| Line | Branch | Exec | Source |
|---|---|---|---|
| 1 | /* | ||
| 2 | * VP9 compatible video decoder | ||
| 3 | * | ||
| 4 | * Copyright (C) 2013 Ronald S. Bultje <rsbultje gmail com> | ||
| 5 | * Copyright (C) 2013 Clément Bœsch <u pkh me> | ||
| 6 | * | ||
| 7 | * This file is part of FFmpeg. | ||
| 8 | * | ||
| 9 | * FFmpeg is free software; you can redistribute it and/or | ||
| 10 | * modify it under the terms of the GNU Lesser General Public | ||
| 11 | * License as published by the Free Software Foundation; either | ||
| 12 | * version 2.1 of the License, or (at your option) any later version. | ||
| 13 | * | ||
| 14 | * FFmpeg is distributed in the hope that it will be useful, | ||
| 15 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | ||
| 16 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | ||
| 17 | * Lesser General Public License for more details. | ||
| 18 | * | ||
| 19 | * You should have received a copy of the GNU Lesser General Public | ||
| 20 | * License along with FFmpeg; if not, write to the Free Software | ||
| 21 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA | ||
| 22 | */ | ||
| 23 | |||
| 24 | #include "libavutil/common.h" | ||
| 25 | #include "bit_depth_template.c" | ||
| 26 | #include "vp9dsp.h" | ||
| 27 | |||
| 28 | #if BIT_DEPTH != 12 | ||
| 29 | |||
| 30 | // FIXME see whether we can merge parts of this (perhaps at least 4x4 and 8x8) | ||
| 31 | // back with h264pred.[ch] | ||
| 32 | |||
| 33 | 127788 | static void vert_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 34 | const uint8_t *left, const uint8_t *_top) | ||
| 35 | { | ||
| 36 | 127788 | pixel *dst = (pixel *) _dst; | |
| 37 | 127788 | const pixel *top = (const pixel *) _top; | |
| 38 | 127788 | pixel4 p4 = AV_RN4PA(top); | |
| 39 | |||
| 40 | 127788 | stride /= sizeof(pixel); | |
| 41 | 127788 | AV_WN4PA(dst + stride * 0, p4); | |
| 42 | 127788 | AV_WN4PA(dst + stride * 1, p4); | |
| 43 | 127788 | AV_WN4PA(dst + stride * 2, p4); | |
| 44 | 127788 | AV_WN4PA(dst + stride * 3, p4); | |
| 45 | 127788 | } | |
| 46 | |||
| 47 | 27099 | static void vert_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 48 | const uint8_t *left, const uint8_t *_top) | ||
| 49 | { | ||
| 50 | 27099 | pixel *dst = (pixel *) _dst; | |
| 51 | 27099 | const pixel *top = (const pixel *) _top; | |
| 52 | 27099 | pixel4 p4a = AV_RN4PA(top + 0); | |
| 53 | 27099 | pixel4 p4b = AV_RN4PA(top + 4); | |
| 54 | int y; | ||
| 55 | |||
| 56 | 27099 | stride /= sizeof(pixel); | |
| 57 |
2/2✓ Branch 0 taken 216792 times.
✓ Branch 1 taken 27099 times.
|
243891 | for (y = 0; y < 8; y++) { |
| 58 | 216792 | AV_WN4PA(dst + 0, p4a); | |
| 59 | 216792 | AV_WN4PA(dst + 4, p4b); | |
| 60 | 216792 | dst += stride; | |
| 61 | } | ||
| 62 | 27099 | } | |
| 63 | |||
| 64 | 6390 | static void vert_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 65 | const uint8_t *left, const uint8_t *_top) | ||
| 66 | { | ||
| 67 | 6390 | pixel *dst = (pixel *) _dst; | |
| 68 | 6390 | const pixel *top = (const pixel *) _top; | |
| 69 | 6390 | pixel4 p4a = AV_RN4PA(top + 0); | |
| 70 | 6390 | pixel4 p4b = AV_RN4PA(top + 4); | |
| 71 | 6390 | pixel4 p4c = AV_RN4PA(top + 8); | |
| 72 | 6390 | pixel4 p4d = AV_RN4PA(top + 12); | |
| 73 | int y; | ||
| 74 | |||
| 75 | 6390 | stride /= sizeof(pixel); | |
| 76 |
2/2✓ Branch 0 taken 102240 times.
✓ Branch 1 taken 6390 times.
|
108630 | for (y = 0; y < 16; y++) { |
| 77 | 102240 | AV_WN4PA(dst + 0, p4a); | |
| 78 | 102240 | AV_WN4PA(dst + 4, p4b); | |
| 79 | 102240 | AV_WN4PA(dst + 8, p4c); | |
| 80 | 102240 | AV_WN4PA(dst + 12, p4d); | |
| 81 | 102240 | dst += stride; | |
| 82 | } | ||
| 83 | 6390 | } | |
| 84 | |||
| 85 | 1042 | static void vert_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 86 | const uint8_t *left, const uint8_t *_top) | ||
| 87 | { | ||
| 88 | 1042 | pixel *dst = (pixel *) _dst; | |
| 89 | 1042 | const pixel *top = (const pixel *) _top; | |
| 90 | 1042 | pixel4 p4a = AV_RN4PA(top + 0); | |
| 91 | 1042 | pixel4 p4b = AV_RN4PA(top + 4); | |
| 92 | 1042 | pixel4 p4c = AV_RN4PA(top + 8); | |
| 93 | 1042 | pixel4 p4d = AV_RN4PA(top + 12); | |
| 94 | 1042 | pixel4 p4e = AV_RN4PA(top + 16); | |
| 95 | 1042 | pixel4 p4f = AV_RN4PA(top + 20); | |
| 96 | 1042 | pixel4 p4g = AV_RN4PA(top + 24); | |
| 97 | 1042 | pixel4 p4h = AV_RN4PA(top + 28); | |
| 98 | int y; | ||
| 99 | |||
| 100 | 1042 | stride /= sizeof(pixel); | |
| 101 |
2/2✓ Branch 0 taken 33344 times.
✓ Branch 1 taken 1042 times.
|
34386 | for (y = 0; y < 32; y++) { |
| 102 | 33344 | AV_WN4PA(dst + 0, p4a); | |
| 103 | 33344 | AV_WN4PA(dst + 4, p4b); | |
| 104 | 33344 | AV_WN4PA(dst + 8, p4c); | |
| 105 | 33344 | AV_WN4PA(dst + 12, p4d); | |
| 106 | 33344 | AV_WN4PA(dst + 16, p4e); | |
| 107 | 33344 | AV_WN4PA(dst + 20, p4f); | |
| 108 | 33344 | AV_WN4PA(dst + 24, p4g); | |
| 109 | 33344 | AV_WN4PA(dst + 28, p4h); | |
| 110 | 33344 | dst += stride; | |
| 111 | } | ||
| 112 | 1042 | } | |
| 113 | |||
| 114 | 278319 | static void hor_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 115 | const uint8_t *_left, const uint8_t *top) | ||
| 116 | { | ||
| 117 | 278319 | pixel *dst = (pixel *) _dst; | |
| 118 | 278319 | const pixel *left = (const pixel *) _left; | |
| 119 | |||
| 120 | 278319 | stride /= sizeof(pixel); | |
| 121 | 278319 | AV_WN4PA(dst + stride * 0, PIXEL_SPLAT_X4(left[3])); | |
| 122 | 278319 | AV_WN4PA(dst + stride * 1, PIXEL_SPLAT_X4(left[2])); | |
| 123 | 278319 | AV_WN4PA(dst + stride * 2, PIXEL_SPLAT_X4(left[1])); | |
| 124 | 278319 | AV_WN4PA(dst + stride * 3, PIXEL_SPLAT_X4(left[0])); | |
| 125 | 278319 | } | |
| 126 | |||
| 127 | 74390 | static void hor_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 128 | const uint8_t *_left, const uint8_t *top) | ||
| 129 | { | ||
| 130 | 74390 | pixel *dst = (pixel *) _dst; | |
| 131 | 74390 | const pixel *left = (const pixel *) _left; | |
| 132 | int y; | ||
| 133 | |||
| 134 | 74390 | stride /= sizeof(pixel); | |
| 135 |
2/2✓ Branch 0 taken 595120 times.
✓ Branch 1 taken 74390 times.
|
669510 | for (y = 0; y < 8; y++) { |
| 136 | 595120 | pixel4 p4 = PIXEL_SPLAT_X4(left[7 - y]); | |
| 137 | |||
| 138 | 595120 | AV_WN4PA(dst + 0, p4); | |
| 139 | 595120 | AV_WN4PA(dst + 4, p4); | |
| 140 | 595120 | dst += stride; | |
| 141 | } | ||
| 142 | 74390 | } | |
| 143 | |||
| 144 | 13695 | static void hor_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 145 | const uint8_t *_left, const uint8_t *top) | ||
| 146 | { | ||
| 147 | 13695 | pixel *dst = (pixel *) _dst; | |
| 148 | 13695 | const pixel *left = (const pixel *) _left; | |
| 149 | int y; | ||
| 150 | |||
| 151 | 13695 | stride /= sizeof(pixel); | |
| 152 |
2/2✓ Branch 0 taken 219120 times.
✓ Branch 1 taken 13695 times.
|
232815 | for (y = 0; y < 16; y++) { |
| 153 | 219120 | pixel4 p4 = PIXEL_SPLAT_X4(left[15 - y]); | |
| 154 | |||
| 155 | 219120 | AV_WN4PA(dst + 0, p4); | |
| 156 | 219120 | AV_WN4PA(dst + 4, p4); | |
| 157 | 219120 | AV_WN4PA(dst + 8, p4); | |
| 158 | 219120 | AV_WN4PA(dst + 12, p4); | |
| 159 | 219120 | dst += stride; | |
| 160 | } | ||
| 161 | 13695 | } | |
| 162 | |||
| 163 | 1474 | static void hor_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 164 | const uint8_t *_left, const uint8_t *top) | ||
| 165 | { | ||
| 166 | 1474 | pixel *dst = (pixel *) _dst; | |
| 167 | 1474 | const pixel *left = (const pixel *) _left; | |
| 168 | int y; | ||
| 169 | |||
| 170 | 1474 | stride /= sizeof(pixel); | |
| 171 |
2/2✓ Branch 0 taken 47168 times.
✓ Branch 1 taken 1474 times.
|
48642 | for (y = 0; y < 32; y++) { |
| 172 | 47168 | pixel4 p4 = PIXEL_SPLAT_X4(left[31 - y]); | |
| 173 | |||
| 174 | 47168 | AV_WN4PA(dst + 0, p4); | |
| 175 | 47168 | AV_WN4PA(dst + 4, p4); | |
| 176 | 47168 | AV_WN4PA(dst + 8, p4); | |
| 177 | 47168 | AV_WN4PA(dst + 12, p4); | |
| 178 | 47168 | AV_WN4PA(dst + 16, p4); | |
| 179 | 47168 | AV_WN4PA(dst + 20, p4); | |
| 180 | 47168 | AV_WN4PA(dst + 24, p4); | |
| 181 | 47168 | AV_WN4PA(dst + 28, p4); | |
| 182 | 47168 | dst += stride; | |
| 183 | } | ||
| 184 | 1474 | } | |
| 185 | |||
| 186 | #endif /* BIT_DEPTH != 12 */ | ||
| 187 | |||
| 188 | 78331 | static void tm_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 189 | const uint8_t *_left, const uint8_t *_top) | ||
| 190 | { | ||
| 191 | 78331 | pixel *dst = (pixel *) _dst; | |
| 192 | 78331 | const pixel *left = (const pixel *) _left; | |
| 193 | 78331 | const pixel *top = (const pixel *) _top; | |
| 194 | 78331 | int y, tl = top[-1]; | |
| 195 | |||
| 196 | 78331 | stride /= sizeof(pixel); | |
| 197 |
2/2✓ Branch 0 taken 313324 times.
✓ Branch 1 taken 78331 times.
|
391655 | for (y = 0; y < 4; y++) { |
| 198 | 313324 | int l_m_tl = left[3 - y] - tl; | |
| 199 | |||
| 200 | 313324 | dst[0] = av_clip_pixel(top[0] + l_m_tl); | |
| 201 | 313324 | dst[1] = av_clip_pixel(top[1] + l_m_tl); | |
| 202 | 313324 | dst[2] = av_clip_pixel(top[2] + l_m_tl); | |
| 203 | 313324 | dst[3] = av_clip_pixel(top[3] + l_m_tl); | |
| 204 | 313324 | dst += stride; | |
| 205 | } | ||
| 206 | 78331 | } | |
| 207 | |||
| 208 | 20021 | static void tm_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 209 | const uint8_t *_left, const uint8_t *_top) | ||
| 210 | { | ||
| 211 | 20021 | pixel *dst = (pixel *) _dst; | |
| 212 | 20021 | const pixel *left = (const pixel *) _left; | |
| 213 | 20021 | const pixel *top = (const pixel *) _top; | |
| 214 | 20021 | int y, tl = top[-1]; | |
| 215 | |||
| 216 | 20021 | stride /= sizeof(pixel); | |
| 217 |
2/2✓ Branch 0 taken 160168 times.
✓ Branch 1 taken 20021 times.
|
180189 | for (y = 0; y < 8; y++) { |
| 218 | 160168 | int l_m_tl = left[7 - y] - tl; | |
| 219 | |||
| 220 | 160168 | dst[0] = av_clip_pixel(top[0] + l_m_tl); | |
| 221 | 160168 | dst[1] = av_clip_pixel(top[1] + l_m_tl); | |
| 222 | 160168 | dst[2] = av_clip_pixel(top[2] + l_m_tl); | |
| 223 | 160168 | dst[3] = av_clip_pixel(top[3] + l_m_tl); | |
| 224 | 160168 | dst[4] = av_clip_pixel(top[4] + l_m_tl); | |
| 225 | 160168 | dst[5] = av_clip_pixel(top[5] + l_m_tl); | |
| 226 | 160168 | dst[6] = av_clip_pixel(top[6] + l_m_tl); | |
| 227 | 160168 | dst[7] = av_clip_pixel(top[7] + l_m_tl); | |
| 228 | 160168 | dst += stride; | |
| 229 | } | ||
| 230 | 20021 | } | |
| 231 | |||
| 232 | 2568 | static void tm_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 233 | const uint8_t *_left, const uint8_t *_top) | ||
| 234 | { | ||
| 235 | 2568 | pixel *dst = (pixel *) _dst; | |
| 236 | 2568 | const pixel *left = (const pixel *) _left; | |
| 237 | 2568 | const pixel *top = (const pixel *) _top; | |
| 238 | 2568 | int y, tl = top[-1]; | |
| 239 | |||
| 240 | 2568 | stride /= sizeof(pixel); | |
| 241 |
2/2✓ Branch 0 taken 41088 times.
