Mercurial > mplayer.hg
annotate liba52/imdct.c @ 13981:3e82594ec231
correct encoding
author | diego |
---|---|
date | Fri, 19 Nov 2004 18:34:33 +0000 |
parents | f881c918739b |
children | 07f1e7669772 |
rev | line source |
---|---|
3394 | 1 /* |
2 * imdct.c | |
3 * Copyright (C) 2000-2001 Michel Lespinasse <walken@zoy.org> | |
4 * Copyright (C) 1999-2000 Aaron Holtzman <aholtzma@ess.engr.uvic.ca> | |
5 * | |
6 * This file is part of a52dec, a free ATSC A-52 stream decoder. | |
7 * See http://liba52.sourceforge.net/ for updates. | |
8 * | |
9 * a52dec is free software; you can redistribute it and/or modify | |
10 * it under the terms of the GNU General Public License as published by | |
11 * the Free Software Foundation; either version 2 of the License, or | |
12 * (at your option) any later version. | |
13 * | |
14 * a52dec 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 | |
17 * GNU General Public License for more details. | |
18 * | |
19 * You should have received a copy of the GNU General Public License | |
20 * along with this program; if not, write to the Free Software | |
21 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA | |
3579 | 22 * |
23 * SSE optimizations from Michael Niedermayer (michaelni@gmx.at) | |
3884 | 24 * 3DNOW optimizations from Nick Kurshev <nickols_k@mail.ru> |
25 * michael did port them from libac3 (untested, perhaps totally broken) | |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
26 * AltiVec optimizations from Romain Dolbeau (romain@dolbeau.org) |
3394 | 27 */ |
28 | |
29 #include "config.h" | |
30 | |
31 #include <math.h> | |
32 #include <stdio.h> | |
33 #ifndef M_PI | |
34 #define M_PI 3.1415926535897932384626433832795029 | |
35 #endif | |
36 #include <inttypes.h> | |
37 | |
38 #include "a52.h" | |
39 #include "a52_internal.h" | |
40 #include "mm_accel.h" | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
41 #include "mangle.h" |
3394 | 42 |
3884 | 43 #ifdef RUNTIME_CPUDETECT |
44 #undef HAVE_3DNOWEX | |
45 #endif | |
46 | |
47 #define USE_AC3_C | |
48 | |
3394 | 49 void (* imdct_256) (sample_t data[], sample_t delay[], sample_t bias); |
50 void (* imdct_512) (sample_t data[], sample_t delay[], sample_t bias); | |
51 | |
52 typedef struct complex_s { | |
53 sample_t real; | |
54 sample_t imag; | |
55 } complex_t; | |
56 | |
3884 | 57 static void fft_128p(complex_t *a); |
58 | |
12303
f881c918739b
attribute_used patch by (VMiklos <mamajom at axelero dot hu>)
michael
parents:
9122
diff
changeset
|
59 static const int pm128[128] attribute_used __attribute__((aligned(16))) = |
3884 | 60 { |
61 0, 16, 32, 48, 64, 80, 96, 112, 8, 40, 72, 104, 24, 56, 88, 120, | |
62 4, 20, 36, 52, 68, 84, 100, 116, 12, 28, 44, 60, 76, 92, 108, 124, | |
63 2, 18, 34, 50, 66, 82, 98, 114, 10, 42, 74, 106, 26, 58, 90, 122, | |
64 6, 22, 38, 54, 70, 86, 102, 118, 14, 46, 78, 110, 30, 62, 94, 126, | |
65 1, 17, 33, 49, 65, 81, 97, 113, 9, 41, 73, 105, 25, 57, 89, 121, | |
66 5, 21, 37, 53, 69, 85, 101, 117, 13, 29, 45, 61, 77, 93, 109, 125, | |
67 3, 19, 35, 51, 67, 83, 99, 115, 11, 43, 75, 107, 27, 59, 91, 123, | |
68 7, 23, 39, 55, 71, 87, 103, 119, 15, 31, 47, 63, 79, 95, 111, 127 | |
69 }; | |
3394 | 70 |
71 /* 128 point bit-reverse LUT */ | |
12303
f881c918739b
attribute_used patch by (VMiklos <mamajom at axelero dot hu>)
michael
parents:
9122
diff
changeset
|
72 static uint8_t attribute_used bit_reverse_512[] = { |
3394 | 73 0x00, 0x40, 0x20, 0x60, 0x10, 0x50, 0x30, 0x70, |
74 0x08, 0x48, 0x28, 0x68, 0x18, 0x58, 0x38, 0x78, | |
75 0x04, 0x44, 0x24, 0x64, 0x14, 0x54, 0x34, 0x74, | |
76 0x0c, 0x4c, 0x2c, 0x6c, 0x1c, 0x5c, 0x3c, 0x7c, | |
77 0x02, 0x42, 0x22, 0x62, 0x12, 0x52, 0x32, 0x72, | |
78 0x0a, 0x4a, 0x2a, 0x6a, 0x1a, 0x5a, 0x3a, 0x7a, | |
79 0x06, 0x46, 0x26, 0x66, 0x16, 0x56, 0x36, 0x76, | |
80 0x0e, 0x4e, 0x2e, 0x6e, 0x1e, 0x5e, 0x3e, 0x7e, | |
81 0x01, 0x41, 0x21, 0x61, 0x11, 0x51, 0x31, 0x71, | |
82 0x09, 0x49, 0x29, 0x69, 0x19, 0x59, 0x39, 0x79, | |
83 0x05, 0x45, 0x25, 0x65, 0x15, 0x55, 0x35, 0x75, | |
84 0x0d, 0x4d, 0x2d, 0x6d, 0x1d, 0x5d, 0x3d, 0x7d, | |
85 0x03, 0x43, 0x23, 0x63, 0x13, 0x53, 0x33, 0x73, | |
86 0x0b, 0x4b, 0x2b, 0x6b, 0x1b, 0x5b, 0x3b, 0x7b, | |
87 0x07, 0x47, 0x27, 0x67, 0x17, 0x57, 0x37, 0x77, | |
88 0x0f, 0x4f, 0x2f, 0x6f, 0x1f, 0x5f, 0x3f, 0x7f}; | |
89 | |
90 static uint8_t bit_reverse_256[] = { | |
91 0x00, 0x20, 0x10, 0x30, 0x08, 0x28, 0x18, 0x38, | |
92 0x04, 0x24, 0x14, 0x34, 0x0c, 0x2c, 0x1c, 0x3c, | |
93 0x02, 0x22, 0x12, 0x32, 0x0a, 0x2a, 0x1a, 0x3a, | |
94 0x06, 0x26, 0x16, 0x36, 0x0e, 0x2e, 0x1e, 0x3e, | |
95 0x01, 0x21, 0x11, 0x31, 0x09, 0x29, 0x19, 0x39, | |
96 0x05, 0x25, 0x15, 0x35, 0x0d, 0x2d, 0x1d, 0x3d, | |
97 0x03, 0x23, 0x13, 0x33, 0x0b, 0x2b, 0x1b, 0x3b, | |
98 0x07, 0x27, 0x17, 0x37, 0x0f, 0x2f, 0x1f, 0x3f}; | |
99 | |
3579 | 100 #ifdef ARCH_X86 |
3508 | 101 // NOTE: SSE needs 16byte alignment or it will segfault |
3581 | 102 // |
3508 | 103 static complex_t __attribute__((aligned(16))) buf[128]; |
3581 | 104 static float __attribute__((aligned(16))) sseSinCos1c[256]; |
105 static float __attribute__((aligned(16))) sseSinCos1d[256]; | |
12303
f881c918739b
attribute_used patch by (VMiklos <mamajom at axelero dot hu>)
michael
parents:
9122
diff
changeset
|
106 static float attribute_used __attribute__((aligned(16))) ps111_1[4]={1,1,1,-1}; |
3534 | 107 //static float __attribute__((aligned(16))) sseW0[4]; |
108 static float __attribute__((aligned(16))) sseW1[8]; | |
109 static float __attribute__((aligned(16))) sseW2[16]; | |
110 static float __attribute__((aligned(16))) sseW3[32]; | |
111 static float __attribute__((aligned(16))) sseW4[64]; | |
112 static float __attribute__((aligned(16))) sseW5[128]; | |
113 static float __attribute__((aligned(16))) sseW6[256]; | |
114 static float __attribute__((aligned(16))) *sseW[7]= | |
115 {NULL /*sseW0*/,sseW1,sseW2,sseW3,sseW4,sseW5,sseW6}; | |
3553 | 116 static float __attribute__((aligned(16))) sseWindow[512]; |
3508 | 117 #else |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
118 static complex_t __attribute__((aligned(16))) buf[128]; |
3508 | 119 #endif |
3394 | 120 |
121 /* Twiddle factor LUT */ | |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
122 static complex_t __attribute__((aligned(16))) w_1[1]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
123 static complex_t __attribute__((aligned(16))) w_2[2]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
124 static complex_t __attribute__((aligned(16))) w_4[4]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
125 static complex_t __attribute__((aligned(16))) w_8[8]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
126 static complex_t __attribute__((aligned(16))) w_16[16]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
127 static complex_t __attribute__((aligned(16))) w_32[32]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
128 static complex_t __attribute__((aligned(16))) w_64[64]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
129 static complex_t __attribute__((aligned(16))) * w[7] = {w_1, w_2, w_4, w_8, w_16, w_32, w_64}; |
3394 | 130 |
131 /* Twiddle factors for IMDCT */ | |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
132 static sample_t __attribute__((aligned(16))) xcos1[128]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
133 static sample_t __attribute__((aligned(16))) xsin1[128]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
134 static sample_t __attribute__((aligned(16))) xcos2[64]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
135 static sample_t __attribute__((aligned(16))) xsin2[64]; |
3394 | 136 |
137 /* Windowing function for Modified DCT - Thank you acroread */ | |
138 sample_t imdct_window[] = { | |
139 0.00014, 0.00024, 0.00037, 0.00051, 0.00067, 0.00086, 0.00107, 0.00130, | |
140 0.00157, 0.00187, 0.00220, 0.00256, 0.00297, 0.00341, 0.00390, 0.00443, | |
141 0.00501, 0.00564, 0.00632, 0.00706, 0.00785, 0.00871, 0.00962, 0.01061, | |
142 0.01166, 0.01279, 0.01399, 0.01526, 0.01662, 0.01806, 0.01959, 0.02121, | |
143 0.02292, 0.02472, 0.02662, 0.02863, 0.03073, 0.03294, 0.03527, 0.03770, | |
144 0.04025, 0.04292, 0.04571, 0.04862, 0.05165, 0.05481, 0.05810, 0.06153, | |
145 0.06508, 0.06878, 0.07261, 0.07658, 0.08069, 0.08495, 0.08935, 0.09389, | |
146 0.09859, 0.10343, 0.10842, 0.11356, 0.11885, 0.12429, 0.12988, 0.13563, | |
147 0.14152, 0.14757, 0.15376, 0.16011, 0.16661, 0.17325, 0.18005, 0.18699, | |
148 0.19407, 0.20130, 0.20867, 0.21618, 0.22382, 0.23161, 0.23952, 0.24757, | |
149 0.25574, 0.26404, 0.27246, 0.28100, 0.28965, 0.29841, 0.30729, 0.31626, | |
150 0.32533, 0.33450, 0.34376, 0.35311, 0.36253, 0.37204, 0.38161, 0.39126, | |
151 0.40096, 0.41072, 0.42054, 0.43040, 0.44030, 0.45023, 0.46020, 0.47019, | |
152 0.48020, 0.49022, 0.50025, 0.51028, 0.52031, 0.53033, 0.54033, 0.55031, | |
153 0.56026, 0.57019, 0.58007, 0.58991, 0.59970, 0.60944, 0.61912, 0.62873, | |
154 0.63827, 0.64774, 0.65713, 0.66643, 0.67564, 0.68476, 0.69377, 0.70269, | |
155 0.71150, 0.72019, 0.72877, 0.73723, 0.74557, 0.75378, 0.76186, 0.76981, | |
156 0.77762, 0.78530, 0.79283, 0.80022, 0.80747, 0.81457, 0.82151, 0.82831, | |
157 0.83496, 0.84145, 0.84779, 0.85398, 0.86001, 0.86588, 0.87160, 0.87716, | |
158 0.88257, 0.88782, 0.89291, 0.89785, 0.90264, 0.90728, 0.91176, 0.91610, | |
159 0.92028, 0.92432, 0.92822, 0.93197, 0.93558, 0.93906, 0.94240, 0.94560, | |
