FFmpeg  4.4
h264_loopfilter.c
Go to the documentation of this file.
1 /*
2  * H.26L/H.264/AVC/JVT/14496-10/... loop filter
3  * Copyright (c) 2003 Michael Niedermayer <michaelni@gmx.at>
4  *
5  * This file is part of FFmpeg.
6  *
7  * FFmpeg is free software; you can redistribute it and/or
8  * modify it under the terms of the GNU Lesser General Public
9  * License as published by the Free Software Foundation; either
10  * version 2.1 of the License, or (at your option) any later version.
11  *
12  * FFmpeg is distributed in the hope that it will be useful,
13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15  * Lesser General Public License for more details.
16  *
17  * You should have received a copy of the GNU Lesser General Public
18  * License along with FFmpeg; if not, write to the Free Software
19  * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
20  */
21 
22 /**
23  * @file
24  * H.264 / AVC / MPEG-4 part10 loop filter.
25  * @author Michael Niedermayer <michaelni@gmx.at>
26  */
27 
28 #include "libavutil/internal.h"
29 #include "libavutil/intreadwrite.h"
30 #include "libavutil/mem_internal.h"
31 #include "internal.h"
32 #include "avcodec.h"
33 #include "h264dec.h"
34 #include "h264_ps.h"
35 #include "mathops.h"
36 #include "mpegutils.h"
37 #include "rectangle.h"
38 
39 /* Deblocking filter (p153) */
40 static const uint8_t alpha_table[52*3] = {
41  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
42  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
43  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
44  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
45  0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
46  0, 0, 0, 0, 0, 0, 4, 4, 5, 6,
47  7, 8, 9, 10, 12, 13, 15, 17, 20, 22,
48  25, 28, 32, 36, 40, 45, 50, 56, 63, 71,
49  80, 90,101,113,127,144,162,182,203,226,
50  255,255,
51  255,255,255,255,255,255,255,255,255,255,255,255,255,
52  255,255,255,255,255,255,255,255,255,255,255,255,255,
53  255,255,255,255,255,255,255,255,255,255,255,255,255,
54  255,255,255,255,255,255,255,255,255,255,255,255,255,
55 };
56 static const uint8_t beta_table[52*3] = {
57  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
58  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
59  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
60  0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
61  0, 0, 0, 0, 0, 0, 0, 0, 0, 0,
62  0, 0, 0, 0, 0, 0, 2, 2, 2, 3,
63  3, 3, 3, 4, 4, 4, 6, 6, 7, 7,
64  8, 8, 9, 9, 10, 10, 11, 11, 12, 12,
65  13, 13, 14, 14, 15, 15, 16, 16, 17, 17,
66  18, 18,
67  18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18,
68  18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18,
69  18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18,
70  18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18, 18,
71 };
72 static const uint8_t tc0_table[52*3][4] = {
73  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
74  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
75  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
76  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
77  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
78  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
79  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
80  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
81  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
82  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
83  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 },
84  {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 0 }, {-1, 0, 0, 1 },
85  {-1, 0, 0, 1 }, {-1, 0, 0, 1 }, {-1, 0, 0, 1 }, {-1, 0, 1, 1 }, {-1, 0, 1, 1 }, {-1, 1, 1, 1 },
86  {-1, 1, 1, 1 }, {-1, 1, 1, 1 }, {-1, 1, 1, 1 }, {-1, 1, 1, 2 }, {-1, 1, 1, 2 }, {-1, 1, 1, 2 },
87  {-1, 1, 1, 2 }, {-1, 1, 2, 3 }, {-1, 1, 2, 3 }, {-1, 2, 2, 3 }, {-1, 2, 2, 4 }, {-1, 2, 3, 4 },
88  {-1, 2, 3, 4 }, {-1, 3, 3, 5 }, {-1, 3, 4, 6 }, {-1, 3, 4, 6 }, {-1, 4, 5, 7 }, {-1, 4, 5, 8 },
89  {-1, 4, 6, 9 }, {-1, 5, 7,10 }, {-1, 6, 8,11 }, {-1, 6, 8,13 }, {-1, 7,10,14 }, {-1, 8,11,16 },
90  {-1, 9,12,18 }, {-1,10,13,20 }, {-1,11,15,23 }, {-1,13,17,25 },
91  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
92  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
93  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
94  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
95  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
96  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
97  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
98  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
99  {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 }, {-1,13,17,25 },
100 };
101 
102 /* intra: 0 if this loopfilter call is guaranteed to be inter (bS < 4), 1 if it might be intra (bS == 4) */
104  const int16_t bS[4],
105  unsigned int qp, int a, int b,
106  const H264Context *h, int intra)
107 {
108  const unsigned int index_a = qp + a;
109  const int alpha = alpha_table[index_a];
110  const int beta = beta_table[qp + b];
111  if (alpha ==0 || beta == 0) return;
112 
113  if( bS[0] < 4 || !intra ) {
114  int8_t tc[4];
115  tc[0] = tc0_table[index_a][bS[0]];