✓ Branch 1 taken 2568 times.
|
43656 | for (y = 0; y < 16; y++) { |
| 242 | 41088 | int l_m_tl = left[15 - y] - tl; | |
| 243 | |||
| 244 | 41088 | dst[ 0] = av_clip_pixel(top[ 0] + l_m_tl); | |
| 245 | 41088 | dst[ 1] = av_clip_pixel(top[ 1] + l_m_tl); | |
| 246 | 41088 | dst[ 2] = av_clip_pixel(top[ 2] + l_m_tl); | |
| 247 | 41088 | dst[ 3] = av_clip_pixel(top[ 3] + l_m_tl); | |
| 248 | 41088 | dst[ 4] = av_clip_pixel(top[ 4] + l_m_tl); | |
| 249 | 41088 | dst[ 5] = av_clip_pixel(top[ 5] + l_m_tl); | |
| 250 | 41088 | dst[ 6] = av_clip_pixel(top[ 6] + l_m_tl); | |
| 251 | 41088 | dst[ 7] = av_clip_pixel(top[ 7] + l_m_tl); | |
| 252 | 41088 | dst[ 8] = av_clip_pixel(top[ 8] + l_m_tl); | |
| 253 | 41088 | dst[ 9] = av_clip_pixel(top[ 9] + l_m_tl); | |
| 254 | 41088 | dst[10] = av_clip_pixel(top[10] + l_m_tl); | |
| 255 | 41088 | dst[11] = av_clip_pixel(top[11] + l_m_tl); | |
| 256 | 41088 | dst[12] = av_clip_pixel(top[12] + l_m_tl); | |
| 257 | 41088 | dst[13] = av_clip_pixel(top[13] + l_m_tl); | |
| 258 | 41088 | dst[14] = av_clip_pixel(top[14] + l_m_tl); | |
| 259 | 41088 | dst[15] = av_clip_pixel(top[15] + l_m_tl); | |
| 260 | 41088 | dst += stride; | |
| 261 | } | ||
| 262 | 2568 | } | |
| 263 | |||
| 264 | 386 | static void tm_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 265 | const uint8_t *_left, const uint8_t *_top) | ||
| 266 | { | ||
| 267 | 386 | pixel *dst = (pixel *) _dst; | |
| 268 | 386 | const pixel *left = (const pixel *) _left; | |
| 269 | 386 | const pixel *top = (const pixel *) _top; | |
| 270 | 386 | int y, tl = top[-1]; | |
| 271 | |||
| 272 | 386 | stride /= sizeof(pixel); | |
| 273 |
2/2✓ Branch 0 taken 12352 times.
✓ Branch 1 taken 386 times.
|
12738 | for (y = 0; y < 32; y++) { |
| 274 | 12352 | int l_m_tl = left[31 - y] - tl; | |
| 275 | |||
| 276 | 12352 | dst[ 0] = av_clip_pixel(top[ 0] + l_m_tl); | |
| 277 | 12352 | dst[ 1] = av_clip_pixel(top[ 1] + l_m_tl); | |
| 278 | 12352 | dst[ 2] = av_clip_pixel(top[ 2] + l_m_tl); | |
| 279 | 12352 | dst[ 3] = av_clip_pixel(top[ 3] + l_m_tl); | |
| 280 | 12352 | dst[ 4] = av_clip_pixel(top[ 4] + l_m_tl); | |
| 281 | 12352 | dst[ 5] = av_clip_pixel(top[ 5] + l_m_tl); | |
| 282 | 12352 | dst[ 6] = av_clip_pixel(top[ 6] + l_m_tl); | |
| 283 | 12352 | dst[ 7] = av_clip_pixel(top[ 7] + l_m_tl); | |
| 284 | 12352 | dst[ 8] = av_clip_pixel(top[ 8] + l_m_tl); | |
| 285 | 12352 | dst[ 9] = av_clip_pixel(top[ 9] + l_m_tl); | |
| 286 | 12352 | dst[10] = av_clip_pixel(top[10] + l_m_tl); | |
| 287 | 12352 | dst[11] = av_clip_pixel(top[11] + l_m_tl); | |
| 288 | 12352 | dst[12] = av_clip_pixel(top[12] + l_m_tl); | |
| 289 | 12352 | dst[13] = av_clip_pixel(top[13] + l_m_tl); | |
| 290 | 12352 | dst[14] = av_clip_pixel(top[14] + l_m_tl); | |
| 291 | 12352 | dst[15] = av_clip_pixel(top[15] + l_m_tl); | |
| 292 | 12352 | dst[16] = av_clip_pixel(top[16] + l_m_tl); | |
| 293 | 12352 | dst[17] = av_clip_pixel(top[17] + l_m_tl); | |
| 294 | 12352 | dst[18] = av_clip_pixel(top[18] + l_m_tl); | |
| 295 | 12352 | dst[19] = av_clip_pixel(top[19] + l_m_tl); | |
| 296 | 12352 | dst[20] = av_clip_pixel(top[20] + l_m_tl); | |
| 297 | 12352 | dst[21] = av_clip_pixel(top[21] + l_m_tl); | |
| 298 | 12352 | dst[22] = av_clip_pixel(top[22] + l_m_tl); | |
| 299 | 12352 | dst[23] = av_clip_pixel(top[23] + l_m_tl); | |
| 300 | 12352 | dst[24] = av_clip_pixel(top[24] + l_m_tl); | |
| 301 | 12352 | dst[25] = av_clip_pixel(top[25] + l_m_tl); | |
| 302 | 12352 | dst[26] = av_clip_pixel(top[26] + l_m_tl); | |
| 303 | 12352 | dst[27] = av_clip_pixel(top[27] + l_m_tl); | |
| 304 | 12352 | dst[28] = av_clip_pixel(top[28] + l_m_tl); | |
| 305 | 12352 | dst[29] = av_clip_pixel(top[29] + l_m_tl); | |
| 306 | 12352 | dst[30] = av_clip_pixel(top[30] + l_m_tl); | |
| 307 | 12352 | dst[31] = av_clip_pixel(top[31] + l_m_tl); | |
| 308 | 12352 | dst += stride; | |
| 309 | } | ||
| 310 | 386 | } | |
| 311 | |||
| 312 | #if BIT_DEPTH != 12 | ||
| 313 | |||
| 314 | 353773 | static void dc_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 315 | const uint8_t *_left, const uint8_t *_top) | ||
| 316 | { | ||
| 317 | 353773 | pixel *dst = (pixel *) _dst; | |
| 318 | 353773 | const pixel *left = (const pixel *) _left; | |
| 319 | 353773 | const pixel *top = (const pixel *) _top; | |
| 320 | 353773 | pixel4 dc = PIXEL_SPLAT_X4((left[0] + left[1] + left[2] + left[3] + | |
| 321 | top[0] + top[1] + top[2] + top[3] + 4) >> 3); | ||
| 322 | |||
| 323 | 353773 | stride /= sizeof(pixel); | |
| 324 | 353773 | AV_WN4PA(dst + stride * 0, dc); | |
| 325 | 353773 | AV_WN4PA(dst + stride * 1, dc); | |
| 326 | 353773 | AV_WN4PA(dst + stride * 2, dc); | |
| 327 | 353773 | AV_WN4PA(dst + stride * 3, dc); | |
| 328 | 353773 | } | |
| 329 | |||
| 330 | 142833 | static void dc_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 331 | const uint8_t *_left, const uint8_t *_top) | ||
| 332 | { | ||
| 333 | 142833 | pixel *dst = (pixel *) _dst; | |
| 334 | 142833 | const pixel *left = (const pixel *) _left; | |
| 335 | 142833 | const pixel *top = (const pixel *) _top; | |
| 336 | 142833 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 337 | ((left[0] + left[1] + left[2] + left[3] + left[4] + left[5] + | ||
| 338 | left[6] + left[7] + top[0] + top[1] + top[2] + top[3] + | ||
| 339 | top[4] + top[5] + top[6] + top[7] + 8) >> 4); | ||
| 340 | int y; | ||
| 341 | |||
| 342 | 142833 | stride /= sizeof(pixel); | |
| 343 |
2/2✓ Branch 0 taken 1142664 times.
✓ Branch 1 taken 142833 times.
|
1285497 | for (y = 0; y < 8; y++) { |
| 344 | 1142664 | AV_WN4PA(dst + 0, dc); | |
| 345 | 1142664 | AV_WN4PA(dst + 4, dc); | |
| 346 | 1142664 | dst += stride; | |
| 347 | } | ||
| 348 | 142833 | } | |
| 349 | |||
| 350 | 20481 | static void dc_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 351 | const uint8_t *_left, const uint8_t *_top) | ||
| 352 | { | ||
| 353 | 20481 | pixel *dst = (pixel *) _dst; | |
| 354 | 20481 | const pixel *left = (const pixel *) _left; | |
| 355 | 20481 | const pixel *top = (const pixel *) _top; | |
| 356 | 20481 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 357 | ((left[0] + left[1] + left[2] + left[3] + left[4] + left[5] + left[6] + | ||
| 358 | left[7] + left[8] + left[9] + left[10] + left[11] + left[12] + | ||
| 359 | left[13] + left[14] + left[15] + top[0] + top[1] + top[2] + top[3] + | ||
| 360 | top[4] + top[5] + top[6] + top[7] + top[8] + top[9] + top[10] + | ||
| 361 | top[11] + top[12] + top[13] + top[14] + top[15] + 16) >> 5); | ||
| 362 | int y; | ||
| 363 | |||
| 364 | 20481 | stride /= sizeof(pixel); | |
| 365 |
2/2✓ Branch 0 taken 327696 times.
✓ Branch 1 taken 20481 times.
|
348177 | for (y = 0; y < 16; y++) { |
| 366 | 327696 | AV_WN4PA(dst + 0, dc); | |
| 367 | 327696 | AV_WN4PA(dst + 4, dc); | |
| 368 | 327696 | AV_WN4PA(dst + 8, dc); | |
| 369 | 327696 | AV_WN4PA(dst + 12, dc); | |
| 370 | 327696 | dst += stride; | |
| 371 | } | ||
| 372 | 20481 | } | |
| 373 | |||
| 374 | 5642 | static void dc_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 375 | const uint8_t *_left, const uint8_t *_top) | ||
| 376 | { | ||
| 377 | 5642 | pixel *dst = (pixel *) _dst; | |
| 378 | 5642 | const pixel *left = (const pixel *) _left; | |
| 379 | 5642 | const pixel *top = (const pixel *) _top; | |
| 380 | 5642 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 381 | ((left[0] + left[1] + left[2] + left[3] + left[4] + left[5] + left[6] + | ||
| 382 | left[7] + left[8] + left[9] + left[10] + left[11] + left[12] + | ||
| 383 | left[13] + left[14] + left[15] + left[16] + left[17] + left[18] + | ||
| 384 | left[19] + left[20] + left[21] + left[22] + left[23] + left[24] + | ||
| 385 | left[25] + left[26] + left[27] + left[28] + left[29] + left[30] + | ||
| 386 | left[31] + top[0] + top[1] + top[2] + top[3] + top[4] + top[5] + | ||
| 387 | top[6] + top[7] + top[8] + top[9] + top[10] + top[11] + top[12] + | ||
| 388 | top[13] + top[14] + top[15] + top[16] + top[17] + top[18] + top[19] + | ||
| 389 | top[20] + top[21] + top[22] + top[23] + top[24] + top[25] + top[26] + | ||
| 390 | top[27] + top[28] + top[29] + top[30] + top[31] + 32) >> 6); | ||
| 391 | int y; | ||
| 392 | |||
| 393 | 5642 | stride /= sizeof(pixel); | |
| 394 |
2/2✓ Branch 0 taken 180544 times.
✓ Branch 1 taken 5642 times.
|
186186 | for (y = 0; y < 32; y++) { |
| 395 | 180544 | AV_WN4PA(dst + 0, dc); | |
| 396 | 180544 | AV_WN4PA(dst + 4, dc); | |
| 397 | 180544 | AV_WN4PA(dst + 8, dc); | |
| 398 | 180544 | AV_WN4PA(dst + 12, dc); | |
| 399 | 180544 | AV_WN4PA(dst + 16, dc); | |
| 400 | 180544 | AV_WN4PA(dst + 20, dc); | |
| 401 | 180544 | AV_WN4PA(dst + 24, dc); | |
| 402 | 180544 | AV_WN4PA(dst + 28, dc); | |
| 403 | 180544 | dst += stride; | |
| 404 | } | ||
| 405 | 5642 | } | |
| 406 | |||
| 407 | 9204 | static void dc_left_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 408 | const uint8_t *_left, const uint8_t *top) | ||
| 409 | { | ||
| 410 | 9204 | pixel *dst = (pixel *) _dst; | |
| 411 | 9204 | const pixel *left = (const pixel *) _left; | |
| 412 | 9204 | pixel4 dc = PIXEL_SPLAT_X4((left[0] + left[1] + left[2] + left[3] + 2) >> 2); | |
| 413 | |||
| 414 | 9204 | stride /= sizeof(pixel); | |
| 415 | 9204 | AV_WN4PA(dst + stride * 0, dc); | |
| 416 | 9204 | AV_WN4PA(dst + stride * 1, dc); | |
| 417 | 9204 | AV_WN4PA(dst + stride * 2, dc); | |
| 418 | 9204 | AV_WN4PA(dst + stride * 3, dc); | |
| 419 | 9204 | } | |
| 420 | |||
| 421 | 3370 | static void dc_left_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 422 | const uint8_t *_left, const uint8_t *top) | ||
| 423 | { | ||
| 424 | 3370 | pixel *dst = (pixel *) _dst; | |
| 425 | 3370 | const pixel *left = (const pixel *) _left; | |
| 426 | 3370 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 427 | ((left[0] + left[1] + left[2] + left[3] + | ||
| 428 | left[4] + left[5] + left[6] + left[7] + 4) >> 3); | ||
| 429 | int y; | ||
| 430 | |||
| 431 | 3370 | stride /= sizeof(pixel); | |
| 432 |
2/2✓ Branch 0 taken 26960 times.
✓ Branch 1 taken 3370 times.
|
30330 | for (y = 0; y < 8; y++) { |
| 433 | 26960 | AV_WN4PA(dst + 0, dc); | |
| 434 | 26960 | AV_WN4PA(dst + 4, dc); | |
| 435 | 26960 | dst += stride; | |
| 436 | } | ||
| 437 | 3370 | } | |
| 438 | |||
| 439 | 1189 | static void dc_left_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 440 | const uint8_t *_left, const uint8_t *top) | ||
| 441 | { | ||
| 442 | 1189 | pixel *dst = (pixel *) _dst; | |
| 443 | 1189 | const pixel *left = (const pixel *) _left; | |
| 444 | 1189 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 445 | ((left[0] + left[1] + left[2] + left[3] + left[4] + left[5] + | ||
| 446 | left[6] + left[7] + left[8] + left[9] + left[10] + left[11] + | ||
| 447 | left[12] + left[13] + left[14] + left[15] + 8) >> 4); | ||
| 448 | int y; | ||
| 449 | |||
| 450 | 1189 | stride /= sizeof(pixel); | |
| 451 |
2/2✓ Branch 0 taken 19024 times.
✓ Branch 1 taken 1189 times.
|
20213 | for (y = 0; y < 16; y++) { |
| 452 | 19024 | AV_WN4PA(dst + 0, dc); | |
| 453 | 19024 | AV_WN4PA(dst + 4, dc); | |
| 454 | 19024 | AV_WN4PA(dst + 8, dc); | |
| 455 | 19024 | AV_WN4PA(dst + 12, dc); | |
| 456 | 19024 | dst += stride; | |
| 457 | } | ||
| 458 | 1189 | } | |
| 459 | |||
| 460 | 1350 | static void dc_left_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 461 | const uint8_t *_left, const uint8_t *top) | ||
| 462 | { | ||
| 463 | 1350 | pixel *dst = (pixel *) _dst; | |
| 464 | 1350 | const pixel *left = (const pixel *) _left; | |
| 465 | 1350 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 466 | ((left[0] + left[1] + left[2] + left[3] + left[4] + left[5] + | ||
| 467 | left[6] + left[7] + left[8] + left[9] + left[10] + left[11] + | ||
| 468 | left[12] + left[13] + left[14] + left[15] + left[16] + left[17] + | ||
| 469 | left[18] + left[19] + left[20] + left[21] + left[22] + left[23] + | ||
| 470 | left[24] + left[25] + left[26] + left[27] + left[28] + left[29] + | ||
| 471 | left[30] + left[31] + 16) >> 5); | ||
| 472 | int y; | ||
| 473 | |||
| 474 | 1350 | stride /= sizeof(pixel); | |
| 475 |
2/2✓ Branch 0 taken 43200 times.
