GCC Code Coverage Report
Directory: ../../../ffmpeg/ Exec Total Coverage
File: src/libavcodec/h264dsp_template.c Lines: 162 162 100.0 %
Date: 2019-11-22 03:34:36 Branches: 70 70 100.0 %

Line Branch Exec Source
1
/*
2
 * H.26L/H.264/AVC/JVT/14496-10/... encoder/decoder
3
 * Copyright (c) 2003-2011 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 DSP functions.
25
 * @author Michael Niedermayer <michaelni@gmx.at>
26
 */
27
28
#include "bit_depth_template.c"
29
30
#define op_scale1(x)  block[x] = av_clip_pixel( (block[x]*weight + offset) >> log2_denom )
31
#define op_scale2(x)  dst[x] = av_clip_pixel( (src[x]*weights + dst[x]*weightd + offset) >> (log2_denom+1))
32
#define H264_WEIGHT(W) \
33
static void FUNCC(weight_h264_pixels ## W)(uint8_t *_block, ptrdiff_t stride, int height, \
34
                                           int log2_denom, int weight, int offset) \
35
{ \
36
    int y; \
37
    pixel *block = (pixel*)_block; \
38
    stride >>= sizeof(pixel)-1; \
39
    offset = (unsigned)offset << (log2_denom + (BIT_DEPTH-8)); \
40
    if(log2_denom) offset += 1<<(log2_denom-1); \
41
    for (y = 0; y < height; y++, block += stride) { \
42
        op_scale1(0); \
43
        op_scale1(1); \
44
        if(W==2) continue; \
45
        op_scale1(2); \
46
        op_scale1(3); \
47
        if(W==4) continue; \
48
        op_scale1(4); \
49
        op_scale1(5); \
50
        op_scale1(6); \
51
        op_scale1(7); \
52
        if(W==8) continue; \
53
        op_scale1(8); \
54
        op_scale1(9); \
55
        op_scale1(10); \
56
        op_scale1(11); \
57
        op_scale1(12); \
58
        op_scale1(13); \
59
        op_scale1(14); \
60
        op_scale1(15); \
61
    } \
62
} \
63
static void FUNCC(biweight_h264_pixels ## W)(uint8_t *_dst, uint8_t *_src, ptrdiff_t stride, int height, \
64
                                             int log2_denom, int weightd, int weights, int offset) \
65
{ \
66
    int y; \
67
    pixel *dst = (pixel*)_dst; \
68
    pixel *src = (pixel*)_src; \
69
    stride >>= sizeof(pixel)-1; \
70
    offset = (unsigned)offset << (BIT_DEPTH-8); \
71
    offset = (unsigned)((offset + 1) | 1) << log2_denom; \
72
    for (y = 0; y < height; y++, dst += stride, src += stride) { \
73
        op_scale2(0); \
74
        op_scale2(1); \
75
        if(W==2) continue; \
76
        op_scale2(2); \
77
        op_scale2(3); \
78
        if(W==4) continue; \
79
        op_scale2(4); \
80
        op_scale2(5); \
81
        op_scale2(6); \
82
        op_scale2(7); \
83
        if(W==8) continue; \
84
        op_scale2(8); \
85
        op_scale2(9); \
86
        op_scale2(10); \
87
        op_scale2(11); \
88
        op_scale2(12); \
89
        op_scale2(13); \
90
        op_scale2(14); \
91
        op_scale2(15); \
92
    } \
93
}
94
95

9871254
H264_WEIGHT(16)
96

3912460
H264_WEIGHT(8)
97

3763998
H264_WEIGHT(4)
98

1708896
H264_WEIGHT(2)
99
100
#undef op_scale1
101
#undef op_scale2
102
#undef H264_WEIGHT
103
104
30770990
static av_always_inline av_flatten void FUNCC(h264_loop_filter_luma)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta, int8_t *tc0)
105
{
106
30770990
    pixel *pix = (pixel*)p_pix;
107
    int i, d;
108
30770990
    xstride >>= sizeof(pixel)-1;
109
30770990
    ystride >>= sizeof(pixel)-1;
110
30770990
    alpha <<= BIT_DEPTH - 8;
111
30770990
    beta  <<= BIT_DEPTH - 8;
112
153854950
    for( i = 0; i < 4; i++ ) {
113
123083960
        const int tc_orig = tc0[i] * (1 << (BIT_DEPTH - 8));
114
123083960