160 0.94867, 0.95162, 0.95444, 0.95713, 0.95971, 0.96217, 0.96451, 0.96674, | |
161 0.96887, 0.97089, 0.97281, 0.97463, 0.97635, 0.97799, 0.97953, 0.98099, | |
162 0.98236, 0.98366, 0.98488, 0.98602, 0.98710, 0.98811, 0.98905, 0.98994, | |
163 0.99076, 0.99153, 0.99225, 0.99291, 0.99353, 0.99411, 0.99464, 0.99513, | |
164 0.99558, 0.99600, 0.99639, 0.99674, 0.99706, 0.99736, 0.99763, 0.99788, | |
165 0.99811, 0.99831, 0.99850, 0.99867, 0.99882, 0.99895, 0.99908, 0.99919, | |
166 0.99929, 0.99938, 0.99946, 0.99953, 0.99959, 0.99965, 0.99969, 0.99974, | |
167 0.99978, 0.99981, 0.99984, 0.99986, 0.99988, 0.99990, 0.99992, 0.99993, | |
168 0.99994, 0.99995, 0.99996, 0.99997, 0.99998, 0.99998, 0.99998, 0.99999, | |
169 0.99999, 0.99999, 0.99999, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000, | |
170 1.00000, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000, 1.00000 }; | |
171 | |
172 | |
173 static inline void swap_cmplx(complex_t *a, complex_t *b) | |
174 { | |
175 complex_t tmp; | |
176 | |
177 tmp = *a; | |
178 *a = *b; | |
179 *b = tmp; | |
180 } | |
181 | |
182 | |
183 | |
184 static inline complex_t cmplx_mult(complex_t a, complex_t b) | |
185 { | |
186 complex_t ret; | |
187 | |
188 ret.real = a.real * b.real - a.imag * b.imag; | |
189 ret.imag = a.real * b.imag + a.imag * b.real; | |
190 | |
191 return ret; | |
192 } | |
193 | |
194 void | |
195 imdct_do_512(sample_t data[],sample_t delay[], sample_t bias) | |
196 { | |
8254
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
197 int i; |
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
198 #ifndef USE_AC3_C |
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
199 int k; |
3394 | 200 int p,q; |
201 int m; | |
202 int two_m; | |
203 int two_m_plus_one; | |
204 | |
8254
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
205 sample_t tmp_b_i; |
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
206 sample_t tmp_b_r; |
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
207 #endif |
3394 | 208 sample_t tmp_a_i; |
209 sample_t tmp_a_r; | |
210 | |
211 sample_t *data_ptr; | |
212 sample_t *delay_ptr; | |
213 sample_t *window_ptr; | |
214 | |
215 /* 512 IMDCT with source and dest data in 'data' */ | |
216 | |
3884 | 217 /* Pre IFFT complex multiply plus IFFT cmplx conjugate & reordering*/ |
3394 | 218 for( i=0; i < 128; i++) { |
219 /* z[i] = (X[256-2*i-1] + j * X[2*i]) * (xcos1[i] + j * xsin1[i]) ; */ | |
3884 | 220 #ifdef USE_AC3_C |
221 int j= pm128[i]; | |
222 #else | |
223 int j= bit_reverse_512[i]; | |
224 #endif | |
225 buf[i].real = (data[256-2*j-1] * xcos1[j]) - (data[2*j] * xsin1[j]); | |
226 buf[i].imag = -1.0 * ((data[2*j] * xcos1[j]) + (data[256-2*j-1] * xsin1[j])); | |
3394 | 227 } |
228 | |
229 /* FFT Merge */ | |
3549 | 230 /* unoptimized variant |
231 for (m=1; m < 7; m++) { | |
232 if(m) | |
233 two_m = (1 << m); | |
234 else | |
235 two_m = 1; | |
236 | |
237 two_m_plus_one = (1 << (m+1)); | |
238 | |
239 for(i = 0; i < 128; i += two_m_plus_one) { | |
240 for(k = 0; k < two_m; k++) { | |
241 p = k + i; | |
242 q = p + two_m; | |
3508 | 243 tmp_a_r = buf[p].real; |
244 tmp_a_i = buf[p].imag; | |
3549 | 245 tmp_b_r = buf[q].real * w[m][k].real - buf[q].imag * w[m][k].imag; |
246 tmp_b_i = buf[q].imag * w[m][k].real + buf[q].real * w[m][k].imag; | |
3508 | 247 buf[p].real = tmp_a_r + tmp_b_r; |
248 buf[p].imag = tmp_a_i + tmp_b_i; | |
249 buf[q].real = tmp_a_r - tmp_b_r; | |
250 buf[q].imag = tmp_a_i - tmp_b_i; | |
3549 | 251 } |
252 } | |
253 } | |
254 */ | |
3884 | 255 #ifdef USE_AC3_C |
256 fft_128p (&buf[0]); | |
257 #else | |
258 | |
3623 | 259 /* 1. iteration */ |
3579 | 260 for(i = 0; i < 128; i += 2) { |
261 tmp_a_r = buf[i].real; | |
262 tmp_a_i = buf[i].imag; | |
263 tmp_b_r = buf[i+1].real; | |
264 tmp_b_i = buf[i+1].imag; | |
265 buf[i].real = tmp_a_r + tmp_b_r; | |
266 buf[i].imag = tmp_a_i + tmp_b_i; | |
267 buf[i+1].real = tmp_a_r - tmp_b_r; | |
268 buf[i+1].imag = tmp_a_i - tmp_b_i; | |
269 } | |
270 | |
3623 | 271 /* 2. iteration */ |
3579 | 272 // Note w[1]={{1,0}, {0,-1}} |
273 for(i = 0; i < 128; i += 4) { | |
274 tmp_a_r = buf[i].real; | |
275 tmp_a_i = buf[i].imag; | |
276 tmp_b_r = buf[i+2].real; | |
277 tmp_b_i = buf[i+2].imag; | |
278 buf[i].real = tmp_a_r + tmp_b_r; | |
279 buf[i].imag = tmp_a_i + tmp_b_i; | |
280 buf[i+2].real = tmp_a_r - tmp_b_r; | |
281 buf[i+2].imag = tmp_a_i - tmp_b_i; | |
282 tmp_a_r = buf[i+1].real; | |
283 tmp_a_i = buf[i+1].imag; | |
284 tmp_b_r = buf[i+3].imag; | |
285 tmp_b_i = buf[i+3].real; | |
286 buf[i+1].real = tmp_a_r + tmp_b_r; | |
287 buf[i+1].imag = tmp_a_i - tmp_b_i; | |
288 buf[i+3].real = tmp_a_r - tmp_b_r; | |
289 buf[i+3].imag = tmp_a_i + tmp_b_i; | |
290 } | |
291 | |
3623 | 292 /* 3. iteration */ |
3579 | 293 for(i = 0; i < 128; i += 8) { |
294 tmp_a_r = buf[i].real; | |
295 tmp_a_i = buf[i].imag; | |
296 tmp_b_r = buf[i+4].real; | |
297 tmp_b_i = buf[i+4].imag; | |
298 buf[i].real = tmp_a_r + tmp_b_r; | |
299 buf[i].imag = tmp_a_i + tmp_b_i; | |
300 buf[i+4].real = tmp_a_r - tmp_b_r; | |
301 buf[i+4].imag = tmp_a_i - tmp_b_i; | |
302 tmp_a_r = buf[1+i].real; | |
303 tmp_a_i = buf[1+i].imag; | |
304 tmp_b_r = (buf[i+5].real + buf[i+5].imag) * w[2][1].real; | |
305 tmp_b_i = (buf[i+5].imag - buf[i+5].real) * w[2][1].real; | |
306 buf[1+i].real = tmp_a_r + tmp_b_r; | |
307 buf[1+i].imag = tmp_a_i + tmp_b_i; | |
308 buf[i+5].real = tmp_a_r - tmp_b_r; | |
309 buf[i+5].imag = tmp_a_i - tmp_b_i; | |
310 tmp_a_r = buf[i+2].real; | |
311 tmp_a_i = buf[i+2].imag; | |
312 tmp_b_r = buf[i+6].imag; | |
313 tmp_b_i = - buf[i+6].real; | |
314 buf[i+2].real = tmp_a_r + tmp_b_r; | |
315 buf[i+2].imag = tmp_a_i + tmp_b_i; | |
316 buf[i+6].real = tmp_a_r - tmp_b_r; | |
317 buf[i+6].imag = tmp_a_i - tmp_b_i; | |
318 tmp_a_r = buf[i+3].real; | |
319 tmp_a_i = buf[i+3].imag; | |
320 tmp_b_r = (buf[i+7].real - buf[i+7].imag) * w[2][3].imag; | |
321 tmp_b_i = (buf[i+7].imag + buf[i+7].real) * w[2][3].imag; | |
322 buf[i+3].real = tmp_a_r + tmp_b_r; | |
323 buf[i+3].imag = tmp_a_i + tmp_b_i; | |
324 buf[i+7].real = tmp_a_r - tmp_b_r; | |
325 buf[i+7].imag = tmp_a_i - tmp_b_i; | |
326 } | |
327 | |
3623 | 328 /* 4-7. iterations */ |
3579 | 329 for (m=3; m < 7; m++) { |
330 two_m = (1 << m); | |
331 | |
332 two_m_plus_one = two_m<<1; | |
333 | |
334 for(i = 0; i < 128; i += two_m_plus_one) { | |
335 for(k = 0; k < two_m; k++) { | |
336 int p = k + i; | |
337 int q = p + two_m; | |
338 tmp_a_r = buf[p].real; | |
339 tmp_a_i = buf[p].imag; | |
340 tmp_b_r = buf[q].real * w[m][k].real - buf[q].imag * w[m][k].imag; | |
341 tmp_b_i = buf[q].imag * w[m][k].real + buf[q].real * w[m][k].imag; | |
342 buf[p].real = tmp_a_r + tmp_b_r; | |
343 buf[p].imag = tmp_a_i + tmp_b_i; | |
344 buf[q].real = tmp_a_r - tmp_b_r; | |
345 buf[q].imag = tmp_a_i - tmp_b_i; | |
346 } | |
347 } | |
348 } | |
3884 | 349 #endif |
3579 | 350 /* Post IFFT complex multiply plus IFFT complex conjugate*/ |
351 for( i=0; i < 128; i++) { | |
352 /* y[n] = z[n] * (xcos1[n] + j * xsin1[n]) ; */ | |
353 tmp_a_r = buf[i].real; | |
354 tmp_a_i = -1.0 * buf[i].imag; | |
355 buf[i].real =(tmp_a_r * xcos1[i]) - (tmp_a_i * xsin1[i]); | |
356 buf[i].imag =(tmp_a_r * xsin1[i]) + (tmp_a_i * xcos1[i]); | |
357 } | |
358 | |
359 data_ptr = data; | |
360 delay_ptr = delay; | |
361 window_ptr = imdct_window; | |
362 | |
363 /* Window and convert to real valued signal */ | |
364 for(i=0; i< 64; i++) { | |
365 *data_ptr++ = -buf[64+i].imag * *window_ptr++ + *delay_ptr++ + bias; | |
366 *data_ptr++ = buf[64-i-1].real * *window_ptr++ + *delay_ptr++ + bias; | |
367 } | |
368 | |
369 for(i=0; i< 64; i++) { | |
370 *data_ptr++ = -buf[i].real * *window_ptr++ + *delay_ptr++ + bias; | |
371 *data_ptr++ = buf[128-i-1].imag * *window_ptr++ + *delay_ptr++ + bias; | |
372 } | |
3884 | 373 |
3579 | 374 /* The trailing edge of the window goes into the delay line */ |
375 delay_ptr = delay; | |
376 | |
377 for(i=0; i< 64; i++) { | |
378 *delay_ptr++ = -buf[64+i].real * *--window_ptr; | |
379 *delay_ptr++ = buf[64-i-1].imag * *--window_ptr; | |
380 } | |
381 | |
382 for(i=0; i<64; i++) { | |
383 *delay_ptr++ = buf[i].imag * *--window_ptr; | |
384 *delay_ptr++ = -buf[128-i-1].real * *--window_ptr; | |
385 } | |
386 } | |
387 | |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
388 #ifdef HAVE_ALTIVEC |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
389 |
9122 | 390 #ifndef SYS_DARWIN |
391 #include <altivec.h> | |
392 #endif | |
393 | |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
394 // used to build registers permutation vectors (vcprm) |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
395 // the 's' are for words in the _s_econd vector |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
396 #define WORD_0 0x00,0x01,0x02,0x03 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
397 #define WORD_1 0x04,0x05,0x06,0x07 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
398 #define WORD_2 0x08,0x09,0x0a,0x0b |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
399 #define WORD_3 0x0c,0x0d,0x0e,0x0f |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
400 #define WORD_s0 0x10,0x11,0x12,0x13 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
401 #define WORD_s1 0x14,0x15,0x16,0x17 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
402 #define WORD_s2 0x18,0x19,0x1a,0x1b |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
