Line | Branch | Exec | Source |
---|---|---|---|
1 | /* | ||
2 | * simple math operations | ||
3 | * Copyright (c) 2001, 2002 Fabrice Bellard | ||
4 | * Copyright (c) 2006 Michael Niedermayer <michaelni@gmx.at> et al | ||
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 | #ifndef AVCODEC_MATHOPS_H | ||
23 | #define AVCODEC_MATHOPS_H | ||
24 | |||
25 | #include <stdint.h> | ||
26 | |||
27 | #include "libavutil/attributes_internal.h" | ||
28 | #include "libavutil/common.h" | ||
29 | #include "config.h" | ||
30 | |||
31 | #define MAX_NEG_CROP 1024 | ||
32 | |||
33 | extern const uint32_t ff_inverse[257]; | ||
34 | extern const uint8_t ff_log2_run[41]; | ||
35 | EXTERN const uint32_t ff_square_tab[512]; | ||
36 | extern const uint8_t ff_sqrt_tab[256]; | ||
37 | EXTERN const uint8_t ff_crop_tab[256 + 2 * MAX_NEG_CROP]; | ||
38 | extern const uint8_t ff_zigzag_direct[64]; | ||
39 | extern const uint8_t ff_zigzag_scan[16+1]; | ||
40 | |||
41 | #if ARCH_ARM | ||
42 | # include "arm/mathops.h" | ||
43 | #elif ARCH_MIPS | ||
44 | # include "mips/mathops.h" | ||
45 | #elif ARCH_PPC | ||
46 | # include "ppc/mathops.h" | ||
47 | #elif ARCH_X86 | ||
48 | # include "x86/mathops.h" | ||
49 | #endif | ||
50 | |||
51 | /* generic implementation */ | ||
52 | |||
53 | #ifndef MUL64 | ||
54 | # define MUL64(a,b) ((int64_t)(a) * (int64_t)(b)) | ||
55 | #endif | ||
56 | |||
57 | #ifndef MULL | ||
58 | # define MULL(a,b,s) (MUL64(a, b) >> (s)) | ||
59 | #endif | ||
60 | |||
61 | #ifndef MULH | ||
62 | 47930349 | static av_always_inline int MULH(int a, int b){ | |
63 | 47930349 | return MUL64(a, b) >> 32; | |
64 | } | ||
65 | #endif | ||
66 | |||
67 | #ifndef UMULH | ||
68 | 499 | static av_always_inline unsigned UMULH(unsigned a, unsigned b){ | |
69 | 499 | return ((uint64_t)(a) * (uint64_t)(b))>>32; | |
70 | } | ||
71 | #endif | ||
72 | |||
73 | #ifndef MAC64 | ||
74 | # define MAC64(d, a, b) ((d) += MUL64(a, b)) | ||
75 | #endif | ||
76 | |||
77 | #ifndef MLS64 | ||
78 | # define MLS64(d, a, b) ((d) -= MUL64(a, b)) | ||
79 | #endif | ||
80 | |||
81 | /* signed 16x16 -> 32 multiply add accumulate */ | ||
82 | #ifndef MAC16 | ||
83 | # define MAC16(rt, ra, rb) rt += (ra) * (rb) | ||
84 | #endif | ||
85 | |||
86 | /* signed 16x16 -> 32 multiply */ | ||
87 | #ifndef MUL16 | ||
88 | # define MUL16(ra, rb) ((ra) * (rb)) | ||
89 | #endif | ||
90 | |||
91 | #ifndef MLS16 | ||
92 | # define MLS16(rt, ra, rb) ((rt) -= (ra) * (rb)) | ||
93 | #endif | ||
94 | |||
95 | /* median of 3 */ | ||
96 | #ifndef mid_pred | ||
97 | #define mid_pred mid_pred | ||
98 | static inline av_const int mid_pred(int a, int b, int c) | ||
99 | { | ||
100 | if(a>b){ | ||
101 | if(c>b){ | ||
102 | if(c>a) b=a; | ||
103 | else b=c; | ||
104 | } | ||
105 | }else{ | ||
106 | if(b>c){ | ||
107 | if(c>a) b=c; | ||
108 | else b=a; | ||
109 | } | ||
110 | } | ||
111 | return b; | ||
112 | } | ||
113 | #endif | ||
114 | |||
115 | #ifndef median4 | ||
116 | #define median4 median4 | ||
117 | 74348 | static inline av_const int median4(int a, int b, int c, int d) | |
118 | { | ||
119 |
2/2✓ Branch 0 taken 10403 times.
