1 /*****************************************************************************
2 * ac3_srfft.c: ac3 FFT in C
3 *****************************************************************************
4 * Copyright (C) 1999-2001 VideoLAN
5 * $Id: ac3_srfft_c.c,v 1.5 2002/06/01 12:31:59 sam Exp $
7 * Authors: Renaud Dartus <reno@videolan.org>
8 * Aaron Holtzman <aholtzma@engr.uvic.ca>
10 * This program is free software; you can redistribute it and/or modify
11 * it under the terms of the GNU General Public License as published by
12 * the Free Software Foundation; either version 2 of the License, or
13 * (at your option) any later version.
15 * This program is distributed in the hope that it will be useful,
16 * but WITHOUT ANY WARRANTY; without even the implied warranty of
17 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
18 * GNU General Public License for more details.
20 * You should have received a copy of the GNU General Public License
21 * along with this program; if not, write to the Free Software
22 * Foundation, Inc., 59 Temple Place - Suite 330, Boston, MA 02111, USA.
23 *****************************************************************************/
25 /*****************************************************************************
27 *****************************************************************************/
33 #include "ac3_imdct.h"
34 #include "ac3_srfft.h"
36 static void fft_8 (complex_t *x);
38 static void fft_4(complex_t *x)
40 /* delta_p = 1 here */
41 /* x[k] = sum_{i=0..3} x[i] * w^{i*k}, w=e^{-2*pi/4}
44 register float yt_r, yt_i, yb_r, yb_i, u_r, u_i, vi_r, vi_i;
47 yb_r = yt_r - x[2].real;
51 vi_i = x[3].real - u_r;
55 vi_r = u_i - x[3].imag;
70 yb_i = yt_i - x[2].imag;
86 static void fft_8 (complex_t *x)
88 /* delta_p = diag{1, sqrt(i)} here */
89 /* x[k] = sum_{i=0..7} x[i] * w^{i*k}, w=e^{-2*pi/8}
91 register float wT1_r, wT1_i, wB1_r, wB1_i, wT2_r, wT2_i, wB2_r, wB2_i;
111 wT2_i = x[0].real - wT2_i;
121 wT2_i = x[0].imag - wT2_i;
132 wT2_i = x[2].real - wT2_i;
142 wT2_i = x[2].imag - wT2_i;
164 wT2_r = x[1].real - wT2_r;
172 wT2_i = x[1].imag - wT2_i;
180 wB1_r = wB2_i - x[7].imag;
181 wB1_i = wB2_r + x[7].real;
184 wB1_r = wT1_r + wT1_i;
189 wB2_i = wB2_r + wT1_i;
194 wB2_r = wB2_i + wB1_r;
201 static void fft_asmb(int k, complex_t *x, complex_t *wTB,
202 const complex_t *d, const complex_t *d_3)
204 register complex_t *x2k, *x3k, *x4k, *wB;
205 register float a_r, a_i, a1_r, a1_i, u_r, u_i, v_r, v_i;
212 TRANSZERO(x[0],x2k[0],x3k[0],x4k[0]);
213 TRANS(x[1],x2k[1],x3k[1],x4k[1],wTB[1],wB[1],d[1],d_3[1]);
217 TRANS(x[2],x2k[2],x3k[2],x4k[2],wTB[2],wB[2],d[2],d_3[2]);
218 TRANS(x[3],x2k[3],x3k[3],x4k[3],wTB[3],wB[3],d[3],d_3[3]);
232 static void fft_asmb16(complex_t *x, complex_t *wTB)
234 register float a_r, a_i, a1_r, a1_i, u_r, u_i, v_r, v_i;
237 /* transform x[0], x[8], x[4], x[12] */
238 TRANSZERO(x[0],x[4],x[8],x[12]);
240 /* transform x[1], x[9], x[5], x[13] */
241 TRANS(x[1],x[5],x[9],x[13],wTB[1],wTB[5],delta16[1],delta16_3[1]);
243 /* transform x[2], x[10], x[6], x[14] */
244 TRANSHALF_16(x[2],x[6],x[10],x[14]);
246 /* transform x[3], x[11], x[7], x[15] */
247 TRANS(x[3],x[7],x[11],x[15],wTB[3],wTB[7],delta16[3],delta16_3[3]);
252 void _M( fft_64p ) ( complex_t *a )
254 fft_8(&a[0]); fft_4(&a[8]); fft_4(&a[12]);
255 fft_asmb16(&a[0], &a[8]);
257 fft_8(&a[16]), fft_8(&a[24]);
258 fft_asmb(4, &a[0], &a[16],&delta32[0], &delta32_3[0]);
260 fft_8(&a[32]); fft_4(&a[40]); fft_4(&a[44]);
261 fft_asmb16(&a[32], &a[40]);
263 fft_8(&a[48]); fft_4(&a[56]); fft_4(&a[60]);
264 fft_asmb16(&a[48], &a[56]);
266 fft_asmb(8, &a[0], &a[32],&delta64[0], &delta64_3[0]);
270 void _M( fft_128p ) ( complex_t *a )
272 fft_8(&a[0]); fft_4(&a[8]); fft_4(&a[12]);
273 fft_asmb16(&a[0], &a[8]);
275 fft_8(&a[16]), fft_8(&a[24]);
276 fft_asmb(4, &a[0], &a[16],&delta32[0], &delta32_3[0]);
278 fft_8(&a[32]); fft_4(&a[40]); fft_4(&a[44]);
279 fft_asmb16(&a[32], &a[40]);
281 fft_8(&a[48]); fft_4(&a[56]); fft_4(&a[60]);
282 fft_asmb16(&a[48], &a[56]);
284 fft_asmb(8, &a[0], &a[32],&delta64[0], &delta64_3[0]);
286 fft_8(&a[64]); fft_4(&a[72]); fft_4(&a[76]);
287 /* fft_16(&a[64]); */
288 fft_asmb16(&a[64], &a[72]);
290 fft_8(&a[80]); fft_8(&a[88]);
292 /* fft_32(&a[64]); */
293 fft_asmb(4, &a[64], &a[80],&delta32[0], &delta32_3[0]);
295 fft_8(&a[96]); fft_4(&a[104]), fft_4(&a[108]);
296 /* fft_16(&a[96]); */
297 fft_asmb16(&a[96], &a[104]);
299 fft_8(&a[112]), fft_8(&a[120]);
300 /* fft_32(&a[96]); */
301 fft_asmb(4, &a[96], &a[112], &delta32[0], &delta32_3[0]);
303 /* fft_128(&a[0]); */
304 fft_asmb(16, &a[0], &a[64], &delta128[0], &delta128_3[0]);