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/common.h"
28 #include "libavutil/reverse.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_sqrt_tab[256];
35 extern const uint8_t ff_crop_tab[256 + 2 * MAX_NEG_CROP];
36 extern const uint8_t ff_zigzag_direct[64];
37 extern const uint8_t ff_zigzag_scan[16+1];
38
39 #if ARCH_ARM
40 # include "arm/mathops.h"
41 #elif ARCH_AVR32
42 # include "avr32/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
MULH(int a,int b)62 static av_always_inline int MULH(int a, int b){
63 return MUL64(a, b) >> 32;
64 }
65 #endif
66
67 #ifndef UMULH
UMULH(unsigned a,unsigned b)68 static av_always_inline unsigned UMULH(unsigned a, unsigned b){
69 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
mid_pred(int a,int b,int c)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
median4(int a,int b,int c,int d)117 static inline av_const int median4(int a, int b, int c, int d)
118 {
119 if (a < b) {
120 if (c < d) return (FFMIN(b, d) + FFMAX(a, c)) / 2;
121 else return (FFMIN(b, c) + FFMAX(a, d)) / 2;
122 } else {
123 if (c < d) return (FFMIN(a, d) + FFMAX(b, c)) / 2;
124 else return (FFMIN(a, c) + FFMAX(b, d)) / 2;
125 }
126 }
127 #endif
128
129 #ifndef sign_extend
sign_extend(int val,unsigned bits)130 static inline av_const int sign_extend(int val, unsigned bits)
131 {
132 unsigned shift = 8 * sizeof(int) - bits;
133 union { unsigned u; int s; } v = { (unsigned) val << shift };
134 return v.s >> shift;
135 }
136 #endif
137
138 #ifndef zero_extend
zero_extend(unsigned val,unsigned bits)139 static inline av_const unsigned zero_extend(unsigned val, unsigned bits)
140 {
141 return (val << ((8 * sizeof(int)) - bits)) >> ((8 * sizeof(int)) - bits);
142 }
143 #endif
144
145 #ifndef COPY3_IF_LT
146 #define COPY3_IF_LT(x, y, a, b, c, d)\
147 if ((y) < (x)) {\
148 (x) = (y);\
149 (a) = (b);\
150 (c) = (d);\
151 }
152 #endif
153
154 #ifndef MASK_ABS
155 #define MASK_ABS(mask, level) do { \
156 mask = level >> 31; \
157 level = (level ^ mask) - mask; \
158 } while (0)
159 #endif
160
161 #ifndef NEG_SSR32
162 # define NEG_SSR32(a,s) ((( int32_t)(a))>>(32-(s)))
163 #endif
164
165 #ifndef NEG_USR32
166 # define NEG_USR32(a,s) (((uint32_t)(a))>>(32-(s)))
167 #endif
168
169 #if HAVE_BIGENDIAN
170 # ifndef PACK_2U8
171 # define PACK_2U8(a,b) (((a) << 8) | (b))
172 # endif
173 # ifndef PACK_4U8
174 # define PACK_4U8(a,b,c,d) (((a) << 24) | ((b) << 16) | ((c) << 8) | (d))
175 # endif
176 # ifndef PACK_2U16
177 # define PACK_2U16(a,b) (((a) << 16) | (b))
178 # endif
179 #else
180 # ifndef PACK_2U8
181 # define PACK_2U8(a,b) (((b) << 8) | (a))
182 # endif
183 # ifndef PACK_4U2
184 # define PACK_4U8(a,b,c,d) (((d) << 24) | ((c) << 16) | ((b) << 8) | (a))
185 # endif
186 # ifndef PACK_2U16
187 # define PACK_2U16(a,b) (((b) << 16) | (a))
188 # endif
189 #endif
190
191 #ifndef PACK_2S8
192 # define PACK_2S8(a,b) PACK_2U8((a)&255, (b)&255)
193 #endif
194 #ifndef PACK_4S8
195 # define PACK_4S8(a,b,c,d) PACK_4U8((a)&255, (b)&255, (c)&255, (d)&255)
196 #endif
197 #ifndef PACK_2S16
198 # define PACK_2S16(a,b) PACK_2U16((a)&0xffff, (b)&0xffff)
199 #endif
200
201 #ifndef FASTDIV
202 # define FASTDIV(a,b) ((uint32_t)((((uint64_t)a) * ff_inverse[b]) >> 32))
203 #endif /* FASTDIV */
204
205 #ifndef ff_sqrt
206 #define ff_sqrt ff_sqrt
ff_sqrt(unsigned int a)207 static inline av_const unsigned int ff_sqrt(unsigned int a)
208 {
209 unsigned int b;
210
211 if (a < 255) return (ff_sqrt_tab[a + 1] - 1) >> 4;
212 else if (a < (1 << 12)) b = ff_sqrt_tab[a >> 4] >> 2;
213 #if !CONFIG_SMALL
214 else if (a < (1 << 14)) b = ff_sqrt_tab[a >> 6] >> 1;
215 else if (a < (1 << 16)) b = ff_sqrt_tab[a >> 8] ;
216 #endif
217 else {
218 int s = av_log2_16bit(a >> 16) >> 1;
219 unsigned int c = a >> (s + 2);
220 b = ff_sqrt_tab[c >> (s + 8)];
221 b = FASTDIV(c,b) + (b << s);
222 }
223
224 return b - (a < b * b);
225 }
226 #endif
227
ff_sqrf(float a)228 static inline av_const float ff_sqrf(float a)
229 {
230 return a*a;
231 }
232
ff_u8_to_s8(uint8_t a)233 static inline int8_t ff_u8_to_s8(uint8_t a)
234 {
235 union {
236 uint8_t u8;
237 int8_t s8;
238 } b;
239 b.u8 = a;
240 return b.s8;
241 }
242
bitswap_32(uint32_t x)243 static av_always_inline uint32_t bitswap_32(uint32_t x)
244 {
245 return (uint32_t)ff_reverse[ x & 0xFF] << 24 |
246 (uint32_t)ff_reverse[(x >> 8) & 0xFF] << 16 |
247 (uint32_t)ff_reverse[(x >> 16) & 0xFF] << 8 |
248 (uint32_t)ff_reverse[ x >> 24];
249 }
250
251 #endif /* AVCODEC_MATHOPS_H */
252