403 #define WORD_s3 0x1c,0x1d,0x1e,0x1f |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
404 |
9122 | 405 #ifdef SYS_DARWIN |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
406 #define vcprm(a,b,c,d) (const vector unsigned char)(WORD_ ## a, WORD_ ## b, WORD_ ## c, WORD_ ## d) |
9122 | 407 #else |
408 #define vcprm(a,b,c,d) (const vector unsigned char){WORD_ ## a, WORD_ ## b, WORD_ ## c, WORD_ ## d} | |
409 #endif | |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
410 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
411 // vcprmle is used to keep the same index as in the SSE version. |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
412 // it's the same as vcprm, with the index inversed |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
413 // ('le' is Little Endian) |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
414 #define vcprmle(a,b,c,d) vcprm(d,c,b,a) |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
415 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
416 // used to build inverse/identity vectors (vcii) |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
417 // n is _n_egative, p is _p_ositive |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
418 #define FLOAT_n -1. |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
419 #define FLOAT_p 1. |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
420 |
9122 | 421 #ifdef SYS_DARWIN |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
422 #define vcii(a,b,c,d) (const vector float)(FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d) |
9122 | 423 #else |
424 #define vcii(a,b,c,d) (const vector float){FLOAT_ ## a, FLOAT_ ## b, FLOAT_ ## c, FLOAT_ ## d} | |
425 #endif | |
426 | |
427 #ifdef SYS_DARWIN | |
428 #define FOUROF(a) (a) | |
429 #else | |
430 #define FOUROF(a) {a,a,a,a} | |
431 #endif | |
432 | |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
433 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
434 void |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
435 imdct_do_512_altivec(sample_t data[],sample_t delay[], sample_t bias) |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
436 { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
437 int i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
438 int k; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
439 int p,q; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
440 int m; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
441 int two_m; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
442 int two_m_plus_one; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
443 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
444 sample_t tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
445 sample_t tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
446 sample_t tmp_a_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
447 sample_t tmp_a_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
448 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
449 sample_t *data_ptr; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
450 sample_t *delay_ptr; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
451 sample_t *window_ptr; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
452 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
453 /* 512 IMDCT with source and dest data in 'data' */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
454 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
455 /* Pre IFFT complex multiply plus IFFT cmplx conjugate & reordering*/ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
456 for( i=0; i < 128; i++) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
457 /* z[i] = (X[256-2*i-1] + j * X[2*i]) * (xcos1[i] + j * xsin1[i]) ; */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
458 int j= bit_reverse_512[i]; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
459 buf[i].real = (data[256-2*j-1] * xcos1[j]) - (data[2*j] * xsin1[j]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
460 buf[i].imag = -1.0 * ((data[2*j] * xcos1[j]) + (data[256-2*j-1] * xsin1[j])); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
461 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
462 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
463 /* 1. iteration */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
464 for(i = 0; i < 128; i += 2) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
465 #if 0 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
466 tmp_a_r = buf[i].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
467 tmp_a_i = buf[i].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
468 tmp_b_r = buf[i+1].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
469 tmp_b_i = buf[i+1].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
470 buf[i].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
471 buf[i].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
472 buf[i+1].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
473 buf[i+1].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
474 #else |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
475 vector float temp, bufv; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
476 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
477 bufv = vec_ld(i << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
478 temp = vec_perm(bufv, bufv, vcprm(2,3,0,1)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
479 bufv = vec_madd(bufv, vcii(p,p,n,n), temp); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
480 vec_st(bufv, i << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
481 #endif |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
482 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
483 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
484 /* 2. iteration */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
485 // Note w[1]={{1,0}, {0,-1}} |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
486 for(i = 0; i < 128; i += 4) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
487 #if 0 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
488 tmp_a_r = buf[i].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
489 tmp_a_i = buf[i].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
490 tmp_b_r = buf[i+2].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
491 tmp_b_i = buf[i+2].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
492 buf[i].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
493 buf[i].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
494 buf[i+2].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
495 buf[i+2].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
496 tmp_a_r = buf[i+1].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
497 tmp_a_i = buf[i+1].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
498 /* WARNING: im <-> re here ! */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
499 tmp_b_r = buf[i+3].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
500 tmp_b_i = buf[i+3].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
501 buf[i+1].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
502 buf[i+1].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
503 buf[i+3].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
504 buf[i+3].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
505 #else |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
506 vector float buf01, buf23, temp1, temp2; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
507 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
508 buf01 = vec_ld((i + 0) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
509 buf23 = vec_ld((i + 2) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
510 buf23 = vec_perm(buf23,buf23,vcprm(0,1,3,2)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
511 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
512 temp1 = vec_madd(buf23, vcii(p,p,p,n), buf01); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
513 temp2 = vec_madd(buf23, vcii(n,n,n,p), buf01); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
514 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
515 vec_st(temp1, (i + 0) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
516 vec_st(temp2, (i + 2) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
517 #endif |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
518 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
519 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
520 /* 3. iteration */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
521 for(i = 0; i < 128; i += 8) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
522 #if 0 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
523 tmp_a_r = buf[i].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
524 tmp_a_i = buf[i].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
525 tmp_b_r = buf[i+4].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
526 tmp_b_i = buf[i+4].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
527 buf[i].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
528 buf[i].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
529 buf[i+4].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
530 buf[i+4].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
531 tmp_a_r = buf[1+i].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