116  tc[1] = tc0_table[index_a][bS[1]];
117  tc[2] = tc0_table[index_a][bS[2]];
118  tc[3] = tc0_table[index_a][bS[3]];
119  h->h264dsp.h264_h_loop_filter_luma(pix, stride, alpha, beta, tc);
120  } else {
121  h->h264dsp.h264_h_loop_filter_luma_intra(pix, stride, alpha, beta);
122  }
123 }
124 
126  const int16_t bS[4],
127  unsigned int qp, int a, int b,
128  const H264Context *h, int intra)
129 {
130  const unsigned int index_a = qp + a;
131  const int alpha = alpha_table[index_a];
132  const int beta = beta_table[qp + b];
133  if (alpha ==0 || beta == 0) return;
134 
135  if( bS[0] < 4 || !intra ) {
136  int8_t tc[4];
137  tc[0] = tc0_table[index_a][bS[0]]+1;
138  tc[1] = tc0_table[index_a][bS[1]]+1;
139  tc[2] = tc0_table[index_a][bS[2]]+1;
140  tc[3] = tc0_table[index_a][bS[3]]+1;
141  h->h264dsp.h264_h_loop_filter_chroma(pix, stride, alpha, beta, tc);
142  } else {
143  h->h264dsp.h264_h_loop_filter_chroma_intra(pix, stride, alpha, beta);
144  }
145 }
146 
148  int stride,
149  const int16_t bS[7], int bsi,
150  int qp, int a, int b,
151  int intra)
152 {
153  const unsigned int index_a = qp + a;
154  const int alpha = alpha_table[index_a];
155  const int beta = beta_table[qp + b];
156  if (alpha ==0 || beta == 0) return;
157 
158  if( bS[0] < 4 || !intra ) {
159  int8_t tc[4];
160  tc[0] = tc0_table[index_a][bS[0*bsi]];
161  tc[1] = tc0_table[index_a][bS[1*bsi]];
162  tc[2] = tc0_table[index_a][bS[2*bsi]];
163  tc[3] = tc0_table[index_a][bS[3*bsi]];
164  h->h264dsp.h264_h_loop_filter_luma_mbaff(pix, stride, alpha, beta, tc);
165  } else {
166  h->h264dsp.h264_h_loop_filter_luma_mbaff_intra(pix, stride, alpha, beta);
167  }
168 }
169 
171  uint8_t *pix, int stride,
172  const int16_t bS[7],
173  int bsi, int qp, int a,
174  int b, int intra)
175 {
176  const unsigned int index_a = qp + a;
177  const int alpha = alpha_table[index_a];
178  const int beta = beta_table[qp + b];
179  if (alpha ==0 || beta == 0) return;
180 
181  if( bS[0] < 4 || !intra ) {
182  int8_t tc[4];
183  tc[0] = tc0_table[index_a][bS[0*bsi]] + 1;
184  tc[1] = tc0_table[index_a][bS[1*bsi]] + 1;
185  tc[2] = tc0_table[index_a][bS[2*bsi]] + 1;
186  tc[3] = tc0_table[index_a][bS[3*bsi]] + 1;
187  h->h264dsp.h264_h_loop_filter_chroma_mbaff(pix, stride, alpha, beta, tc);
188  } else {
189  h->h264dsp.h264_h_loop_filter_chroma_mbaff_intra(pix, stride, alpha, beta);
190  }
191 }
192 
194  const int16_t bS[4],
195  unsigned int qp, int a, int b,
196  const H264Context *h, int intra)
197 {
198  const unsigned int index_a = qp + a;
199  const int alpha = alpha_table[index_a];
200  const int beta = beta_table[qp + b];
201  if (alpha ==0 || beta == 0) return;
202 
203  if( bS[0] < 4 || !intra ) {
204  int8_t tc[4];
205  tc[0] = tc0_table[index_a][bS[0]];
206  tc[1] = tc0_table[index_a][bS[1]];
207  tc[2] = tc0_table[index_a][bS[2]];
208  tc[3] = tc0_table[index_a][bS[3]];
209  h->h264dsp.h264_v_loop_filter_luma(pix, stride, alpha, beta, tc);
210  } else {
211  h->h264dsp.h264_v_loop_filter_luma_intra(pix, stride, alpha, beta);
212  }
213 }
214 
216  const int16_t bS[4],
217  unsigned int qp, int a, int b,
218  const H264Context *h, int intra)
219 {
220  const unsigned int index_a = qp + a;
221  const int alpha = alpha_table[index_a];
222  const int beta = beta_table[qp + b];
223  if (alpha ==0 || beta == 0) return;
224 
225  if( bS[0] < 4 || !intra ) {
226  int8_t tc[4];
227  tc[0] = tc0_table[index_a][bS[0]]+1;
228  tc[1] = tc0_table[index_a][bS[1]]+1;
229  tc[2] = tc0_table[index_a][bS[2]]+1;
230  tc[3] = tc0_table[index_a][bS[3]]+1;
231  h->h264dsp.h264_v_loop_filter_chroma(pix, stride, alpha, beta, tc);
232  } else {
233  h->h264dsp.h264_v_loop_filter_chroma_intra(pix, stride, alpha, beta);
234  }
235 }
236 
238  H264SliceContext *sl,
239  int mb_x, int mb_y,
240  uint8_t *img_y,
241  uint8_t *img_cb,
242  uint8_t *img_cr,
243  unsigned int linesize,
244  unsigned int uvlinesize,
245  int pixel_shift)
246 {
247  int chroma = CHROMA(h) && !(CONFIG_GRAY && (h->flags & AV_CODEC_FLAG_GRAY));
248  int chroma444 = CHROMA444(h);
249  int chroma422 = CHROMA422(h);
250 
251  int mb_xy = sl->mb_xy;
252  int left_type = sl->left_type[LTOP];
253  int top_type = sl->top_type;
254 
255  int qp_bd_offset = 6 * (h->ps.sps->bit_depth_luma - 8);
256  int a = 52 + sl->slice_alpha_c0_offset - qp_bd_offset;
257  int b = 52 + sl->slice_beta_offset - qp_bd_offset;
258 
259  int mb_type = h->cur_pic.mb_type[mb_xy];
260  int qp = h->cur_pic.qscale_table[mb_xy];
261  int qp0 = h->cur_pic.qscale_table[mb_xy - 1];
262  int qp1 = h->cur_pic.qscale_table[sl->top_mb_xy];
263  int qpc = get_chroma_qp(h->ps.pps, 0, qp);
264  int qpc0 = get_chroma_qp(h->ps.pps, 0, qp0);
265  int qpc1 = get_chroma_qp(h->ps.pps, 0, qp1);
266  qp0 = (qp + qp0 + 1) >> 1;
267  qp1 = (qp + qp1 + 1) >> 1;
268  qpc0 = (qpc + qpc0 + 1) >> 1;
269  qpc1 = (qpc + qpc1 + 1) >> 1;
270 
271  if( IS_INTRA(mb_type) ) {
272  static const int16_t bS4[4] = {4,4,4,4};
273  static const int16_t bS3[4] = {3,3,3,3};
274  const int16_t *bSH = FIELD_PICTURE(h) ? bS3 : bS4;
275  if(left_type)
276  filter_mb_edgev( &img_y[4*0<<pixel_shift], linesize, bS4, qp0, a, b, h, 1);
277  if( IS_8x8DCT(mb_type) ) {