✓ Branch 1 taken 1350 times.
|
44550 | for (y = 0; y < 32; y++) { |
| 476 | 43200 | AV_WN4PA(dst + 0, dc); | |
| 477 | 43200 | AV_WN4PA(dst + 4, dc); | |
| 478 | 43200 | AV_WN4PA(dst + 8, dc); | |
| 479 | 43200 | AV_WN4PA(dst + 12, dc); | |
| 480 | 43200 | AV_WN4PA(dst + 16, dc); | |
| 481 | 43200 | AV_WN4PA(dst + 20, dc); | |
| 482 | 43200 | AV_WN4PA(dst + 24, dc); | |
| 483 | 43200 | AV_WN4PA(dst + 28, dc); | |
| 484 | 43200 | dst += stride; | |
| 485 | } | ||
| 486 | 1350 | } | |
| 487 | |||
| 488 | 11040 | static void dc_top_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 489 | const uint8_t *left, const uint8_t *_top) | ||
| 490 | { | ||
| 491 | 11040 | pixel *dst = (pixel *) _dst; | |
| 492 | 11040 | const pixel *top = (const pixel *) _top; | |
| 493 | 11040 | pixel4 dc = PIXEL_SPLAT_X4((top[0] + top[1] + top[2] + top[3] + 2) >> 2); | |
| 494 | |||
| 495 | 11040 | stride /= sizeof(pixel); | |
| 496 | 11040 | AV_WN4PA(dst + stride * 0, dc); | |
| 497 | 11040 | AV_WN4PA(dst + stride * 1, dc); | |
| 498 | 11040 | AV_WN4PA(dst + stride * 2, dc); | |
| 499 | 11040 | AV_WN4PA(dst + stride * 3, dc); | |
| 500 | 11040 | } | |
| 501 | |||
| 502 | 7211 | static void dc_top_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 503 | const uint8_t *left, const uint8_t *_top) | ||
| 504 | { | ||
| 505 | 7211 | pixel *dst = (pixel *) _dst; | |
| 506 | 7211 | const pixel *top = (const pixel *) _top; | |
| 507 | 7211 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 508 | ((top[0] + top[1] + top[2] + top[3] + | ||
| 509 | top[4] + top[5] + top[6] + top[7] + 4) >> 3); | ||
| 510 | int y; | ||
| 511 | |||
| 512 | 7211 | stride /= sizeof(pixel); | |
| 513 |
2/2✓ Branch 0 taken 57688 times.
✓ Branch 1 taken 7211 times.
|
64899 | for (y = 0; y < 8; y++) { |
| 514 | 57688 | AV_WN4PA(dst + 0, dc); | |
| 515 | 57688 | AV_WN4PA(dst + 4, dc); | |
| 516 | 57688 | dst += stride; | |
| 517 | } | ||
| 518 | 7211 | } | |
| 519 | |||
| 520 | 2608 | static void dc_top_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 521 | const uint8_t *left, const uint8_t *_top) | ||
| 522 | { | ||
| 523 | 2608 | pixel *dst = (pixel *) _dst; | |
| 524 | 2608 | const pixel *top = (const pixel *) _top; | |
| 525 | 2608 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 526 | ((top[0] + top[1] + top[2] + top[3] + top[4] + top[5] + | ||
| 527 | top[6] + top[7] + top[8] + top[9] + top[10] + top[11] + | ||
| 528 | top[12] + top[13] + top[14] + top[15] + 8) >> 4); | ||
| 529 | int y; | ||
| 530 | |||
| 531 | 2608 | stride /= sizeof(pixel); | |
| 532 |
2/2✓ Branch 0 taken 41728 times.
✓ Branch 1 taken 2608 times.
|
44336 | for (y = 0; y < 16; y++) { |
| 533 | 41728 | AV_WN4PA(dst + 0, dc); | |
| 534 | 41728 | AV_WN4PA(dst + 4, dc); | |
| 535 | 41728 | AV_WN4PA(dst + 8, dc); | |
| 536 | 41728 | AV_WN4PA(dst + 12, dc); | |
| 537 | 41728 | dst += stride; | |
| 538 | } | ||
| 539 | 2608 | } | |
| 540 | |||
| 541 | 820 | static void dc_top_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 542 | const uint8_t *left, const uint8_t *_top) | ||
| 543 | { | ||
| 544 | 820 | pixel *dst = (pixel *) _dst; | |
| 545 | 820 | const pixel *top = (const pixel *) _top; | |
| 546 | 820 | pixel4 dc = PIXEL_SPLAT_X4 | |
| 547 | ((top[0] + top[1] + top[2] + top[3] + top[4] + top[5] + | ||
| 548 | top[6] + top[7] + top[8] + top[9] + top[10] + top[11] + | ||
| 549 | top[12] + top[13] + top[14] + top[15] + top[16] + top[17] + | ||
| 550 | top[18] + top[19] + top[20] + top[21] + top[22] + top[23] + | ||
| 551 | top[24] + top[25] + top[26] + top[27] + top[28] + top[29] + | ||
| 552 | top[30] + top[31] + 16) >> 5); | ||
| 553 | int y; | ||
| 554 | |||
| 555 | 820 | stride /= sizeof(pixel); | |
| 556 |
2/2✓ Branch 0 taken 26240 times.
✓ Branch 1 taken 820 times.
|
27060 | for (y = 0; y < 32; y++) { |
| 557 | 26240 | AV_WN4PA(dst + 0, dc); | |
| 558 | 26240 | AV_WN4PA(dst + 4, dc); | |
| 559 | 26240 | AV_WN4PA(dst + 8, dc); | |
| 560 | 26240 | AV_WN4PA(dst + 12, dc); | |
| 561 | 26240 | AV_WN4PA(dst + 16, dc); | |
| 562 | 26240 | AV_WN4PA(dst + 20, dc); | |
| 563 | 26240 | AV_WN4PA(dst + 24, dc); | |
| 564 | 26240 | AV_WN4PA(dst + 28, dc); | |
| 565 | 26240 | dst += stride; | |
| 566 | } | ||
| 567 | 820 | } | |
| 568 | |||
| 569 | #endif /* BIT_DEPTH != 12 */ | ||
| 570 | |||
| 571 | 386 | static void dc_128_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 572 | const uint8_t *left, const uint8_t *top) | ||
| 573 | { | ||
| 574 | 386 | pixel *dst = (pixel *) _dst; | |
| 575 | 386 | pixel4 val = PIXEL_SPLAT_X4(128 << (BIT_DEPTH - 8)); | |
| 576 | |||
| 577 | 386 | stride /= sizeof(pixel); | |
| 578 | 386 | AV_WN4PA(dst + stride * 0, val); | |
| 579 | 386 | AV_WN4PA(dst + stride * 1, val); | |
| 580 | 386 | AV_WN4PA(dst + stride * 2, val); | |
| 581 | 386 | AV_WN4PA(dst + stride * 3, val); | |
| 582 | 386 | } | |
| 583 | |||
| 584 | 223 | static void dc_128_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 585 | const uint8_t *left, const uint8_t *top) | ||
| 586 | { | ||
| 587 | 223 | pixel *dst = (pixel *) _dst; | |
| 588 | 223 | pixel4 val = PIXEL_SPLAT_X4(128 << (BIT_DEPTH - 8)); | |
| 589 | int y; | ||
| 590 | |||
| 591 | 223 | stride /= sizeof(pixel); | |
| 592 |
2/2✓ Branch 0 taken 1784 times.
✓ Branch 1 taken 223 times.
|
2007 | for (y = 0; y < 8; y++) { |
| 593 | 1784 | AV_WN4PA(dst + 0, val); | |
| 594 | 1784 | AV_WN4PA(dst + 4, val); | |
| 595 | 1784 | dst += stride; | |
| 596 | } | ||
| 597 | 223 | } | |
| 598 | |||
| 599 | 131 | static void dc_128_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 600 | const uint8_t *left, const uint8_t *top) | ||
| 601 | { | ||
| 602 | 131 | pixel *dst = (pixel *) _dst; | |
| 603 | 131 | pixel4 val = PIXEL_SPLAT_X4(128 << (BIT_DEPTH - 8)); | |
| 604 | int y; | ||
| 605 | |||
| 606 | 131 | stride /= sizeof(pixel); | |
| 607 |
2/2✓ Branch 0 taken 2096 times.
✓ Branch 1 taken 131 times.
|
2227 | for (y = 0; y < 16; y++) { |
| 608 | 2096 | AV_WN4PA(dst + 0, val); | |
| 609 | 2096 | AV_WN4PA(dst + 4, val); | |
| 610 | 2096 | AV_WN4PA(dst + 8, val); | |
| 611 | 2096 | AV_WN4PA(dst + 12, val); | |
| 612 | 2096 | dst += stride; | |
| 613 | } | ||
| 614 | 131 | } | |
| 615 | |||
| 616 | 127 | static void dc_128_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 617 | const uint8_t *left, const uint8_t *top) | ||
| 618 | { | ||
| 619 | 127 | pixel *dst = (pixel *) _dst; | |
| 620 | 127 | pixel4 val = PIXEL_SPLAT_X4(128 << (BIT_DEPTH - 8)); | |
| 621 | int y; | ||
| 622 | |||
| 623 | 127 | stride /= sizeof(pixel); | |
| 624 |
2/2✓ Branch 0 taken 4064 times.
✓ Branch 1 taken 127 times.
|
4191 | for (y = 0; y < 32; y++) { |
| 625 | 4064 | AV_WN4PA(dst + 0, val); | |
| 626 | 4064 | AV_WN4PA(dst + 4, val); | |
| 627 | 4064 | AV_WN4PA(dst + 8, val); | |
| 628 | 4064 | AV_WN4PA(dst + 12, val); | |
| 629 | 4064 | AV_WN4PA(dst + 16, val); | |
| 630 | 4064 | AV_WN4PA(dst + 20, val); | |
| 631 | 4064 | AV_WN4PA(dst + 24, val); | |
| 632 | 4064 | AV_WN4PA(dst + 28, val); | |
| 633 | 4064 | dst += stride; | |
| 634 | } | ||
| 635 | 127 | } | |
| 636 | |||
| 637 | 2528 | static void dc_127_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 638 | const uint8_t *left, const uint8_t *top) | ||
| 639 | { | ||
| 640 | 2528 | pixel *dst = (pixel *) _dst; | |
| 641 | 2528 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) - 1); | |
| 642 | |||
| 643 | 2528 | stride /= sizeof(pixel); | |
| 644 | 2528 | AV_WN4PA(dst + stride * 0, val); | |
| 645 | 2528 | AV_WN4PA(dst + stride * 1, val); | |
| 646 | 2528 | AV_WN4PA(dst + stride * 2, val); | |
| 647 | 2528 | AV_WN4PA(dst + stride * 3, val);} | |
| 648 | |||
| 649 | 334 | static void dc_127_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 650 | const uint8_t *left, const uint8_t *top) | ||
| 651 | { | ||
| 652 | 334 | pixel *dst = (pixel *) _dst; | |
| 653 | 334 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) - 1); | |
| 654 | int y; | ||
| 655 | |||
| 656 | 334 | stride /= sizeof(pixel); | |
| 657 |
2/2✓ Branch 0 taken 2672 times.
✓ Branch 1 taken 334 times.
|
3006 | for (y = 0; y < 8; y++) { |
| 658 | 2672 | AV_WN4PA(dst + 0, val); | |
| 659 | 2672 | AV_WN4PA(dst + 4, val); | |
| 660 | 2672 | dst += stride; | |
| 661 | } | ||
| 662 | 334 | } | |
| 663 | |||
| 664 | 178 | static void dc_127_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 665 | const uint8_t *left, const uint8_t *top) | ||
| 666 | { | ||
| 667 | 178 | pixel *dst = (pixel *) _dst; | |
| 668 | 178 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) - 1); | |
| 669 | int y; | ||
| 670 | |||
| 671 | 178 | stride /= sizeof(pixel); | |
| 672 |
2/2✓ Branch 0 taken 2848 times.
✓ Branch 1 taken 178 times.
|
3026 | for (y = 0; y < 16; y++) { |
| 673 | 2848 | AV_WN4PA(dst + 0, val); | |
| 674 | 2848 | AV_WN4PA(dst + 4, val); | |
| 675 | 2848 | AV_WN4PA(dst + 8, val); | |
| 676 | 2848 | AV_WN4PA(dst + 12, val); | |
| 677 | 2848 | dst += stride; | |
| 678 | } | ||
| 679 | 178 | } | |
| 680 | |||
| 681 | 83 | static void dc_127_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 682 | const uint8_t *left, const uint8_t *top) | ||
| 683 | { | ||
| 684 | 83 | pixel *dst = (pixel *) _dst; | |
| 685 | 83 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) - 1); | |
| 686 | int y; | ||
| 687 | |||
| 688 | 83 | stride /= sizeof(pixel); | |
| 689 |
2/2✓ Branch 0 taken 2656 times.
✓ Branch 1 taken 83 times.
|
2739 | for (y = 0; y < 32; y++) { |
| 690 | 2656 | AV_WN4PA(dst + 0, val); | |
| 691 | 2656 | AV_WN4PA(dst + 4, val); | |
| 692 | 2656 | AV_WN4PA(dst + 8, val); | |
| 693 | 2656 | AV_WN4PA(dst + 12, val); | |
| 694 | 2656 | AV_WN4PA(dst + 16, val); | |
| 695 | 2656 | AV_WN4PA(dst + 20, val); | |
| 696 | 2656 | AV_WN4PA(dst + 24, val); | |
| 697 | 2656 | AV_WN4PA(dst + 28, val); | |
| 698 | 2656 | dst += stride; | |
| 699 | } | ||
| 700 | 83 | } | |
| 701 | |||
| 702 | 2229 | static void dc_129_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 703 | const uint8_t *left, const uint8_t *top) | ||
| 704 | { | ||
| 705 | 2229 | pixel *dst = (pixel *) _dst; | |
| 706 | 2229 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) + 1); | |
| 707 | |||
| 708 | 2229 | stride /= sizeof(pixel); | |
| 709 | 2229 | AV_WN4PA(dst + stride * 0, val); | |
| 710 | 2229 | AV_WN4PA(dst + stride * 1, val); | |
| 711 | 2229 | AV_WN4PA(dst + stride * 2, val); | |
| 712 | 2229 | AV_WN4PA(dst + stride * 3, val); | |
| 713 | 2229 | } | |
| 714 | |||
| 715 | 348 | static void dc_129_8x8_c(uint8_t *_dst, ptrdiff_t stride, | |
| 716 | const uint8_t *left, const uint8_t *top) | ||
| 717 | { | ||
| 718 | 348 | pixel *dst = (pixel *) _dst; | |
| 719 | 348 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) + 1); | |
| 720 | int y; | ||
| 721 | |||
| 722 | 348 | stride /= sizeof(pixel); | |
| 723 |
2/2✓ Branch 0 taken 2784 times.
✓ Branch 1 taken 348 times.
|
3132 | for (y = 0; y < 8; y++) { |
| 724 | 2784 | AV_WN4PA(dst + 0, val); | |
| 725 | 2784 | AV_WN4PA(dst + 4, val); | |
| 726 | 2784 | dst += stride; | |
| 727 | } | ||
| 728 | 348 | } | |
| 729 | |||
| 730 | 258 | static void dc_129_16x16_c(uint8_t *_dst, ptrdiff_t stride, | |
| 731 | const uint8_t *left, const uint8_t *top) | ||
| 732 | { | ||
| 733 | 258 | pixel *dst = (pixel *) _dst; | |
| 734 | 258 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) + 1); | |
| 735 | int y; | ||
| 736 | |||
| 737 | 258 | stride /= sizeof(pixel); | |
| 738 |
2/2✓ Branch 0 taken 4128 times.
