| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | /* | ||
| 2 | * Copyright (c) 2012 | ||
| 3 | * MIPS Technologies, Inc., California. | ||
| 4 | * | ||
| 5 | * Redistribution and use in source and binary forms, with or without | ||
| 6 | * modification, are permitted provided that the following conditions | ||
| 7 | * are met: | ||
| 8 | * 1. Redistributions of source code must retain the above copyright | ||
| 9 | * notice, this list of conditions and the following disclaimer. | ||
| 10 | * 2. Redistributions in binary form must reproduce the above copyright | ||
| 11 | * notice, this list of conditions and the following disclaimer in the | ||
| 12 | * documentation and/or other materials provided with the distribution. | ||
| 13 | * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its | ||
| 14 | * contributors may be used to endorse or promote products derived from | ||
| 15 | * this software without specific prior written permission. | ||
| 16 | * | ||
| 17 | * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND | ||
| 18 | * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE | ||
| 19 | * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE | ||
| 20 | * ARE DISCLAIMED. IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE | ||
| 21 | * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL | ||
| 22 | * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS | ||
| 23 | * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) | ||
| 24 | * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT | ||
| 25 | * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY | ||
| 26 | * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF | ||
| 27 | * SUCH DAMAGE. | ||
| 28 | * | ||
| 29 | * Author: Nedeljko Babic (nbabic@mips.com) | ||
| 30 | * | ||
| 31 | * This file is part of FFmpeg. | ||
| 32 | * | ||
| 33 | * FFmpeg is free software; you can redistribute it and/or | ||
| 34 | * modify it under the terms of the GNU Lesser General Public | ||
| 35 | * License as published by the Free Software Foundation; either | ||
| 36 | * version 2.1 of the License, or (at your option) any later version. | ||
| 37 | * | ||
| 38 | * FFmpeg is distributed in the hope that it will be useful, | ||
| 39 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 40 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
| 41 | * Lesser General Public License for more details. | ||
| 42 | * | ||
| 43 | * You should have received a copy of the GNU Lesser General Public | ||
| 44 | * License along with FFmpeg; if not, write to the Free Software | ||
| 45 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||
| 46 | */ | ||
| 47 | |||
| 48 | #ifndef AVUTIL_FIXED_DSP_H | ||
| 49 | #define AVUTIL_FIXED_DSP_H | ||
| 50 | |||
| 51 | #include <stdint.h> | ||
| 52 | #include "attributes.h" | ||
| 53 | #include "libavcodec/mathops.h" | ||
| 54 | |||
| 55 | typedef struct AVFixedDSPContext { | ||
| 56 | /* Assume len is a multiple of 16, and arrays are 32-byte aligned */ | ||
| 57 | /* Results of multiplications are scaled down by 31 bit (and rounded) if not | ||
| 58 | * stated otherwise */ | ||
| 59 | |||
| 60 | /** | ||
| 61 | * Overlap/add with window function. | ||
| 62 | * Result is scaled down by "bits" bits. | ||
| 63 | * Used primarily by MDCT-based audio codecs. | ||
| 64 | * Source and destination vectors must overlap exactly or not at all. | ||
| 65 | * | ||
| 66 | * @param dst result vector | ||
| 67 | * constraints: 16-byte aligned | ||
| 68 | * @param src0 first source vector | ||
| 69 | * constraints: 16-byte aligned | ||
| 70 | * @param src1 second source vector | ||
| 71 | * constraints: 16-byte aligned | ||
| 72 | * @param win half-window vector | ||
| 73 | * constraints: 16-byte aligned | ||
| 74 | * @param len length of vector | ||
| 75 | * constraints: multiple of 4 | ||
| 76 | * @param bits scaling parameter | ||
| 77 | * | ||
| 78 | */ | ||
| 79 | void (*vector_fmul_window_scaled)(int16_t *dst, const int32_t *src0, const int32_t *src1, const int32_t *win, int len, uint8_t bits); | ||
| 80 | |||
| 81 | /** | ||
| 82 | * Overlap/add with window function. | ||
| 83 | * Used primarily by MDCT-based audio codecs. | ||
| 84 | * Source and destination vectors must overlap exactly or not at all. | ||
| 85 | * | ||
| 86 | * @param dst result vector | ||
| 87 | * constraints: 32-byte aligned | ||
| 88 | * @param src0 first source vector | ||
| 89 | * constraints: 16-byte aligned | ||
| 90 | * @param src1 second source vector | ||
| 91 | * constraints: 16-byte aligned | ||
| 92 | * @param win half-window vector | ||
| 93 | * constraints: 16-byte aligned | ||
| 94 | * @param len length of vector | ||
| 95 | * constraints: multiple of 4 | ||
| 96 | */ | ||