278  filter_mb_edgev( &img_y[4*2<<pixel_shift], linesize, bS3, qp, a, b, h, 0);
279  if(top_type){
280  filter_mb_edgeh( &img_y[4*0*linesize], linesize, bSH, qp1, a, b, h, 1);
281  }
282  filter_mb_edgeh( &img_y[4*2*linesize], linesize, bS3, qp, a, b, h, 0);
283  } else {
284  filter_mb_edgev( &img_y[4*1<<pixel_shift], linesize, bS3, qp, a, b, h, 0);
285  filter_mb_edgev( &img_y[4*2<<pixel_shift], linesize, bS3, qp, a, b, h, 0);
286  filter_mb_edgev( &img_y[4*3<<pixel_shift], linesize, bS3, qp, a, b, h, 0);
287  if(top_type){
288  filter_mb_edgeh( &img_y[4*0*linesize], linesize, bSH, qp1, a, b, h, 1);
289  }
290  filter_mb_edgeh( &img_y[4*1*linesize], linesize, bS3, qp, a, b, h, 0);
291  filter_mb_edgeh( &img_y[4*2*linesize], linesize, bS3, qp, a, b, h, 0);
292  filter_mb_edgeh( &img_y[4*3*linesize], linesize, bS3, qp, a, b, h, 0);
293  }
294  if(chroma){
295  if(chroma444){
296  if(left_type){
297  filter_mb_edgev( &img_cb[4*0<<pixel_shift], linesize, bS4, qpc0, a, b, h, 1);
298  filter_mb_edgev( &img_cr[4*0<<pixel_shift], linesize, bS4, qpc0, a, b, h, 1);
299  }
300  if( IS_8x8DCT(mb_type) ) {
301  filter_mb_edgev( &img_cb[4*2<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
302  filter_mb_edgev( &img_cr[4*2<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
303  if(top_type){
304  filter_mb_edgeh( &img_cb[4*0*linesize], linesize, bSH, qpc1, a, b, h, 1 );
305  filter_mb_edgeh( &img_cr[4*0*linesize], linesize, bSH, qpc1, a, b, h, 1 );
306  }
307  filter_mb_edgeh( &img_cb[4*2*linesize], linesize, bS3, qpc, a, b, h, 0);
308  filter_mb_edgeh( &img_cr[4*2*linesize], linesize, bS3, qpc, a, b, h, 0);
309  } else {
310  filter_mb_edgev( &img_cb[4*1<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
311  filter_mb_edgev( &img_cr[4*1<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
312  filter_mb_edgev( &img_cb[4*2<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
313  filter_mb_edgev( &img_cr[4*2<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
314  filter_mb_edgev( &img_cb[4*3<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
315  filter_mb_edgev( &img_cr[4*3<<pixel_shift], linesize, bS3, qpc, a, b, h, 0);
316  if(top_type){
317  filter_mb_edgeh( &img_cb[4*0*linesize], linesize, bSH, qpc1, a, b, h, 1);
318  filter_mb_edgeh( &img_cr[4*0*linesize], linesize, bSH, qpc1, a, b, h, 1);
319  }
320  filter_mb_edgeh( &img_cb[4*1*linesize], linesize, bS3, qpc, a, b, h, 0);
321  filter_mb_edgeh( &img_cr[4*1*linesize], linesize, bS3, qpc, a, b, h, 0);
322  filter_mb_edgeh( &img_cb[4*2*linesize], linesize, bS3, qpc, a, b, h, 0);
323  filter_mb_edgeh( &img_cr[4*2*linesize], linesize, bS3, qpc, a, b, h, 0);
324  filter_mb_edgeh( &img_cb[4*3*linesize], linesize, bS3, qpc, a, b, h, 0);
325  filter_mb_edgeh( &img_cr[4*3*linesize], linesize, bS3, qpc, a, b, h, 0);
326  }
327  }else if(chroma422){
328  if(left_type){
329  filter_mb_edgecv(&img_cb[2*0<<pixel_shift], uvlinesize, bS4, qpc0, a, b, h, 1);
330  filter_mb_edgecv(&img_cr[2*0<<pixel_shift], uvlinesize, bS4, qpc0, a, b, h, 1);
331  }
332  filter_mb_edgecv(&img_cb[2*2<<pixel_shift], uvlinesize, bS3, qpc, a, b, h, 0);
333  filter_mb_edgecv(&img_cr[2*2<<pixel_shift], uvlinesize, bS3, qpc, a, b, h, 0);
334  if(top_type){
335  filter_mb_edgech(&img_cb[4*0*uvlinesize], uvlinesize, bSH, qpc1, a, b, h, 1);
336  filter_mb_edgech(&img_cr[4*0*uvlinesize], uvlinesize, bSH, qpc1, a, b, h, 1);
337  }
338  filter_mb_edgech(&img_cb[4*1*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
339  filter_mb_edgech(&img_cr[4*1*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
340  filter_mb_edgech(&img_cb[4*2*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
341  filter_mb_edgech(&img_cr[4*2*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
342  filter_mb_edgech(&img_cb[4*3*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
343  filter_mb_edgech(&img_cr[4*3*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
344  }else{
345  if(left_type){
346  filter_mb_edgecv( &img_cb[2*0<<pixel_shift], uvlinesize, bS4, qpc0, a, b, h, 1);
347  filter_mb_edgecv( &img_cr[2*0<<pixel_shift], uvlinesize, bS4, qpc0, a, b, h, 1);
348  }
349  filter_mb_edgecv( &img_cb[2*2<<pixel_shift], uvlinesize, bS3, qpc, a, b, h, 0);
350  filter_mb_edgecv( &img_cr[2*2<<pixel_shift], uvlinesize, bS3, qpc, a, b, h, 0);
351  if(top_type){
352  filter_mb_edgech( &img_cb[2*0*uvlinesize], uvlinesize, bSH, qpc1, a, b, h, 1);
353  filter_mb_edgech( &img_cr[2*0*uvlinesize], uvlinesize, bSH, qpc1, a, b, h, 1);
354  }
355  filter_mb_edgech( &img_cb[2*2*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
356  filter_mb_edgech( &img_cr[2*2*uvlinesize], uvlinesize, bS3, qpc, a, b, h, 0);
357  }
358  }
359  return;
360  } else {
361  LOCAL_ALIGNED(8, int16_t, bS, [2], [4][4]);
362  int edges;
363  if( IS_8x8DCT(mb_type) && (sl->cbp&7) == 7 && !chroma444 ) {
364  edges = 4;
365  AV_WN64A(bS[0][0], 0x0002000200020002ULL);
366  AV_WN64A(bS[0][2], 0x0002000200020002ULL);
367  AV_WN64A(bS[1][0], 0x0002000200020002ULL);
368  AV_WN64A(bS[1][2], 0x0002000200020002ULL);
369  } else {
370  int mask_edge1 = (3*(((5*mb_type)>>5)&1)) | (mb_type>>4); //(mb_type & (MB_TYPE_16x16 | MB_TYPE_8x16)) ? 3 : (mb_type & MB_TYPE_16x8) ? 1 : 0;