✓ Branch 1 taken 63945 times.
|
74348 | if (a < b) { |
120 |
2/2✓ Branch 0 taken 2971 times.
✓ Branch 1 taken 7432 times.
|
10403 | if (c < d) return (FFMIN(b, d) + FFMAX(a, c)) / 2; |
121 | 7432 | else return (FFMIN(b, c) + FFMAX(a, d)) / 2; | |
122 | } else { | ||
123 |
2/2✓ Branch 0 taken 5325 times.
✓ Branch 1 taken 58620 times.
|
63945 | if (c < d) return (FFMIN(a, d) + FFMAX(b, c)) / 2; |
124 | 58620 | else return (FFMIN(a, c) + FFMAX(b, d)) / 2; | |
125 | } | ||
126 | } | ||
127 | #endif | ||
128 | |||
129 | #define FF_SIGNBIT(x) ((x) >> CHAR_BIT * sizeof(x) - 1) | ||
130 | |||
131 | #ifndef sign_extend | ||
132 | 416906076 | static inline av_const int sign_extend(int val, unsigned bits) | |
133 | { | ||
134 | 416906076 | unsigned shift = 8 * sizeof(int) - bits; | |
135 | 416906076 | union { unsigned u; int s; } v = { (unsigned) val << shift }; | |
136 | 416906076 | return v.s >> shift; | |
137 | } | ||
138 | #endif | ||
139 | |||
140 | #ifndef sign_extend64 | ||
141 | 5648 | static inline av_const int64_t sign_extend64(int64_t val, unsigned bits) | |
142 | { | ||
143 | 5648 | unsigned shift = 8 * sizeof(int64_t) - bits; | |
144 | 5648 | union { uint64_t u; int64_t s; } v = { (uint64_t) val << shift }; | |
145 | 5648 | return v.s >> shift; | |
146 | } | ||
147 | #endif | ||
148 | |||
149 | #ifndef zero_extend | ||
150 | 80323576 | static inline av_const unsigned zero_extend(unsigned val, unsigned bits) | |
151 | { | ||
152 | 80323576 | return (val << ((8 * sizeof(int)) - bits)) >> ((8 * sizeof(int)) - bits); | |
153 | } | ||
154 | #endif | ||
155 | |||
156 | #ifndef COPY3_IF_LT | ||
157 | #define COPY3_IF_LT(x, y, a, b, c, d)\ | ||
158 | if ((y) < (x)) {\ | ||
159 | (x) = (y);\ | ||
160 | (a) = (b);\ | ||
161 | (c) = (d);\ | ||
162 | } | ||
163 | #endif | ||
164 | |||
165 | #ifndef MASK_ABS | ||
166 | #define MASK_ABS(mask, level) do { \ | ||
167 | mask = level >> 31; \ | ||
168 | level = (level ^ mask) - mask; \ | ||
169 | } while (0) | ||
170 | #endif | ||
171 | |||
172 | #ifndef NEG_SSR32 | ||
173 | # define NEG_SSR32(a,s) ((( int32_t)(a))>>(32-(s))) | ||
174 | #endif | ||
175 | |||
176 | #ifndef NEG_USR32 | ||
177 | # define NEG_USR32(a,s) (((uint32_t)(a))>>(32-(s))) | ||
178 | #endif | ||
179 | |||
180 | #if HAVE_BIGENDIAN | ||
181 | # ifndef PACK_2U8 | ||
182 | # define PACK_2U8(a,b) (((a) << 8) | (b)) | ||
183 | # endif | ||
184 | # ifndef PACK_4U8 | ||
185 | # define PACK_4U8(a,b,c,d) (((a) << 24) | ((b) << 16) | ((c) << 8) | (d)) | ||
186 | # endif | ||
187 | # ifndef PACK_2U16 | ||