✓ Branch 1 taken 258 times.
|
4386 | for (y = 0; y < 16; y++) { |
| 739 | 4128 | AV_WN4PA(dst + 0, val); | |
| 740 | 4128 | AV_WN4PA(dst + 4, val); | |
| 741 | 4128 | AV_WN4PA(dst + 8, val); | |
| 742 | 4128 | AV_WN4PA(dst + 12, val); | |
| 743 | 4128 | dst += stride; | |
| 744 | } | ||
| 745 | 258 | } | |
| 746 | |||
| 747 | 140 | static void dc_129_32x32_c(uint8_t *_dst, ptrdiff_t stride, | |
| 748 | const uint8_t *left, const uint8_t *top) | ||
| 749 | { | ||
| 750 | 140 | pixel *dst = (pixel *) _dst; | |
| 751 | 140 | pixel4 val = PIXEL_SPLAT_X4((128 << (BIT_DEPTH - 8)) + 1); | |
| 752 | int y; | ||
| 753 | |||
| 754 | 140 | stride /= sizeof(pixel); | |
| 755 |
2/2✓ Branch 0 taken 4480 times.
✓ Branch 1 taken 140 times.
|
4620 | for (y = 0; y < 32; y++) { |
| 756 | 4480 | AV_WN4PA(dst + 0, val); | |
| 757 | 4480 | AV_WN4PA(dst + 4, val); | |
| 758 | 4480 | AV_WN4PA(dst + 8, val); | |
| 759 | 4480 | AV_WN4PA(dst + 12, val); | |
| 760 | 4480 | AV_WN4PA(dst + 16, val); | |
| 761 | 4480 | AV_WN4PA(dst + 20, val); | |
| 762 | 4480 | AV_WN4PA(dst + 24, val); | |
| 763 | 4480 | AV_WN4PA(dst + 28, val); | |
| 764 | 4480 | dst += stride; | |
| 765 | } | ||
| 766 | 140 | } | |
| 767 | |||
| 768 | #if BIT_DEPTH != 12 | ||
| 769 | |||
| 770 | #if BIT_DEPTH == 8 | ||
| 771 | #define memset_bpc memset | ||
| 772 | #else | ||
| 773 | 8040 | static inline void memset_bpc(uint16_t *dst, int val, int len) { | |
| 774 | int n; | ||
| 775 |
2/2✓ Branch 0 taken 45272 times.
✓ Branch 1 taken 8040 times.
|
53312 | for (n = 0; n < len; n++) { |
| 776 | 45272 | dst[n] = val; | |
| 777 | } | ||
| 778 | 8040 | } | |
| 779 | #endif | ||
| 780 | |||
| 781 | #define DST(x, y) dst[(x) + (y) * stride] | ||
| 782 | |||
| 783 | 20364 | static void diag_downleft_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 784 | const uint8_t *left, const uint8_t *_top) | ||
| 785 | { | ||
| 786 | 20364 | pixel *dst = (pixel *) _dst; | |
| 787 | 20364 | const pixel *top = (const pixel *) _top; | |
| 788 | 20364 | int a0 = top[0], a1 = top[1], a2 = top[2], a3 = top[3], | |
| 789 | 20364 | a4 = top[4], a5 = top[5], a6 = top[6], a7 = top[7]; | |
| 790 | |||
| 791 | 20364 | stride /= sizeof(pixel); | |
| 792 | 20364 | DST(0,0) = (a0 + a1 * 2 + a2 + 2) >> 2; | |
| 793 | 20364 | DST(1,0) = DST(0,1) = (a1 + a2 * 2 + a3 + 2) >> 2; | |
| 794 | 20364 | DST(2,0) = DST(1,1) = DST(0,2) = (a2 + a3 * 2 + a4 + 2) >> 2; | |
| 795 | 20364 | DST(3,0) = DST(2,1) = DST(1,2) = DST(0,3) = (a3 + a4 * 2 + a5 + 2) >> 2; | |
| 796 | 20364 | DST(3,1) = DST(2,2) = DST(1,3) = (a4 + a5 * 2 + a6 + 2) >> 2; | |
| 797 | 20364 | DST(3,2) = DST(2,3) = (a5 + a6 * 2 + a7 + 2) >> 2; | |
| 798 | 20364 | DST(3,3) = a7; // note: this is different from vp8 and such | |
| 799 | 20364 | } | |
| 800 | |||
| 801 | #define def_diag_downleft(size) \ | ||
| 802 | static void diag_downleft_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \ | ||
| 803 | const uint8_t *left, const uint8_t *_top) \ | ||
| 804 | { \ | ||
| 805 | pixel *dst = (pixel *) _dst; \ | ||
| 806 | const pixel *top = (const pixel *) _top; \ | ||
| 807 | int i, j; \ | ||
| 808 | pixel v[size - 1]; \ | ||
| 809 | \ | ||
| 810 | stride /= sizeof(pixel); \ | ||
| 811 | for (i = 0; i < size - 2; i++) \ | ||
| 812 | v[i] = (top[i] + top[i + 1] * 2 + top[i + 2] + 2) >> 2; \ | ||
| 813 | v[size - 2] = (top[size - 2] + top[size - 1] * 3 + 2) >> 2; \ | ||
| 814 | \ | ||
| 815 | for (j = 0; j < size; j++) { \ | ||
| 816 | memcpy(dst + j*stride, v + j, (size - 1 - j) * sizeof(pixel)); \ | ||
| 817 | memset_bpc(dst + j*stride + size - 1 - j, top[size - 1], j + 1); \ | ||
| 818 | } \ | ||
| 819 | } | ||
| 820 | |||
| 821 |
5/5✓ Branch 0 taken 34014 times.
✓ Branch 1 taken 5669 times.
✓ Branch 2 taken 44112 times.
✓ Branch 3 taken 6754 times.
✓ Branch 4 taken 155 times.
|
85035 | def_diag_downleft(8) |
| 822 |
5/5✓ Branch 0 taken 15792 times.
✓ Branch 1 taken 1128 times.
✓ Branch 2 taken 17856 times.
✓ Branch 3 taken 1308 times.
✓ Branch 4 taken 12 times.
|
34968 | def_diag_downleft(16) |
| 823 |
5/5✓ Branch 0 taken 4560 times.
✓ Branch 1 taken 152 times.
✓ Branch 2 taken 4288 times.
✓ Branch 3 taken 710 times.
✓ Branch 4 taken 18 times.
|
9576 | def_diag_downleft(32) |
| 824 | |||
| 825 | 57827 | static void diag_downright_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 826 | const uint8_t *_left, const uint8_t *_top) | ||
| 827 | { | ||
| 828 | 57827 | pixel *dst = (pixel *) _dst; | |
| 829 | 57827 | const pixel *top = (const pixel *) _top; | |
| 830 | 57827 | const pixel *left = (const pixel *) _left; | |
| 831 | 57827 | int tl = top[-1], a0 = top[0], a1 = top[1], a2 = top[2], a3 = top[3], | |
| 832 | 57827 | l0 = left[3], l1 = left[2], l2 = left[1], l3 = left[0]; | |
| 833 | |||
| 834 | 57827 | stride /= sizeof(pixel); | |
| 835 | 57827 | DST(0,3) = (l1 + l2 * 2 + l3 + 2) >> 2; | |
| 836 | 57827 | DST(0,2) = DST(1,3) = (l0 + l1 * 2 + l2 + 2) >> 2; | |
| 837 | 57827 | DST(0,1) = DST(1,2) = DST(2,3) = (tl + l0 * 2 + l1 + 2) >> 2; | |
| 838 | 57827 | DST(0,0) = DST(1,1) = DST(2,2) = DST(3,3) = (l0 + tl * 2 + a0 + 2) >> 2; | |
| 839 | 57827 | DST(1,0) = DST(2,1) = DST(3,2) = (tl + a0 * 2 + a1 + 2) >> 2; | |
| 840 | 57827 | DST(2,0) = DST(3,1) = (a0 + a1 * 2 + a2 + 2) >> 2; | |
| 841 | 57827 | DST(3,0) = (a1 + a2 * 2 + a3 + 2) >> 2; | |
| 842 | 57827 | } | |
| 843 | |||
| 844 | #define def_diag_downright(size) \ | ||
| 845 | static void diag_downright_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \ | ||
| 846 | const uint8_t *_left, const uint8_t *_top) \ | ||
| 847 | { \ | ||
| 848 | pixel *dst = (pixel *) _dst; \ | ||
| 849 | const pixel *top = (const pixel *) _top; \ | ||
| 850 | const pixel *left = (const pixel *) _left; \ | ||
| 851 | int i, j; \ | ||
| 852 | pixel v[size + size - 1]; \ | ||
| 853 | \ | ||
| 854 | stride /= sizeof(pixel); \ | ||
| 855 | for (i = 0; i < size - 2; i++) { \ | ||
| 856 | v[i ] = (left[i] + left[i + 1] * 2 + left[i + 2] + 2) >> 2; \ | ||
| 857 | v[size + 1 + i] = (top[i] + top[i + 1] * 2 + top[i + 2] + 2) >> 2; \ | ||
| 858 | } \ | ||
| 859 | v[size - 2] = (left[size - 2] + left[size - 1] * 2 + top[-1] + 2) >> 2; \ | ||
| 860 | v[size - 1] = (left[size - 1] + top[-1] * 2 + top[ 0] + 2) >> 2; \ | ||
| 861 | v[size ] = (top[-1] + top[0] * 2 + top[ 1] + 2) >> 2; \ | ||
| 862 | \ | ||
| 863 | for (j = 0; j < size; j++) \ | ||
| 864 | memcpy(dst + j*stride, v + size - 1 - j, size * sizeof(pixel)); \ | ||
| 865 | } | ||
| 866 | |||
| 867 |
4/4✓ Branch 0 taken 41100 times.
✓ Branch 1 taken 6850 times.
✓ Branch 2 taken 54800 times.
✓ Branch 3 taken 6850 times.
|
102750 | def_diag_downright(8) |
| 868 |
4/4✓ Branch 0 taken 12754 times.
✓ Branch 1 taken 911 times.
✓ Branch 2 taken 14576 times.
✓ Branch 3 taken 911 times.
|
28241 | def_diag_downright(16) |
| 869 |
4/4✓ Branch 0 taken 5100 times.
✓ Branch 1 taken 170 times.
✓ Branch 2 taken 5440 times.
✓ Branch 3 taken 170 times.
|
10710 | def_diag_downright(32) |
| 870 | |||
| 871 | 44165 | static void vert_right_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 872 | const uint8_t *_left, const uint8_t *_top) | ||
| 873 | { | ||
| 874 | 44165 | pixel *dst = (pixel *) _dst; | |
| 875 | 44165 | const pixel *top = (const pixel *) _top; | |
| 876 | 44165 | const pixel *left = (const pixel *) _left; | |
| 877 | 44165 | int tl = top[-1], a0 = top[0], a1 = top[1], a2 = top[2], a3 = top[3], | |
| 878 | 44165 | l0 = left[3], l1 = left[2], l2 = left[1]; | |
| 879 | |||
| 880 | 44165 | stride /= sizeof(pixel); | |
| 881 | 44165 | DST(0,3) = (l0 + l1 * 2 + l2 + 2) >> 2; | |
| 882 | 44165 | DST(0,2) = (tl + l0 * 2 + l1 + 2) >> 2; | |
| 883 | 44165 | DST(0,0) = DST(1,2) = (tl + a0 + 1) >> 1; | |
| 884 | 44165 | DST(0,1) = DST(1,3) = (l0 + tl * 2 + a0 + 2) >> 2; | |
| 885 | 44165 | DST(1,0) = DST(2,2) = (a0 + a1 + 1) >> 1; | |
| 886 | 44165 | DST(1,1) = DST(2,3) = (tl + a0 * 2 + a1 + 2) >> 2; | |
| 887 | 44165 | DST(2,0) = DST(3,2) = (a1 + a2 + 1) >> 1; | |
| 888 | 44165 | DST(2,1) = DST(3,3) = (a0 + a1 * 2 + a2 + 2) >> 2; | |
| 889 | 44165 | DST(3,0) = (a2 + a3 + 1) >> 1; | |
| 890 | 44165 | DST(3,1) = (a1 + a2 * 2 + a3 + 2) >> 2; | |
| 891 | 44165 | } | |
| 892 | |||
| 893 | #define def_vert_right(size) \ | ||
| 894 | static void vert_right_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \ | ||
| 895 | const uint8_t *_left, const uint8_t *_top) \ | ||
| 896 | { \ | ||
| 897 | pixel *dst = (pixel *) _dst; \ | ||
| 898 | const pixel *top = (const pixel *) _top; \ | ||
| 899 | const pixel *left = (const pixel *) _left; \ | ||
| 900 | int i, j; \ | ||
| 901 | pixel ve[size + size/2 - 1], vo[size + size/2 - 1]; \ | ||
| 902 | \ | ||
| 903 | stride /= sizeof(pixel); \ | ||
| 904 | for (i = 0; i < size/2 - 2; i++) { \ | ||
| 905 | vo[i] = (left[i*2 + 3] + left[i*2 + 2] * 2 + left[i*2 + 1] + 2) >> 2; \ | ||
| 906 | ve[i] = (left[i*2 + 4] + left[i*2 + 3] * 2 + left[i*2 + 2] + 2) >> 2; \ | ||
| 907 | } \ | ||
| 908 | vo[size/2 - 2] = (left[size - 1] + left[size - 2] * 2 + left[size - 3] + 2) >> 2; \ | ||
| 909 | ve[size/2 - 2] = (top[-1] + left[size - 1] * 2 + left[size - 2] + 2) >> 2; \ | ||
| 910 | \ | ||
| 911 | ve[size/2 - 1] = (top[-1] + top[0] + 1) >> 1; \ | ||
| 912 | vo[size/2 - 1] = (left[size - 1] + top[-1] * 2 + top[0] + 2) >> 2; \ | ||
| 913 | for (i = 0; i < size - 1; i++) { \ | ||
| 914 | ve[size/2 + i] = (top[i] + top[i + 1] + 1) >> 1; \ | ||
| 915 | vo[size/2 + i] = (top[i - 1] + top[i] * 2 + top[i + 1] + 2) >> 2; \ | ||
| 916 | } \ | ||
| 917 | \ | ||
| 918 | for (j = 0; j < size / 2; j++) { \ | ||
| 919 | memcpy(dst + j*2 *stride, ve + size/2 - 1 - j, size * sizeof(pixel)); \ | ||
| 920 | memcpy(dst + (j*2 + 1)*stride, vo + size/2 - 1 - j, size * sizeof(pixel)); \ | ||
| 921 | } \ | ||
| 922 | } | ||
| 923 | |||
| 924 |
6/6✓ Branch 0 taken 9738 times.
✓ Branch 1 taken 4869 times.
✓ Branch 2 taken 34083 times.
✓ Branch 3 taken 4869 times.
✓ Branch 4 taken 19476 times.
✓ Branch 5 taken 4869 times.
|
68166 | def_vert_right(8) |
| 925 |
6/6✓ Branch 0 taken 5610 times.
✓ Branch 1 taken 935 times.
✓ Branch 2 taken 14025 times.
✓ Branch 3 taken 935 times.
✓ Branch 4 taken 7480 times.
✓ Branch 5 taken 935 times.
|
28050 | def_vert_right(16) |
| 926 |
6/6✓ Branch 0 taken 2646 times.
✓ Branch 1 taken 189 times.
✓ Branch 2 taken 5859 times.
✓ Branch 3 taken 189 times.
✓ Branch 4 taken 3024 times.
