| 1 | /* |
| 2 | * thirdpel DSP functions |
| 3 | * |
| 4 | * This file is part of FFmpeg. |
| 5 | * |
| 6 | * FFmpeg is free software; you can redistribute it and/or |
| 7 | * modify it under the terms of the GNU Lesser General Public |
| 8 | * License as published by the Free Software Foundation; either |
| 9 | * version 2.1 of the License, or (at your option) any later version. |
| 10 | * |
| 11 | * FFmpeg is distributed in the hope that it will be useful, |
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU |
| 14 | * Lesser General Public License for more details. |
| 15 | * |
| 16 | * You should have received a copy of the GNU Lesser General Public |
| 17 | * License along with FFmpeg; if not, write to the Free Software |
| 18 | * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA |
| 19 | */ |
| 20 | |
| 21 | /** |
| 22 | * @file |
| 23 | * thirdpel DSP functions |
| 24 | */ |
| 25 | |
| 26 | #include <stdint.h> |
| 27 | |
| 28 | #include "libavutil/attributes.h" |
| 29 | #include "tpeldsp.h" |
| 30 | |
| 31 | #define BIT_DEPTH 8 |
| 32 | #include "pel_template.c" |
| 33 | |
| 34 | static inline void put_tpel_pixels_mc00_c(uint8_t *dst, const uint8_t *src, |
| 35 | int stride, int width, int height) |
| 36 | { |
| 37 | switch (width) { |
| 38 | case 2: |
| 39 | put_pixels2_8_c(dst, src, stride, height); |
| 40 | break; |
| 41 | case 4: |
| 42 | put_pixels4_8_c(dst, src, stride, height); |
| 43 | break; |
| 44 | case 8: |
| 45 | put_pixels8_8_c(dst, src, stride, height); |
| 46 | break; |
| 47 | case 16: |
| 48 | put_pixels16_8_c(dst, src, stride, height); |
| 49 | break; |
| 50 | } |
| 51 | } |
| 52 | |
| 53 | static inline void put_tpel_pixels_mc10_c(uint8_t *dst, const uint8_t *src, |
| 54 | int stride, int width, int height) |
| 55 | { |
| 56 | int i, j; |
| 57 | |
| 58 | for (i = 0; i < height; i++) { |
| 59 | for (j = 0; j < width; j++) |
| 60 | dst[j] = ((2 * src[j] + src[j + 1] + 1) * |
| 61 | 683) >> 11; |
| 62 | src += stride; |
| 63 | dst += stride; |
| 64 | } |
| 65 | } |
| 66 | |
| 67 | static inline void put_tpel_pixels_mc20_c(uint8_t *dst, const uint8_t *src, |
| 68 | int stride, int width, int height) |
| 69 | { |
| 70 | int i, j; |
| 71 | |
| 72 | for (i = 0; i < height; i++) { |
| 73 | for (j = 0; j < width; j++) |
| 74 | dst[j] = ((src[j] + 2 * src[j + 1] + 1) * |
| 75 | 683) >> 11; |
| 76 | src += stride; |
| 77 | dst += stride; |
| 78 | } |
| 79 | } |
| 80 | |
| 81 | static inline void put_tpel_pixels_mc01_c(uint8_t *dst, const uint8_t *src, |
| 82 | int stride, int width, int height) |
| 83 | { |
| 84 | int i, j; |
| 85 | |
| 86 | for (i = 0; i < height; i++) { |
| 87 | for (j = 0; j < width; j++) |
| 88 | dst[j] = ((2 * src[j] + src[j + stride] + 1) * |
| 89 | 683) >> 11; |
| 90 | src += stride; |
| 91 | dst += stride; |
| 92 | } |
| 93 | } |
| 94 | |
| 95 | static inline void put_tpel_pixels_mc11_c(uint8_t *dst, const uint8_t *src, |
| 96 | int stride, int width, int height) |
| 97 | { |
| 98 | int i, j; |
| 99 | |
| 100 | for (i = 0; i < height; i++) { |
| 101 | for (j = 0; j < width; j++) |
| 102 | dst[j] = ((4 * src[j] + 3 * src[j + 1] + |
| 103 | 3 * src[j + stride] + 2 * src[j + stride + 1] + 6) * |
| 104 | 2731) >> 15; |
| 105 | src += stride; |
| 106 | dst += stride; |
| 107 | } |
| 108 | } |
| 109 | |
| 110 | static inline void put_tpel_pixels_mc12_c(uint8_t *dst, const uint8_t *src, |
| 111 | int stride, int width, int height) |
| 112 | { |
| 113 | int i, j; |
| 114 | |
| 115 | for (i = 0; i < height; i++) { |
| 116 | for (j = 0; j < width; j++) |