532 tmp_a_i = buf[1+i].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
533 tmp_b_r = (buf[i+5].real + buf[i+5].imag) * w[2][1].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
534 tmp_b_i = (buf[i+5].imag - buf[i+5].real) * w[2][1].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
535 buf[1+i].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
536 buf[1+i].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
537 buf[i+5].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
538 buf[i+5].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
539 tmp_a_r = buf[i+2].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
540 tmp_a_i = buf[i+2].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
541 /* WARNING re <-> im & sign */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
542 tmp_b_r = buf[i+6].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
543 tmp_b_i = - buf[i+6].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
544 buf[i+2].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
545 buf[i+2].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
546 buf[i+6].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
547 buf[i+6].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
548 tmp_a_r = buf[i+3].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
549 tmp_a_i = buf[i+3].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
550 tmp_b_r = (buf[i+7].real - buf[i+7].imag) * w[2][3].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
551 tmp_b_i = (buf[i+7].imag + buf[i+7].real) * w[2][3].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
552 buf[i+3].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
553 buf[i+3].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
554 buf[i+7].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
555 buf[i+7].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
556 #else |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
557 vector float buf01, buf23, buf45, buf67; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
558 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
559 buf01 = vec_ld((i + 0) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
560 buf23 = vec_ld((i + 2) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
561 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
562 tmp_b_r = (buf[i+5].real + buf[i+5].imag) * w[2][1].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
563 tmp_b_i = (buf[i+5].imag - buf[i+5].real) * w[2][1].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
564 buf[i+5].real = tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
565 buf[i+5].imag = tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
566 tmp_b_r = (buf[i+7].real - buf[i+7].imag) * w[2][3].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
567 tmp_b_i = (buf[i+7].imag + buf[i+7].real) * w[2][3].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
568 buf[i+7].real = tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
569 buf[i+7].imag = tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
570 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
571 buf23 = vec_ld((i + 2) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
572 buf45 = vec_ld((i + 4) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
573 buf67 = vec_ld((i + 6) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
574 buf67 = vec_perm(buf67, buf67, vcprm(1,0,2,3)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
575 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
576 vec_st(vec_add(buf01, buf45), (i + 0) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
577 vec_st(vec_madd(buf67, vcii(p,n,p,p), buf23), (i + 2) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
578 vec_st(vec_sub(buf01, buf45), (i + 4) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
579 vec_st(vec_nmsub(buf67, vcii(p,n,p,p), buf23), (i + 6) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
580 #endif |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
581 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
582 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
583 /* 4-7. iterations */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
584 for (m=3; m < 7; m++) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
585 two_m = (1 << m); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
586 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
587 two_m_plus_one = two_m<<1; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
588 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
589 for(i = 0; i < 128; i += two_m_plus_one) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
590 for(k = 0; k < two_m; k+=2) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
591 #if 0 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
592 int p = k + i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
593 int q = p + two_m; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
594 tmp_a_r = buf[p].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
595 tmp_a_i = buf[p].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
596 tmp_b_r = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
597 buf[q].real * w[m][k].real - |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
598 buf[q].imag * w[m][k].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
599 tmp_b_i = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
600 buf[q].imag * w[m][k].real + |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
601 buf[q].real * w[m][k].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
602 buf[p].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
603 buf[p].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
604 buf[q].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
605 buf[q].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
606 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
607 tmp_a_r = buf[(p + 1)].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
608 tmp_a_i = buf[(p + 1)].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
609 tmp_b_r = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
610 buf[(q + 1)].real * w[m][(k + 1)].real - |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
611 buf[(q + 1)].imag * w[m][(k + 1)].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
612 tmp_b_i = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
613 buf[(q + 1)].imag * w[m][(k + 1)].real + |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
614 buf[(q + 1)].real * w[m][(k + 1)].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
615 buf[(p + 1)].real = tmp_a_r + tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
616 buf[(p + 1)].imag = tmp_a_i + tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
617 buf[(q + 1)].real = tmp_a_r - tmp_b_r; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
618 buf[(q + 1)].imag = tmp_a_i - tmp_b_i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
619 #else |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
620 int p = k + i; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
621 int q = p + two_m; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
622 vector float vecp, vecq, vecw, temp1, temp2, temp3, temp4; |
9122 | 623 const vector float vczero = (const vector float)FOUROF(0.); |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
624 // first compute buf[q] and buf[q+1] |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
625 vecq = vec_ld(q << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
626 vecw = vec_ld(0, (float*)&(w[m][k])); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
627 temp1 = vec_madd(vecq, vecw, vczero); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
628 temp2 = vec_perm(vecq, vecq, vcprm(1,0,3,2)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
629 temp2 = vec_madd(temp2, vecw, vczero); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
630 temp3 = vec_perm(temp1, temp2, vcprm(0,s0,2,s2)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
631 temp4 = vec_perm(temp1, temp2, vcprm(1,s1,3,s3)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
632 vecq = vec_madd(temp4, vcii(n,p,n,p), temp3); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
633 // then butterfly with buf[p] and buf[p+1] |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
634 vecp = vec_ld(p << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
635 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
636 temp1 = vec_add(vecp, vecq); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
637 temp2 = vec_sub(vecp, vecq); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
638 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
639 vec_st(temp1, p << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
640 vec_st(temp2, q << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
641 #endif |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
642 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
643 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
644 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
645 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
646 /* Post IFFT complex multiply plus IFFT complex conjugate*/ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
647 for( i=0; i < 128; i+=4) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