| 97 | void (*vector_fmul_window)(int32_t *dst, const int32_t *src0, const int32_t *src1, const int32_t *win, int len); | ||
| 98 | |||
| 99 | /** | ||
| 100 | * Fixed-point multiplication that calculates the entry wise product of two | ||
| 101 | * vectors of integers and stores the result in a vector of integers. | ||
| 102 | * | ||
| 103 | * @param dst output vector | ||
| 104 | * constraints: 32-byte aligned | ||
| 105 | * @param src0 first input vector | ||
| 106 | * constraints: 32-byte aligned | ||
| 107 | * @param src1 second input vector | ||
| 108 | * constraints: 32-byte aligned | ||
| 109 | * @param len number of elements in the input | ||
| 110 | * constraints: multiple of 16 | ||
| 111 | */ | ||
| 112 | void (*vector_fmul)(int *dst, const int *src0, const int *src1, | ||
| 113 | int len); | ||
| 114 | |||
| 115 | void (*vector_fmul_reverse)(int *dst, const int *src0, const int *src1, int len); | ||
| 116 | /** | ||
| 117 | * Calculate the entry wise product of two vectors of integers, add a third vector of | ||
| 118 | * integers and store the result in a vector of integers. | ||
| 119 | * | ||
| 120 | * @param dst output vector | ||
| 121 | * constraints: 32-byte aligned | ||
| 122 | * @param src0 first input vector | ||
| 123 | * constraints: 32-byte aligned | ||
| 124 | * @param src1 second input vector | ||
| 125 | * constraints: 32-byte aligned | ||
| 126 | * @param src2 third input vector | ||
| 127 | * constraints: 32-byte aligned | ||
| 128 | * @param len number of elements in the input | ||
| 129 | * constraints: multiple of 16 | ||
| 130 | */ | ||
| 131 | void (*vector_fmul_add)(int *dst, const int *src0, const int *src1, | ||
| 132 | const int *src2, int len); | ||
| 133 | |||
| 134 | /** | ||
| 135 | * Calculate the scalar product of two vectors of integers. | ||
| 136 | * | ||
| 137 | * @param v1 first vector, 16-byte aligned | ||
| 138 | * @param v2 second vector, 16-byte aligned | ||
| 139 | * @param len length of vectors, multiple of 4 | ||
| 140 | * | ||
| 141 | * @return sum of elementwise products | ||
| 142 | */ | ||
| 143 | int (*scalarproduct_fixed)(const int *v1, const int *v2, int len); | ||
| 144 | |||
| 145 | /** | ||
| 146 | * Calculate the sum and difference of two vectors of integers. | ||
| 147 | * | ||
| 148 | * @param v1 first input vector, sum output, 16-byte aligned | ||
| 149 | * @param v2 second input vector, difference output, 16-byte aligned | ||
| 150 | * @param len length of vectors, multiple of 4 | ||
| 151 | */ | ||
| 152 | void (*butterflies_fixed)(int *restrict v1, int *restrict v2, int len); | ||
| 153 | } AVFixedDSPContext; | ||
| 154 | |||
| 155 | /** | ||
| 156 | * Allocate and initialize a fixed DSP context. | ||
| 157 | * note: should be freed with a av_free call when no longer needed. | ||
| 158 | * | ||
| 159 | * @param strict setting to non-zero avoids using functions which may not be IEEE-754 compliant | ||
| 160 | */ | ||
| 161 | AVFixedDSPContext * avpriv_alloc_fixed_dsp(int strict); | ||
| 162 | |||
| 163 | void ff_fixed_dsp_init_riscv(AVFixedDSPContext *fdsp); | ||
| 164 | void ff_fixed_dsp_init_x86(AVFixedDSPContext *fdsp); | ||
| 165 | |||
| 166 | /** | ||
| 167 | * Calculate the square root | ||
| 168 | * | ||
| 169 | * @param x input fixed point number | ||
| 170 | * | ||
| 171 | * @param bits format of fixed point number (32 - bits).bits | ||
| 172 | * | ||
| 173 | * note: input is normalized to (0, 1) fixed point value | ||
| 174 | */ | ||
| 175 | |||
| 176 | 137931 | static av_always_inline int fixed_sqrt(int x, int bits) | |
| 177 | { | ||
| 178 | int retval, bit_mask, guess, square, i; | ||
| 179 | int64_t accu; | ||
| 180 | 137931 | int shift1 = 30 - bits; | |
| 181 | 137931 | int shift2 = bits - 15; | |
| 182 | |||
| 183 |
1/2✗ Branch 0 not taken.
✓ Branch 1 taken 137931 times.
|
137931 | if (shift1 > 0) retval = ff_sqrt(x << shift1); |
| 184 | 137931 | else retval = ff_sqrt(x >> -shift1); | |
| 185 | |||
| 186 |
1/2✓ Branch 0 taken 137931 times.
✗ Branch 1 not taken.
|
137931 | if (shift2 > 0) { |
| 187 | 137931 | retval = retval << shift2; | |
| 188 | 137931 | bit_mask = (1 << (shift2 - 1)); | |
| 189 | |||
| 190 |
2/2✓ Branch 0 taken 2206896 times.
✓ Branch 1 taken 137931 times.
|
2344827 | for (i=0; i<shift2; i++){ |
| 191 | 2206896 | guess = retval + bit_mask; | |
| 192 | 2206896 | accu = (int64_t)guess * guess; | |
| 193 | 2206896 | square = (int)((accu + bit_mask) >> bits); | |
| 194 |
2/2✓ Branch 0 taken 1102901 times.
✓ Branch 1 taken 1103995 times.
|
2206896 | if (x >= square) |
| 195 | 1102901 | retval += bit_mask; | |
| 196 | 2206896 | bit_mask >>= 1; | |
| 197 | } | ||
| 198 | |||
| 199 | } | ||
| 200 | ✗ | else retval >>= (-shift2); | |
| 201 | |||
| 202 | 137931 | return retval; | |
| 203 | } | ||
| 204 | |||
| 205 | #endif /* AVUTIL_FIXED_DSP_H */ | ||
| 206 |