        if( tc_orig < 0 ) {
115
23823054
            pix += inner_iters*ystride;
116
23823054
            continue;
117
        }
118
493904370
        for( d = 0; d < inner_iters; d++ ) {
119
394643464
            const int p0 = pix[-1*xstride];
120
394643464
            const int p1 = pix[-2*xstride];
121
394643464
            const int p2 = pix[-3*xstride];
122
394643464
            const int q0 = pix[0];
123
394643464
            const int q1 = pix[1*xstride];
124
394643464
            const int q2 = pix[2*xstride];
125
126
394643464
            if( FFABS( p0 - q0 ) < alpha &&
127
336526583
                FFABS( p1 - p0 ) < beta &&
128
253824429
                FFABS( q1 - q0 ) < beta ) {
129
130
214018079
                int tc = tc_orig;
131
                int i_delta;
132
133
214018079
                if( FFABS( p2 - p0 ) < beta ) {
134
183221832
                    if(tc_orig)
135
171143706
                    pix[-2*xstride] = p1 + av_clip( (( p2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - p1, -tc_orig, tc_orig );
136
183221832
                    tc++;
137
                }
138
214018079
                if( FFABS( q2 - q0 ) < beta ) {
139
182180689
                    if(tc_orig)
140
170118800
                    pix[   xstride] = q1 + av_clip( (( q2 + ( ( p0 + q0 + 1 ) >> 1 ) ) >> 1) - q1, -tc_orig, tc_orig );
141
182180689
                    tc++;
142
                }
143
144
214018079
                i_delta = av_clip( (((q0 - p0 ) * 4) + (p1 - q1) + 4) >> 3, -tc, tc );
145
214018079
                pix[-xstride] = av_clip_pixel( p0 + i_delta );    /* p0' */
146
214018079
                pix[0]        = av_clip_pixel( q0 - i_delta );    /* q0' */
147
            }
148
394643464
            pix += ystride;
149
        }
150
    }
151
30770990
}
152
15556154
static void FUNCC(h264_v_loop_filter_luma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
153
{
154
15556154
    FUNCC(h264_loop_filter_luma)(pix, stride, sizeof(pixel), 4, alpha, beta, tc0);
155
15556154
}
156
14914816
static void FUNCC(h264_h_loop_filter_luma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
157
{
158
14914816
    FUNCC(h264_loop_filter_luma)(pix, sizeof(pixel), stride, 4, alpha, beta, tc0);
159
14914816
}
160
300020
static void FUNCC(h264_h_loop_filter_luma_mbaff)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
161
{
162
300020
    FUNCC(h264_loop_filter_luma)(pix, sizeof(pixel), stride, 2, alpha, beta, tc0);
163
300020
}
164
165
2856848
static av_always_inline av_flatten void FUNCC(h264_loop_filter_luma_intra)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta)
166
{
167
2856848
    pixel *pix = (pixel*)p_pix;
168
    int d;
169
2856848
    xstride >>= sizeof(pixel)-1;
170
2856848
    ystride >>= sizeof(pixel)-1;
171
2856848
    alpha <<= BIT_DEPTH - 8;
172
2856848
    beta  <<= BIT_DEPTH - 8;
173
47346824
    for( d = 0; d < 4 * inner_iters; d++ ) {
174
44489976
        const int p2 = pix[-3*xstride];
175
44489976
        const int p1 = pix[-2*xstride];
176
44489976
        const int p0 = pix[-1*xstride];
177
178
44489976
        const int q0 = pix[ 0*xstride];
179
44489976
        const int q1 = pix[ 1*xstride];
180
44489976
        const int q2 = pix[ 2*xstride];
181
182
44489976
        if( FFABS( p0 - q0 ) < alpha &&
183
39032243
            FFABS( p1 - p0 ) < beta &&
184
33371982
            FFABS( q1 - q0 ) < beta ) {
185
186
30068495
            if(FFABS( p0 - q0 ) < (( alpha >> 2 ) + 2 )){
187
26815380
                if( FFABS( p2 - p0 ) < beta)