648 /* y[n] = z[n] * (xcos1[n] + j * xsin1[n]) ; */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
649 #if 0 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
650 tmp_a_r = buf[(i + 0)].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
651 tmp_a_i = -1.0 * buf[(i + 0)].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
652 buf[(i + 0)].real = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
653 (tmp_a_r * xcos1[(i + 0)]) - (tmp_a_i * xsin1[(i + 0)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
654 buf[(i + 0)].imag = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
655 (tmp_a_r * xsin1[(i + 0)]) + (tmp_a_i * xcos1[(i + 0)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
656 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
657 tmp_a_r = buf[(i + 1)].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
658 tmp_a_i = -1.0 * buf[(i + 1)].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
659 buf[(i + 1)].real = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
660 (tmp_a_r * xcos1[(i + 1)]) - (tmp_a_i * xsin1[(i + 1)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
661 buf[(i + 1)].imag = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
662 (tmp_a_r * xsin1[(i + 1)]) + (tmp_a_i * xcos1[(i + 1)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
663 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
664 tmp_a_r = buf[(i + 2)].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
665 tmp_a_i = -1.0 * buf[(i + 2)].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
666 buf[(i + 2)].real = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
667 (tmp_a_r * xcos1[(i + 2)]) - (tmp_a_i * xsin1[(i + 2)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
668 buf[(i + 2)].imag = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
669 (tmp_a_r * xsin1[(i + 2)]) + (tmp_a_i * xcos1[(i + 2)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
670 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
671 tmp_a_r = buf[(i + 3)].real; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
672 tmp_a_i = -1.0 * buf[(i + 3)].imag; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
673 buf[(i + 3)].real = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
674 (tmp_a_r * xcos1[(i + 3)]) - (tmp_a_i * xsin1[(i + 3)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
675 buf[(i + 3)].imag = |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
676 (tmp_a_r * xsin1[(i + 3)]) + (tmp_a_i * xcos1[(i + 3)]); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
677 #else |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
678 vector float bufv_0, bufv_2, cosv, sinv, temp1, temp2; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
679 vector float temp0022, temp1133, tempCS01; |
9122 | 680 const vector float vczero = (const vector float)FOUROF(0.); |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
681 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
682 bufv_0 = vec_ld((i + 0) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
683 bufv_2 = vec_ld((i + 2) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
684 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
685 cosv = vec_ld(i << 2, xcos1); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
686 sinv = vec_ld(i << 2, xsin1); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
687 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
688 temp0022 = vec_perm(bufv_0, bufv_0, vcprm(0,0,2,2)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
689 temp1133 = vec_perm(bufv_0, bufv_0, vcprm(1,1,3,3)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
690 tempCS01 = vec_perm(cosv, sinv, vcprm(0,s0,1,s1)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
691 temp1 = vec_madd(temp0022, tempCS01, vczero); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
692 tempCS01 = vec_perm(cosv, sinv, vcprm(s0,0,s1,1)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
693 temp2 = vec_madd(temp1133, tempCS01, vczero); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
694 bufv_0 = vec_madd(temp2, vcii(p,n,p,n), temp1); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
695 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
696 vec_st(bufv_0, (i + 0) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
697 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
698 /* idem with bufv_2 and high-order cosv/sinv */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
699 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
700 temp0022 = vec_perm(bufv_2, bufv_2, vcprm(0,0,2,2)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
701 temp1133 = vec_perm(bufv_2, bufv_2, vcprm(1,1,3,3)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
702 tempCS01 = vec_perm(cosv, sinv, vcprm(2,s2,3,s3)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
703 temp1 = vec_madd(temp0022, tempCS01, vczero); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
704 tempCS01 = vec_perm(cosv, sinv, vcprm(s2,2,s3,3)); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
705 temp2 = vec_madd(temp1133, tempCS01, vczero); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
706 bufv_2 = vec_madd(temp2, vcii(p,n,p,n), temp1); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
707 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
708 vec_st(bufv_2, (i + 2) << 3, (float*)buf); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
709 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
710 #endif |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
711 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
712 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
713 data_ptr = data; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
714 delay_ptr = delay; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
715 window_ptr = imdct_window; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
716 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
717 /* Window and convert to real valued signal */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
718 for(i=0; i< 64; i++) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
719 *data_ptr++ = -buf[64+i].imag * *window_ptr++ + *delay_ptr++ + bias; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
720 *data_ptr++ = buf[64-i-1].real * *window_ptr++ + *delay_ptr++ + bias; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
721 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
722 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
723 for(i=0; i< 64; i++) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
724 *data_ptr++ = -buf[i].real * *window_ptr++ + *delay_ptr++ + bias; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
725 *data_ptr++ = buf[128-i-1].imag * *window_ptr++ + *delay_ptr++ + bias; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
726 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
727 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
728 /* The trailing edge of the window goes into the delay line */ |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
729 delay_ptr = delay; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
730 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
731 for(i=0; i< 64; i++) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
732 *delay_ptr++ = -buf[64+i].real * *--window_ptr; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
733 *delay_ptr++ = buf[64-i-1].imag * *--window_ptr; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
734 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
735 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
736 for(i=0; i<64; i++) { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
737 *delay_ptr++ = buf[i].imag * *--window_ptr; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
738 *delay_ptr++ = -buf[128-i-1].real * *--window_ptr; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
739 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
740 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
741 #endif |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
742 |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
743 |
4497 | 744 // Stuff below this line is borrowed from libac3 |
745 #include "srfftp.h" | |
3579 | 746 #ifdef ARCH_X86 |
4497 | 747 #ifndef HAVE_3DNOW |
748 #define HAVE_3DNOW 1 | |
749 #endif | |
3884 | 750 #include "srfftp_3dnow.h" |
751 | |
8451 | 752 const i_cmplx_t x_plus_minus_3dnow __attribute__ ((aligned (8))) = {{ 0x00000000UL, 0x80000000UL }}; |
753 const i_cmplx_t x_minus_plus_3dnow __attribute__ ((aligned (8))) = {{ 0x80000000UL, 0x00000000UL }}; | |
3884 | 754 const complex_t HSQRT2_3DNOW __attribute__ ((aligned (8))) = { 0.707106781188, 0.707106781188 }; |
755 | |
4497 | 756 #undef HAVE_3DNOWEX |
757 #include "imdct_3dnow.h" | |
758 #define HAVE_3DNOWEX | |
759 #include "imdct_3dnow.h" | |
3884 | 760 |
3579 | 761 void |
762 imdct_do_512_sse(sample_t data[],sample_t delay[], sample_t bias) | |
763 { | |
8254
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
764 /* int i,k; |
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
765 int p,q;*/ |
3579 | 766 int m; |
767 int two_m; | |
768 int two_m_plus_one; | |
769 | |
8254
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
770 /* sample_t tmp_a_i; |
3579 | 771 sample_t tmp_a_r; |
772 sample_t tmp_b_i; | |
8254
772d6d27fd66
warning patch by (Dominik Mierzejewski <dominik at rangers dot eu dot org>)
michael
parents:
4497
diff
changeset
|
773 sample_t tmp_b_r;*/ |
3579 | 774 |
775 sample_t *data_ptr; | |
776 sample_t *delay_ptr; | |
777 sample_t *window_ptr; | |
778 | |
779 /* 512 IMDCT with source and dest data in 'data' */ | |
3623 | 780 /* see the c version (dct_do_512()), its allmost identical, just in C */ |
781 | |
3579 | 782 /* Pre IFFT complex multiply plus IFFT cmplx conjugate */ |
783 /* Bit reversed shuffling */ | |
784 asm volatile( | |
785 "xorl %%esi, %%esi \n\t" | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
786 "leal "MANGLE(bit_reverse_512)", %%eax \n\t" |
3579 | 787 "movl $1008, %%edi \n\t" |
788 "pushl %%ebp \n\t" //use ebp without telling gcc | |
789 ".balign 16 \n\t" | |
790 "1: \n\t" | |
3584 | 791 "movlps (%0, %%esi), %%xmm0 \n\t" // XXXI |
792 "movhps 8(%0, %%edi), %%xmm0 \n\t" // RXXI | |
793 "movlps 8(%0, %%esi), %%xmm1 \n\t" // XXXi | |
794 "movhps (%0, %%edi), %%xmm1 \n\t" // rXXi | |
795 "shufps $0x33, %%xmm1, %%xmm0 \n\t" // irIR | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
796 "movaps "MANGLE(sseSinCos1c)"(%%esi), %%xmm2\n\t" |
3584 | 797 "mulps %%xmm0, %%xmm2 \n\t" |
798 "shufps $0xB1, %%xmm0, %%xmm0 \n\t" // riRI | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
799 "mulps "MANGLE(sseSinCos1d)"(%%esi), %%xmm0\n\t" |
3584 | 800 "subps %%xmm0, %%xmm2 \n\t" |
3579 | 801 "movzbl (%%eax), %%edx \n\t" |
802 "movzbl 1(%%eax), %%ebp \n\t" | |
3584 | 803 "movlps %%xmm2, (%1, %%edx,8) \n\t" |
804 "movhps %%xmm2, (%1, %%ebp,8) \n\t" | |
3579 | 805 "addl $16, %%esi \n\t" |
806 "addl $2, %%eax \n\t" // avoid complex addressing for P4 crap | |
807 "subl $16, %%edi \n\t" | |
808 " jnc 1b \n\t" | |
809 "popl %%ebp \n\t"//no we didnt touch ebp *g* | |
810 :: "b" (data), "c" (buf) | |
811 : "%esi", "%edi", "%eax", "%edx" | |
812 ); | |
813 | |
814 | |
815 /* FFT Merge */ | |
816 /* unoptimized variant | |
817 for (m=1; m < 7; m++) { | |
818 if(m) | |
819 two_m = (1 << m); | |
820 else | |
821 two_m = 1; | |
822 | |
823 two_m_plus_one = (1 << (m+1)); | |
824 | |
825 for(i = 0; i < 128; i += two_m_plus_one) { | |
826 for(k = 0; k < two_m; k++) { | |
827 p = k + i; | |
828 q = p + two_m; | |
829 tmp_a_r = buf[p].real; | |
830 tmp_a_i = buf[p].imag; | |
831 tmp_b_r = buf[q].real * w[m][k].real - buf[q].imag * w[m][k].imag; | |
832 tmp_b_i = buf[q].imag * w[m][k].real + buf[q].real * w[m][k].imag; | |
833 buf[p].real = tmp_a_r + tmp_b_r; | |
834 buf[p].imag = tmp_a_i + tmp_b_i; | |
835 buf[q].real = tmp_a_r - tmp_b_r; | |
836 buf[q].imag = tmp_a_i - tmp_b_i; | |
837 } | |
838 } | |
839 } | |
840 */ | |
841 | |
3623 | 842 /* 1. iteration */ |
3549 | 843 // Note w[0][0]={1,0} |
3508 | 844 asm volatile( |
845 "xorps %%xmm1, %%xmm1 \n\t" | |
846 "xorps %%xmm2, %%xmm2 \n\t" | |
847 "movl %0, %%esi \n\t" | |
3529 | 848 ".balign 16 \n\t" |
3508 | 849 "1: \n\t" |
850 "movlps (%%esi), %%xmm0 \n\t" //buf[p] | |
851 "movlps 8(%%esi), %%xmm1\n\t" //buf[q] | |
852 "movhps (%%esi), %%xmm0 \n\t" //buf[p] | |
853 "movhps 8(%%esi), %%xmm2\n\t" //buf[q] | |
854 "addps %%xmm1, %%xmm0 \n\t" | |
855 "subps %%xmm2, %%xmm0 \n\t" | |
856 "movaps %%xmm0, (%%esi) \n\t" | |
857 "addl $16, %%esi \n\t" | |
858 "cmpl %1, %%esi \n\t" | |
859 " jb 1b \n\t" | |
860 :: "g" (buf), "r" (buf + 128) | |
861 : "%esi" | |
862 ); | |
3549 | 863 |
3623 | 864 /* 2. iteration */ |
3512 | 865 // Note w[1]={{1,0}, {0,-1}} |
866 asm volatile( | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
867 "movaps "MANGLE(ps111_1)", %%xmm7\n\t" // 1,1,1,-1 |
3512 | 868 "movl %0, %%esi \n\t" |
3529 | 869 ".balign 16 \n\t" |
3512 | 870 "1: \n\t" |
871 "movaps 16(%%esi), %%xmm2 \n\t" //r2,i2,r3,i3 | |
872 "shufps $0xB4, %%xmm2, %%xmm2 \n\t" //r2,i2,i3,r3 | |
873 "mulps %%xmm7, %%xmm2 \n\t" //r2,i2,i3,-r3 | |
874 "movaps (%%esi), %%xmm0 \n\t" //r0,i0,r1,i1 | |
875 "movaps (%%esi), %%xmm1 \n\t" //r0,i0,r1,i1 | |
876 "addps %%xmm2, %%xmm0 \n\t" | |
877 "subps %%xmm2, %%xmm1 \n\t" | |
878 "movaps %%xmm0, (%%esi) \n\t" | |
879 "movaps %%xmm1, 16(%%esi) \n\t" | |
880 "addl $32, %%esi \n\t" | |
881 "cmpl %1, %%esi \n\t" | |
882 " jb 1b \n\t" | |
883 :: "g" (buf), "r" (buf + 128) | |
884 : "%esi" | |
885 ); | |
3549 | 886 |
3623 | 887 /* 3. iteration */ |
3534 | 888 /* |
889 Note sseW2+0={1,1,sqrt(2),sqrt(2)) | |
890 Note sseW2+16={0,0,sqrt(2),-sqrt(2)) | |
891 Note sseW2+32={0,0,-sqrt(2),-sqrt(2)) | |
892 Note sseW2+48={1,-1,sqrt(2),-sqrt(2)) | |
893 */ | |
894 asm volatile( | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
895 "movaps 48+"MANGLE(sseW2)", %%xmm6\n\t" |
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
896 "movaps 16+"MANGLE(sseW2)", %%xmm7\n\t" |
3534 | 897 "xorps %%xmm5, %%xmm5 \n\t" |
898 "xorps %%xmm2, %%xmm2 \n\t" | |
899 "movl %0, %%esi \n\t" | |
900 ".balign 16 \n\t" | |
901 "1: \n\t" | |
3537 | 902 "movaps 32(%%esi), %%xmm2 \n\t" //r4,i4,r5,i5 |
3534 | 903 "movaps 48(%%esi), %%xmm3 \n\t" //r6,i6,r7,i7 |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
904 "movaps "MANGLE(sseW2)", %%xmm4 \n\t" //r4,i4,r5,i5 |
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
905 "movaps 32+"MANGLE(sseW2)", %%xmm5\n\t" //r6,i6,r7,i7 |
3537 | 906 "mulps %%xmm2, %%xmm4 \n\t" |
907 "mulps %%xmm3, %%xmm5 \n\t" | |
3534 | 908 "shufps $0xB1, %%xmm2, %%xmm2 \n\t" //i4,r4,i5,r5 |
909 "shufps $0xB1, %%xmm3, %%xmm3 \n\t" //i6,r6,i7,r7 | |
3537 | 910 "mulps %%xmm6, %%xmm3 \n\t" |
3534 | 911 "mulps %%xmm7, %%xmm2 \n\t" |
912 "movaps (%%esi), %%xmm0 \n\t" //r0,i0,r1,i1 | |
913 "movaps 16(%%esi), %%xmm1 \n\t" //r2,i2,r3,i3 | |
914 "addps %%xmm4, %%xmm2 \n\t" | |
915 "addps %%xmm5, %%xmm3 \n\t" | |
916 "movaps %%xmm2, %%xmm4 \n\t" | |
917 "movaps %%xmm3, %%xmm5 \n\t" | |
918 "addps %%xmm0, %%xmm2 \n\t" | |
919 "addps %%xmm1, %%xmm3 \n\t" | |
920 "subps %%xmm4, %%xmm0 \n\t" | |
921 "subps %%xmm5, %%xmm1 \n\t" | |
922 "movaps %%xmm2, (%%esi) \n\t" | |
923 "movaps %%xmm3, 16(%%esi) \n\t" | |
924 "movaps %%xmm0, 32(%%esi) \n\t" | |
925 "movaps %%xmm1, 48(%%esi) \n\t" | |
926 "addl $64, %%esi \n\t" | |
927 "cmpl %1, %%esi \n\t" | |
928 " jb 1b \n\t" | |
929 :: "g" (buf), "r" (buf + 128) | |
930 : "%esi" | |
931 ); | |
3508 | 932 |
3623 | 933 /* 4-7. iterations */ |
3546 | 934 for (m=3; m < 7; m++) { |
935 two_m = (1 << m); | |
936 two_m_plus_one = two_m<<1; | |
937 asm volatile( | |
938 "movl %0, %%esi \n\t" | |
939 ".balign 16 \n\t" | |
940 "1: \n\t" | |
941 "xorl %%edi, %%edi \n\t" // k | |
942 "leal (%%esi, %3), %%edx \n\t" | |
943 "2: \n\t" | |
944 "movaps (%%edx, %%edi), %%xmm1 \n\t" | |
945 "movaps (%4, %%edi, 2), %%xmm2 \n\t" | |
946 "mulps %%xmm1, %%xmm2 \n\t" | |
947 "shufps $0xB1, %%xmm1, %%xmm1 \n\t" | |
948 "mulps 16(%4, %%edi, 2), %%xmm1 \n\t" | |
949 "movaps (%%esi, %%edi), %%xmm0 \n\t" | |
950 "addps %%xmm2, %%xmm1 \n\t" | |
951 "movaps %%xmm1, %%xmm2 \n\t" | |
952 "addps %%xmm0, %%xmm1 \n\t" | |
953 "subps %%xmm2, %%xmm0 \n\t" | |
954 "movaps %%xmm1, (%%esi, %%edi) \n\t" | |
955 "movaps %%xmm0, (%%edx, %%edi) \n\t" | |
956 "addl $16, %%edi \n\t" | |
957 "cmpl %3, %%edi \n\t" //FIXME (opt) count against 0 | |
958 " jb 2b \n\t" | |
959 "addl %2, %%esi \n\t" | |
960 "cmpl %1, %%esi \n\t" | |
961 " jb 1b \n\t" | |
962 :: "g" (buf), "m" (buf+128), "m" (two_m_plus_one<<3), "r" (two_m<<3), | |
963 "r" (sseW[m]) | |
964 : "%esi", "%edi", "%edx" | |
965 ); | |
966 } | |
967 | |
3623 | 968 /* Post IFFT complex multiply plus IFFT complex conjugate*/ |
3581 | 969 asm volatile( |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
970 "movl $-1024, %%esi \n\t" |
3581 | 971 ".balign 16 \n\t" |
972 "1: \n\t" | |
973 "movaps (%0, %%esi), %%xmm0 \n\t" | |
974 "movaps (%0, %%esi), %%xmm1 \n\t" | |
975 "shufps $0xB1, %%xmm0, %%xmm0 \n\t" | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
976 "mulps 1024+"MANGLE(sseSinCos1c)"(%%esi), %%xmm1\n\t" |
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
977 "mulps 1024+"MANGLE(sseSinCos1d)"(%%esi), %%xmm0\n\t" |
3581 | 978 "addps %%xmm1, %%xmm0 \n\t" |
979 "movaps %%xmm0, (%0, %%esi) \n\t" | |
980 "addl $16, %%esi \n\t" | |
981 " jnz 1b \n\t" | |
982 :: "r" (buf+128) | |
983 : "%esi" | |
984 ); | |
985 | |
3394 | 986 |
987 data_ptr = data; | |
988 delay_ptr = delay; | |
989 window_ptr = imdct_window; | |
990 | |
991 /* Window and convert to real valued signal */ | |
3552 | 992 asm volatile( |
993 "xorl %%edi, %%edi \n\t" // 0 | |
994 "xorl %%esi, %%esi \n\t" // 0 | |
995 "movss %3, %%xmm2 \n\t" // bias | |
996 "shufps $0x00, %%xmm2, %%xmm2 \n\t" // bias, bias, ... | |
997 ".balign 16 \n\t" | |
998 "1: \n\t" | |
999 "movlps (%0, %%esi), %%xmm0 \n\t" // ? ? A ? | |
1000 "movlps 8(%0, %%esi), %%xmm1 \n\t" // ? ? C ? | |
1001 "movhps -16(%0, %%edi), %%xmm1 \n\t" // ? D C ? | |
1002 "movhps -8(%0, %%edi), %%xmm0 \n\t" // ? B A ? | |
1003 "shufps $0x99, %%xmm1, %%xmm0 \n\t" // D C B A | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
1004 "mulps "MANGLE(sseWindow)"(%%esi), %%xmm0\n\t" |
3552 | 1005 "addps (%2, %%esi), %%xmm0 \n\t" |
1006 "addps %%xmm2, %%xmm0 \n\t" | |
1007 "movaps %%xmm0, (%1, %%esi) \n\t" | |
1008 "addl $16, %%esi \n\t" | |
1009 "subl $16, %%edi \n\t" | |
1010 "cmpl $512, %%esi \n\t" | |
1011 " jb 1b \n\t" | |
1012 :: "r" (buf+64), "r" (data_ptr), "r" (delay_ptr), "m" (bias) | |
1013 : "%esi", "%edi" | |
1014 ); | |
1015 data_ptr+=128; | |
1016 delay_ptr+=128; | |
3553 | 1017 // window_ptr+=128; |
3579 | 1018 |
3552 | 1019 asm volatile( |
1020 "movl $1024, %%edi \n\t" // 512 | |