| 117 | dst[j] = ((3 * src[j] + 2 * src[j + 1] + |
| 118 | 4 * src[j + stride] + 3 * src[j + stride + 1] + 6) * |
| 119 | 2731) >> 15; |
| 120 | src += stride; |
| 121 | dst += stride; |
| 122 | } |
| 123 | } |
| 124 | |
| 125 | static inline void put_tpel_pixels_mc02_c(uint8_t *dst, const uint8_t *src, |
| 126 | int stride, int width, int height) |
| 127 | { |
| 128 | int i, j; |
| 129 | |
| 130 | for (i = 0; i < height; i++) { |
| 131 | for (j = 0; j < width; j++) |
| 132 | dst[j] = ((src[j] + 2 * src[j + stride] + 1) * |
| 133 | 683) >> 11; |
| 134 | src += stride; |
| 135 | dst += stride; |
| 136 | } |
| 137 | } |
| 138 | |
| 139 | static inline void put_tpel_pixels_mc21_c(uint8_t *dst, const uint8_t *src, |
| 140 | int stride, int width, int height) |
| 141 | { |
| 142 | int i, j; |
| 143 | |
| 144 | for (i = 0; i < height; i++) { |
| 145 | for (j = 0; j < width; j++) |
| 146 | dst[j] = ((3 * src[j] + 4 * src[j + 1] + |
| 147 | 2 * src[j + stride] + 3 * src[j + stride + 1] + 6) * |
| 148 | 2731) >> 15; |
| 149 | src += stride; |
| 150 | dst += stride; |
| 151 | } |
| 152 | } |
| 153 | |
| 154 | static inline void put_tpel_pixels_mc22_c(uint8_t *dst, const uint8_t *src, |
| 155 | int stride, int width, int height) |
| 156 | { |
| 157 | int i, j; |
| 158 | |
| 159 | for (i = 0; i < height; i++) { |
| 160 | for (j = 0; j < width; j++) |
| 161 | dst[j] = ((2 * src[j] + 3 * src[j + 1] + |
| 162 | 3 * src[j + stride] + 4 * src[j + stride + 1] + 6) * |
| 163 | 2731) >> 15; |
| 164 | src += stride; |
| 165 | dst += stride; |
| 166 | } |
| 167 | } |
| 168 | |
| 169 | static inline void avg_tpel_pixels_mc00_c(uint8_t *dst, const uint8_t *src, |
| 170 | int stride, int width, int height) |
| 171 | { |
| 172 | switch (width) { |
| 173 | case 2: |
| 174 | avg_pixels2_8_c(dst, src, stride, height); |
| 175 | break; |
| 176 | case 4: |
| 177 | avg_pixels4_8_c(dst, src, stride, height); |
| 178 | break; |
| 179 | case 8: |
| 180 | avg_pixels8_8_c(dst, src, stride, height); |
| 181 | break; |
| 182 | case 16: |
| 183 | avg_pixels16_8_c(dst, src, stride, height); |
| 184 | break; |
| 185 | } |
| 186 | } |
| 187 | |
| 188 | static inline void avg_tpel_pixels_mc10_c(uint8_t *dst, const uint8_t *src, |
| 189 | int stride, int width, int height) |
| 190 | { |
| 191 | int i, j; |
| 192 | |
| 193 | for (i = 0; i < height; i++) { |
| 194 | for (j = 0; j < width; j++) |
| 195 | dst[j] = (dst[j] + |
| 196 | (((2 * src[j] + src[j + 1] + 1) * |
| 197 | 683) >> 11) + 1) >> 1; |
| 198 | src += stride; |
| 199 | dst += stride; |
| 200 | } |
| 201 | } |
| 202 | |
| 203 | static inline void avg_tpel_pixels_mc20_c(uint8_t *dst, const uint8_t *src, |
| 204 | int stride, int width, int height) |
| 205 | { |
| 206 | int i, j; |
| 207 | |
| 208 | for (i = 0; i < height; i++) { |
| 209 | for (j = 0; j < width; j++) |
| 210 | dst[j] = (dst[j] + |
| 211 | (((src[j] + 2 * src[j + 1] + 1) * |
| 212 | 683) >> 11) + 1) >> 1; |
| 213 | src += stride; |
| 214 | dst += stride; |
| 215 | } |
| 216 | } |
| 217 | |
| 218 | static inline void avg_tpel_pixels_mc01_c(uint8_t *dst, const uint8_t *src, |
| 219 | int stride, int width, int height) |
| 220 | { |
| 221 | int i, j; |
| 222 | |
| 223 | for (i = 0; i < height; i++) { |
| 224 | for (j = 0; j < width; j++) |
| 225 | dst[j] = (dst[j] + |
| 226 | (((2 * src[j] + src[j + stride] + 1) * |
| 227 | 683) >> 11) + 1) >> 1; |
| 228 | src += stride; |
| 229 | dst += stride; |
| 230 | } |
| 231 | } |
| 232 | |
| 233 | static inline void avg_tpel_pixels_mc11_c(uint8_t *dst, const uint8_t *src, |
| 234 | int stride, int width, int height) |