188
                {
189
24691239
                    const int p3 = pix[-4*xstride];
190
                    /* p0', p1', p2' */
191
24691239
                    pix[-1*xstride] = ( p2 + 2*p1 + 2*p0 + 2*q0 + q1 + 4 ) >> 3;
192
24691239
                    pix[-2*xstride] = ( p2 + p1 + p0 + q0 + 2 ) >> 2;
193
24691239
                    pix[-3*xstride] = ( 2*p3 + 3*p2 + p1 + p0 + q0 + 4 ) >> 3;
194
                } else {
195
                    /* p0' */
196
2124141
                    pix[-1*xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;
197
                }
198
26815380
                if( FFABS( q2 - q0 ) < beta)
199
                {
200
24603840
                    const int q3 = pix[3*xstride];
201
                    /* q0', q1', q2' */
202
24603840
                    pix[0*xstride] = ( p1 + 2*p0 + 2*q0 + 2*q1 + q2 + 4 ) >> 3;
203
24603840
                    pix[1*xstride] = ( p0 + q0 + q1 + q2 + 2 ) >> 2;
204
24603840
                    pix[2*xstride] = ( 2*q3 + 3*q2 + q1 + q0 + p0 + 4 ) >> 3;
205
                } else {
206
                    /* q0' */
207
2211540
                    pix[0*xstride] = ( 2*q1 + q0 + p1 + 2 ) >> 2;
208
                }
209
            }else{
210
                /* p0', q0' */
211
3253115
                pix[-1*xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;
212
3253115
                pix[ 0*xstride] = ( 2*q1 + q0 + p1 + 2 ) >> 2;
213
            }
214
        }
215
44489976
        pix += ystride;
216
    }
217
2856848
}
218
1088902
static void FUNCC(h264_v_loop_filter_luma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
219
{
220
1088902
    FUNCC(h264_loop_filter_luma_intra)(pix, stride, sizeof(pixel), 4, alpha, beta);
221
1088902
}
222
1615497
static void FUNCC(h264_h_loop_filter_luma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
223
{
224
1615497
    FUNCC(h264_loop_filter_luma_intra)(pix, sizeof(pixel), stride, 4, alpha, beta);
225
1615497
}
226
152449
static void FUNCC(h264_h_loop_filter_luma_mbaff_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
227
{
228
152449
    FUNCC(h264_loop_filter_luma_intra)(pix, sizeof(pixel), stride, 2, alpha, beta);
229
152449
}
230
231
34407435
static av_always_inline av_flatten void FUNCC(h264_loop_filter_chroma)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta, int8_t *tc0)
232
{
233
34407435
    pixel *pix = (pixel*)p_pix;
234
    int i, d;
235
34407435
    alpha <<= BIT_DEPTH - 8;
236
34407435
    beta  <<= BIT_DEPTH - 8;
237
34407435
    xstride >>= sizeof(pixel)-1;
238
34407435
    ystride >>= sizeof(pixel)-1;
239
172037175
    for( i = 0; i < 4; i++ ) {
240
137629740
        const int tc = ((tc0[i] - 1U) << (BIT_DEPTH - 8)) + 1;
241
137629740
        if( tc <= 0 ) {
242
22350203
            pix += inner_iters*ystride;
243
22350203
            continue;
244
        }
245
343505007
        for( d = 0; d < inner_iters; d++ ) {
246
228225470
            const int p0 = pix[-1*xstride];
247
228225470
            const int p1 = pix[-2*xstride];
248
228225470
            const int q0 = pix[0];
249
228225470
            const int q1 = pix[1*xstride];
250
251
228225470
            if( FFABS( p0 - q0 ) < alpha &&
252
223188902
                FFABS( p1 - p0 ) < beta &&
253
210420583
                FFABS( q1 - q0 ) < beta ) {
254
255
201962775
                int delta = av_clip( ((q0 - p0) * 4 + (p1 - q1) + 4) >> 3, -tc, tc );
256
257
201962775
                pix[-xstride] = av_clip_pixel( p0 + delta );    /* p0' */