1021 "xorl %%esi, %%esi \n\t" // 0 | |
1022 "movss %3, %%xmm2 \n\t" // bias | |
1023 "shufps $0x00, %%xmm2, %%xmm2 \n\t" // bias, bias, ... | |
1024 ".balign 16 \n\t" | |
1025 "1: \n\t" | |
1026 "movlps (%0, %%esi), %%xmm0 \n\t" // ? ? ? A | |
1027 "movlps 8(%0, %%esi), %%xmm1 \n\t" // ? ? ? C | |
1028 "movhps -16(%0, %%edi), %%xmm1 \n\t" // D ? ? C | |
1029 "movhps -8(%0, %%edi), %%xmm0 \n\t" // B ? ? A | |
1030 "shufps $0xCC, %%xmm1, %%xmm0 \n\t" // D C B A | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
1031 "mulps 512+"MANGLE(sseWindow)"(%%esi), %%xmm0\n\t" |
3552 | 1032 "addps (%2, %%esi), %%xmm0 \n\t" |
1033 "addps %%xmm2, %%xmm0 \n\t" | |
1034 "movaps %%xmm0, (%1, %%esi) \n\t" | |
1035 "addl $16, %%esi \n\t" | |
1036 "subl $16, %%edi \n\t" | |
1037 "cmpl $512, %%esi \n\t" | |
1038 " jb 1b \n\t" | |
1039 :: "r" (buf), "r" (data_ptr), "r" (delay_ptr), "m" (bias) | |
1040 : "%esi", "%edi" | |
1041 ); | |
1042 data_ptr+=128; | |
3553 | 1043 // window_ptr+=128; |
3394 | 1044 |
1045 /* The trailing edge of the window goes into the delay line */ | |
1046 delay_ptr = delay; | |
1047 | |
3553 | 1048 asm volatile( |
1049 "xorl %%edi, %%edi \n\t" // 0 | |
1050 "xorl %%esi, %%esi \n\t" // 0 | |
1051 ".balign 16 \n\t" | |
1052 "1: \n\t" | |
1053 "movlps (%0, %%esi), %%xmm0 \n\t" // ? ? ? A | |
1054 "movlps 8(%0, %%esi), %%xmm1 \n\t" // ? ? ? C | |
1055 "movhps -16(%0, %%edi), %%xmm1 \n\t" // D ? ? C | |
1056 "movhps -8(%0, %%edi), %%xmm0 \n\t" // B ? ? A | |
1057 "shufps $0xCC, %%xmm1, %%xmm0 \n\t" // D C B A | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
1058 "mulps 1024+"MANGLE(sseWindow)"(%%esi), %%xmm0\n\t" |
3553 | 1059 "movaps %%xmm0, (%1, %%esi) \n\t" |
1060 "addl $16, %%esi \n\t" | |
1061 "subl $16, %%edi \n\t" | |
1062 "cmpl $512, %%esi \n\t" | |
1063 " jb 1b \n\t" | |
1064 :: "r" (buf+64), "r" (delay_ptr) | |
1065 : "%esi", "%edi" | |
1066 ); | |
1067 delay_ptr+=128; | |
1068 // window_ptr-=128; | |
3579 | 1069 |
3553 | 1070 asm volatile( |
1071 "movl $1024, %%edi \n\t" // 1024 | |
1072 "xorl %%esi, %%esi \n\t" // 0 | |
1073 ".balign 16 \n\t" | |
1074 "1: \n\t" | |
1075 "movlps (%0, %%esi), %%xmm0 \n\t" // ? ? A ? | |
1076 "movlps 8(%0, %%esi), %%xmm1 \n\t" // ? ? C ? | |
1077 "movhps -16(%0, %%edi), %%xmm1 \n\t" // ? D C ? | |
1078 "movhps -8(%0, %%edi), %%xmm0 \n\t" // ? B A ? | |
1079 "shufps $0x99, %%xmm1, %%xmm0 \n\t" // D C B A | |
4247
2dbd637ffe05
mangle for win32 in liba52 (includes dummy mangle.h pointing to the one in main)
atmos4
parents:
3908
diff
changeset
|
1080 "mulps 1536+"MANGLE(sseWindow)"(%%esi), %%xmm0\n\t" |
3553 | 1081 "movaps %%xmm0, (%1, %%esi) \n\t" |
1082 "addl $16, %%esi \n\t" | |
1083 "subl $16, %%edi \n\t" | |
1084 "cmpl $512, %%esi \n\t" | |
1085 " jb 1b \n\t" | |
1086 :: "r" (buf), "r" (delay_ptr) | |
1087 : "%esi", "%edi" | |
1088 ); | |
3394 | 1089 } |
3579 | 1090 #endif //arch_x86 |
3394 | 1091 |
1092 void | |
1093 imdct_do_256(sample_t data[],sample_t delay[],sample_t bias) | |
1094 { | |
1095 int i,k; | |
1096 int p,q; | |
1097 int m; | |
1098 int two_m; | |
1099 int two_m_plus_one; | |
1100 | |
1101 sample_t tmp_a_i; | |
1102 sample_t tmp_a_r; | |
1103 sample_t tmp_b_i; | |
1104 sample_t tmp_b_r; | |
1105 | |
1106 sample_t *data_ptr; | |
1107 sample_t *delay_ptr; | |
1108 sample_t *window_ptr; | |
1109 | |
1110 complex_t *buf_1, *buf_2; | |
1111 | |
1112 buf_1 = &buf[0]; | |
1113 buf_2 = &buf[64]; | |
1114 | |
1115 /* Pre IFFT complex multiply plus IFFT cmplx conjugate */ | |
1116 for(k=0; k<64; k++) { | |
1117 /* X1[k] = X[2*k] */ | |
1118 /* X2[k] = X[2*k+1] */ | |
1119 | |
1120 p = 2 * (128-2*k-1); | |
1121 q = 2 * (2 * k); | |
1122 | |
1123 /* Z1[k] = (X1[128-2*k-1] + j * X1[2*k]) * (xcos2[k] + j * xsin2[k]); */ | |
1124 buf_1[k].real = data[p] * xcos2[k] - data[q] * xsin2[k]; | |
1125 buf_1[k].imag = -1.0f * (data[q] * xcos2[k] + data[p] * xsin2[k]); | |
1126 /* Z2[k] = (X2[128-2*k-1] + j * X2[2*k]) * (xcos2[k] + j * xsin2[k]); */ | |
1127 buf_2[k].real = data[p + 1] * xcos2[k] - data[q + 1] * xsin2[k]; | |
1128 buf_2[k].imag = -1.0f * ( data[q + 1] * xcos2[k] + data[p + 1] * xsin2[k]); | |
1129 } | |
1130 | |
1131 /* IFFT Bit reversed shuffling */ | |
1132 for(i=0; i<64; i++) { | |
1133 k = bit_reverse_256[i]; | |
1134 if (k < i) { | |
1135 swap_cmplx(&buf_1[i],&buf_1[k]); | |
1136 swap_cmplx(&buf_2[i],&buf_2[k]); | |
1137 } | |
1138 } | |
1139 | |
1140 /* FFT Merge */ | |
1141 for (m=0; m < 6; m++) { | |
1142 two_m = (1 << m); | |
1143 two_m_plus_one = (1 << (m+1)); | |
1144 | |
1145 /* FIXME */ | |
1146 if(m) | |
1147 two_m = (1 << m); | |
1148 else | |
1149 two_m = 1; | |
1150 | |
1151 for(k = 0; k < two_m; k++) { | |
1152 for(i = 0; i < 64; i += two_m_plus_one) { | |
1153 p = k + i; | |
1154 q = p + two_m; | |
1155 /* Do block 1 */ | |
1156 tmp_a_r = buf_1[p].real; | |
1157 tmp_a_i = buf_1[p].imag; | |
1158 tmp_b_r = buf_1[q].real * w[m][k].real - buf_1[q].imag * w[m][k].imag; | |
1159 tmp_b_i = buf_1[q].imag * w[m][k].real + buf_1[q].real * w[m][k].imag; | |
1160 buf_1[p].real = tmp_a_r + tmp_b_r; | |
1161 buf_1[p].imag = tmp_a_i + tmp_b_i; | |
1162 buf_1[q].real = tmp_a_r - tmp_b_r; | |
1163 buf_1[q].imag = tmp_a_i - tmp_b_i; | |
1164 | |
1165 /* Do block 2 */ | |
1166 tmp_a_r = buf_2[p].real; | |
1167 tmp_a_i = buf_2[p].imag; | |
1168 tmp_b_r = buf_2[q].real * w[m][k].real - buf_2[q].imag * w[m][k].imag; | |
1169 tmp_b_i = buf_2[q].imag * w[m][k].real + buf_2[q].real * w[m][k].imag; | |
1170 buf_2[p].real = tmp_a_r + tmp_b_r; | |
1171 buf_2[p].imag = tmp_a_i + tmp_b_i; | |
1172 buf_2[q].real = tmp_a_r - tmp_b_r; | |
1173 buf_2[q].imag = tmp_a_i - tmp_b_i; | |
1174 } | |
1175 } | |
1176 } | |
1177 | |
1178 /* Post IFFT complex multiply */ | |
1179 for( i=0; i < 64; i++) { | |
1180 /* y1[n] = z1[n] * (xcos2[n] + j * xs in2[n]) ; */ | |
1181 tmp_a_r = buf_1[i].real; | |
1182 tmp_a_i = -buf_1[i].imag; | |
1183 buf_1[i].real =(tmp_a_r * xcos2[i]) - (tmp_a_i * xsin2[i]); | |
1184 buf_1[i].imag =(tmp_a_r * xsin2[i]) + (tmp_a_i * xcos2[i]); | |
1185 /* y2[n] = z2[n] * (xcos2[n] + j * xsin2[n]) ; */ | |
1186 tmp_a_r = buf_2[i].real; | |
1187 tmp_a_i = -buf_2[i].imag; | |
1188 buf_2[i].real =(tmp_a_r * xcos2[i]) - (tmp_a_i * xsin2[i]); | |
1189 buf_2[i].imag =(tmp_a_r * xsin2[i]) + (tmp_a_i * xcos2[i]); | |
1190 } | |
1191 | |
1192 data_ptr = data; | |
1193 delay_ptr = delay; | |
1194 window_ptr = imdct_window; | |
1195 | |
1196 /* Window and convert to real valued signal */ | |
1197 for(i=0; i< 64; i++) { | |
1198 *data_ptr++ = -buf_1[i].imag * *window_ptr++ + *delay_ptr++ + bias; | |
1199 *data_ptr++ = buf_1[64-i-1].real * *window_ptr++ + *delay_ptr++ + bias; | |
1200 } | |
1201 | |
1202 for(i=0; i< 64; i++) { | |
1203 *data_ptr++ = -buf_1[i].real * *window_ptr++ + *delay_ptr++ + bias; | |
1204 *data_ptr++ = buf_1[64-i-1].imag * *window_ptr++ + *delay_ptr++ + bias; | |
1205 } | |
1206 | |
1207 delay_ptr = delay; | |
1208 | |
1209 for(i=0; i< 64; i++) { | |
1210 *delay_ptr++ = -buf_2[i].real * *--window_ptr; | |
1211 *delay_ptr++ = buf_2[64-i-1].imag * *--window_ptr; | |
1212 } | |
1213 | |
1214 for(i=0; i< 64; i++) { | |
1215 *delay_ptr++ = buf_2[i].imag * *--window_ptr; | |
1216 *delay_ptr++ = -buf_2[64-i-1].real * *--window_ptr; | |
1217 } | |
1218 } | |
1219 | |
1220 void imdct_init (uint32_t mm_accel) | |
1221 { | |
1222 #ifdef LIBA52_MLIB | |
1223 if (mm_accel & MM_ACCEL_MLIB) { | |
1224 fprintf (stderr, "Using mlib for IMDCT transform\n"); | |
1225 imdct_512 = imdct_do_512_mlib; | |
1226 imdct_256 = imdct_do_256_mlib; | |
1227 } else | |
1228 #endif | |
1229 { | |
1230 int i, j, k; | |
1231 | |
1232 /* Twiddle factors to turn IFFT into IMDCT */ | |
1233 for (i = 0; i < 128; i++) { | |
1234 xcos1[i] = -cos ((M_PI / 2048) * (8 * i + 1)); | |
1235 xsin1[i] = -sin ((M_PI / 2048) * (8 * i + 1)); | |
1236 } | |
3579 | 1237 #ifdef ARCH_X86 |
3527 | 1238 for (i = 0; i < 128; i++) { |
3581 | 1239 sseSinCos1c[2*i+0]= xcos1[i]; |
1240 sseSinCos1c[2*i+1]= -xcos1[i]; | |
1241 sseSinCos1d[2*i+0]= xsin1[i]; | |
1242 sseSinCos1d[2*i+1]= xsin1[i]; | |
3527 | 1243 } |
1244 #endif | |
3394 | 1245 |
1246 /* More twiddle factors to turn IFFT into IMDCT */ | |
1247 for (i = 0; i < 64; i++) { | |
1248 xcos2[i] = -cos ((M_PI / 1024) * (8 * i + 1)); | |
1249 xsin2[i] = -sin ((M_PI / 1024) * (8 * i + 1)); | |
1250 } | |
1251 | |
1252 for (i = 0; i < 7; i++) { | |
1253 j = 1 << i; | |
1254 for (k = 0; k < j; k++) { | |
1255 w[i][k].real = cos (-M_PI * k / j); | |
1256 w[i][k].imag = sin (-M_PI * k / j); | |
1257 } | |
1258 } | |
3579 | 1259 #ifdef ARCH_X86 |
3534 | 1260 for (i = 1; i < 7; i++) { |
1261 j = 1 << i; | |
1262 for (k = 0; k < j; k+=2) { | |
1263 | |
1264 sseW[i][4*k + 0] = w[i][k+0].real; | |
1265 sseW[i][4*k + 1] = w[i][k+0].real; | |
1266 sseW[i][4*k + 2] = w[i][k+1].real; | |
1267 sseW[i][4*k + 3] = w[i][k+1].real; | |
1268 | |
1269 sseW[i][4*k + 4] = -w[i][k+0].imag; | |
1270 sseW[i][4*k + 5] = w[i][k+0].imag; | |
1271 sseW[i][4*k + 6] = -w[i][k+1].imag; | |
1272 sseW[i][4*k + 7] = w[i][k+1].imag; | |
1273 | |
1274 //we multiply more or less uninitalized numbers so we need to use exactly 0.0 | |
1275 if(k==0) | |
1276 { | |
1277 // sseW[i][4*k + 0]= sseW[i][4*k + 1]= 1.0; | |
1278 sseW[i][4*k + 4]= sseW[i][4*k + 5]= 0.0; | |
1279 } | |
1280 | |
1281 if(2*k == j) | |
1282 { | |
1283 sseW[i][4*k + 0]= sseW[i][4*k + 1]= 0.0; | |