✓ Branch 5 taken 189 times.
|
11718 | def_vert_right(32) |
| 927 | |||
| 928 | 47549 | static void hor_down_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 929 | const uint8_t *_left, const uint8_t *_top) | ||
| 930 | { | ||
| 931 | 47549 | pixel *dst = (pixel *) _dst; | |
| 932 | 47549 | const pixel *top = (const pixel *) _top; | |
| 933 | 47549 | const pixel *left = (const pixel *) _left; | |
| 934 | 47549 | int l0 = left[3], l1 = left[2], l2 = left[1], l3 = left[0], | |
| 935 | 47549 | tl = top[-1], a0 = top[0], a1 = top[1], a2 = top[2]; | |
| 936 | |||
| 937 | 47549 | stride /= sizeof(pixel); | |
| 938 | 47549 | DST(2,0) = (tl + a0 * 2 + a1 + 2) >> 2; | |
| 939 | 47549 | DST(3,0) = (a0 + a1 * 2 + a2 + 2) >> 2; | |
| 940 | 47549 | DST(0,0) = DST(2,1) = (tl + l0 + 1) >> 1; | |
| 941 | 47549 | DST(1,0) = DST(3,1) = (a0 + tl * 2 + l0 + 2) >> 2; | |
| 942 | 47549 | DST(0,1) = DST(2,2) = (l0 + l1 + 1) >> 1; | |
| 943 | 47549 | DST(1,1) = DST(3,2) = (tl + l0 * 2 + l1 + 2) >> 2; | |
| 944 | 47549 | DST(0,2) = DST(2,3) = (l1 + l2 + 1) >> 1; | |
| 945 | 47549 | DST(1,2) = DST(3,3) = (l0 + l1 * 2 + l2 + 2) >> 2; | |
| 946 | 47549 | DST(0,3) = (l2 + l3 + 1) >> 1; | |
| 947 | 47549 | DST(1,3) = (l1 + l2 * 2 + l3 + 2) >> 2; | |
| 948 | 47549 | } | |
| 949 | |||
| 950 | #define def_hor_down(size) \ | ||
| 951 | static void hor_down_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \ | ||
| 952 | const uint8_t *_left, const uint8_t *_top) \ | ||
| 953 | { \ | ||
| 954 | pixel *dst = (pixel *) _dst; \ | ||
| 955 | const pixel *top = (const pixel *) _top; \ | ||
| 956 | const pixel *left = (const pixel *) _left; \ | ||
| 957 | int i, j; \ | ||
| 958 | pixel v[size * 3 - 2]; \ | ||
| 959 | \ | ||
| 960 | stride /= sizeof(pixel); \ | ||
| 961 | for (i = 0; i < size - 2; i++) { \ | ||
| 962 | v[i*2 ] = (left[i + 1] + left[i + 0] + 1) >> 1; \ | ||
| 963 | v[i*2 + 1] = (left[i + 2] + left[i + 1] * 2 + left[i + 0] + 2) >> 2; \ | ||
| 964 | v[size*2 + i] = (top[i - 1] + top[i] * 2 + top[i + 1] + 2) >> 2; \ | ||
| 965 | } \ | ||
| 966 | v[size*2 - 2] = (top[-1] + left[size - 1] + 1) >> 1; \ | ||
| 967 | v[size*2 - 4] = (left[size - 1] + left[size - 2] + 1) >> 1; \ | ||
| 968 | v[size*2 - 1] = (top[0] + top[-1] * 2 + left[size - 1] + 2) >> 2; \ | ||
| 969 | v[size*2 - 3] = (top[-1] + left[size - 1] * 2 + left[size - 2] + 2) >> 2; \ | ||
| 970 | \ | ||
| 971 | for (j = 0; j < size; j++) \ | ||
| 972 | memcpy(dst + j*stride, v + size*2 - 2 - j*2, size * sizeof(pixel)); \ | ||
| 973 | } | ||
| 974 | |||
| 975 |
4/4✓ Branch 0 taken 30348 times.
✓ Branch 1 taken 5058 times.
✓ Branch 2 taken 40464 times.
✓ Branch 3 taken 5058 times.
|
75870 | def_hor_down(8) |
| 976 |
4/4✓ Branch 0 taken 8148 times.
✓ Branch 1 taken 582 times.
✓ Branch 2 taken 9312 times.
✓ Branch 3 taken 582 times.
|
18042 | def_hor_down(16) |
| 977 |
4/4✓ Branch 0 taken 2310 times.
✓ Branch 1 taken 77 times.
✓ Branch 2 taken 2464 times.
✓ Branch 3 taken 77 times.
|
4851 | def_hor_down(32) |
| 978 | |||
| 979 | 36962 | static void vert_left_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 980 | const uint8_t *left, const uint8_t *_top) | ||
| 981 | { | ||
| 982 | 36962 | pixel *dst = (pixel *) _dst; | |
| 983 | 36962 | const pixel *top = (const pixel *) _top; | |
| 984 | 36962 | int a0 = top[0], a1 = top[1], a2 = top[2], a3 = top[3], | |
| 985 | 36962 | a4 = top[4], a5 = top[5], a6 = top[6]; | |
| 986 | |||
| 987 | 36962 | stride /= sizeof(pixel); | |
| 988 | 36962 | DST(0,0) = (a0 + a1 + 1) >> 1; | |
| 989 | 36962 | DST(0,1) = (a0 + a1 * 2 + a2 + 2) >> 2; | |
| 990 | 36962 | DST(1,0) = DST(0,2) = (a1 + a2 + 1) >> 1; | |
| 991 | 36962 | DST(1,1) = DST(0,3) = (a1 + a2 * 2 + a3 + 2) >> 2; | |
| 992 | 36962 | DST(2,0) = DST(1,2) = (a2 + a3 + 1) >> 1; | |
| 993 | 36962 | DST(2,1) = DST(1,3) = (a2 + a3 * 2 + a4 + 2) >> 2; | |
| 994 | 36962 | DST(3,0) = DST(2,2) = (a3 + a4 + 1) >> 1; | |
| 995 | 36962 | DST(3,1) = DST(2,3) = (a3 + a4 * 2 + a5 + 2) >> 2; | |
| 996 | 36962 | DST(3,2) = (a4 + a5 + 1) >> 1; | |
| 997 | 36962 | DST(3,3) = (a4 + a5 * 2 + a6 + 2) >> 2; | |
| 998 | 36962 | } | |
| 999 | |||
| 1000 | #define def_vert_left(size) \ | ||
| 1001 | static void vert_left_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \ | ||
| 1002 | const uint8_t *left, const uint8_t *_top) \ | ||
| 1003 | { \ | ||
| 1004 | pixel *dst = (pixel *) _dst; \ | ||
| 1005 | const pixel *top = (const pixel *) _top; \ | ||
| 1006 | int i, j; \ | ||
| 1007 | pixel ve[size - 1], vo[size - 1]; \ | ||
| 1008 | \ | ||
| 1009 | stride /= sizeof(pixel); \ | ||
| 1010 | for (i = 0; i < size - 2; i++) { \ | ||
| 1011 | ve[i] = (top[i] + top[i + 1] + 1) >> 1; \ | ||
| 1012 | vo[i] = (top[i] + top[i + 1] * 2 + top[i + 2] + 2) >> 2; \ | ||
| 1013 | } \ | ||
| 1014 | ve[size - 2] = (top[size - 2] + top[size - 1] + 1) >> 1; \ | ||
| 1015 | vo[size - 2] = (top[size - 2] + top[size - 1] * 3 + 2) >> 2; \ | ||
| 1016 | \ | ||
| 1017 | for (j = 0; j < size / 2; j++) { \ | ||
| 1018 | memcpy(dst + j*2 * stride, ve + j, (size - j - 1) * sizeof(pixel)); \ | ||
| 1019 | memset_bpc(dst + j*2 * stride + size - j - 1, top[size - 1], j + 1); \ | ||
| 1020 | memcpy(dst + (j*2 + 1) * stride, vo + j, (size - j - 1) * sizeof(pixel)); \ | ||
| 1021 | memset_bpc(dst + (j*2 + 1) * stride + size - j - 1, top[size - 1], j + 1); \ | ||
| 1022 | } \ | ||
| 1023 | } | ||
| 1024 | |||
| 1025 |
6/6✓ Branch 0 taken 39852 times.
✓ Branch 1 taken 6642 times.
✓ Branch 2 taken 25396 times.
✓ Branch 3 taken 6349 times.
✓ Branch 4 taken 1172 times.
✓ Branch 5 taken 293 times.
|
73062 | def_vert_left(8) |
| 1026 |
6/6✓ Branch 0 taken 19390 times.
✓ Branch 1 taken 1385 times.
✓ Branch 2 taken 10376 times.
✓ Branch 3 taken 1297 times.
✓ Branch 4 taken 704 times.
✓ Branch 5 taken 88 times.
|
31855 | def_vert_left(16) |
| 1027 |
6/6✓ Branch 0 taken 5760 times.
✓ Branch 1 taken 192 times.
✓ Branch 2 taken 2912 times.
✓ Branch 3 taken 182 times.
✓ Branch 4 taken 160 times.
✓ Branch 5 taken 10 times.
|
9024 | def_vert_left(32) |
| 1028 | |||
| 1029 | 51466 | static void hor_up_4x4_c(uint8_t *_dst, ptrdiff_t stride, | |
| 1030 | const uint8_t *_left, const uint8_t *top) | ||
| 1031 | { | ||
| 1032 | 51466 | pixel *dst = (pixel *) _dst; | |
| 1033 | 51466 | const pixel *left = (const pixel *) _left; | |
| 1034 | 51466 | int l0 = left[0], l1 = left[1], l2 = left[2], l3 = left[3]; | |
| 1035 | |||
| 1036 | 51466 | stride /= sizeof(pixel); | |
| 1037 | 51466 | DST(0,0) = (l0 + l1 + 1) >> 1; | |
| 1038 | 51466 | DST(1,0) = (l0 + l1 * 2 + l2 + 2) >> 2; | |
| 1039 | 51466 | DST(0,1) = DST(2,0) = (l1 + l2 + 1) >> 1; | |
| 1040 | 51466 | DST(1,1) = DST(3,0) = (l1 + l2 * 2 + l3 + 2) >> 2; | |
| 1041 | 51466 | DST(0,2) = DST(2,1) = (l2 + l3 + 1) >> 1; | |
| 1042 | 51466 | DST(1,2) = DST(3,1) = (l2 + l3 * 3 + 2) >> 2; | |
| 1043 | 51466 | DST(0,3) = DST(1,3) = DST(2,2) = DST(2,3) = DST(3,2) = DST(3,3) = l3; | |
| 1044 | 51466 | } | |
| 1045 | |||
| 1046 | #define def_hor_up(size) \ | ||
| 1047 | static void hor_up_##size##x##size##_c(uint8_t *_dst, ptrdiff_t stride, \ | ||
| 1048 | const uint8_t *_left, const uint8_t *top) \ | ||
| 1049 | { \ | ||
| 1050 | pixel *dst = (pixel *) _dst; \ | ||
| 1051 | const pixel *left = (const pixel *) _left; \ | ||
| 1052 | int i, j; \ | ||
| 1053 | pixel v[size*2 - 2]; \ | ||
| 1054 | \ | ||
| 1055 | stride /= sizeof(pixel); \ | ||
| 1056 | for (i = 0; i < size - 2; i++) { \ | ||
| 1057 | v[i*2 ] = (left[i] + left[i + 1] + 1) >> 1; \ | ||
| 1058 | v[i*2 + 1] = (left[i] + left[i + 1] * 2 + left[i + 2] + 2) >> 2; \ | ||
| 1059 | } \ | ||
| 1060 | v[size*2 - 4] = (left[size - 2] + left[size - 1] + 1) >> 1; \ | ||
| 1061 | v[size*2 - 3] = (left[size - 2] + left[size - 1] * 3 + 2) >> 2; \ | ||
| 1062 | \ | ||
| 1063 | for (j = 0; j < size / 2; j++) \ | ||
| 1064 | memcpy(dst + j*stride, v + j*2, size * sizeof(pixel)); \ | ||
| 1065 | for (j = size / 2; j < size; j++) { \ | ||
| 1066 | memcpy(dst + j*stride, v + j*2, (size*2 - 2 - j*2) * sizeof(pixel)); \ | ||
| 1067 | memset_bpc(dst + j*stride + size*2 - 2 - j*2, left[size - 1], \ | ||
| 1068 | 2 + j*2 - size); \ | ||
| 1069 | } \ | ||
| 1070 | } | ||
| 1071 | |||
| 1072 |
7/7✓ Branch 0 taken 58146 times.
✓ Branch 1 taken 9691 times.
✓ Branch 2 taken 38764 times.
✓ Branch 3 taken 9691 times.
✓ Branch 4 taken 37252 times.
✓ Branch 5 taken 10825 times.
✓ Branch 6 taken 378 times.
|
145365 | def_hor_up(8) |
| 1073 |
7/7✓ Branch 0 taken 24976 times.
✓ Branch 1 taken 1784 times.
✓ Branch 2 taken 14272 times.
✓ Branch 3 taken 1784 times.
✓ Branch 4 taken 14080 times.
✓ Branch 5 taken 1952 times.
✓ Branch 6 taken 24 times.
|
55304 | def_hor_up(16) |
| 1074 |
7/7✓ Branch 0 taken 5040 times.
✓ Branch 1 taken 168 times.
✓ Branch 2 taken 2688 times.
✓ Branch 3 taken 168 times.
✓ Branch 4 taken 2432 times.
✓ Branch 5 taken 408 times.
✓ Branch 6 taken 16 times.