258
201962775
                pix[0]        = av_clip_pixel( q0 - delta );    /* q0' */
259
            }
260
228225470
            pix += ystride;
261
        }
262
    }
263
34407435
}
264
17556720
static void FUNCC(h264_v_loop_filter_chroma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
265
{
266
17556720
    FUNCC(h264_loop_filter_chroma)(pix, stride, sizeof(pixel), 2, alpha, beta, tc0);
267
17556720
}
268
16243177
static void FUNCC(h264_h_loop_filter_chroma)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
269
{
270
16243177
    FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 2, alpha, beta, tc0);
271
16243177
}
272
599438
static void FUNCC(h264_h_loop_filter_chroma_mbaff)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
273
{
274
599438
    FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 1, alpha, beta, tc0);
275
599438
}
276
8028
static void FUNCC(h264_h_loop_filter_chroma422)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
277
{
278
8028
    FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 4, alpha, beta, tc0);
279
8028
}
280
72
static void FUNCC(h264_h_loop_filter_chroma422_mbaff)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta, int8_t *tc0)
281
{
282
72
    FUNCC(h264_loop_filter_chroma)(pix, sizeof(pixel), stride, 2, alpha, beta, tc0);
283
72
}
284
285
5513143
static av_always_inline av_flatten void FUNCC(h264_loop_filter_chroma_intra)(uint8_t *p_pix, ptrdiff_t xstride, ptrdiff_t ystride, int inner_iters, int alpha, int beta)
286
{
287
5513143
    pixel *pix = (pixel*)p_pix;
288
    int d;
289
5513143
    xstride >>= sizeof(pixel)-1;
290
5513143
    ystride >>= sizeof(pixel)-1;
291
5513143
    alpha <<= BIT_DEPTH - 8;
292
5513143
    beta  <<= BIT_DEPTH - 8;
293
48460863
    for( d = 0; d < 4 * inner_iters; d++ ) {
294
42947720
        const int p0 = pix[-1*xstride];
295
42947720
        const int p1 = pix[-2*xstride];
296
42947720
        const int q0 = pix[0];
297
42947720
        const int q1 = pix[1*xstride];
298
299
42947720
        if( FFABS( p0 - q0 ) < alpha &&
300
40946920
            FFABS( p1 - p0 ) < beta &&
301
39388845
            FFABS( q1 - q0 ) < beta ) {
302
303
38302635
            pix[-xstride] = ( 2*p1 + p0 + q1 + 2 ) >> 2;   /* p0' */
304
38302635
            pix[0]        = ( 2*q1 + q0 + p1 + 2 ) >> 2;   /* q0' */
305
        }
306
42947720
        pix += ystride;
307
    }
308
5513143
}
309
2077810
static void FUNCC(h264_v_loop_filter_chroma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
310
{
311
2077810
    FUNCC(h264_loop_filter_chroma_intra)(pix, stride, sizeof(pixel), 2, alpha, beta);
312
2077810
}
313
3122901
static void FUNCC(h264_h_loop_filter_chroma_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
314
{
315
3122901
    FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 2, alpha, beta);
316
3122901
}
317
304692
static void FUNCC(h264_h_loop_filter_chroma_mbaff_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
318
{
319
304692
    FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 1, alpha, beta);
320
304692
}
321
7668
static void FUNCC(h264_h_loop_filter_chroma422_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
322
{
323
7668
    FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 4, alpha, beta);
324
7668
}
325
72
static void FUNCC(h264_h_loop_filter_chroma422_mbaff_intra)(uint8_t *pix, ptrdiff_t stride, int alpha, int beta)
326
{
327
72
    FUNCC(h264_loop_filter_chroma_intra)(pix, sizeof(pixel), stride, 2, alpha, beta);
328
72
}