1284 // sseW[i][4*k + 4]= -(sseW[i][4*k + 5]= -1.0); | |
1285 } | |
1286 } | |
1287 } | |
3552 | 1288 |
1289 for(i=0; i<128; i++) | |
1290 { | |
1291 sseWindow[2*i+0]= -imdct_window[2*i+0]; | |
3553 | 1292 sseWindow[2*i+1]= imdct_window[2*i+1]; |
3552 | 1293 } |
3553 | 1294 |
1295 for(i=0; i<64; i++) | |
1296 { | |
1297 sseWindow[256 + 2*i+0]= -imdct_window[254 - 2*i+1]; | |
1298 sseWindow[256 + 2*i+1]= imdct_window[254 - 2*i+0]; | |
1299 sseWindow[384 + 2*i+0]= imdct_window[126 - 2*i+1]; | |
1300 sseWindow[384 + 2*i+1]= -imdct_window[126 - 2*i+0]; | |
1301 } | |
3579 | 1302 #endif // arch_x86 |
1303 | |
3720
120ac80f13c2
Fixed #ifdef discrepancy that was breaking compilation on PPC platform
melanson
parents:
3623
diff
changeset
|
1304 imdct_512 = imdct_do_512; |
120ac80f13c2
Fixed #ifdef discrepancy that was breaking compilation on PPC platform
melanson
parents:
3623
diff
changeset
|
1305 #ifdef ARCH_X86 |
4497 | 1306 if(mm_accel & MM_ACCEL_X86_SSE) |
1307 { | |
1308 fprintf (stderr, "Using SSE optimized IMDCT transform\n"); | |
1309 imdct_512 = imdct_do_512_sse; | |
1310 } | |
1311 else | |
1312 if(mm_accel & MM_ACCEL_X86_3DNOWEXT) | |
1313 { | |
1314 fprintf (stderr, "Using 3DNowEx optimized IMDCT transform\n"); | |
1315 imdct_512 = imdct_do_512_3dnowex; | |
1316 } | |
1317 else | |
1318 if(mm_accel & MM_ACCEL_X86_3DNOW) | |
1319 { | |
1320 fprintf (stderr, "Using 3DNow optimized IMDCT transform\n"); | |
1321 imdct_512 = imdct_do_512_3dnow; | |
1322 } | |
1323 else | |
3720
120ac80f13c2
Fixed #ifdef discrepancy that was breaking compilation on PPC platform
melanson
parents:
3623
diff
changeset
|
1324 #endif // arch_x86 |
9001
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1325 #ifdef HAVE_ALTIVEC |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1326 if (mm_accel & MM_ACCEL_PPC_ALTIVEC) |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1327 { |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1328 fprintf(stderr, "Using AltiVec optimized IMDCT transform\n"); |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1329 imdct_512 = imdct_do_512_altivec; |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1330 } |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1331 else |
01a9cf43074c
An AltiVec-enhanced IMDCT for liba52 (liba52/imdct.c)
arpi
parents:
8451
diff
changeset
|
1332 #endif |
4497 | 1333 fprintf (stderr, "No accelerated IMDCT transform found\n"); |
3394 | 1334 imdct_256 = imdct_do_256; |
1335 } | |
1336 } | |
3884 | 1337 |
1338 static void fft_asmb(int k, complex_t *x, complex_t *wTB, | |
1339 const complex_t *d, const complex_t *d_3) | |
1340 { | |
1341 register complex_t *x2k, *x3k, *x4k, *wB; | |
1342 register float a_r, a_i, a1_r, a1_i, u_r, u_i, v_r, v_i; | |
1343 | |
1344 x2k = x + 2 * k; | |
1345 x3k = x2k + 2 * k; | |
1346 x4k = x3k + 2 * k; | |
1347 wB = wTB + 2 * k; | |
1348 | |
1349 TRANSZERO(x[0],x2k[0],x3k[0],x4k[0]); | |
1350 TRANS(x[1],x2k[1],x3k[1],x4k[1],wTB[1],wB[1],d[1],d_3[1]); | |
1351 | |
1352 --k; | |
1353 for(;;) { | |
1354 TRANS(x[2],x2k[2],x3k[2],x4k[2],wTB[2],wB[2],d[2],d_3[2]); | |
1355 TRANS(x[3],x2k[3],x3k[3],x4k[3],wTB[3],wB[3],d[3],d_3[3]); | |
1356 if (!--k) break; | |
1357 x += 2; | |
1358 x2k += 2; | |
1359 x3k += 2; | |
1360 x4k += 2; | |
1361 d += 2; | |
1362 d_3 += 2; | |
1363 wTB += 2; | |
1364 wB += 2; | |
1365 } | |
1366 | |
1367 } | |
1368 | |
1369 static void fft_asmb16(complex_t *x, complex_t *wTB) | |
1370 { | |
1371 register float a_r, a_i, a1_r, a1_i, u_r, u_i, v_r, v_i; | |
1372 int k = 2; | |
1373 | |
1374 /* transform x[0], x[8], x[4], x[12] */ | |
1375 TRANSZERO(x[0],x[4],x[8],x[12]); | |
1376 | |
1377 /* transform x[1], x[9], x[5], x[13] */ | |
1378 TRANS(x[1],x[5],x[9],x[13],wTB[1],wTB[5],delta16[1],delta16_3[1]); | |
1379 | |
1380 /* transform x[2], x[10], x[6], x[14] */ | |
1381 TRANSHALF_16(x[2],x[6],x[10],x[14]); | |
1382 | |
1383 /* transform x[3], x[11], x[7], x[15] */ | |
1384 TRANS(x[3],x[7],x[11],x[15],wTB[3],wTB[7],delta16[3],delta16_3[3]); | |
1385 | |
1386 } | |
1387 | |
1388 static void fft_4(complex_t *x) | |
1389 { | |
1390 /* delta_p = 1 here */ | |
1391 /* x[k] = sum_{i=0..3} x[i] * w^{i*k}, w=e^{-2*pi/4} | |
1392 */ | |
1393 | |
1394 register float yt_r, yt_i, yb_r, yb_i, u_r, u_i, vi_r, vi_i; | |
1395 | |
1396 yt_r = x[0].real; | |
1397 yb_r = yt_r - x[2].real; | |
1398 yt_r += x[2].real; | |
1399 | |
1400 u_r = x[1].real; | |
1401 vi_i = x[3].real - u_r; | |
1402 u_r += x[3].real; | |
1403 | |
1404 u_i = x[1].imag; | |
1405 vi_r = u_i - x[3].imag; | |
1406 u_i += x[3].imag; | |
1407 | |
1408 yt_i = yt_r; | |
1409 yt_i += u_r; | |
1410 x[0].real = yt_i; | |
1411 yt_r -= u_r; | |
1412 x[2].real = yt_r; | |
1413 yt_i = yb_r; | |
1414 yt_i += vi_r; | |
1415 x[1].real = yt_i; | |
1416 yb_r -= vi_r; | |
1417 x[3].real = yb_r; | |
1418 | |
1419 yt_i = x[0].imag; | |
1420 yb_i = yt_i - x[2].imag; | |
1421 yt_i += x[2].imag; | |
1422 | |
1423 yt_r = yt_i; | |
1424 yt_r += u_i; | |
1425 x[0].imag = yt_r; | |
1426 yt_i -= u_i; | |
1427 x[2].imag = yt_i; | |
1428 yt_r = yb_i; | |
1429 yt_r += vi_i; | |
1430 x[1].imag = yt_r; | |
1431 yb_i -= vi_i; | |
1432 x[3].imag = yb_i; | |
1433 } | |
1434 | |
1435 | |
1436 static void fft_8(complex_t *x) | |
1437 { | |
1438 /* delta_p = diag{1, sqrt(i)} here */ | |
1439 /* x[k] = sum_{i=0..7} x[i] * w^{i*k}, w=e^{-2*pi/8} | |
1440 */ | |
1441 register float wT1_r, wT1_i, wB1_r, wB1_i, wT2_r, wT2_i, wB2_r, wB2_i; | |
1442 | |
1443 wT1_r = x[1].real; | |
1444 wT1_i = x[1].imag; | |
1445 wB1_r = x[3].real; | |
1446 wB1_i = x[3].imag; | |
1447 | |
1448 x[1] = x[2]; | |
1449 x[2] = x[4]; | |
1450 x[3] = x[6]; | |
1451 fft_4(&x[0]); | |
1452 | |
1453 | |
1454 /* x[0] x[4] */ | |
1455 wT2_r = x[5].real; | |
1456 wT2_r += x[7].real; | |
1457 wT2_r += wT1_r; | |
1458 wT2_r += wB1_r; | |
1459 wT2_i = wT2_r; | |
1460 wT2_r += x[0].real; | |
1461 wT2_i = x[0].real - wT2_i; | |
1462 x[0].real = wT2_r; | |
1463 x[4].real = wT2_i; | |
1464 | |
1465 wT2_i = x[5].imag; | |
1466 wT2_i += x[7].imag; | |
1467 wT2_i += wT1_i; | |
1468 wT2_i += wB1_i; | |
1469 wT2_r = wT2_i; | |
1470 wT2_r += x[0].imag; | |
1471 wT2_i = x[0].imag - wT2_i; | |
1472 x[0].imag = wT2_r; | |
1473 x[4].imag = wT2_i; | |
1474 | |
1475 /* x[2] x[6] */ | |
1476 wT2_r = x[5].imag; | |
1477 wT2_r -= x[7].imag; | |
1478 wT2_r += wT1_i; | |
1479 wT2_r -= wB1_i; | |
1480 wT2_i = wT2_r; | |
1481 wT2_r += x[2].real; | |
1482 wT2_i = x[2].real - wT2_i; | |
1483 x[2].real = wT2_r; | |
1484 x[6].real = wT2_i; | |
1485 | |
1486 wT2_i = x[5].real; | |
1487 wT2_i -= x[7].real; | |
1488 wT2_i += wT1_r; | |
1489 wT2_i -= wB1_r; | |
1490 wT2_r = wT2_i; | |
1491 wT2_r += x[2].imag; | |
1492 wT2_i = x[2].imag - wT2_i; | |
1493 x[2].imag = wT2_i; | |
1494 x[6].imag = wT2_r; | |
1495 | |
1496 | |
1497 /* x[1] x[5] */ | |
1498 wT2_r = wT1_r; | |
1499 wT2_r += wB1_i; | |
1500 wT2_r -= x[5].real; | |
1501 wT2_r -= x[7].imag; | |
1502 wT2_i = wT1_i; | |
1503 wT2_i -= wB1_r; | |
1504 wT2_i -= x[5].imag; | |
1505 wT2_i += x[7].real; | |
1506 | |
1507 wB2_r = wT2_r; | |
1508 wB2_r += wT2_i; | |
1509 wT2_i -= wT2_r; | |
1510 wB2_r *= HSQRT2; | |
1511 wT2_i *= HSQRT2; | |
1512 wT2_r = wB2_r; | |
1513 wB2_r += x[1].real; | |
1514 wT2_r = x[1].real - wT2_r; | |
1515 | |
1516 wB2_i = x[5].real; | |
1517 x[1].real = wB2_r; | |
1518 x[5].real = wT2_r; | |
1519 | |
1520 wT2_r = wT2_i; | |
1521 wT2_r += x[1].imag; | |
1522 wT2_i = x[1].imag - wT2_i; | |
1523 wB2_r = x[5].imag; | |
1524 x[1].imag = wT2_r; | |
1525 x[5].imag = wT2_i; | |
1526 | |
1527 /* x[3] x[7] */ | |
1528 wT1_r -= wB1_i; | |
1529 wT1_i += wB1_r; | |
1530 wB1_r = wB2_i - x[7].imag; | |
1531 wB1_i = wB2_r + x[7].real; | |
1532 wT1_r -= wB1_r; | |
1533 wT1_i -= wB1_i; | |
1534 wB1_r = wT1_r + wT1_i; | |
1535 wB1_r *= HSQRT2; | |
1536 wT1_i -= wT1_r; | |
1537 wT1_i *= HSQRT2; | |
1538 wB2_r = x[3].real; | |
1539 wB2_i = wB2_r + wT1_i; | |
1540 wB2_r -= wT1_i; | |
1541 x[3].real = wB2_i; | |
1542 x[7].real = wB2_r; | |
1543 wB2_i = x[3].imag; | |
1544 wB2_r = wB2_i + wB1_r; | |
1545 wB2_i -= wB1_r; | |
1546 x[3].imag = wB2_i; | |
1547 x[7].imag = wB2_r; | |
1548 } | |
1549 | |
1550 | |
1551 static void fft_128p(complex_t *a) | |
1552 { | |
1553 fft_8(&a[0]); fft_4(&a[8]); fft_4(&a[12]); | |
1554 fft_asmb16(&a[0], &a[8]); | |
1555 | |
1556 fft_8(&a[16]), fft_8(&a[24]); | |
1557 fft_asmb(4, &a[0], &a[16],&delta32[0], &delta32_3[0]); | |
1558 | |
1559 fft_8(&a[32]); fft_4(&a[40]); fft_4(&a[44]); | |
1560 fft_asmb16(&a[32], &a[40]); | |
1561 | |
1562 fft_8(&a[48]); fft_4(&a[56]); fft_4(&a[60]); | |
1563 fft_asmb16(&a[48], &a[56]); | |
1564 | |
1565 fft_asmb(8, &a[0], &a[32],&delta64[0], &delta64_3[0]); | |
1566 | |
1567 fft_8(&a[64]); fft_4(&a[72]); fft_4(&a[76]); | |
1568 /* fft_16(&a[64]); */ | |
1569 fft_asmb16(&a[64], &a[72]); | |
1570 | |
1571 fft_8(&a[80]); fft_8(&a[88]); | |
1572 | |
1573 /* fft_32(&a[64]); */ | |
1574 fft_asmb(4, &a[64], &a[80],&delta32[0], &delta32_3[0]); | |
1575 | |
1576 fft_8(&a[96]); fft_4(&a[104]), fft_4(&a[108]); | |
1577 /* fft_16(&a[96]); */ | |
1578 fft_asmb16(&a[96], &a[104]); | |
1579 | |
1580 fft_8(&a[112]), fft_8(&a[120]); | |
1581 /* fft_32(&a[96]); */ | |
1582 fft_asmb(4, &a[96], &a[112], &delta32[0], &delta32_3[0]); | |
1583 | |
1584 /* fft_128(&a[0]); */ | |
1585 fft_asmb(16, &a[0], &a[64], &delta128[0], &delta128_3[0]); | |
1586 } | |
1587 | |
1588 | |
1589 |