|
10584 | def_hor_up(32) |
| 1075 | |||
| 1076 | #undef DST | ||
| 1077 | |||
| 1078 | #endif /* BIT_DEPTH != 12 */ | ||
| 1079 | |||
| 1080 | #if BIT_DEPTH != 8 | ||
| 1081 | void ff_vp9dsp_intrapred_init_10(VP9DSPContext *dsp); | ||
| 1082 | #endif | ||
| 1083 | #if BIT_DEPTH != 10 | ||
| 1084 | static | ||
| 1085 | #endif | ||
| 1086 | 746 | av_cold void FUNC(ff_vp9dsp_intrapred_init)(VP9DSPContext *dsp) | |
| 1087 | { | ||
| 1088 | #define init_intra_pred_bd_aware(tx, sz) \ | ||
| 1089 | dsp->intra_pred[tx][TM_VP8_PRED] = tm_##sz##_c; \ | ||
| 1090 | dsp->intra_pred[tx][DC_128_PRED] = dc_128_##sz##_c; \ | ||
| 1091 | dsp->intra_pred[tx][DC_127_PRED] = dc_127_##sz##_c; \ | ||
| 1092 | dsp->intra_pred[tx][DC_129_PRED] = dc_129_##sz##_c | ||
| 1093 | |||
| 1094 | #if BIT_DEPTH == 12 | ||
| 1095 | 75 | ff_vp9dsp_intrapred_init_10(dsp); | |
| 1096 | #define init_intra_pred(tx, sz) \ | ||
| 1097 | init_intra_pred_bd_aware(tx, sz) | ||
| 1098 | #else | ||
| 1099 | #define init_intra_pred(tx, sz) \ | ||
| 1100 | dsp->intra_pred[tx][VERT_PRED] = vert_##sz##_c; \ | ||
| 1101 | dsp->intra_pred[tx][HOR_PRED] = hor_##sz##_c; \ | ||
| 1102 | dsp->intra_pred[tx][DC_PRED] = dc_##sz##_c; \ | ||
| 1103 | dsp->intra_pred[tx][DIAG_DOWN_LEFT_PRED] = diag_downleft_##sz##_c; \ | ||
| 1104 | dsp->intra_pred[tx][DIAG_DOWN_RIGHT_PRED] = diag_downright_##sz##_c; \ | ||
| 1105 | dsp->intra_pred[tx][VERT_RIGHT_PRED] = vert_right_##sz##_c; \ | ||
| 1106 | dsp->intra_pred[tx][HOR_DOWN_PRED] = hor_down_##sz##_c; \ | ||
| 1107 | dsp->intra_pred[tx][VERT_LEFT_PRED] = vert_left_##sz##_c; \ | ||
| 1108 | dsp->intra_pred[tx][HOR_UP_PRED] = hor_up_##sz##_c; \ | ||
| 1109 | dsp->intra_pred[tx][LEFT_DC_PRED] = dc_left_##sz##_c; \ | ||
| 1110 | dsp->intra_pred[tx][TOP_DC_PRED] = dc_top_##sz##_c; \ | ||
| 1111 | init_intra_pred_bd_aware(tx, sz) | ||
| 1112 | #endif | ||
| 1113 | |||
| 1114 | 746 | init_intra_pred(TX_4X4, 4x4); | |
| 1115 | 746 | init_intra_pred(TX_8X8, 8x8); | |
| 1116 | 746 | init_intra_pred(TX_16X16, 16x16); | |
| 1117 | 746 | init_intra_pred(TX_32X32, 32x32); | |
| 1118 | |||
| 1119 | #undef init_intra_pred | ||
| 1120 | #undef init_intra_pred_bd_aware | ||
| 1121 | 746 | } | |
| 1122 | |||
| 1123 | #define itxfm_wrapper(type_a, type_b, sz, bits, has_dconly) \ | ||
| 1124 | static void type_a##_##type_b##_##sz##x##sz##_add_c(uint8_t *_dst, \ | ||
| 1125 | ptrdiff_t stride, \ | ||
| 1126 | int16_t *_block, int eob) \ | ||
| 1127 | { \ | ||
| 1128 | int i, j; \ | ||
| 1129 | pixel *dst = (pixel *) _dst; \ | ||
| 1130 | dctcoef *block = (dctcoef *) _block, tmp[sz * sz], out[sz]; \ | ||
| 1131 | \ | ||
| 1132 | stride /= sizeof(pixel); \ | ||
| 1133 | if (has_dconly && eob == 1) { \ | ||
| 1134 | const int t = ((((dctint) block[0] * 11585 + (1 << 13)) >> 14) \ | ||
| 1135 | * 11585 + (1 << 13)) >> 14; \ | ||
| 1136 | block[0] = 0; \ | ||
| 1137 | for (i = 0; i < sz; i++) { \ | ||
| 1138 | for (j = 0; j < sz; j++) \ | ||
| 1139 | dst[j * stride] = av_clip_pixel(dst[j * stride] + \ | ||
| 1140 | (bits ? \ | ||
| 1141 | (int)(t + (1U << (bits - 1))) >> bits : \ | ||
| 1142 | t)); \ | ||
| 1143 | dst++; \ | ||
| 1144 | } \ | ||
| 1145 | return; \ | ||
| 1146 | } \ | ||
| 1147 | \ | ||
| 1148 | for (i = 0; i < sz; i++) \ | ||
| 1149 | type_a##sz##_1d(block + i, sz, tmp + i * sz, 0); \ | ||
| 1150 | memset(block, 0, sz * sz * sizeof(*block)); \ | ||
| 1151 | for (i = 0; i < sz; i++) { \ | ||
| 1152 | type_b##sz##_1d(tmp + i, sz, out, 1); \ | ||
| 1153 | for (j = 0; j < sz; j++) \ | ||
| 1154 | dst[j * stride] = av_clip_pixel(dst[j * stride] + \ | ||
| 1155 | (bits ? \ | ||
| 1156 | (int)(out[j] + (1U << (bits - 1))) >> bits : \ | ||
| 1157 | out[j])); \ | ||
| 1158 | dst++; \ | ||
| 1159 | } \ | ||
| 1160 | } | ||
| 1161 | |||
| 1162 | #define itxfm_wrap(sz, bits) \ | ||
| 1163 | itxfm_wrapper(idct, idct, sz, bits, 1) \ | ||
| 1164 | itxfm_wrapper(iadst, idct, sz, bits, 0) \ | ||
| 1165 | itxfm_wrapper(idct, iadst, sz, bits, 0) \ | ||
| 1166 | itxfm_wrapper(iadst, iadst, sz, bits, 0) | ||
| 1167 | |||
| 1168 | #define IN(x) ((dctint) in[(x) * stride]) | ||
| 1169 | |||
| 1170 | 3507364 | static av_always_inline void idct4_1d(const dctcoef *in, ptrdiff_t stride, | |
| 1171 | dctcoef *out, int pass) | ||
| 1172 | { | ||
| 1173 | dctint t0, t1, t2, t3; | ||
| 1174 | |||
| 1175 | 3507364 | t0 = ((IN(0) + IN(2)) * 11585 + (1 << 13)) >> 14; | |
| 1176 | 3507364 | t1 = ((IN(0) - IN(2)) * 11585 + (1 << 13)) >> 14; | |
| 1177 | 3507364 | t2 = (IN(1) * 6270 - IN(3) * 15137 + (1 << 13)) >> 14; | |
| 1178 | 3507364 | t3 = (IN(1) * 15137 + IN(3) * 6270 + (1 << 13)) >> 14; | |
| 1179 | |||
| 1180 | 3507364 | out[0] = t0 + t3; | |
| 1181 | 3507364 | out[1] = t1 + t2; | |
| 1182 | 3507364 | out[2] = t1 - t2; | |
| 1183 | 3507364 | out[3] = t0 - t3; | |
| 1184 | 3507364 | } | |
| 1185 | |||
| 1186 | 1519020 | static av_always_inline void iadst4_1d(const dctcoef *in, ptrdiff_t stride, | |
| 1187 | dctcoef *out, int pass) | ||
| 1188 | { | ||
| 1189 | dctint t0, t1, t2, t3; | ||
| 1190 | |||
| 1191 | 1519020 | t0 = 5283 * IN(0) + 15212 * IN(2) + 9929 * IN(3); | |
| 1192 | 1519020 | t1 = 9929 * IN(0) - 5283 * IN(2) - 15212 * IN(3); | |
| 1193 | 1519020 | t2 = 13377 * (IN(0) - IN(2) + IN(3)); | |
| 1194 | 1519020 | t3 = 13377 * IN(1); | |
| 1195 | |||
| 1196 | 1519020 | out[0] = (t0 + t3 + (1 << 13)) >> 14; | |
| 1197 | 1519020 | out[1] = (t1 + t3 + (1 << 13)) >> 14; | |
| 1198 | 1519020 | out[2] = (t2 + (1 << 13)) >> 14; | |
| 1199 | 1519020 | out[3] = (t0 + t1 - t3 + (1 << 13)) >> 14; | |
| 1200 | 1519020 | } | |
| 1201 | |||
| 1202 |
13/13✓ Branch 0 taken 113652 times.
✓ Branch 1 taken 1514444 times.
✓ Branch 2 taken 2113814 times.
✓ Branch 3 taken 454608 times.
✓ Branch 4 taken 5180720 times.
✓ Branch 5 taken 1295180 times.
✓ Branch 6 taken 1181528 times.
✓ Branch 7 taken 1627046 times.
✓ Branch 8 taken 332916 times.
✓ Branch 10 taken 5326656 times.
✓ Branch 11 taken 1331664 times.
✓ Branch 12 taken 1331664 times.
✓ Branch 13 taken 332916 times.
|
36188284 | itxfm_wrap(4, 4) |
| 1203 | |||
| 1204 | 2969888 | static av_always_inline void idct8_1d(const dctcoef *in, ptrdiff_t stride, | |
| 1205 | dctcoef *out, int pass) | ||
| 1206 | { | ||
| 1207 | dctint t0, t0a, t1, t1a, t2, t2a, t3, t3a, t4, t4a, t5, t5a, t6, t6a, t7, t7a; | ||
| 1208 | |||
| 1209 | 2969888 | t0a = ((IN(0) + IN(4)) * 11585 + (1 << 13)) >> 14; | |
| 1210 | 2969888 | t1a = ((IN(0) - IN(4)) * 11585 + (1 << 13)) >> 14; | |
| 1211 | 2969888 | t2a = (IN(2) * 6270 - IN(6) * 15137 + (1 << 13)) >> 14; | |
| 1212 | 2969888 | t3a = (IN(2) * 15137 + IN(6) * 6270 + (1 << 13)) >> 14; | |
| 1213 | 2969888 | t4a = (IN(1) * 3196 - IN(7) * 16069 + (1 << 13)) >> 14; | |
| 1214 | 2969888 | t5a = (IN(5) * 13623 - IN(3) * 9102 + (1 << 13)) >> 14; | |
| 1215 | 2969888 | t6a = (IN(5) * 9102 + IN(3) * 13623 + (1 << 13)) >> 14; | |
| 1216 | 2969888 | t7a = (IN(1) * 16069 + IN(7) * 3196 + (1 << 13)) >> 14; | |
| 1217 | |||
| 1218 | 2969888 | t0 = t0a + t3a; | |
| 1219 | 2969888 | t1 = t1a + t2a; | |
| 1220 | 2969888 | t2 = t1a - t2a; | |
| 1221 | 2969888 | t3 = t0a - t3a; | |
| 1222 | 2969888 | t4 = t4a + t5a; | |
| 1223 | 2969888 | t5a = t4a - t5a; | |
| 1224 | 2969888 | t7 = t7a + t6a; | |
| 1225 | 2969888 | t6a = t7a - t6a; | |
| 1226 | |||
| 1227 | 2969888 | t5 = ((t6a - t5a) * 11585 + (1 << 13)) >> 14; | |
| 1228 | 2969888 | t6 = ((t6a + t5a) * 11585 + (1 << 13)) >> 14; | |
| 1229 | |||
| 1230 | 2969888 | out[0] = t0 + t7; | |
| 1231 | 2969888 | out[1] = t1 + t6; | |
| 1232 | 2969888 | out[2] = t2 + t5; | |
| 1233 | 2969888 | out[3] = t3 + t4; | |
| 1234 | 2969888 | out[4] = t3 - t4; | |
| 1235 | 2969888 | out[5] = t2 - t5; | |
| 1236 | 2969888 | out[6] = t1 - t6; | |
| 1237 | 2969888 | out[7] = t0 - t7; | |
| 1238 | 2969888 | } | |
| 1239 | |||
| 1240 | 761408 | static av_always_inline void iadst8_1d(const dctcoef *in, ptrdiff_t stride, | |
| 1241 | dctcoef *out, int pass) | ||
| 1242 | { | ||
| 1243 | dctint t0, t0a, t1, t1a, t2, t2a, t3, t3a, t4, t4a, t5, t5a, t6, t6a, t7, t7a; | ||
| 1244 | |||
| 1245 | 761408 | t0a = 16305 * IN(7) + 1606 * IN(0); | |
| 1246 | 761408 | t1a = 1606 * IN(7) - 16305 * IN(0); | |
| 1247 | 761408 | t2a = 14449 * IN(5) + 7723 * IN(2); | |
| 1248 | 761408 | t3a = 7723 * IN(5) - 14449 * IN(2); | |
| 1249 | 761408 | t4a = 10394 * IN(3) + 12665 * IN(4); | |
| 1250 | 761408 | t5a = 12665 * IN(3) - 10394 * IN(4); | |
| 1251 | 761408 | t6a = 4756 * IN(1) + 15679 * IN(6); | |
| 1252 | 761408 | t7a = 15679 * IN(1) - 4756 * IN(6); | |
| 1253 | |||
| 1254 | 761408 | t0 = (t0a + t4a + (1 << 13)) >> 14; | |
| 1255 | 761408 | t1 = (t1a + t5a + (1 << 13)) >> 14; | |
| 1256 | 761408 | t2 = (t2a + t6a + (1 << 13)) >> 14; | |
| 1257 | 761408 | t3 = (t3a + t7a + (1 << 13)) >> 14; | |
| 1258 | 761408 | t4 = (t0a - t4a + (1 << 13)) >> 14; | |
| 1259 | 761408 | t5 = (t1a - t5a + (1 << 13)) >> 14; | |
| 1260 | 761408 | t6 = (t2a - t6a + (1 << 13)) >> 14; | |
| 1261 | 761408 | t7 = (t3a - t7a + (1 << 13)) >> 14; | |
| 1262 | |||
| 1263 | 761408 | t4a = 15137U * t4 + 6270U * t5; | |
| 1264 | 761408 | t5a = 6270U * t4 - 15137U * t5; | |
| 1265 | 761408 | t6a = 15137U * t7 - 6270U * t6; | |
| 1266 | 761408 | t7a = 6270U * t7 + 15137U * t6; | |
| 1267 | |||
| 1268 | 761408 | out[0] = t0 + t2; | |
| 1269 | 761408 | out[7] = -(t1 + t3); | |
| 1270 | 761408 | t2 = t0 - t2; | |
| 1271 | 761408 | t3 = t1 - t3; | |
| 1272 | |||
| 1273 | 761408 | out[1] = -((dctint)((1U << 13) + t4a + t6a) >> 14); | |
| 1274 | 761408 | out[6] = (dctint)((1U << 13) + t5a + t7a) >> 14; | |
| 1275 | 761408 | t6 = (dctint)((1U << 13) + t4a - t6a) >> 14; | |
| 1276 | 761408 | t7 = (dctint)((1U << 13) + t5a - t7a) >> 14; | |
| 1277 | |||
| 1278 | 761408 | out[3] = -((dctint)((t2 + t3) * 11585U + (1 << 13)) >> 14); | |
| 1279 | 761408 | out[4] = (dctint)((t2 - t3) * 11585U + (1 << 13)) >> 14; | |
| 1280 | 761408 | out[2] = (dctint)((t6 + t7) * 11585U + (1 << 13)) >> 14; | |
| 1281 | 761408 | out[5] = -((dctint)((t6 - t7) * 11585U + (1 << 13)) >> 14); | |
| 1282 | 761408 | } | |
| 1283 | |||
| 1284 |
13/13✓ Branch 0 taken 34570 times.
✓ Branch 1 taken 755189 times.
✓ Branch 2 taken 2287049 times.
✓ Branch 3 taken 276560 times.
✓ Branch 4 taken 5048976 times.
✓ Branch 5 taken 631122 times.
✓ Branch 6 taken 596552 times.
✓ Branch 7 taken 1343665 times.
✓ Branch 8 taken 158637 times.
✓ Branch 10 taken 10152768 times.
✓ Branch 11 taken 1269096 times.
✓ Branch 12 taken 1269096 times.