371  int mask_edge0 = 3*((mask_edge1>>1) & ((5*left_type)>>5)&1); // (mb_type & (MB_TYPE_16x16 | MB_TYPE_8x16)) && (h->left_type[LTOP] & (MB_TYPE_16x16 | MB_TYPE_8x16)) ? 3 : 0;
372  int step = 1+(mb_type>>24); //IS_8x8DCT(mb_type) ? 2 : 1;
373  edges = 4 - 3*((mb_type>>3) & !(sl->cbp & 15)); //(mb_type & MB_TYPE_16x16) && !(h->cbp & 15) ? 1 : 4;
374  h->h264dsp.h264_loop_filter_strength(bS, sl->non_zero_count_cache, sl->ref_cache, sl->mv_cache,
375  sl->list_count==2, edges, step, mask_edge0, mask_edge1, FIELD_PICTURE(h));
376  }
377  if( IS_INTRA(left_type) )
378  AV_WN64A(bS[0][0], 0x0004000400040004ULL);
379  if( IS_INTRA(top_type) )
380  AV_WN64A(bS[1][0], FIELD_PICTURE(h) ? 0x0003000300030003ULL : 0x0004000400040004ULL);
381 
382 #define FILTER(hv,dir,edge,intra)\
383  if(AV_RN64A(bS[dir][edge])) { \
384  filter_mb_edge##hv( &img_y[4*edge*(dir?linesize:1<<pixel_shift)], linesize, bS[dir][edge], edge ? qp : qp##dir, a, b, h, intra );\
385  if(chroma){\
386  if(chroma444){\
387  filter_mb_edge##hv( &img_cb[4*edge*(dir?linesize:1<<pixel_shift)], linesize, bS[dir][edge], edge ? qpc : qpc##dir, a, b, h, intra );\
388  filter_mb_edge##hv( &img_cr[4*edge*(dir?linesize:1<<pixel_shift)], linesize, bS[dir][edge], edge ? qpc : qpc##dir, a, b, h, intra );\
389  } else if(!(edge&1)) {\
390  filter_mb_edgec##hv( &img_cb[2*edge*(dir?uvlinesize:1<<pixel_shift)], uvlinesize, bS[dir][edge], edge ? qpc : qpc##dir, a, b, h, intra );\
391  filter_mb_edgec##hv( &img_cr[2*edge*(dir?uvlinesize:1<<pixel_shift)], uvlinesize, bS[dir][edge], edge ? qpc : qpc##dir, a, b, h, intra );\
392  }\
393  }\
394  }
395  if(left_type)
396  FILTER(v,0,0,1);
397  if( edges == 1 ) {
398  if(top_type)
399  FILTER(h,1,0,1);
400  } else if( IS_8x8DCT(mb_type) ) {
401  FILTER(v,0,2,0);
402  if(top_type)
403  FILTER(h,1,0,1);
404  FILTER(h,1,2,0);
405  } else {
406  FILTER(v,0,1,0);
407  FILTER(v,0,2,0);
408  FILTER(v,0,3,0);
409  if(top_type)
410  FILTER(h,1,0,1);
411  FILTER(h,1,1,0);
412  FILTER(h,1,2,0);
413  FILTER(h,1,3,0);
414  }
415 #undef FILTER
416  }
417 }
418 
420  int mb_x, int mb_y, uint8_t *img_y,
421  uint8_t *img_cb, uint8_t *img_cr,
422  unsigned int linesize, unsigned int uvlinesize)
423 {
425  if(!h->h264dsp.h264_loop_filter_strength || h->ps.pps->chroma_qp_diff) {
426  ff_h264_filter_mb(h, sl, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize);
427  return;
428  }
429 
430 #if CONFIG_SMALL
431  h264_filter_mb_fast_internal(h, sl, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, h->pixel_shift);
432 #else
433  if(h->pixel_shift){
434  h264_filter_mb_fast_internal(h, sl, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, 1);
435  }else{
436  h264_filter_mb_fast_internal(h, sl, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, 0);
437  }
438 #endif
439 }
440 
441 static int check_mv(H264SliceContext *sl, long b_idx, long bn_idx, int mvy_limit)
442 {
443  int v;
444 
445  v = sl->ref_cache[0][b_idx] != sl->ref_cache[0][bn_idx];
446  if (!v && sl->ref_cache[0][b_idx] != -1)
447  v = sl->mv_cache[0][b_idx][0] - sl->mv_cache[0][bn_idx][0] + 3 >= 7U |
448  FFABS(sl->mv_cache[0][b_idx][1] - sl->mv_cache[0][bn_idx][1]) >= mvy_limit;
449 
450  if (sl->list_count == 2) {
451  if(!v)
452  v = sl->ref_cache[1][b_idx] != sl->ref_cache[1][bn_idx] |
453  sl->mv_cache[1][b_idx][0] - sl->mv_cache[1][bn_idx][0] + 3 >= 7U |
454  FFABS(sl->mv_cache[1][b_idx][1] - sl->mv_cache[1][bn_idx][1]) >= mvy_limit;
455 
456  if(v){
457  if (sl->ref_cache[0][b_idx] != sl->ref_cache[1][bn_idx] |
458  sl->ref_cache[1][b_idx] != sl->ref_cache[0][bn_idx])
459  return 1;
460  return
461  sl->mv_cache[0][b_idx][0] - sl->mv_cache[1][bn_idx][0] + 3 >= 7U |
462  FFABS(sl->mv_cache[0][b_idx][1] - sl->mv_cache[1][bn_idx][1]) >= mvy_limit |
463  sl->mv_cache[1][b_idx][0] - sl->mv_cache[0][bn_idx][0] + 3 >= 7U |
464  FFABS(sl->mv_cache[1][b_idx][1] - sl->mv_cache[0][bn_idx][1]) >= mvy_limit;
465  }
466  }
467 
468  return v;
469 }
470 
472  int mb_x, int mb_y,
473  uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr,
474  unsigned int linesize, unsigned int uvlinesize,
475  int mb_xy, int mb_type, int mvy_limit,
476  int first_vertical_edge_done, int a, int b,
477  int chroma, int dir)
478 {
479  int edge;
480  int chroma_qp_avg[2];
481  int chroma444 = CHROMA444(h);
482  int chroma422 = CHROMA422(h);
483  const int mbm_xy = dir == 0 ? mb_xy -1 : sl->top_mb_xy;
484  const int mbm_type = dir == 0 ? sl->left_type[LTOP] : sl->top_type;
485 
486  // how often to recheck mv-based bS when iterating between edges
487  static const uint8_t mask_edge_tab[2][8]={{0,3,3,3,1,1,1,1},
488  {0,3,1,1,3,3,3,3}};
489  const int mask_edge = mask_edge_tab[dir][(mb_type>>3)&7];
490  const int edges = mask_edge== 3 && !(sl->cbp&15) ? 1 : 4;
491 
492  // how often to recheck mv-based bS when iterating along each edge
493  const int mask_par0 = mb_type & (MB_TYPE_16x16 | (MB_TYPE_8x16 >> dir));
494 
495  if(mbm_type && !first_vertical_edge_done){
496 
497  if (FRAME_MBAFF(h) && (dir == 1) && ((mb_y&1) == 0)
498  && IS_INTERLACED(mbm_type&~mb_type)
499  ) {
500  // This is a special case in the norm where the filtering must
501  // be done twice (one each of the field) even if we are in a
502  // frame macroblock.
