FFmpeg coverage

Directory: ../../../ffmpeg/
File: src/libavcodec/x86/dirac_dwt_init.c
Date: 2026-01-23 19:11:46
Exec Total Coverage
Lines: 4 42 9.5%
Functions: 1 9 11.1%
Branches: 1 17 5.9%
Line Branch Exec Source
1 /*
2 * x86 optimized discrete wavelet transform
3 * Copyright (c) 2002-2004 Michael Niedermayer <michaelni@gmx.at>
4 * Copyright (c) 2010 David Conrad
5 *
6 * This file is part of FFmpeg.
7 *
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
12 *
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
17 *
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
21 */
22
23 #include "libavutil/x86/cpu.h"
24 #include "libavcodec/dirac_dwt.h"
25
26 #define COMPOSE_VERTICAL(ext, align) \
27 void ff_vertical_compose53iL0##ext(int16_t *b0, int16_t *b1, int16_t *b2, int width); \
28 void ff_vertical_compose_dirac53iH0##ext(int16_t *b0, int16_t *b1, int16_t *b2, int width); \
29 void ff_vertical_compose_dd137iL0##ext(int16_t *b0, int16_t *b1, int16_t *b2, int16_t *b3, int16_t *b4, int width); \
30 void ff_vertical_compose_dd97iH0##ext(int16_t *b0, int16_t *b1, int16_t *b2, int16_t *b3, int16_t *b4, int width); \
31 void ff_vertical_compose_haar##ext(int16_t *b0, int16_t *b1, int width); \
32 void ff_horizontal_compose_haar0i##ext(int16_t *b, int16_t *tmp, int w);\
33 void ff_horizontal_compose_haar1i##ext(int16_t *b, int16_t *tmp, int w);\
34 \
35 static void vertical_compose53iL0##ext(uint8_t *_b0, uint8_t *_b1, uint8_t *_b2, int width) \
36 { \
37 int i, width_align = width&~(align-1); \
38 int16_t *b0 = (int16_t *)_b0; \
39 int16_t *b1 = (int16_t *)_b1; \
40 int16_t *b2 = (int16_t *)_b2; \
41 \
42 for(i=width_align; i<width; i++) \
43 b1[i] = COMPOSE_53iL0(b0[i], b1[i], b2[i]); \
44 \
45 ff_vertical_compose53iL0##ext(b0, b1, b2, width_align); \
46 } \
47 \
48 static void vertical_compose_dirac53iH0##ext(uint8_t *_b0, uint8_t *_b1, uint8_t *_b2, int width) \
49 { \
50 int i, width_align = width&~(align-1); \
51 int16_t *b0 = (int16_t *)_b0; \
52 int16_t *b1 = (int16_t *)_b1; \
53 int16_t *b2 = (int16_t *)_b2; \
54 \
55 for(i=width_align; i<width; i++) \
56 b1[i] = COMPOSE_DIRAC53iH0(b0[i], b1[i], b2[i]); \
57 \
58 ff_vertical_compose_dirac53iH0##ext(b0, b1, b2, width_align); \
59 } \
60 \
61 static void vertical_compose_dd137iL0##ext(uint8_t *_b0, uint8_t *_b1, uint8_t *_b2, \
62 uint8_t *_b3, uint8_t *_b4, int width) \
63 { \
64 int i, width_align = width&~(align-1); \
65 int16_t *b0 = (int16_t *)_b0; \
66 int16_t *b1 = (int16_t *)_b1; \
67 int16_t *b2 = (int16_t *)_b2; \
68 int16_t *b3 = (int16_t *)_b3; \
69 int16_t *b4 = (int16_t *)_b4; \
70 \
71 for(i=width_align; i<width; i++) \
72 b2[i] = COMPOSE_DD137iL0(b0[i], b1[i], b2[i], b3[i], b4[i]); \
73 \
74 ff_vertical_compose_dd137iL0##ext(b0, b1, b2, b3, b4, width_align); \
75 } \
76 \
77 static void vertical_compose_dd97iH0##ext(uint8_t *_b0, uint8_t *_b1, uint8_t *_b2, \
78 uint8_t *_b3, uint8_t *_b4, int width) \
79 { \
80 int i, width_align = width&~(align-1); \
81 int16_t *b0 = (int16_t *)_b0; \
82 int16_t *b1 = (int16_t *)_b1; \
83 int16_t *b2 = (int16_t *)_b2; \
84 int16_t *b3 = (int16_t *)_b3; \
85 int16_t *b4 = (int16_t *)_b4; \
86 \
87 for(i=width_align; i<width; i++) \