| 235 | { |
| 236 | int i, j; |
| 237 | |
| 238 | for (i = 0; i < height; i++) { |
| 239 | for (j = 0; j < width; j++) |
| 240 | dst[j] = (dst[j] + |
| 241 | (((4 * src[j] + 3 * src[j + 1] + |
| 242 | 3 * src[j + stride] + 2 * src[j + stride + 1] + 6) * |
| 243 | 2731) >> 15) + 1) >> 1; |
| 244 | src += stride; |
| 245 | dst += stride; |
| 246 | } |
| 247 | } |
| 248 | |
| 249 | static inline void avg_tpel_pixels_mc12_c(uint8_t *dst, const uint8_t *src, |
| 250 | int stride, int width, int height) |
| 251 | { |
| 252 | int i, j; |
| 253 | |
| 254 | for (i = 0; i < height; i++) { |
| 255 | for (j = 0; j < width; j++) |
| 256 | dst[j] = (dst[j] + |
| 257 | (((3 * src[j] + 2 * src[j + 1] + |
| 258 | 4 * src[j + stride] + 3 * src[j + stride + 1] + 6) * |
| 259 | 2731) >> 15) + 1) >> 1; |
| 260 | src += stride; |
| 261 | dst += stride; |
| 262 | } |
| 263 | } |
| 264 | |
| 265 | static inline void avg_tpel_pixels_mc02_c(uint8_t *dst, const uint8_t *src, |
| 266 | int stride, int width, int height) |
| 267 | { |
| 268 | int i, j; |
| 269 | |
| 270 | for (i = 0; i < height; i++) { |
| 271 | for (j = 0; j < width; j++) |
| 272 | dst[j] = (dst[j] + |
| 273 | (((src[j] + 2 * src[j + stride] + 1) * |
| 274 | 683) >> 11) + 1) >> 1; |
| 275 | src += stride; |
| 276 | dst += stride; |
| 277 | } |
| 278 | } |
| 279 | |
| 280 | static inline void avg_tpel_pixels_mc21_c(uint8_t *dst, const uint8_t *src, |
| 281 | int stride, int width, int height) |
| 282 | { |
| 283 | int i, j; |
| 284 | |
| 285 | for (i = 0; i < height; i++) { |
| 286 | for (j = 0; j < width; j++) |
| 287 | dst[j] = (dst[j] + |
| 288 | (((3 * src[j] + 4 * src[j + 1] + |
| 289 | 2 * src[j + stride] + 3 * src[j + stride + 1] + 6) * |
| 290 | 2731) >> 15) + 1) >> 1; |
| 291 | src += stride; |
| 292 | dst += stride; |
| 293 | } |
| 294 | } |
| 295 | |
| 296 | static inline void avg_tpel_pixels_mc22_c(uint8_t *dst, const uint8_t *src, |
| 297 | int stride, int width, int height) |
| 298 | { |
| 299 | int i, j; |
| 300 | |
| 301 | for (i = 0; i < height; i++) { |
| 302 | for (j = 0; j < width; j++) |
| 303 | dst[j] = (dst[j] + |
| 304 | (((2 * src[j] + 3 * src[j + 1] + |
| 305 | 3 * src[j + stride] + 4 * src[j + stride + 1] + 6) * |
| 306 | 2731) >> 15) + 1) >> 1; |
| 307 | src += stride; |
| 308 | dst += stride; |
| 309 | } |
| 310 | } |
| 311 | |
| 312 | av_cold void ff_tpeldsp_init(TpelDSPContext *c) |
| 313 | { |
| 314 | c->put_tpel_pixels_tab[ 0] = put_tpel_pixels_mc00_c; |
| 315 | c->put_tpel_pixels_tab[ 1] = put_tpel_pixels_mc10_c; |
| 316 | c->put_tpel_pixels_tab[ 2] = put_tpel_pixels_mc20_c; |
| 317 | c->put_tpel_pixels_tab[ 4] = put_tpel_pixels_mc01_c; |
| 318 | c->put_tpel_pixels_tab[ 5] = put_tpel_pixels_mc11_c; |
| 319 | c->put_tpel_pixels_tab[ 6] = put_tpel_pixels_mc21_c; |
| 320 | c->put_tpel_pixels_tab[ 8] = put_tpel_pixels_mc02_c; |
| 321 | c->put_tpel_pixels_tab[ 9] = put_tpel_pixels_mc12_c; |
| 322 | c->put_tpel_pixels_tab[10] = put_tpel_pixels_mc22_c; |
| 323 | |
| 324 | c->avg_tpel_pixels_tab[ 0] = avg_tpel_pixels_mc00_c; |
| 325 | c->avg_tpel_pixels_tab[ 1] = avg_tpel_pixels_mc10_c; |
| 326 | c->avg_tpel_pixels_tab[ 2] = avg_tpel_pixels_mc20_c; |
| 327 | c->avg_tpel_pixels_tab[ 4] = avg_tpel_pixels_mc01_c; |
| 328 | c->avg_tpel_pixels_tab[ 5] = avg_tpel_pixels_mc11_c; |
| 329 | c->avg_tpel_pixels_tab[ 6] = avg_tpel_pixels_mc21_c; |
| 330 | c->avg_tpel_pixels_tab[ 8] = avg_tpel_pixels_mc02_c; |
| 331 | c->avg_tpel_pixels_tab[ 9] = avg_tpel_pixels_mc12_c; |
| 332 | c->avg_tpel_pixels_tab[10] = avg_tpel_pixels_mc22_c; |
| 333 | } |