503  //
504  unsigned int tmp_linesize = 2 * linesize;
505  unsigned int tmp_uvlinesize = 2 * uvlinesize;
506  int mbn_xy = mb_xy - 2 * h->mb_stride;
507  int j;
508 
509  for(j=0; j<2; j++, mbn_xy += h->mb_stride){
510  LOCAL_ALIGNED(8, int16_t, bS, [4]);
511  int qp;
512  if (IS_INTRA(mb_type | h->cur_pic.mb_type[mbn_xy])) {
513  AV_WN64A(bS, 0x0003000300030003ULL);
514  } else {
515  if (!CABAC(h) && IS_8x8DCT(h->cur_pic.mb_type[mbn_xy])) {
516  bS[0]= 1+((h->cbp_table[mbn_xy] & 0x4000) || sl->non_zero_count_cache[scan8[0]+0]);
517  bS[1]= 1+((h->cbp_table[mbn_xy] & 0x4000) || sl->non_zero_count_cache[scan8[0]+1]);
518  bS[2]= 1+((h->cbp_table[mbn_xy] & 0x8000) || sl->non_zero_count_cache[scan8[0]+2]);
519  bS[3]= 1+((h->cbp_table[mbn_xy] & 0x8000) || sl->non_zero_count_cache[scan8[0]+3]);
520  }else{
521  const uint8_t *mbn_nnz = h->non_zero_count[mbn_xy] + 3*4;
522  int i;
523  for( i = 0; i < 4; i++ ) {
524  bS[i] = 1 + !!(sl->non_zero_count_cache[scan8[0]+i] | mbn_nnz[i]);
525  }
526  }
527  }
528  // Do not use s->qscale as luma quantizer because it has not the same
529  // value in IPCM macroblocks.
530  qp = (h->cur_pic.qscale_table[mb_xy] + h->cur_pic.qscale_table[mbn_xy] + 1) >> 1;
531  ff_tlog(h->avctx, "filter mb:%d/%d dir:%d, QPy:%d ls:%d uvls:%d", mb_x, mb_y, dir, qp, tmp_linesize, tmp_uvlinesize);
532  { int i; for (i = 0; i < 4; i++) ff_tlog(h->avctx, " bS[%d]:%d", i, bS[i]); ff_tlog(h->avctx, "\n"); }
533  filter_mb_edgeh( &img_y[j*linesize], tmp_linesize, bS, qp, a, b, h, 0 );
534  chroma_qp_avg[0] = (sl->chroma_qp[0] + get_chroma_qp(h->ps.pps, 0, h->cur_pic.qscale_table[mbn_xy]) + 1) >> 1;
535  chroma_qp_avg[1] = (sl->chroma_qp[1] + get_chroma_qp(h->ps.pps, 1, h->cur_pic.qscale_table[mbn_xy]) + 1) >> 1;
536  if (chroma) {
537  if (chroma444) {
538  filter_mb_edgeh (&img_cb[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[0], a, b, h, 0);
539  filter_mb_edgeh (&img_cr[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[1], a, b, h, 0);
540  } else {
541  filter_mb_edgech(&img_cb[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[0], a, b, h, 0);
542  filter_mb_edgech(&img_cr[j*uvlinesize], tmp_uvlinesize, bS, chroma_qp_avg[1], a, b, h, 0);
543  }
544  }
545  }
546  }else{
547  LOCAL_ALIGNED(8, int16_t, bS, [4]);
548  int qp;
549 
550  if( IS_INTRA(mb_type|mbm_type)) {
551  AV_WN64A(bS, 0x0003000300030003ULL);
552  if ( (!IS_INTERLACED(mb_type|mbm_type))
553  || ((FRAME_MBAFF(h) || (h->picture_structure != PICT_FRAME)) && (dir == 0))
554  )
555  AV_WN64A(bS, 0x0004000400040004ULL);
556  } else {
557  int i;
558  int mv_done;
559 
560  if( dir && FRAME_MBAFF(h) && IS_INTERLACED(mb_type ^ mbm_type)) {
561  AV_WN64A(bS, 0x0001000100010001ULL);
562  mv_done = 1;
563  }
564  else if( mask_par0 && ((mbm_type & (MB_TYPE_16x16 | (MB_TYPE_8x16 >> dir)))) ) {
565  int b_idx= 8 + 4;
566  int bn_idx= b_idx - (dir ? 8:1);
567 
568  bS[0] = bS[1] = bS[2] = bS[3] = check_mv(sl, 8 + 4, bn_idx, mvy_limit);
569  mv_done = 1;
570  }
571  else
572  mv_done = 0;
573 
574  for( i = 0; i < 4; i++ ) {
575  int x = dir == 0 ? 0 : i;
576  int y = dir == 0 ? i : 0;
577  int b_idx= 8 + 4 + x + 8*y;
578  int bn_idx= b_idx - (dir ? 8:1);
579 
580  if (sl->non_zero_count_cache[b_idx] |
581  sl->non_zero_count_cache[bn_idx]) {
582  bS[i] = 2;
583  }
584  else if(!mv_done)
585  {
586  bS[i] = check_mv(sl, b_idx, bn_idx, mvy_limit);
587  }
588  }
589  }
590 
591  /* Filter edge */
592  // Do not use s->qscale as luma quantizer because it has not the same
593  // value in IPCM macroblocks.
594  if(bS[0]+bS[1]+bS[2]+bS[3]){
595  qp = (h->cur_pic.qscale_table[mb_xy] + h->cur_pic.qscale_table[mbm_xy] + 1) >> 1;
596  //ff_tlog(h->avctx, "filter mb:%d/%d dir:%d edge:%d, QPy:%d, QPc:%d, QPcn:%d\n", mb_x, mb_y, dir, edge, qp, h->chroma_qp[0], h->cur_pic.qscale_table[mbn_xy]);
597  ff_tlog(h->avctx, "filter mb:%d/%d dir:%d edge:%d, QPy:%d ls:%d uvls:%d", mb_x, mb_y, dir, edge, qp, linesize, uvlinesize);
598  //{ int i; for (i = 0; i < 4; i++) ff_tlog(h->avctx, " bS[%d]:%d", i, bS[i]); ff_tlog(h->avctx, "\n"); }
599  chroma_qp_avg[0] = (sl->chroma_qp[0] + get_chroma_qp(h->ps.pps, 0, h->cur_pic.qscale_table[mbm_xy]) + 1) >> 1;
600  chroma_qp_avg[1] = (sl->chroma_qp[1] + get_chroma_qp(h->ps.pps, 1, h->cur_pic.qscale_table[mbm_xy]) + 1) >> 1;
601  if( dir == 0 ) {
602  filter_mb_edgev( &img_y[0], linesize, bS, qp, a, b, h, 1 );
603  if (chroma) {
604  if (chroma444) {
605  filter_mb_edgev ( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], a, b, h, 1);
606  filter_mb_edgev ( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], a, b, h, 1);
607  } else {
608  filter_mb_edgecv( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], a, b, h, 1);
609  filter_mb_edgecv( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], a, b, h, 1);
610  }
611  }
612  } else {
613  filter_mb_edgeh( &img_y[0], linesize, bS, qp, a, b, h, 1 );
614  if (chroma) {
615  if (chroma444) {
616  filter_mb_edgeh ( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], a, b, h, 1);
617  filter_mb_edgeh ( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], a, b, h, 1);
618  } else {
619  filter_mb_edgech( &img_cb[0], uvlinesize, bS, chroma_qp_avg[0], a, b, h, 1);
620  filter_mb_edgech( &img_cr[0], uvlinesize, bS, chroma_qp_avg[1], a, b, h, 1);
621  }
622  }
623  }
624  }
625  }
