 int p4 = pack_pixel_1to2(F1(l3,l4), F2(l2,l3,l4));
811 int p5 = pack_pixel_1to2(F1(l2,l3), F2(l1,l2,l3));
812 int p6 = pack_pixel_1to2(F1(l1,l2), F2(l0,l1,l2));
813 int p7 = pack_pixel_1to2(F1(l0,l1), F2(lt,l0,l1));
814 int p8 = pack_pixel_1to2(F1(lt,l0), F2(l0,lt,t0));
815 int p9 = pack_pixel_1to2(F2(t1,t0,lt), F2(t2,t1,t0));
816 int p10 = pack_pixel_1to2(F2(t3,t2,t1), F2(t4,t3,t2));
817 int p11 = pack_pixel_1to2(F2(t5,t4,t3), F2(t6,t5,t4));
818 SRC_X4(0,7)= pack_pixel_2to4(p1,p2);
819 SRC_X4(0,6)= pack_pixel_2to4(p2,p3);
820 SRC_X4(4,7)=SRC_X4(0,5)= pack_pixel_2to4(p3,p4);
821 SRC_X4(4,6)=SRC_X4(0,4)= pack_pixel_2to4(p4,p5);
822 SRC_X4(4,5)=SRC_X4(0,3)= pack_pixel_2to4(p5,p6);
823 SRC_X4(4,4)=SRC_X4(0,2)= pack_pixel_2to4(p6,p7);
824 SRC_X4(4,3)=SRC_X4(0,1)= pack_pixel_2to4(p7,p8);
825 SRC_X4(4,2)=SRC_X4(0,0)= pack_pixel_2to4(p8,p9);
826 SRC_X4(4,1)= pack_pixel_2to4(p9,p10);
827 SRC_X4(4,0)= pack_pixel_2to4(p10,p11);
829 static void x264_predict_8x8_vl_c( pixel *src, pixel edge[36] )
832 PREDICT_8x8_LOAD_TOPRIGHT
834 SRC(0,1)= F2(t0,t1,t2);
835 SRC(0,2)=SRC(1,0)= F1(t1,t2);
836 SRC(0,3)=SRC(1,1)= F2(t1,t2,t3);
837 SRC(0,4)=SRC(1,2)=SRC(2,0)= F1(t2,t3);
838 SRC(0,5)=SRC(1,3)=SRC(2,1)= F2(t2,t3,t4);
839 SRC(0,6)=SRC(1,4)=SRC(2,2)=SRC(3,0)= F1(t3,t4);
840 SRC(0,7)=SRC(1,5)=SRC(2,3)=SRC(3,1)= F2(t3,t4,t5);
841 SRC(1,6)=SRC(2,4)=SRC(3,2)=SRC(4,0)= F1(t4,t5);
842 SRC(1,7)=SRC(2,5)=SRC(3,3)=SRC(4,1)= F2(t4,t5,t6);
843 SRC(2,6)=SRC(3,4)=SRC(4,2)=SRC(5,0)= F1(t5,t6);
844 SRC(2,7)=SRC(3,5)=SRC(4,3)=SRC(5,1)= F2(t5,t6,t7);
845 SRC(3,6)=SRC(4,4)=SRC(5,2)=SRC(6,0)= F1(t6,t7);
846 SRC(3,7)=SRC(4,5)=SRC(5,3)=SRC(6,1)= F2(t6,t7,t8);
847 SRC(4,6)=SRC(5,4)=SRC(6,2)=SRC(7,0)= F1(t7,t8);
848 SRC(4,7)=SRC(5,5)=SRC(6,3)=SRC(7,1)= F2(t7,t8,t9);
849 SRC(5,6)=SRC(6,4)=SRC(7,2)= F1(t8,t9);
850 SRC(5,7)=SRC(6,5)=SRC(7,3)= F2(t8,t9,t10);
851 SRC(6,6)=SRC(7,4)= F1(t9,t10);
852 SRC(6,7)=SRC(7,5)= F2(t9,t10,t11);
853 SRC(7,6)= F1(t10,t11);
854 SRC(7,7)= F2(t10,t11,t12);
856 static void x264_predict_8x8_hu_c( pixel *src, pixel edge[36] )
858 PREDICT_8x8_LOAD_LEFT
859 int p1 = pack_pixel_1to2(F1(l0,l1), F2(l0,l1,l2));