88 b2[i] = COMPOSE_DD97iH0(b0[i], b1[i], b2[i], b3[i], b4[i]); \
89 \
90 ff_vertical_compose_dd97iH0##ext(b0, b1, b2, b3, b4, width_align); \
91 } \
92 static void vertical_compose_haar##ext(uint8_t *_b0, uint8_t *_b1, int width) \
93 { \
94 int i, width_align = width&~(align-1); \
95 int16_t *b0 = (int16_t *)_b0; \
96 int16_t *b1 = (int16_t *)_b1; \
97 \
98 for(i=width_align; i<width; i++) { \
99 b0[i] = COMPOSE_HAARiL0(b0[i], b1[i]); \
100 b1[i] = COMPOSE_HAARiH0(b1[i], b0[i]); \
101 } \
102 \
103 ff_vertical_compose_haar##ext(b0, b1, width_align); \
104 } \
105 static void horizontal_compose_haar0i##ext(uint8_t *_b, uint8_t *_tmp, int w)\
106 {\
107 int w2= w>>1;\
108 int x= w2 - (w2&(align-1));\
109 int16_t *b = (int16_t *)_b; \
110 int16_t *tmp = (int16_t *)_tmp; \
111 \
112 ff_horizontal_compose_haar0i##ext(b, tmp, w);\
113 \
114 for (; x < w2; x++) {\
115 b[2*x ] = tmp[x];\
116 b[2*x+1] = COMPOSE_HAARiH0(b[x+w2], tmp[x]);\
117 }\
118 }\
119 static void horizontal_compose_haar1i##ext(uint8_t *_b, uint8_t *_tmp, int w)\
120 {\
121 int w2= w>>1;\
122 int x= w2 - (w2&(align-1));\
123 int16_t *b = (int16_t *)_b; \
124 int16_t *tmp = (int16_t *)_tmp; \
125 \
126 ff_horizontal_compose_haar1i##ext(b, tmp, w);\
127 \
128 for (; x < w2; x++) {\
129 b[2*x ] = (tmp[x] + 1)>>1;\
130 b[2*x+1] = (COMPOSE_HAARiH0(b[x+w2], tmp[x]) + 1)>>1;\
131 }\
132 }\
133 \
134
135 ✗ COMPOSE_VERTICAL(_sse2, 8)
136
137 void ff_horizontal_compose_dd97i_ssse3(int16_t *_b, int16_t *_tmp, int w);
138
139 ✗ static void horizontal_compose_dd97i_ssse3(uint8_t *_b, uint8_t *_tmp, int w)
140 {
141 ✗ int w2= w>>1;
142 ✗ int x= w2 - (w2&7);
143 ✗ int16_t *b = (int16_t *)_b;
144 ✗ int16_t *tmp = (int16_t *)_tmp;
145
146 ✗ ff_horizontal_compose_dd97i_ssse3(b, tmp, w);
147
148 ✗ for (; x < w2; x++) {
149 ✗ b[2*x ] = (tmp[x] + 1)>>1;
150 ✗ b[2*x+1] = (COMPOSE_DD97iH0(tmp[x-1], tmp[x], b[x+w2], tmp[x+1], tmp[x+2]) + 1)>>1;
151 }
152 ✗ }
153
154 348 void ff_spatial_idwt_init_x86(DWTContext *d, enum dwt_type type)
155 {
156 348 int mm_flags = av_get_cpu_flags();
157
158
1/2
✓ Branch 0 taken 348 times.
✗ Branch 1 not taken.
 348 if (!(mm_flags & AV_CPU_FLAG_SSE2))
159 348 return;
160
161 ✗ switch (type) {
162 ✗ case DWT_DIRAC_DD9_7:
163 ✗ d->vertical_compose_l0.tap3 = vertical_compose53iL0_sse2;
164 ✗ d->vertical_compose_h0.tap5 = vertical_compose_dd97iH0_sse2;
165 ✗ break;
166 ✗ case DWT_DIRAC_LEGALL5_3:
167 ✗ d->vertical_compose_l0.tap3 = vertical_compose53iL0_sse2;
168 ✗ d->vertical_compose_h0.tap3 = vertical_compose_dirac53iH0_sse2;
169 ✗ break;
170 ✗ case DWT_DIRAC_DD13_7:
171 ✗ d->vertical_compose_l0.tap5 = vertical_compose_dd137iL0_sse2;
172 ✗ d->vertical_compose_h0.tap5 = vertical_compose_dd97iH0_sse2;
173 ✗ break;
174 ✗ case DWT_DIRAC_HAAR0:
175 ✗ d->vertical_compose = vertical_compose_haar_sse2;
176 ✗ d->horizontal_compose = horizontal_compose_haar0i_sse2;
177 ✗ break;
178 ✗ case DWT_DIRAC_HAAR1:
179 ✗ d->vertical_compose = vertical_compose_haar_sse2;
180 ✗ d->horizontal_compose = horizontal_compose_haar1i_sse2;
181 ✗ break;
182 }
183
184 ✗ if (!(mm_flags & AV_CPU_FLAG_SSSE3))
185 ✗ return;
186
187 ✗ switch (type) {
188 ✗ case DWT_DIRAC_DD9_7:
189 ✗ d->horizontal_compose = horizontal_compose_dd97i_ssse3;
190 ✗ break;
191 }
192 }
193