626  }
627 
628  /* Calculate bS */
629  for( edge = 1; edge < edges; edge++ ) {
630  LOCAL_ALIGNED(8, int16_t, bS, [4]);
631  int qp;
632  const int deblock_edge = !IS_8x8DCT(mb_type & (edge<<24)); // (edge&1) && IS_8x8DCT(mb_type)
633 
634  if (!deblock_edge && (!chroma422 || dir == 0))
635  continue;
636 
637  if( IS_INTRA(mb_type)) {
638  AV_WN64A(bS, 0x0003000300030003ULL);
639  } else {
640  int i;
641  int mv_done;
642 
643  if( edge & mask_edge ) {
644  AV_ZERO64(bS);
645  mv_done = 1;
646  }
647  else if( mask_par0 ) {
648  int b_idx= 8 + 4 + edge * (dir ? 8:1);
649  int bn_idx= b_idx - (dir ? 8:1);
650 
651  bS[0] = bS[1] = bS[2] = bS[3] = check_mv(sl, b_idx, bn_idx, mvy_limit);
652  mv_done = 1;
653  }
654  else
655  mv_done = 0;
656 
657  for( i = 0; i < 4; i++ ) {
658  int x = dir == 0 ? edge : i;
659  int y = dir == 0 ? i : edge;
660  int b_idx= 8 + 4 + x + 8*y;
661  int bn_idx= b_idx - (dir ? 8:1);
662 
663  if (sl->non_zero_count_cache[b_idx] |
664  sl->non_zero_count_cache[bn_idx]) {
665  bS[i] = 2;
666  }
667  else if(!mv_done)
668  {
669  bS[i] = check_mv(sl, b_idx, bn_idx, mvy_limit);
670  }
671  }
672 
673  if(bS[0]+bS[1]+bS[2]+bS[3] == 0)
674  continue;
675  }
676 
677  /* Filter edge */
678  // Do not use s->qscale as luma quantizer because it has not the same
679  // value in IPCM macroblocks.
680  qp = h->cur_pic.qscale_table[mb_xy];
681  //ff_tlog(h->avctx, "filter mb:%d/%d dir:%d edge:%d, QPy:%d, QPc:%d, QPcn:%d\n", mb_x, mb_y, dir, edge, qp, h->chroma_qp[0], h->cur_pic.qscale_table[mbn_xy]);
682  ff_tlog(h->avctx, "filter mb:%d/%d dir:%d edge:%d, QPy:%d ls:%d uvls:%d", mb_x, mb_y, dir, edge, qp, linesize, uvlinesize);
683  //{ int i; for (i = 0; i < 4; i++) ff_tlog(h->avctx, " bS[%d]:%d", i, bS[i]); ff_tlog(h->avctx, "\n"); }
684  if( dir == 0 ) {
685  filter_mb_edgev( &img_y[4*edge << h->pixel_shift], linesize, bS, qp, a, b, h, 0 );
686  if (chroma) {
687  if (chroma444) {
688  filter_mb_edgev ( &img_cb[4*edge << h->pixel_shift], uvlinesize, bS, sl->chroma_qp[0], a, b, h, 0);
689  filter_mb_edgev ( &img_cr[4*edge << h->pixel_shift], uvlinesize, bS, sl->chroma_qp[1], a, b, h, 0);
690  } else if( (edge&1) == 0 ) {
691  filter_mb_edgecv( &img_cb[2*edge << h->pixel_shift], uvlinesize, bS, sl->chroma_qp[0], a, b, h, 0);
692  filter_mb_edgecv( &img_cr[2*edge << h->pixel_shift], uvlinesize, bS, sl->chroma_qp[1], a, b, h, 0);
693  }
694  }
695  } else {
696  if (chroma422) {
697  if (deblock_edge)
698  filter_mb_edgeh(&img_y[4*edge*linesize], linesize, bS, qp, a, b, h, 0);
699  if (chroma) {
700  filter_mb_edgech(&img_cb[4*edge*uvlinesize], uvlinesize, bS, sl->chroma_qp[0], a, b, h, 0);
701  filter_mb_edgech(&img_cr[4*edge*uvlinesize], uvlinesize, bS, sl->chroma_qp[1], a, b, h, 0);
702  }
703  } else {
704  filter_mb_edgeh(&img_y[4*edge*linesize], linesize, bS, qp, a, b, h, 0);
705  if (chroma) {
706  if (chroma444) {
707  filter_mb_edgeh (&img_cb[4*edge*uvlinesize], uvlinesize, bS, sl->chroma_qp[0], a, b, h, 0);
708  filter_mb_edgeh (&img_cr[4*edge*uvlinesize], uvlinesize, bS, sl->chroma_qp[1], a, b, h, 0);
709  } else if ((edge&1) == 0) {
710  filter_mb_edgech(&img_cb[2*edge*uvlinesize], uvlinesize, bS, sl->chroma_qp[0], a, b, h, 0);
711  filter_mb_edgech(&img_cr[2*edge*uvlinesize], uvlinesize, bS, sl->chroma_qp[1], a, b, h, 0);
712  }
713  }
714  }
715  }
716  }
717 }
718 
720  int mb_x, int mb_y,
721  uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr,
722  unsigned int linesize, unsigned int uvlinesize)
723 {
724  const int mb_xy= mb_x + mb_y*h->mb_stride;
725  const int mb_type = h->cur_pic.mb_type[mb_xy];
726  const int mvy_limit = IS_INTERLACED(mb_type) ? 2 : 4;
727  int first_vertical_edge_done = 0;
728  int chroma = CHROMA(h) && !(CONFIG_GRAY && (h->flags & AV_CODEC_FLAG_GRAY));
729  int qp_bd_offset = 6 * (h->ps.sps->bit_depth_luma - 8);
730  int a = 52 + sl->slice_alpha_c0_offset - qp_bd_offset;
731  int b = 52 + sl->slice_beta_offset - qp_bd_offset;
732 
733  if (FRAME_MBAFF(h)
734  // and current and left pair do not have the same interlaced type
735  && IS_INTERLACED(mb_type ^ sl->left_type[LTOP])
736  // and left mb is in available to us
737  && sl->left_type[LTOP]) {
738  /* First vertical edge is different in MBAFF frames
739  * There are 8 different bS to compute and 2 different Qp
740  */
741  LOCAL_ALIGNED(8, int16_t, bS, [8]);
742  int qp[2];
743  int bqp[2];
744  int rqp[2];
745  int mb_qp, mbn0_qp, mbn1_qp;
746  int i;
747  first_vertical_edge_done = 1;
748 
749  if( IS_INTRA(mb_type) ) {
750  AV_WN64A(&bS[0], 0x0004000400040004ULL);
751  AV_WN64A(&bS[4], 0x0004000400040004ULL);
752  } else {
753  static const uint8_t offset[2][2][8]={
754  {
755  {3+4*0, 3+4*0, 3+4*0, 3+4*0, 3+4*1, 3+4*1, 3+4*1, 3+4*1},
756  {3+4*2, 3+4*2, 3+4*2, 3+4*2, 3+4*3, 3+4*3, 3+4*3, 3+4*3},
757  },{
758  {3+4*0, 3+4*1, 3+4*2, 3+4*3, 3+4*0, 3+4*1, 3+4*2, 3+4*3},
759  {3+4*0, 3+4*1, 3+4*2, 3+4*3, 3+4*0, 3+4*1, 3+4*2, 3+4*3},
760  }
761  };
762  const uint8_t *off= offset[MB_FIELD(sl)][mb_y&1];
763  for( i = 0; i < 8; i++ ) {
764  int j= MB_FIELD(sl) ? i>>2 : i&1;
765  int mbn_xy = sl->left_mb_xy[LEFT(j)];
766  int mbn_type = sl->left_type[LEFT(j)];
767 
768  if( IS_INTRA( mbn_type ) )
769  bS[i] = 4;
770  else{
771  bS[i] = 1 + !!(sl->non_zero_count_cache[12+8*(i>>1)] |
772  ((!h->ps.pps->cabac && IS_8x8DCT(mbn_type)) ?
