4 * Copyright (c) 2001 Michael Niedermayer <michaelni@gmx.at>
6 * This file is part of FFmpeg.
8 * FFmpeg is free software; you can redistribute it and/or
9 * modify it under the terms of the GNU Lesser General Public
10 * License as published by the Free Software Foundation; either
11 * version 2.1 of the License, or (at your option) any later version.
13 * FFmpeg is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
16 * Lesser General Public License for more details.
18 * You should have received a copy of the GNU Lesser General Public
19 * License along with FFmpeg; if not, write to the Free Software
20 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
28 /* Based upon some commented-out C code from mpeg2dec (idct_mmx.c
29 * written by Aaron Holtzman <aholtzma@ess.engr.uvic.ca>). */
31 #include "simple_idct.h"
33 #include "bit_depth_template.c"
50 #define W1 22725 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
51 #define W2 21407 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
52 #define W3 19266 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
53 #define W4 16383 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
54 #define W5 12873 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
55 #define W6 8867 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
56 #define W7 4520 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
62 #define MUL(a, b) MUL16(a, b)
63 #define MAC(a, b, c) MAC16(a, b, c)
65 #elif BIT_DEPTH == 10 || BIT_DEPTH == 12
68 #define W1 22725 // 90901
69 #define W2 21407 // 85627
70 #define W3 19265 // 77062
71 #define W4 16384 // 65535
72 #define W5 12873 // 51491
73 #define W6 8867 // 35468
74 #define W7 4520 // 18081
100 #define MUL(a, b) ((a) * (b))
101 #define MAC(a, b, c) ((a) += (b) * (c))
105 #error "Unsupported bitdepth"
110 static inline void FUNC(idctRowCondDC_extrashift)(int16_t *row, int extra_shift)
112 static inline void FUNC(idctRowCondDC)(int16_t *row, int extra_shift)
115 SUINT a0, a1, a2, a3, b0, b1, b2, b3;
118 #define ROW0_MASK (0xffffLL << 48 * HAVE_BIGENDIAN)
119 if (((AV_RN64A(row) & ~ROW0_MASK) | AV_RN64A(row+4)) == 0) {
121 if (DC_SHIFT - extra_shift >= 0) {
122 temp = (row[0] * (1 << (DC_SHIFT - extra_shift))) & 0xffff;
124 temp = ((row[0] + (1<<(extra_shift - DC_SHIFT-1))) >> (extra_shift - DC_SHIFT)) & 0xffff;
126 temp += temp * (1 << 16);
127 temp += temp * ((uint64_t) 1 << 32);
129 AV_WN64A(row + 4, temp);
133 if (!(AV_RN32A(row+2) |
138 if (DC_SHIFT - extra_shift >= 0) {
139 temp = (row[0] * (1 << (DC_SHIFT - extra_shift))) & 0xffff;
141 temp = ((row[0] + (1<<(extra_shift - DC_SHIFT-1))) >> (extra_shift - DC_SHIFT)) & 0xffff;
143 temp += temp * (1 << 16);
145 AV_WN32A(row+2, temp);
146 AV_WN32A(row+4, temp);
147 AV_WN32A(row+6, temp);
152 a0 = (W4 * row[0]) + (1 << (ROW_SHIFT + extra_shift - 1));
162 b0 = MUL(W1, row[1]);
164 b1 = MUL(W3, row[1]);
165 MAC(b1, -W7, row[3]);
166 b2 = MUL(W5, row[1]);
167 MAC(b2, -W1, row[3]);
168 b3 = MUL(W7, row[1]);
169 MAC(b3, -W5, row[3]);
171 if (AV_RN64A(row + 4)) {
172 a0 += W4*row[4] + W6*row[6];
173 a1 += - W4*row[4] - W2*row[6];
174 a2 += - W4*row[4] + W2*row[6];
175 a3 += W4*row[4] - W6*row[6];
180 MAC(b1, -W1, row[5]);
181 MAC(b1, -W5, row[7]);
187 MAC(b3, -W1, row[7]);
190 row[0] = (int)(a0 + b0) >> (ROW_SHIFT + extra_shift);
191 row[7] = (int)(a0 - b0) >> (ROW_SHIFT + extra_shift);
192 row[1] = (int)(a1 + b1) >> (ROW_SHIFT + extra_shift);
193 row[6] = (int)(a1 - b1) >> (ROW_SHIFT + extra_shift);