✓ Branch 13 taken 158637 times.
|
42826592 | itxfm_wrap(8, 5) |
| 1285 | |||
| 1286 | 1130576 | static av_always_inline void idct16_1d(const dctcoef *in, ptrdiff_t stride, | |
| 1287 | dctcoef *out, int pass) | ||
| 1288 | { | ||
| 1289 | dctint t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, t10, t11, t12, t13, t14, t15; | ||
| 1290 | dctint t0a, t1a, t2a, t3a, t4a, t5a, t6a, t7a; | ||
| 1291 | dctint t8a, t9a, t10a, t11a, t12a, t13a, t14a, t15a; | ||
| 1292 | |||
| 1293 | 1130576 | t0a = (dctint)((IN(0) + IN(8)) * 11585U + (1 << 13)) >> 14; | |
| 1294 | 1130576 | t1a = (dctint)((IN(0) - IN(8)) * 11585U + (1 << 13)) >> 14; | |
| 1295 | 1130576 | t2a = (dctint)(IN(4) * 6270U - IN(12) * 15137U + (1 << 13)) >> 14; | |
| 1296 | 1130576 | t3a = (dctint)(IN(4) * 15137U + IN(12) * 6270U + (1 << 13)) >> 14; | |
| 1297 | 1130576 | t4a = (dctint)(IN(2) * 3196U - IN(14) * 16069U + (1 << 13)) >> 14; | |
| 1298 | 1130576 | t7a = (dctint)(IN(2) * 16069U + IN(14) * 3196U + (1 << 13)) >> 14; | |
| 1299 | 1130576 | t5a = (dctint)(IN(10) * 13623U - IN(6) * 9102U + (1 << 13)) >> 14; | |
| 1300 | 1130576 | t6a = (dctint)(IN(10) * 9102U + IN(6) * 13623U + (1 << 13)) >> 14; | |
| 1301 | 1130576 | t8a = (dctint)(IN(1) * 1606U - IN(15) * 16305U + (1 << 13)) >> 14; | |
| 1302 | 1130576 | t15a = (dctint)(IN(1) * 16305U + IN(15) * 1606U + (1 << 13)) >> 14; | |
| 1303 | 1130576 | t9a = (dctint)(IN(9) * 12665U - IN(7) * 10394U + (1 << 13)) >> 14; | |
| 1304 | 1130576 | t14a = (dctint)(IN(9) * 10394U + IN(7) * 12665U + (1 << 13)) >> 14; | |
| 1305 | 1130576 | t10a = (dctint)(IN(5) * 7723U - IN(11) * 14449U + (1 << 13)) >> 14; | |
| 1306 | 1130576 | t13a = (dctint)(IN(5) * 14449U + IN(11) * 7723U + (1 << 13)) >> 14; | |
| 1307 | 1130576 | t11a = (dctint)(IN(13) * 15679U - IN(3) * 4756U + (1 << 13)) >> 14; | |
| 1308 | 1130576 | t12a = (dctint)(IN(13) * 4756U + IN(3) * 15679U + (1 << 13)) >> 14; | |
| 1309 | |||
| 1310 | 1130576 | t0 = t0a + t3a; | |
| 1311 | 1130576 | t1 = t1a + t2a; | |
| 1312 | 1130576 | t2 = t1a - t2a; | |
| 1313 | 1130576 | t3 = t0a - t3a; | |
| 1314 | 1130576 | t4 = t4a + t5a; | |
| 1315 | 1130576 | t5 = t4a - t5a; | |
| 1316 | 1130576 | t6 = t7a - t6a; | |
| 1317 | 1130576 | t7 = t7a + t6a; | |
| 1318 | 1130576 | t8 = t8a + t9a; | |
| 1319 | 1130576 | t9 = t8a - t9a; | |
| 1320 | 1130576 | t10 = t11a - t10a; | |
| 1321 | 1130576 | t11 = t11a + t10a; | |
| 1322 | 1130576 | t12 = t12a + t13a; | |
| 1323 | 1130576 | t13 = t12a - t13a; | |
| 1324 | 1130576 | t14 = t15a - t14a; | |
| 1325 | 1130576 | t15 = t15a + t14a; | |
| 1326 | |||
| 1327 | 1130576 | t5a = (dctint)((t6 - t5) * 11585U + (1 << 13)) >> 14; | |
| 1328 | 1130576 | t6a = (dctint)((t6 + t5) * 11585U + (1 << 13)) >> 14; | |
| 1329 | 1130576 | t9a = (dctint)( t14 * 6270U - t9 * 15137U + (1 << 13)) >> 14; | |
| 1330 | 1130576 | t14a = (dctint)( t14 * 15137U + t9 * 6270U + (1 << 13)) >> 14; | |
| 1331 | 1130576 | t10a = (dctint)(-(t13 * 15137U + t10 * 6270U) + (1 << 13)) >> 14; | |
| 1332 | 1130576 | t13a = (dctint)( t13 * 6270U - t10 * 15137U + (1 << 13)) >> 14; | |
| 1333 | |||
| 1334 | 1130576 | t0a = t0 + t7; | |
| 1335 | 1130576 | t1a = t1 + t6a; | |
| 1336 | 1130576 | t2a = t2 + t5a; | |
| 1337 | 1130576 | t3a = t3 + t4; | |
| 1338 | 1130576 | t4 = t3 - t4; | |
| 1339 | 1130576 | t5 = t2 - t5a; | |
| 1340 | 1130576 | t6 = t1 - t6a; | |
| 1341 | 1130576 | t7 = t0 - t7; | |
| 1342 | 1130576 | t8a = t8 + t11; | |
| 1343 | 1130576 | t9 = t9a + t10a; | |
| 1344 | 1130576 | t10 = t9a - t10a; | |
| 1345 | 1130576 | t11a = t8 - t11; | |
| 1346 | 1130576 | t12a = t15 - t12; | |
| 1347 | 1130576 | t13 = t14a - t13a; | |
| 1348 | 1130576 | t14 = t14a + t13a; | |
| 1349 | 1130576 | t15a = t15 + t12; | |
| 1350 | |||
| 1351 | 1130576 | t10a = (dctint)((t13 - t10) * 11585U + (1 << 13)) >> 14; | |
| 1352 | 1130576 | t13a = (dctint)((t13 + t10) * 11585U + (1 << 13)) >> 14; | |
| 1353 | 1130576 | t11 = (dctint)((t12a - t11a) * 11585U + (1 << 13)) >> 14; | |
| 1354 | 1130576 | t12 = (dctint)((t12a + t11a) * 11585U + (1 << 13)) >> 14; | |
| 1355 | |||
| 1356 | 1130576 | out[ 0] = t0a + t15a; | |
| 1357 | 1130576 | out[ 1] = t1a + t14; | |
| 1358 | 1130576 | out[ 2] = t2a + t13a; | |
| 1359 | 1130576 | out[ 3] = t3a + t12; | |
| 1360 | 1130576 | out[ 4] = t4 + t11; | |
| 1361 | 1130576 | out[ 5] = t5 + t10a; | |
| 1362 | 1130576 | out[ 6] = t6 + t9; | |
| 1363 | 1130576 | out[ 7] = t7 + t8a; | |
| 1364 | 1130576 | out[ 8] = t7 - t8a; | |
| 1365 | 1130576 | out[ 9] = t6 - t9; | |
| 1366 | 1130576 | out[10] = t5 - t10a; | |
| 1367 | 1130576 | out[11] = t4 - t11; | |
| 1368 | 1130576 | out[12] = t3a - t12; | |
| 1369 | 1130576 | out[13] = t2a - t13a; | |
| 1370 | 1130576 | out[14] = t1a - t14; | |
| 1371 | 1130576 | out[15] = t0a - t15a; | |
| 1372 | 1130576 | } | |
| 1373 | |||
| 1374 | 334320 | static av_always_inline void iadst16_1d(const dctcoef *in, ptrdiff_t stride, | |
| 1375 | dctcoef *out, int pass) | ||
| 1376 | { | ||
| 1377 | dctint t0, t1, t2, t3, t4, t5, t6, t7, t8, t9, t10, t11, t12, t13, t14, t15; | ||
| 1378 | dctint t0a, t1a, t2a, t3a, t4a, t5a, t6a, t7a; | ||
| 1379 | dctint t8a, t9a, t10a, t11a, t12a, t13a, t14a, t15a; | ||
| 1380 | |||
| 1381 | 334320 | t0 = IN(15) * 16364U + IN(0) * 804U; | |
| 1382 | 334320 | t1 = IN(15) * 804U - IN(0) * 16364U; | |
| 1383 | 334320 | t2 = IN(13) * 15893U + IN(2) * 3981U; | |
| 1384 | 334320 | t3 = IN(13) * 3981U - IN(2) * 15893U; | |
| 1385 | 334320 | t4 = IN(11) * 14811U + IN(4) * 7005U; | |
| 1386 | 334320 | t5 = IN(11) * 7005U - IN(4) * 14811U; | |
| 1387 | 334320 | t6 = IN(9) * 13160U + IN(6) * 9760U; | |
| 1388 | 334320 | t7 = IN(9) * 9760U - IN(6) * 13160U; | |
| 1389 | 334320 | t8 = IN(7) * 11003U + IN(8) * 12140U; | |
| 1390 | 334320 | t9 = IN(7) * 12140U - IN(8) * 11003U; | |
| 1391 | 334320 | t10 = IN(5) * 8423U + IN(10) * 14053U; | |
| 1392 | 334320 | t11 = IN(5) * 14053U - IN(10) * 8423U; | |
| 1393 | 334320 | t12 = IN(3) * 5520U + IN(12) * 15426U; | |
| 1394 | 334320 | t13 = IN(3) * 15426U - IN(12) * 5520U; | |
| 1395 | 334320 | t14 = IN(1) * 2404U + IN(14) * 16207U; | |
| 1396 | 334320 | t15 = IN(1) * 16207U - IN(14) * 2404U; | |
| 1397 | |||
| 1398 | 334320 | t0a = (dctint)((1U << 13) + t0 + t8 ) >> 14; | |
| 1399 | 334320 | t1a = (dctint)((1U << 13) + t1 + t9 ) >> 14; | |
| 1400 | 334320 | t2a = (dctint)((1U << 13) + t2 + t10) >> 14; | |
| 1401 | 334320 | t3a = (dctint)((1U << 13) + t3 + t11) >> 14; | |
| 1402 | 334320 | t4a = (dctint)((1U << 13) + t4 + t12) >> 14; | |
| 1403 | 334320 | t5a = (dctint)((1U << 13) + t5 + t13) >> 14; | |
| 1404 | 334320 | t6a = (dctint)((1U << 13) + t6 + t14) >> 14; | |
| 1405 | 334320 | t7a = (dctint)((1U << 13) + t7 + t15) >> 14; | |
| 1406 | 334320 | t8a = (dctint)((1U << 13) + t0 - t8 ) >> 14; | |
| 1407 | 334320 | t9a = (dctint)((1U << 13) + t1 - t9 ) >> 14; | |
| 1408 | 334320 | t10a = (dctint)((1U << 13) + t2 - t10) >> 14; | |
| 1409 | 334320 | t11a = (dctint)((1U << 13) + t3 - t11) >> 14; | |
| 1410 | 334320 | t12a = (dctint)((1U << 13) + t4 - t12) >> 14; | |
| 1411 | 334320 | t13a = (dctint)((1U << 13) + t5 - t13) >> 14; | |
| 1412 | 334320 | t14a = (dctint)((1U << 13) + t6 - t14) >> 14; | |
| 1413 | 334320 | t15a = (dctint)((1U << 13) + t7 - t15) >> 14; | |
| 1414 | |||
| 1415 | 334320 | t8 = t8a * 16069U + t9a * 3196U; | |
| 1416 | 334320 | t9 = t8a * 3196U - t9a * 16069U; | |
| 1417 | 334320 | t10 = t10a * 9102U + t11a * 13623U; | |
| 1418 | 334320 | t11 = t10a * 13623U - t11a * 9102U; | |
| 1419 | 334320 | t12 = t13a * 16069U - t12a * 3196U; | |
| 1420 | 334320 | t13 = t13a * 3196U + t12a * 16069U; | |
| 1421 | 334320 | t14 = t15a * 9102U - t14a * 13623U; | |
| 1422 | 334320 | t15 = t15a * 13623U + t14a * 9102U; | |
| 1423 | |||
| 1424 | 334320 | t0 = t0a + t4a; | |
| 1425 | 334320 | t1 = t1a + t5a; | |
| 1426 | 334320 | t2 = t2a + t6a; | |
| 1427 | 334320 | t3 = t3a + t7a; | |
| 1428 | 334320 | t4 = t0a - t4a; | |
| 1429 | 334320 | t5 = t1a - t5a; | |
| 1430 | 334320 | t6 = t2a - t6a; | |
| 1431 | 334320 | t7 = t3a - t7a; | |
| 1432 | 334320 | t8a = (dctint)((1U << 13) + t8 + t12) >> 14; | |
| 1433 | 334320 | t9a = (dctint)((1U << 13) + t9 + t13) >> 14; | |
| 1434 | 334320 | t10a = (dctint)((1U << 13) + t10 + t14) >> 14; | |
| 1435 | 334320 | t11a = (dctint)((1U << 13) + t11 + t15) >> 14; | |
| 1436 | 334320 | t12a = (dctint)((1U << 13) + t8 - t12) >> 14; | |
| 1437 | 334320 | t13a = (dctint)((1U << 13) + t9 - t13) >> 14; | |
| 1438 | 334320 | t14a = (dctint)((1U << 13) + t10 - t14) >> 14; | |
| 1439 | 334320 | t15a = (dctint)((1U << 13) + t11 - t15) >> 14; | |
| 1440 | |||
| 1441 | 334320 | t4a = t4 * 15137U + t5 * 6270U; | |
| 1442 | 334320 | t5a = t4 * 6270U - t5 * 15137U; | |
| 1443 | 334320 | t6a = t7 * 15137U - t6 * 6270U; | |
| 1444 | 334320 | t7a = t7 * 6270U + t6 * 15137U; | |
| 1445 | 334320 | t12 = t12a * 15137U + t13a * 6270U; | |
| 1446 | 334320 | t13 = t12a * 6270U - t13a * 15137U; | |
| 1447 | 334320 | t14 = t15a * 15137U - t14a * 6270U; | |
| 1448 | 334320 | t15 = t15a * 6270U + t14a * 15137U; | |
| 1449 | |||
| 1450 | 334320 | out[ 0] = t0 + t2; | |
| 1451 | 334320 | out[15] = -(t1 + t3); | |
| 1452 | 334320 | t2a = t0 - t2; | |
| 1453 | 334320 | t3a = t1 - t3; | |
| 1454 | 334320 | out[ 3] = -((dctint)((1U << 13) + t4a + t6a) >> 14); | |
| 1455 | 334320 | out[12] = (dctint)((1U << 13) + t5a + t7a) >> 14; | |
| 1456 | 334320 | t6 = (dctint)((1U << 13) + t4a - t6a) >> 14; | |
| 1457 | 334320 | t7 = (dctint)((1U << 13) + t5a - t7a) >> 14; | |
| 1458 | 334320 | out[ 1] = -(t8a + t10a); | |
| 1459 | 334320 | out[14] = t9a + t11a; | |
| 1460 | 334320 | t10 = t8a - t10a; | |
| 1461 | 334320 | t11 = t9a - t11a; | |
| 1462 | 334320 | out[ 2] = (dctint)((1U << 13) + t12 + t14) >> 14; | |
| 1463 | 334320 | out[13] = -((dctint)((1U << 13) + t13 + t15) >> 14); | |
| 1464 | 334320 | t14a = (dctint)((1U << 13) + t12 - t14) >> 14; | |
| 1465 | 334320 | t15a = (dctint)((1U << 13) + t13 - t15) >> 14; | |
| 1466 | |||
| 1467 | 334320 | out[ 7] = (dctint)(-(t2a + t3a) * 11585U + (1 << 13)) >> 14; | |
| 1468 | 334320 | out[ 8] = (dctint)( (t2a - t3a) * 11585U + (1 << 13)) >> 14; | |
| 1469 | 334320 | out[ 4] = (dctint)( (t7 + t6) * 11585U + (1 << 13)) >> 14; | |
| 1470 | 334320 | out[11] = (dctint)( (t7 - t6) * 11585U + (1 << 13)) >> 14; | |
| 1471 | 334320 | out[ 6] = (dctint)( (t11 + t10) * 11585U + (1 << 13)) >> 14; | |
| 1472 | 334320 | out[ 9] = (dctint)( (t11 - t10) * 11585U + (1 << 13)) >> 14; | |
| 1473 | 334320 | out[ 5] = (dctint)(-(t14a + t15a) * 11585U + (1 << 13)) >> 14; | |
| 1474 | 334320 | out[10] = (dctint)( (t14a - t15a) * 11585U + (1 << 13)) >> 14; | |
| 1475 | 334320 | } | |
| 1476 | |||
| 1477 |
13/13✓ Branch 0 taken 8406 times.
✓ Branch 1 taken 309838 times.
✓ Branch 2 taken 2169540 times.
✓ Branch 3 taken 134496 times.
✓ Branch 4 taken 4641120 times.
✓ Branch 5 taken 290070 times.
✓ Branch 6 taken 281664 times.
✓ Branch 7 taken 468388 times.
✓ Branch 8 taken 28174 times.
✓ Branch 10 taken 7212544 times.
✓ Branch 11 taken 450784 times.
✓ Branch 12 taken 450784 times.
