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 #include "libavutil/intreadwrite.h"
31 #include "simple_idct.h"
34 #include "simple_idct_template.c"
38 #include "simple_idct_template.c"
42 #include "simple_idct_template.c"
48 #define C_FIX(x) ((int)((x) * (1 << CN_SHIFT) + 0.5))
49 #define C1 C_FIX(0.6532814824)
50 #define C2 C_FIX(0.2705980501)
52 /* row idct is multiple by 16 * sqrt(2.0), col idct4 is normalized,
53 and the butterfly must be multiplied by 0.5 * sqrt(2.0) */
54 #define C_SHIFT (4+1+12)
56 static inline void idct4col_put(uint8_t *dest, int line_size, const int16_t *col)
58 int c0, c1, c2, c3, a0, a1, a2, a3;
64 c0 = ((a0 + a2) << (CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
65 c2 = ((a0 - a2) << (CN_SHIFT - 1)) + (1 << (C_SHIFT - 1));
66 c1 = a1 * C1 + a3 * C2;
67 c3 = a1 * C2 - a3 * C1;
68 dest[0] = av_clip_uint8((c0 + c1) >> C_SHIFT);
70 dest[0] = av_clip_uint8((c2 + c3) >> C_SHIFT);
72 dest[0] = av_clip_uint8((c2 - c3) >> C_SHIFT);
74 dest[0] = av_clip_uint8((c0 - c1) >> C_SHIFT);
83 ptr[8 + k] = a0 - a1;\
86 /* only used by DV codec. The input must be interlaced. 128 is added
87 to the pixels before clamping to avoid systematic error
88 (1024*sqrt(2)) offset would be needed otherwise. */
89 /* XXX: I think a 1.0/sqrt(2) normalization should be needed to
90 compensate the extra butterfly stage - I don't have the full DV
92 void ff_simple_idct248_put(uint8_t *dest, int line_size, int16_t *block)
111 /* IDCT8 on each line */
113 idctRowCondDC_8(block + i*8, 0);
116 /* IDCT4 and store */
118 idct4col_put(dest + i, 2 * line_size, block + i);
119 idct4col_put(dest + line_size + i, 2 * line_size, block + 8 + i);
123 /* 8x4 & 4x8 WMV2 IDCT */
130 #define C_FIX(x) ((int)((x) * 1.414213562 * (1 << CN_SHIFT) + 0.5))
131 #define C1 C_FIX(0.6532814824)
132 #define C2 C_FIX(0.2705980501)
133 #define C3 C_FIX(0.5)
134 #define C_SHIFT (4+1+12)
135 static inline void idct4col_add(uint8_t *dest, int line_size, const int16_t *col)
137 int c0, c1, c2, c3, a0, a1, a2, a3;
143 c0 = (a0 + a2)*C3 + (1 << (C_SHIFT - 1));
144 c2 = (a0 - a2)*C3 + (1 << (C_SHIFT - 1));
145 c1 = a1 * C1 + a3 * C2;
146 c3 = a1 * C2 - a3 * C1;
147 dest[0] = av_clip_uint8(dest[0] + ((c0 + c1) >> C_SHIFT));
149 dest[0] = av_clip_uint8(dest[0] + ((c2 + c3) >> C_SHIFT));
151 dest[0] = av_clip_uint8(dest[0] + ((c2 - c3) >> C_SHIFT));
153 dest[0] = av_clip_uint8(dest[0] + ((c0 - c1) >> C_SHIFT));
157 #define R_FIX(x) ((int)((x) * 1.414213562 * (1 << RN_SHIFT) + 0.5))
158 #define R1 R_FIX(0.6532814824)
159 #define R2 R_FIX(0.2705980501)
160 #define R3 R_FIX(0.5)
162 static inline void idct4row(int16_t *row)
164 int c0, c1, c2, c3, a0, a1, a2, a3;
170 c0 = (a0 + a2)*R3 + (1 << (R_SHIFT - 1));
171 c2 = (a0 - a2)*R3 + (1 << (R_SHIFT - 1));
172 c1 = a1 * R1 + a3 * R2;
173 c3 = a1 * R2 - a3 * R1;
174 row[0]= (c0 + c1) >> R_SHIFT;
175 row[1]= (c2 + c3) >> R_SHIFT;
176 row[2]= (c2 - c3) >> R_SHIFT;
177 row[3]= (c0 - c1) >> R_SHIFT;
180 void ff_simple_idct84_add(uint8_t *dest, int line_size, int16_t *block)
184 /* IDCT8 on each line */
186 idctRowCondDC_8(block + i*8, 0);
189 /* IDCT4 and store */
191 idct4col_add(dest + i, line_size, block + i);
195 void ff_simple_idct48_add(uint8_t *dest, int line_size, int16_t *block)
199 /* IDCT4 on each line */
201 idct4row(block + i*8);
204 /* IDCT8 and store */
206 idctSparseColAdd_8(dest + i, line_size, block + i);
210 void ff_simple_idct44_add(uint8_t *dest, int line_size, int16_t *block)
214 /* IDCT4 on each line */
216 idct4row(block + i*8);
219 /* IDCT4 and store */
221 idct4col_add(dest + i, line_size, block + i);
225 void ff_prores_idct(int16_t *block, const int16_t *qmat)
229 for (i = 0; i < 64; i++)
232 for (i = 0; i < 8; i++)
233 idctRowCondDC_10(block + i*8, 2);
235 for (i = 0; i < 8; i++) {
237 idctSparseCol_10(block + i);