188 | # define PACK_2U16(a,b) (((a) << 16) | (b)) | ||
189 | # endif | ||
190 | #else | ||
191 | # ifndef PACK_2U8 | ||
192 | # define PACK_2U8(a,b) (((b) << 8) | (a)) | ||
193 | # endif | ||
194 | # ifndef PACK_4U2 | ||
195 | # define PACK_4U8(a,b,c,d) (((d) << 24) | ((c) << 16) | ((b) << 8) | (a)) | ||
196 | # endif | ||
197 | # ifndef PACK_2U16 | ||
198 | # define PACK_2U16(a,b) (((b) << 16) | (a)) | ||
199 | # endif | ||
200 | #endif | ||
201 | |||
202 | #ifndef PACK_2S8 | ||
203 | # define PACK_2S8(a,b) PACK_2U8((a)&255, (b)&255) | ||
204 | #endif | ||
205 | #ifndef PACK_4S8 | ||
206 | # define PACK_4S8(a,b,c,d) PACK_4U8((a)&255, (b)&255, (c)&255, (d)&255) | ||
207 | #endif | ||
208 | #ifndef PACK_2S16 | ||
209 | # define PACK_2S16(a,b) PACK_2U16((a)&0xffff, (b)&0xffff) | ||
210 | #endif | ||
211 | |||
212 | #ifndef FASTDIV | ||
213 | # define FASTDIV(a,b) ((uint32_t)((((uint64_t)a) * ff_inverse[b]) >> 32)) | ||
214 | #endif /* FASTDIV */ | ||
215 | |||
216 | #ifndef ff_sqrt | ||
217 | #define ff_sqrt ff_sqrt | ||
218 | 7071544 | static inline av_const unsigned int ff_sqrt(unsigned int a) | |
219 | { | ||
220 | unsigned int b; | ||
221 | |||
222 |
2/2✓ Branch 0 taken 596832 times.
✓ Branch 1 taken 6474712 times.
|
7071544 | if (a < 255) return (ff_sqrt_tab[a + 1] - 1) >> 4; |
223 |
2/2✓ Branch 0 taken 848799 times.
✓ Branch 1 taken 5625913 times.
|
6474712 | else if (a < (1 << 12)) b = ff_sqrt_tab[a >> 4] >> 2; |
224 | #if !CONFIG_SMALL | ||
225 |
2/2✓ Branch 0 taken 1055793 times.
✓ Branch 1 taken 4570120 times.
|
5625913 | else if (a < (1 << 14)) b = ff_sqrt_tab[a >> 6] >> 1; |
226 |
2/2✓ Branch 0 taken 1053436 times.
✓ Branch 1 taken 3516684 times.
|
4570120 | else if (a < (1 << 16)) b = ff_sqrt_tab[a >> 8] ; |
227 | #endif | ||
228 | else { | ||
229 | 3516684 | int s = av_log2_16bit(a >> 16) >> 1; | |
230 | 3516684 | unsigned int c = a >> (s + 2); | |
231 | 3516684 | b = ff_sqrt_tab[c >> (s + 8)]; | |
232 | 3516684 | b = FASTDIV(c,b) + (b << s); | |
233 | } | ||
234 | |||
235 | 6474712 | return b - (a < b * b); | |
236 | } | ||
237 | #endif | ||
238 | |||
239 | 30243 | static inline av_const float ff_sqrf(float a) | |
240 | { | ||
241 | 30243 | return a*a; | |
242 | } | ||
243 | |||
244 | 393216 | static inline int8_t ff_u8_to_s8(uint8_t a) | |
245 | { | ||
246 | union { | ||
247 | uint8_t u8; | ||
248 | int8_t s8; | ||
249 | } b; | ||
250 | 393216 | b.u8 = a; | |
251 | 393216 | return b.s8; | |
252 | } | ||
253 | |||
254 | #endif /* AVCODEC_MATHOPS_H */ | ||
255 |