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
29 based upon some outcommented c code from mpeg2dec (idct_mmx.c
30 written by Aaron Holtzman <aholtzma@ess.engr.uvic.ca>)
33 #include "simple_idct.h"
35 #include "bit_depth_template.c"
52 #define W1 22725 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
53 #define W2 21407 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
54 #define W3 19266 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
55 #define W4 16383 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
56 #define W5 12873 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
57 #define W6 8867 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
58 #define W7 4520 //cos(i*M_PI/16)*sqrt(2)*(1<<14) + 0.5
64 #define MUL(a, b) MUL16(a, b)
65 #define MAC(a, b, c) MAC16(a, b, c)
67 #elif BIT_DEPTH == 10 || BIT_DEPTH == 12
70 #define W1 (22725*4) // 90901
71 #define W2 (21407*4) // 85627
72 #define W3 (19265*4) // 77062
73 #define W4 (16384*4) // 65535
74 #define W5 (12873*4) // 51491
75 #define W6 ( 8867*4) // 35468
76 #define W7 ( 4520*4) // 18081
95 #define MUL(a, b) ((a) * (b))
96 #define MAC(a, b, c) ((a) += (b) * (c))
100 #error "Unsupported bitdepth"
104 static inline void FUNC(idctRowCondDC)(int16_t *row, int extra_shift)
106 int a0, a1, a2, a3, b0, b1, b2, b3;
109 #define ROW0_MASK (0xffffLL << 48 * HAVE_BIGENDIAN)
110 if (((((uint64_t *)row)[0] & ~ROW0_MASK) | ((uint64_t *)row)[1]) == 0) {
112 if (DC_SHIFT - extra_shift >= 0) {
113 temp = (row[0] << (DC_SHIFT - extra_shift)) & 0xffff;
115 temp = ((row[0] + (1<<(extra_shift - DC_SHIFT-1))) >> (extra_shift - DC_SHIFT)) & 0xffff;
119 ((uint64_t *)row)[0] = temp;
120 ((uint64_t *)row)[1] = temp;
124 if (!(((uint32_t*)row)[1] |
125 ((uint32_t*)row)[2] |
126 ((uint32_t*)row)[3] |
129 if (DC_SHIFT - extra_shift >= 0) {
130 temp = (row[0] << (DC_SHIFT - extra_shift)) & 0xffff;
132 temp = ((row[0] + (1<<(extra_shift - DC_SHIFT-1))) >> (extra_shift - DC_SHIFT)) & 0xffff;
135 ((uint32_t*)row)[0]=((uint32_t*)row)[1] =
136 ((uint32_t*)row)[2]=((uint32_t*)row)[3] = temp;
141 a0 = (W4 * row[0]) + (1 << (ROW_SHIFT + extra_shift - 1));
151 b0 = MUL(W1, row[1]);
153 b1 = MUL(W3, row[1]);
154 MAC(b1, -W7, row[3]);
155 b2 = MUL(W5, row[1]);
156 MAC(b2, -W1, row[3]);
157 b3 = MUL(W7, row[1]);
158 MAC(b3, -W5, row[3]);
160 if (AV_RN64A(row + 4)) {
161 a0 += W4*row[4] + W6*row[6];
162 a1 += - W4*row[4] - W2*row[6];
163 a2 += - W4*row[4] + W2*row[6];
164 a3 += W4*row[4] - W6*row[6];
169 MAC(b1, -W1, row[5]);
170 MAC(b1, -W5, row[7]);
176 MAC(b3, -W1, row[7]);
179 row[0] = (a0 + b0) >> (ROW_SHIFT + extra_shift);
180 row[7] = (a0 - b0) >> (ROW_SHIFT + extra_shift);
181 row[1] = (a1 + b1) >> (ROW_SHIFT + extra_shift);
182 row[6] = (a1 - b1) >> (ROW_SHIFT + extra_shift);
183 row[2] = (a2 + b2) >> (ROW_SHIFT + extra_shift);
184 row[5] = (a2 - b2) >> (ROW_SHIFT + extra_shift);
185 row[3] = (a3 + b3) >> (ROW_SHIFT + extra_shift);
186 row[4] = (a3 - b3) >> (ROW_SHIFT + extra_shift);
189 #define IDCT_COLS do { \