194 row[2] = (int)(a2 + b2) >> (ROW_SHIFT + extra_shift);
195 row[5] = (int)(a2 - b2) >> (ROW_SHIFT + extra_shift);
196 row[3] = (int)(a3 + b3) >> (ROW_SHIFT + extra_shift);
197 row[4] = (int)(a3 - b3) >> (ROW_SHIFT + extra_shift);
200 #define IDCT_COLS do { \
201 a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); \
208 a2 += -W6*col[8*2]; \
209 a3 += -W2*col[8*2]; \
211 b0 = MUL(W1, col[8*1]); \
212 b1 = MUL(W3, col[8*1]); \
213 b2 = MUL(W5, col[8*1]); \
214 b3 = MUL(W7, col[8*1]); \
216 MAC(b0, W3, col[8*3]); \
217 MAC(b1, -W7, col[8*3]); \
218 MAC(b2, -W1, col[8*3]); \
219 MAC(b3, -W5, col[8*3]); \
223 a1 += -W4*col[8*4]; \
224 a2 += -W4*col[8*4]; \
229 MAC(b0, W5, col[8*5]); \
230 MAC(b1, -W1, col[8*5]); \
231 MAC(b2, W7, col[8*5]); \
232 MAC(b3, W3, col[8*5]); \
237 a1 += -W2*col[8*6]; \
239 a3 += -W6*col[8*6]; \
243 MAC(b0, W7, col[8*7]); \
244 MAC(b1, -W5, col[8*7]); \
245 MAC(b2, W3, col[8*7]); \
246 MAC(b3, -W1, col[8*7]); \
251 static inline void FUNC(idctSparseCol_extrashift)(int16_t *col)
253 static inline void FUNC(idctSparseColPut)(pixel *dest, int line_size,
256 SUINT a0, a1, a2, a3, b0, b1, b2, b3;
260 dest[0] = av_clip_pixel((int)(a0 + b0) >> COL_SHIFT);
262 dest[0] = av_clip_pixel((int)(a1 + b1) >> COL_SHIFT);
264 dest[0] = av_clip_pixel((int)(a2 + b2) >> COL_SHIFT);
266 dest[0] = av_clip_pixel((int)(a3 + b3) >> COL_SHIFT);
268 dest[0] = av_clip_pixel((int)(a3 - b3) >> COL_SHIFT);
270 dest[0] = av_clip_pixel((int)(a2 - b2) >> COL_SHIFT);
272 dest[0] = av_clip_pixel((int)(a1 - b1) >> COL_SHIFT);
274 dest[0] = av_clip_pixel((int)(a0 - b0) >> COL_SHIFT);
277 static inline void FUNC(idctSparseColAdd)(pixel *dest, int line_size,
280 int a0, a1, a2, a3, b0, b1, b2, b3;
284 dest[0] = av_clip_pixel(dest[0] + ((a0 + b0) >> COL_SHIFT));
286 dest[0] = av_clip_pixel(dest[0] + ((a1 + b1) >> COL_SHIFT));
288 dest[0] = av_clip_pixel(dest[0] + ((a2 + b2) >> COL_SHIFT));
290 dest[0] = av_clip_pixel(dest[0] + ((a3 + b3) >> COL_SHIFT));
292 dest[0] = av_clip_pixel(dest[0] + ((a3 - b3) >> COL_SHIFT));
294 dest[0] = av_clip_pixel(dest[0] + ((a2 - b2) >> COL_SHIFT));
296 dest[0] = av_clip_pixel(dest[0] + ((a1 - b1) >> COL_SHIFT));
298 dest[0] = av_clip_pixel(dest[0] + ((a0 - b0) >> COL_SHIFT));
301 static inline void FUNC(idctSparseCol)(int16_t *col)
304 int a0, a1, a2, a3, b0, b1, b2, b3;
308 col[0 ] = ((a0 + b0) >> COL_SHIFT);
309 col[8 ] = ((a1 + b1) >> COL_SHIFT);
310 col[16] = ((a2 + b2) >> COL_SHIFT);
311 col[24] = ((a3 + b3) >> COL_SHIFT);
312 col[32] = ((a3 - b3) >> COL_SHIFT);
313 col[40] = ((a2 - b2) >> COL_SHIFT);
314 col[48] = ((a1 - b1) >> COL_SHIFT);
315 col[56] = ((a0 - b0) >> COL_SHIFT);
319 void FUNC(ff_simple_idct_put)(uint8_t *dest_, int line_size, int16_t *block)
321 pixel *dest = (pixel *)dest_;
324 line_size /= sizeof(pixel);
326 for (i = 0; i < 8; i++)
327 FUNC(idctRowCondDC)(block + i*8, 0);
329 for (i = 0; i < 8; i++)
330 FUNC(idctSparseColPut)(dest + i, line_size, block + i);
333 void FUNC(ff_simple_idct_add)(uint8_t *dest_, int line_size, int16_t *block)
335 pixel *dest = (pixel *)dest_;
338 line_size /= sizeof(pixel);
340 for (i = 0; i < 8; i++)
341 FUNC(idctRowCondDC)(block + i*8, 0);
343 for (i = 0; i < 8; i++)
344 FUNC(idctSparseColAdd)(dest + i, line_size, block + i);
347 void FUNC(ff_simple_idct)(int16_t *block)
351 for (i = 0; i < 8; i++)
352 FUNC(idctRowCondDC)(block + i*8, 0);
354 for (i = 0; i < 8; i++)
355 FUNC(idctSparseCol)(block + i);