860 int p2 = pack_pixel_1to2(F1(l1,l2), F2(l1,l2,l3));
861 int p3 = pack_pixel_1to2(F1(l2,l3), F2(l2,l3,l4));
862 int p4 = pack_pixel_1to2(F1(l3,l4), F2(l3,l4,l5));
863 int p5 = pack_pixel_1to2(F1(l4,l5), F2(l4,l5,l6));
864 int p6 = pack_pixel_1to2(F1(l5,l6), F2(l5,l6,l7));
865 int p7 = pack_pixel_1to2(F1(l6,l7), F2(l6,l7,l7));
866 int p8 = pack_pixel_1to2(l7,l7);
867 SRC_X4(0,0)= pack_pixel_2to4(p1,p2);
868 SRC_X4(0,1)= pack_pixel_2to4(p2,p3);
869 SRC_X4(4,0)=SRC_X4(0,2)= pack_pixel_2to4(p3,p4);
870 SRC_X4(4,1)=SRC_X4(0,3)= pack_pixel_2to4(p4,p5);
871 SRC_X4(4,2)=SRC_X4(0,4)= pack_pixel_2to4(p5,p6);
872 SRC_X4(4,3)=SRC_X4(0,5)= pack_pixel_2to4(p6,p7);
873 SRC_X4(4,4)=SRC_X4(0,6)= pack_pixel_2to4(p7,p8);
874 SRC_X4(4,5)=SRC_X4(4,6)= SRC_X4(0,7) = SRC_X4(4,7) = pack_pixel_2to4(p8,p8);
877 /****************************************************************************
878 * Exported functions:
879 ****************************************************************************/
880 void x264_predict_16x16_init( int cpu, x264_predict_t pf[7] )
882 pf[I_PRED_16x16_V ] = x264_predict_16x16_v_c;
883 pf[I_PRED_16x16_H ] = x264_predict_16x16_h_c;
884 pf[I_PRED_16x16_DC] = x264_predict_16x16_dc_c;
885 pf[I_PRED_16x16_P ] = x264_predict_16x16_p_c;
886 pf[I_PRED_16x16_DC_LEFT]= x264_predict_16x16_dc_left_c;
887 pf[I_PRED_16x16_DC_TOP ]= x264_predict_16x16_dc_top_c;
888 pf[I_PRED_16x16_DC_128 ]= x264_predict_16x16_dc_128_c;
891 x264_predict_16x16_init_mmx( cpu, pf );
895 if( cpu&X264_CPU_ALTIVEC )
896 x264_predict_16x16_init_altivec( pf );
900 x264_predict_16x16_init_arm( cpu, pf );
904 void x264_predict_8x8c_init( int cpu, x264_predict_t pf[7] )
906 pf[I_PRED_CHROMA_V ] = x264_predict_8x8c_v_c;
907 pf[I_PRED_CHROMA_H ] = x264_predict_8x8c_h_c;
908 pf[I_PRED_CHROMA_DC] = x264_predict_8x8c_dc_c;
909 pf[I_PRED_CHROMA_P ] = x264_predict_8x8c_p_c;
910 pf[I_PRED_CHROMA_DC_LEFT]= x264_predict_8x8c_dc_left_c;
911 pf[I_PRED_CHROMA_DC_TOP ]= x264_predict_8x8c_dc_top_c;
912 pf[I_PRED_CHROMA_DC_128 ]= x264_predict_8x8c_dc_128_c;
915 x264_predict_8x8c_init_mmx( cpu, pf );
919 if( cpu&X264_CPU_ALTIVEC )
920 x264_predict_8x8c_init_altivec( pf );