✓ Branch 13 taken 28174 times.
|
31049360 | itxfm_wrap(16, 6) |
| 1478 | |||
| 1479 | 649536 | static av_always_inline void idct32_1d(const dctcoef *in, ptrdiff_t stride, | |
| 1480 | dctcoef *out, int pass) | ||
| 1481 | { | ||
| 1482 | 649536 | dctint t0a = (dctint)((IN(0) + IN(16)) * 11585U + (1 << 13)) >> 14; | |
| 1483 | 649536 | dctint t1a = (dctint)((IN(0) - IN(16)) * 11585U + (1 << 13)) >> 14; | |
| 1484 | 649536 | dctint t2a = (dctint)(IN( 8) * 6270U - IN(24) * 15137U + (1 << 13)) >> 14; | |
| 1485 | 649536 | dctint t3a = (dctint)(IN( 8) * 15137U + IN(24) * 6270U + (1 << 13)) >> 14; | |
| 1486 | 649536 | dctint t4a = (dctint)(IN( 4) * 3196U - IN(28) * 16069U + (1 << 13)) >> 14; | |
| 1487 | 649536 | dctint t7a = (dctint)(IN( 4) * 16069U + IN(28) * 3196U + (1 << 13)) >> 14; | |
| 1488 | 649536 | dctint t5a = (dctint)(IN(20) * 13623U - IN(12) * 9102U + (1 << 13)) >> 14; | |
| 1489 | 649536 | dctint t6a = (dctint)(IN(20) * 9102U + IN(12) * 13623U + (1 << 13)) >> 14; | |
| 1490 | 649536 | dctint t8a = (dctint)(IN( 2) * 1606U - IN(30) * 16305U + (1 << 13)) >> 14; | |
| 1491 | 649536 | dctint t15a = (dctint)(IN( 2) * 16305U + IN(30) * 1606U + (1 << 13)) >> 14; | |
| 1492 | 649536 | dctint t9a = (dctint)(IN(18) * 12665U - IN(14) * 10394U + (1 << 13)) >> 14; | |
| 1493 | 649536 | dctint t14a = (dctint)(IN(18) * 10394U + IN(14) * 12665U + (1 << 13)) >> 14; | |
| 1494 | 649536 | dctint t10a = (dctint)(IN(10) * 7723U - IN(22) * 14449U + (1 << 13)) >> 14; | |
| 1495 | 649536 | dctint t13a = (dctint)(IN(10) * 14449U + IN(22) * 7723U + (1 << 13)) >> 14; | |
| 1496 | 649536 | dctint t11a = (dctint)(IN(26) * 15679U - IN( 6) * 4756U + (1 << 13)) >> 14; | |
| 1497 | 649536 | dctint t12a = (dctint)(IN(26) * 4756U + IN( 6) * 15679U + (1 << 13)) >> 14; | |
| 1498 | 649536 | dctint t16a = (dctint)(IN( 1) * 804U - IN(31) * 16364U + (1 << 13)) >> 14; | |
| 1499 | 649536 | dctint t31a = (dctint)(IN( 1) * 16364U + IN(31) * 804U + (1 << 13)) >> 14; | |
| 1500 | 649536 | dctint t17a = (dctint)(IN(17) * 12140U - IN(15) * 11003U + (1 << 13)) >> 14; | |
| 1501 | 649536 | dctint t30a = (dctint)(IN(17) * 11003U + IN(15) * 12140U + (1 << 13)) >> 14; | |
| 1502 | 649536 | dctint t18a = (dctint)(IN( 9) * 7005U - IN(23) * 14811U + (1 << 13)) >> 14; | |
| 1503 | 649536 | dctint t29a = (dctint)(IN( 9) * 14811U + IN(23) * 7005U + (1 << 13)) >> 14; | |
| 1504 | 649536 | dctint t19a = (dctint)(IN(25) * 15426U - IN( 7) * 5520U + (1 << 13)) >> 14; | |
| 1505 | 649536 | dctint t28a = (dctint)(IN(25) * 5520U + IN( 7) * 15426U + (1 << 13)) >> 14; | |
| 1506 | 649536 | dctint t20a = (dctint)(IN( 5) * 3981U - IN(27) * 15893U + (1 << 13)) >> 14; | |
| 1507 | 649536 | dctint t27a = (dctint)(IN( 5) * 15893U + IN(27) * 3981U + (1 << 13)) >> 14; | |
| 1508 | 649536 | dctint t21a = (dctint)(IN(21) * 14053U - IN(11) * 8423U + (1 << 13)) >> 14; | |
| 1509 | 649536 | dctint t26a = (dctint)(IN(21) * 8423U + IN(11) * 14053U + (1 << 13)) >> 14; | |
| 1510 | 649536 | dctint t22a = (dctint)(IN(13) * 9760U - IN(19) * 13160U + (1 << 13)) >> 14; | |
| 1511 | 649536 | dctint t25a = (dctint)(IN(13) * 13160U + IN(19) * 9760U + (1 << 13)) >> 14; | |
| 1512 | 649536 | dctint t23a = (dctint)(IN(29) * 16207U - IN( 3) * 2404U + (1 << 13)) >> 14; | |
| 1513 | 649536 | dctint t24a = (dctint)(IN(29) * 2404U + IN( 3) * 16207U + (1 << 13)) >> 14; | |
| 1514 | |||
| 1515 | 649536 | dctint t0 = t0a + t3a; | |
| 1516 | 649536 | dctint t1 = t1a + t2a; | |
| 1517 | 649536 | dctint t2 = t1a - t2a; | |
| 1518 | 649536 | dctint t3 = t0a - t3a; | |
| 1519 | 649536 | dctint t4 = t4a + t5a; | |
| 1520 | 649536 | dctint t5 = t4a - t5a; | |
| 1521 | 649536 | dctint t6 = t7a - t6a; | |
| 1522 | 649536 | dctint t7 = t7a + t6a; | |
| 1523 | 649536 | dctint t8 = t8a + t9a; | |
| 1524 | 649536 | dctint t9 = t8a - t9a; | |
| 1525 | 649536 | dctint t10 = t11a - t10a; | |
| 1526 | 649536 | dctint t11 = t11a + t10a; | |
| 1527 | 649536 | dctint t12 = t12a + t13a; | |
| 1528 | 649536 | dctint t13 = t12a - t13a; | |
| 1529 | 649536 | dctint t14 = t15a - t14a; | |
| 1530 | 649536 | dctint t15 = t15a + t14a; | |
| 1531 | 649536 | dctint t16 = t16a + t17a; | |
| 1532 | 649536 | dctint t17 = t16a - t17a; | |
| 1533 | 649536 | dctint t18 = t19a - t18a; | |
| 1534 | 649536 | dctint t19 = t19a + t18a; | |
| 1535 | 649536 | dctint t20 = t20a + t21a; | |
| 1536 | 649536 | dctint t21 = t20a - t21a; | |
| 1537 | 649536 | dctint t22 = t23a - t22a; | |
| 1538 | 649536 | dctint t23 = t23a + t22a; | |
| 1539 | 649536 | dctint t24 = t24a + t25a; | |
| 1540 | 649536 | dctint t25 = t24a - t25a; | |
| 1541 | 649536 | dctint t26 = t27a - t26a; | |
| 1542 | 649536 | dctint t27 = t27a + t26a; | |
| 1543 | 649536 | dctint t28 = t28a + t29a; | |
| 1544 | 649536 | dctint t29 = t28a - t29a; | |
| 1545 | 649536 | dctint t30 = t31a - t30a; | |
| 1546 | 649536 | dctint t31 = t31a + t30a; | |
| 1547 | |||
| 1548 | 649536 | t5a = (dctint)((t6 - t5) * 11585U + (1 << 13)) >> 14; | |
| 1549 | 649536 | t6a = (dctint)((t6 + t5) * 11585U + (1 << 13)) >> 14; | |
| 1550 | 649536 | t9a = (dctint)( t14 * 6270U - t9 * 15137U + (1 << 13)) >> 14; | |
| 1551 | 649536 | t14a = (dctint)( t14 * 15137U + t9 * 6270U + (1 << 13)) >> 14; | |
| 1552 | 649536 | t10a = (dctint)(-(t13 * 15137U + t10 * 6270U) + (1 << 13)) >> 14; | |
| 1553 | 649536 | t13a = (dctint)( t13 * 6270U - t10 * 15137U + (1 << 13)) >> 14; | |
| 1554 | 649536 | t17a = (dctint)( t30 * 3196U - t17 * 16069U + (1 << 13)) >> 14; | |
| 1555 | 649536 | t30a = (dctint)( t30 * 16069U + t17 * 3196U + (1 << 13)) >> 14; | |
| 1556 | 649536 | t18a = (dctint)(-(t29 * 16069U + t18 * 3196U) + (1 << 13)) >> 14; | |
| 1557 | 649536 | t29a = (dctint)( t29 * 3196U - t18 * 16069U + (1 << 13)) >> 14; | |
| 1558 | 649536 | t21a = (dctint)( t26 * 13623U - t21 * 9102U + (1 << 13)) >> 14; | |
| 1559 | 649536 | t26a = (dctint)( t26 * 9102U + t21 * 13623U + (1 << 13)) >> 14; | |
| 1560 | 649536 | t22a = (dctint)(-(t25 * 9102U + t22 * 13623U) + (1 << 13)) >> 14; | |
| 1561 | 649536 | t25a = (dctint)( t25 * 13623U - t22 * 9102U + (1 << 13)) >> 14; | |
| 1562 | |||
| 1563 | 649536 | t0a = t0 + t7; | |
| 1564 | 649536 | t1a = t1 + t6a; | |
| 1565 | 649536 | t2a = t2 + t5a; | |
| 1566 | 649536 | t3a = t3 + t4; | |
| 1567 | 649536 | t4a = t3 - t4; | |
| 1568 | 649536 | t5 = t2 - t5a; | |
| 1569 | 649536 | t6 = t1 - t6a; | |
| 1570 | 649536 | t7a = t0 - t7; | |
| 1571 | 649536 | t8a = t8 + t11; | |
| 1572 | 649536 | t9 = t9a + t10a; | |
| 1573 | 649536 | t10 = t9a - t10a; | |
| 1574 | 649536 | t11a = t8 - t11; | |
| 1575 | 649536 | t12a = t15 - t12; | |
| 1576 | 649536 | t13 = t14a - t13a; | |
| 1577 | 649536 | t14 = t14a + t13a; | |
| 1578 | 649536 | t15a = t15 + t12; | |
| 1579 | 649536 | t16a = t16 + t19; | |
| 1580 | 649536 | t17 = t17a + t18a; | |
| 1581 | 649536 | t18 = t17a - t18a; | |
| 1582 | 649536 | t19a = t16 - t19; | |
| 1583 | 649536 | t20a = t23 - t20; | |
| 1584 | 649536 | t21 = t22a - t21a; | |
| 1585 | 649536 | t22 = t22a + t21a; | |
| 1586 | 649536 | t23a = t23 + t20; | |
| 1587 | 649536 | t24a = t24 + t27; | |
| 1588 | 649536 | t25 = t25a + t26a; | |
| 1589 | 649536 | t26 = t25a - t26a; | |
| 1590 | 649536 | t27a = t24 - t27; | |
| 1591 | 649536 | t28a = t31 - t28; | |
| 1592 | 649536 | t29 = t30a - t29a; | |
| 1593 | 649536 | t30 = t30a + t29a; | |
| 1594 | 649536 | t31a = t31 + t28; | |
| 1595 | |||
| 1596 | 649536 | t10a = (dctint)((t13 - t10) * 11585U + (1 << 13)) >> 14; | |
| 1597 | 649536 | t13a = (dctint)((t13 + t10) * 11585U + (1 << 13)) >> 14; | |
| 1598 | 649536 | t11 = (dctint)((t12a - t11a) * 11585U + (1 << 13)) >> 14; | |
| 1599 | 649536 | t12 = (dctint)((t12a + t11a) * 11585U + (1 << 13)) >> 14; | |
| 1600 | 649536 | t18a = (dctint)( t29 * 6270U - t18 * 15137U + (1 << 13)) >> 14; | |
| 1601 | 649536 | t29a = (dctint)( t29 * 15137U + t18 * 6270U + (1 << 13)) >> 14; | |
| 1602 | 649536 | t19 = (dctint)( t28a * 6270U - t19a * 15137U + (1 << 13)) >> 14; | |
| 1603 | 649536 | t28 = (dctint)( t28a * 15137U + t19a * 6270U + (1 << 13)) >> 14; | |
| 1604 | 649536 | t20 = (dctint)(-(t27a * 15137U + t20a * 6270U) + (1 << 13)) >> 14; | |
| 1605 | 649536 | t27 = (dctint)( t27a * 6270U - t20a * 15137U + (1 << 13)) >> 14; | |
| 1606 | 649536 | t21a = (dctint)(-(t26 * 15137U + t21 * 6270U) + (1 << 13)) >> 14; | |
| 1607 | 649536 | t26a = (dctint)( t26 * 6270U - t21 * 15137U + (1 << 13)) >> 14; | |
| 1608 | |||
| 1609 | 649536 | t0 = t0a + t15a; | |
| 1610 | 649536 | t1 = t1a + t14; | |
| 1611 | 649536 | t2 = t2a + t13a; | |
| 1612 | 649536 | t3 = t3a + t12; | |
| 1613 | 649536 | t4 = t4a + t11; | |
| 1614 | 649536 | t5a = t5 + t10a; | |
| 1615 | 649536 | t6a = t6 + t9; | |
| 1616 | 649536 | t7 = t7a + t8a; | |
| 1617 | 649536 | t8 = t7a - t8a; | |
| 1618 | 649536 | t9a = t6 - t9; | |
| 1619 | 649536 | t10 = t5 - t10a; | |
| 1620 | 649536 | t11a = t4a - t11; | |
| 1621 | 649536 | t12a = t3a - t12; | |
| 1622 | 649536 | t13 = t2a - t13a; | |
| 1623 | 649536 | t14a = t1a - t14; | |
| 1624 | 649536 | t15 = t0a - t15a; | |
| 1625 | 649536 | t16 = t16a + t23a; | |
| 1626 | 649536 | t17a = t17 + t22; | |
| 1627 | 649536 | t18 = t18a + t21a; | |
| 1628 | 649536 | t19a = t19 + t20; | |
| 1629 | 649536 | t20a = t19 - t20; | |
| 1630 | 649536 | t21 = t18a - t21a; | |
| 1631 | 649536 | t22a = t17 - t22; | |
| 1632 | 649536 | t23 = t16a - t23a; | |
| 1633 | 649536 | t24 = t31a - t24a; | |
| 1634 | 649536 | t25a = t30 - t25; | |
| 1635 | 649536 | t26 = t29a - t26a; | |
| 1636 | 649536 | t27a = t28 - t27; | |
| 1637 | 649536 | t28a = t28 + t27; | |
| 1638 | 649536 | t29 = t29a + t26a; | |
| 1639 | 649536 | t30a = t30 + t25; | |
| 1640 | 649536 | t31 = t31a + t24a; | |
| 1641 | |||
| 1642 | 649536 | t20 = (dctint)((t27a - t20a) * 11585U + (1 << 13)) >> 14; | |
| 1643 | 649536 | t27 = (dctint)((t27a + t20a) * 11585U + (1 << 13)) >> 14; | |
| 1644 | 649536 | t21a = (dctint)((t26 - t21 ) * 11585U + (1 << 13)) >> 14; | |
| 1645 | 649536 | t26a = (dctint)((t26 + t21 ) * 11585U + (1 << 13)) >> 14; | |
| 1646 | 649536 | t22 = (dctint)((t25a - t22a) * 11585U + (1 << 13)) >> 14; | |
| 1647 | 649536 | t25 = (dctint)((t25a + t22a) * 11585U + (1 << 13)) >> 14; | |
| 1648 | 649536 | t23a = (dctint)((t24 - t23 ) * 11585U + (1 << 13)) >> 14; | |
| 1649 | 649536 | t24a = (< |