190 a0 = W4 * (col[8*0] + ((1<<(COL_SHIFT-1))/W4)); \
197 a2 += -W6*col[8*2]; \
198 a3 += -W2*col[8*2]; \
200 b0 = MUL(W1, col[8*1]); \
201 b1 = MUL(W3, col[8*1]); \
202 b2 = MUL(W5, col[8*1]); \
203 b3 = MUL(W7, col[8*1]); \
205 MAC(b0, W3, col[8*3]); \
206 MAC(b1, -W7, col[8*3]); \
207 MAC(b2, -W1, col[8*3]); \
208 MAC(b3, -W5, col[8*3]); \
212 a1 += -W4*col[8*4]; \
213 a2 += -W4*col[8*4]; \
218 MAC(b0, W5, col[8*5]); \
219 MAC(b1, -W1, col[8*5]); \
220 MAC(b2, W7, col[8*5]); \
221 MAC(b3, W3, col[8*5]); \
226 a1 += -W2*col[8*6]; \
228 a3 += -W6*col[8*6]; \
232 MAC(b0, W7, col[8*7]); \
233 MAC(b1, -W5, col[8*7]); \
234 MAC(b2, W3, col[8*7]); \
235 MAC(b3, -W1, col[8*7]); \
239 static inline void FUNC(idctSparseColPut)(pixel *dest, int line_size,
242 int a0, a1, a2, a3, b0, b1, b2, b3;
246 dest[0] = av_clip_pixel((a0 + b0) >> COL_SHIFT);
248 dest[0] = av_clip_pixel((a1 + b1) >> COL_SHIFT);
250 dest[0] = av_clip_pixel((a2 + b2) >> COL_SHIFT);
252 dest[0] = av_clip_pixel((a3 + b3) >> COL_SHIFT);
254 dest[0] = av_clip_pixel((a3 - b3) >> COL_SHIFT);
256 dest[0] = av_clip_pixel((a2 - b2) >> COL_SHIFT);
258 dest[0] = av_clip_pixel((a1 - b1) >> COL_SHIFT);
260 dest[0] = av_clip_pixel((a0 - b0) >> COL_SHIFT);
263 static inline void FUNC(idctSparseColAdd)(pixel *dest, int line_size,
266 int a0, a1, a2, a3, b0, b1, b2, b3;
270 dest[0] = av_clip_pixel(dest[0] + ((a0 + b0) >> COL_SHIFT));
272 dest[0] = av_clip_pixel(dest[0] + ((a1 + b1) >> COL_SHIFT));
274 dest[0] = av_clip_pixel(dest[0] + ((a2 + b2) >> COL_SHIFT));
276 dest[0] = av_clip_pixel(dest[0] + ((a3 + b3) >> COL_SHIFT));
278 dest[0] = av_clip_pixel(dest[0] + ((a3 - b3) >> COL_SHIFT));
280 dest[0] = av_clip_pixel(dest[0] + ((a2 - b2) >> COL_SHIFT));
282 dest[0] = av_clip_pixel(dest[0] + ((a1 - b1) >> COL_SHIFT));
284 dest[0] = av_clip_pixel(dest[0] + ((a0 - b0) >> COL_SHIFT));
287 static inline void FUNC(idctSparseCol)(int16_t *col)
289 int a0, a1, a2, a3, b0, b1, b2, b3;
293 col[0 ] = ((a0 + b0) >> COL_SHIFT);
294 col[8 ] = ((a1 + b1) >> COL_SHIFT);
295 col[16] = ((a2 + b2) >> COL_SHIFT);
296 col[24] = ((a3 + b3) >> COL_SHIFT);
297 col[32] = ((a3 - b3) >> COL_SHIFT);
298 col[40] = ((a2 - b2) >> COL_SHIFT);
299 col[48] = ((a1 - b1) >> COL_SHIFT);
300 col[56] = ((a0 - b0) >> COL_SHIFT);
303 void FUNC(ff_simple_idct_put)(uint8_t *dest_, int line_size, int16_t *block)
305 pixel *dest = (pixel *)dest_;
308 line_size /= sizeof(pixel);
310 for (i = 0; i < 8; i++)
311 FUNC(idctRowCondDC)(block + i*8, 0);
313 for (i = 0; i < 8; i++)
314 FUNC(idctSparseColPut)(dest + i, line_size, block + i);
317 void FUNC(ff_simple_idct_add)(uint8_t *dest_, int line_size, int16_t *block)
319 pixel *dest = (pixel *)dest_;
322 line_size /= sizeof(pixel);
324 for (i = 0; i < 8; i++)
325 FUNC(idctRowCondDC)(block + i*8, 0);
327 for (i = 0; i < 8; i++)
328 FUNC(idctSparseColAdd)(dest + i, line_size, block + i);
331 void FUNC(ff_simple_idct)(int16_t *block)
335 for (i = 0; i < 8; i++)
336 FUNC(idctRowCondDC)(block + i*8, 0);
338 for (i = 0; i < 8; i++)
339 FUNC(idctSparseCol)(block + i);