773  (h->cbp_table[mbn_xy] & (((MB_FIELD(sl) ? (i&2) : (mb_y&1)) ? 8 : 2) << 12))
774  :
775  h->non_zero_count[mbn_xy][ off[i] ]));
776  }
777  }
778  }
779 
780  mb_qp = h->cur_pic.qscale_table[mb_xy];
781  mbn0_qp = h->cur_pic.qscale_table[sl->left_mb_xy[0]];
782  mbn1_qp = h->cur_pic.qscale_table[sl->left_mb_xy[1]];
783  qp[0] = ( mb_qp + mbn0_qp + 1 ) >> 1;
784  bqp[0] = (get_chroma_qp(h->ps.pps, 0, mb_qp) +
785  get_chroma_qp(h->ps.pps, 0, mbn0_qp) + 1) >> 1;
786  rqp[0] = (get_chroma_qp(h->ps.pps, 1, mb_qp) +
787  get_chroma_qp(h->ps.pps, 1, mbn0_qp) + 1) >> 1;
788  qp[1] = ( mb_qp + mbn1_qp + 1 ) >> 1;
789  bqp[1] = (get_chroma_qp(h->ps.pps, 0, mb_qp) +
790  get_chroma_qp(h->ps.pps, 0, mbn1_qp) + 1 ) >> 1;
791  rqp[1] = (get_chroma_qp(h->ps.pps, 1, mb_qp) +
792  get_chroma_qp(h->ps.pps, 1, mbn1_qp) + 1 ) >> 1;
793 
794  /* Filter edge */
795  ff_tlog(h->avctx, "filter mb:%d/%d MBAFF, QPy:%d/%d, QPb:%d/%d QPr:%d/%d ls:%d uvls:%d", mb_x, mb_y, qp[0], qp[1], bqp[0], bqp[1], rqp[0], rqp[1], linesize, uvlinesize);
796  { int i; for (i = 0; i < 8; i++) ff_tlog(h->avctx, " bS[%d]:%d", i, bS[i]); ff_tlog(h->avctx, "\n"); }
797  if (MB_FIELD(sl)) {
798  filter_mb_mbaff_edgev ( h, img_y , linesize, bS , 1, qp [0], a, b, 1 );
799  filter_mb_mbaff_edgev ( h, img_y + 8* linesize, linesize, bS+4, 1, qp [1], a, b, 1 );
800  if (chroma){
801  if (CHROMA444(h)) {
802  filter_mb_mbaff_edgev ( h, img_cb, uvlinesize, bS , 1, bqp[0], a, b, 1 );
803  filter_mb_mbaff_edgev ( h, img_cb + 8*uvlinesize, uvlinesize, bS+4, 1, bqp[1], a, b, 1 );
804  filter_mb_mbaff_edgev ( h, img_cr, uvlinesize, bS , 1, rqp[0], a, b, 1 );
805  filter_mb_mbaff_edgev ( h, img_cr + 8*uvlinesize, uvlinesize, bS+4, 1, rqp[1], a, b, 1 );
806  } else if (CHROMA422(h)) {
807  filter_mb_mbaff_edgecv(h, img_cb, uvlinesize, bS , 1, bqp[0], a, b, 1);
808  filter_mb_mbaff_edgecv(h, img_cb + 8*uvlinesize, uvlinesize, bS+4, 1, bqp[1], a, b, 1);
809  filter_mb_mbaff_edgecv(h, img_cr, uvlinesize, bS , 1, rqp[0], a, b, 1);
810  filter_mb_mbaff_edgecv(h, img_cr + 8*uvlinesize, uvlinesize, bS+4, 1, rqp[1], a, b, 1);
811  }else{
812  filter_mb_mbaff_edgecv( h, img_cb, uvlinesize, bS , 1, bqp[0], a, b, 1 );
813  filter_mb_mbaff_edgecv( h, img_cb + 4*uvlinesize, uvlinesize, bS+4, 1, bqp[1], a, b, 1 );
814  filter_mb_mbaff_edgecv( h, img_cr, uvlinesize, bS , 1, rqp[0], a, b, 1 );
815  filter_mb_mbaff_edgecv( h, img_cr + 4*uvlinesize, uvlinesize, bS+4, 1, rqp[1], a, b, 1 );
816  }
817  }
818  }else{
819  filter_mb_mbaff_edgev ( h, img_y , 2* linesize, bS , 2, qp [0], a, b, 1 );
820  filter_mb_mbaff_edgev ( h, img_y + linesize, 2* linesize, bS+1, 2, qp [1], a, b, 1 );
821  if (chroma){
822  if (CHROMA444(h)) {
823  filter_mb_mbaff_edgev ( h, img_cb, 2*uvlinesize, bS , 2, bqp[0], a, b, 1 );
824  filter_mb_mbaff_edgev ( h, img_cb + uvlinesize, 2*uvlinesize, bS+1, 2, bqp[1], a, b, 1 );
825  filter_mb_mbaff_edgev ( h, img_cr, 2*uvlinesize, bS , 2, rqp[0], a, b, 1 );
826  filter_mb_mbaff_edgev ( h, img_cr + uvlinesize, 2*uvlinesize, bS+1, 2, rqp[1], a, b, 1 );
827  }else{
828  filter_mb_mbaff_edgecv( h, img_cb, 2*uvlinesize, bS , 2, bqp[0], a, b, 1 );
829  filter_mb_mbaff_edgecv( h, img_cb + uvlinesize, 2*uvlinesize, bS+1, 2, bqp[1], a, b, 1 );
830  filter_mb_mbaff_edgecv( h, img_cr, 2*uvlinesize, bS , 2, rqp[0], a, b, 1 );
831  filter_mb_mbaff_edgecv( h, img_cr + uvlinesize, 2*uvlinesize, bS+1, 2, rqp[1], a, b, 1 );
832  }
833  }
834  }
835  }
836 
837 #if CONFIG_SMALL
838  {
839  int dir;
840  for (dir = 0; dir < 2; dir++)
841  filter_mb_dir(h, sl, mb_x, mb_y, img_y, img_cb, img_cr, linesize,
842  uvlinesize, mb_xy, mb_type, mvy_limit,
843  dir ? 0 : first_vertical_edge_done, a, b,
844  chroma, dir);
845  }
846 #else
847  filter_mb_dir(h, sl, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, first_vertical_edge_done, a, b, chroma, 0);
848  filter_mb_dir(h, sl, mb_x, mb_y, img_y, img_cb, img_cr, linesize, uvlinesize, mb_xy, mb_type, mvy_limit, 0, a, b, chroma, 1);
849 #endif
850 }
#define av_always_inline
Definition: attributes.h:45
uint8_t
#define av_assert2(cond)
assert() equivalent, that does lie in speed critical code.