924 x264_predict_8x8c_init_arm( cpu, pf );
928 void x264_predict_8x16c_init( int cpu, x264_predict_t pf[7] )
930 pf[I_PRED_CHROMA_V ] = x264_predict_8x16c_v_c;
931 pf[I_PRED_CHROMA_H ] = x264_predict_8x16c_h_c;
932 pf[I_PRED_CHROMA_DC] = x264_predict_8x16c_dc_c;
933 pf[I_PRED_CHROMA_P ] = x264_predict_8x16c_p_c;
934 pf[I_PRED_CHROMA_DC_LEFT]= x264_predict_8x16c_dc_left_c;
935 pf[I_PRED_CHROMA_DC_TOP ]= x264_predict_8x16c_dc_top_c;
936 pf[I_PRED_CHROMA_DC_128 ]= x264_predict_8x16c_dc_128_c;
939 x264_predict_8x16c_init_mmx( cpu, pf );
943 void x264_predict_8x8_init( int cpu, x264_predict8x8_t pf[12], x264_predict_8x8_filter_t *predict_filter )
945 pf[I_PRED_8x8_V] = x264_predict_8x8_v_c;
946 pf[I_PRED_8x8_H] = x264_predict_8x8_h_c;
947 pf[I_PRED_8x8_DC] = x264_predict_8x8_dc_c;
948 pf[I_PRED_8x8_DDL] = x264_predict_8x8_ddl_c;
949 pf[I_PRED_8x8_DDR] = x264_predict_8x8_ddr_c;
950 pf[I_PRED_8x8_VR] = x264_predict_8x8_vr_c;
951 pf[I_PRED_8x8_HD] = x264_predict_8x8_hd_c;
952 pf[I_PRED_8x8_VL] = x264_predict_8x8_vl_c;
953 pf[I_PRED_8x8_HU] = x264_predict_8x8_hu_c;
954 pf[I_PRED_8x8_DC_LEFT]= x264_predict_8x8_dc_left_c;
955 pf[I_PRED_8x8_DC_TOP] = x264_predict_8x8_dc_top_c;
956 pf[I_PRED_8x8_DC_128] = x264_predict_8x8_dc_128_c;
957 *predict_filter = x264_predict_8x8_filter_c;
960 x264_predict_8x8_init_mmx( cpu, pf, predict_filter );
964 x264_predict_8x8_init_arm( cpu, pf, predict_filter );
968 void x264_predict_4x4_init( int cpu, x264_predict_t pf[12] )
970 pf[I_PRED_4x4_V] = x264_predict_4x4_v_c;
971 pf[I_PRED_4x4_H] = x264_predict_4x4_h_c;
972 pf[I_PRED_4x4_DC] = x264_predict_4x4_dc_c;
973 pf[I_PRED_4x4_DDL] = x264_predict_4x4_ddl_c;
974 pf[I_PRED_4x4_DDR] = x264_predict_4x4_ddr_c;
975 pf[I_PRED_4x4_VR] = x264_predict_4x4_vr_c;
976 pf[I_PRED_4x4_HD] = x264_predict_4x4_hd_c;
977 pf[I_PRED_4x4_VL] = x264_predict_4x4_vl_c;
978 pf[I_PRED_4x4_HU] = x264_predict_4x4_hu_c;
979 pf[I_PRED_4x4_DC_LEFT]= x264_predict_4x4_dc_left_c;
980 pf[I_PRED_4x4_DC_TOP] = x264_predict_4x4_dc_top_c;
981 pf[I_PRED_4x4_DC_128] = x264_predict_4x4_dc_128_c;
984 x264_predict_4x4_init_mmx( cpu, pf );
988 x264_predict_4x4_init_arm( cpu, pf );