Definition: avassert.h:64
Libavcodec external API header.
#define LEFT
Definition: cdgraphics.c:166
#define FFABS(a)
Absolute value, Note, INT_MIN / INT64_MIN result in undefined behavior as they are not representable ...
Definition: common.h:72
#define CONFIG_GRAY
Definition: config.h:556
#define AV_CODEC_FLAG_GRAY
Only decode/encode grayscale.
Definition: avcodec.h:308
#define CABAC(h)
Definition: h264_cabac.c:28
static av_always_inline void filter_mb_edgecv(uint8_t *pix, int stride, const int16_t bS[4], unsigned int qp, int a, int b, const H264Context *h, int intra)
static av_always_inline void filter_mb_dir(const H264Context *h, H264SliceContext *sl, int mb_x, int mb_y, uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr, unsigned int linesize, unsigned int uvlinesize, int mb_xy, int mb_type, int mvy_limit, int first_vertical_edge_done, int a, int b, int chroma, int dir)
static av_always_inline void h264_filter_mb_fast_internal(const H264Context *h, H264SliceContext *sl, int mb_x, int mb_y, uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr, unsigned int linesize, unsigned int uvlinesize, int pixel_shift)
static const uint8_t tc0_table[52 *3][4]
static av_always_inline void filter_mb_edgeh(uint8_t *pix, int stride, const int16_t bS[4], unsigned int qp, int a, int b, const H264Context *h, int intra)
static const uint8_t beta_table[52 *3]
static int check_mv(H264SliceContext *sl, long b_idx, long bn_idx, int mvy_limit)
void ff_h264_filter_mb(const H264Context *h, H264SliceContext *sl, int mb_x, int mb_y, uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr, unsigned int linesize, unsigned int uvlinesize)
static av_always_inline void filter_mb_mbaff_edgev(const H264Context *h, uint8_t *pix, int stride, const int16_t bS[7], int bsi, int qp, int a, int b, int intra)
static const uint8_t alpha_table[52 *3]
static av_always_inline void filter_mb_edgev(uint8_t *pix, int stride, const int16_t bS[4], unsigned int qp, int a, int b, const H264Context *h, int intra)
#define FILTER(hv, dir, edge, intra)
static av_always_inline void filter_mb_edgech(uint8_t *pix, int stride, const int16_t bS[4], unsigned int qp, int a, int b, const H264Context *h, int intra)
void ff_h264_filter_mb_fast(const H264Context *h, H264SliceContext *sl, int mb_x, int mb_y, uint8_t *img_y, uint8_t *img_cb, uint8_t *img_cr, unsigned int linesize, unsigned int uvlinesize)
static av_always_inline void filter_mb_mbaff_edgecv(const H264Context *h, uint8_t *pix, int stride, const int16_t bS[7], int bsi, int qp, int a, int b, int intra)
H.264 parameter set handling.
H.264 / AVC / MPEG-4 part10 codec.
static const uint8_t scan8[16 *3+3]
Definition: h264dec.h:651
#define CHROMA444(h)
Definition: h264dec.h:100
#define IS_8x8DCT(a)
Definition: h264dec.h:105
#define MB_FIELD(sl)
Definition: h264dec.h:73
#define FIELD_PICTURE(h)
Definition: h264dec.h:75
static av_always_inline int get_chroma_qp(const PPS *pps, int t, int qscale)
Get the chroma qp.
Definition: h264dec.h:688
#define CHROMA422(h)
Definition: h264dec.h:99
#define FRAME_MBAFF(h)
Definition: h264dec.h:74
#define LTOP
Definition: h264dec.h:77
#define IS_INTRA(x, y)
static const int16_t alpha[]
Definition: ilbcdata.h:55
int i
Definition: input.c:407
#define AV_ZERO64(d)
Definition: intreadwrite.h:633
#define AV_WN64A(p, v)
Definition: intreadwrite.h:542
#define ff_tlog(ctx,...)
Definition: internal.h:91
common internal API header
int stride
Definition: mace.c:144
#define LOCAL_ALIGNED(a, t, v,...)
Definition: mem_internal.h:113
#define MB_TYPE_8x16
Definition: mpegutils.h:56
#define IS_INTERLACED(a)
Definition: mpegutils.h:83
#define MB_TYPE_16x16
Definition: mpegutils.h:54
#define PICT_FRAME
Definition: mpegutils.h:39
useful rectangle filling function
#define tc
Definition: regdef.h:69
H264Context.
Definition: h264dec.h:344
unsigned int list_count
Definition: h264dec.h:275
int8_t ref_cache[2][5 *8]
Definition: h264dec.h:307
int left_type[LEFT_MBS]
Definition: h264dec.h:223
int16_t mv_cache[2][5 *8][2]
Motion vector cache.
Definition: h264dec.h:306
int left_mb_xy[LEFT_MBS]
Definition: h264dec.h:218
int chroma_qp[2]
Definition: h264dec.h:195
int slice_alpha_c0_offset
Definition: h264dec.h:201
uint8_t non_zero_count_cache[15 *8]
non zero coeff count cache.
Definition: h264dec.h:301
int slice_beta_offset
Definition: h264dec.h:202
const char * b
Definition: vf_curves.c:118
static const uint8_t offset[127][2]
Definition: vf_spp.c:107
static av_always_inline void chroma(WaveformContext *s, AVFrame *in, AVFrame *out, int component, int intensity, int offset_y, int offset_x, int column, int mirror, int jobnr, int nb_jobs)
Definition: vf_waveform.c:1624
@ CHROMA
Definition: vf_waveform.c:44