1 /*****************************************************************************
2 * ac3_downmix_3dn.c: accelerated 3D Now! ac3 downmix functions
3 *****************************************************************************
4 * Copyright (C) 1999, 2000, 2001 VideoLAN
5 * $Id: ac3_downmix_3dn.c,v 1.12 2002/07/31 20:56:51 sam Exp $
7 * Authors: Renaud Dartus <reno@videolan.org>
9 * This program 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.
14 * This program 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.
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, USA.
22 *****************************************************************************/
24 /*****************************************************************************
26 *****************************************************************************/
29 #include "ac3_downmix.h"
31 static const float sqrt2_3dn __asm__ ("sqrt2_3dn") = 0.7071068;
33 void E_( downmix_3f_2r_to_2ch ) (float * samples, dm_par_t * dm_par)
35 __asm__ __volatile__ (
38 "movl $128, %%ebx\n" /* loop counter */
40 "movd (%%ecx), %%mm5\n" /* unit */
41 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
43 "movd 4(%%ecx), %%mm6\n" /* clev */
44 "punpckldq %%mm6, %%mm6\n" /* clev | clev */
46 "movd 8(%%ecx), %%mm7\n" /* slev */
47 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
51 "movq (%%eax), %%mm0\n" /* left */
52 "movq 2048(%%eax), %%mm1\n" /* right */
53 "movq 1024(%%eax), %%mm2\n" /* center */
54 "movq 3072(%%eax), %%mm3\n" /* leftsur */
55 "movq 4096(%%eax), %%mm4\n" /* rightsur */
56 "pfmul %%mm5, %%mm0\n"
57 "pfmul %%mm5, %%mm1\n"
58 "pfmul %%mm6, %%mm2\n"
59 "pfadd %%mm2, %%mm0\n"
60 "pfadd %%mm2, %%mm1\n"
61 "pfmul %%mm7, %%mm3\n"
62 "pfmul %%mm7, %%mm4\n"
63 "pfadd %%mm3, %%mm0\n"
64 "pfadd %%mm4, %%mm1\n"
66 "movq %%mm0, (%%eax)\n"
67 "movq %%mm1, 1024(%%eax)\n"
76 : "a" (samples), "c" (dm_par));
79 void E_( downmix_2f_2r_to_2ch ) (float *samples, dm_par_t * dm_par)
81 __asm__ __volatile__ (
84 "movl $128, %%ebx\n" /* loop counter */
86 "movd (%%ecx), %%mm5\n" /* unit */
87 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
89 "movd 8(%%ecx), %%mm7\n" /* slev */
90 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
94 "movq (%%eax), %%mm0\n" /* left */
95 "movq 1024(%%eax), %%mm1\n" /* right */
96 "movq 2048(%%eax), %%mm3\n" /* leftsur */
97 "movq 3072(%%eax), %%mm4\n" /* rightsur */
98 "pfmul %%mm5, %%mm0\n"
99 "pfmul %%mm5, %%mm1\n"
100 "pfmul %%mm7, %%mm3\n"
101 "pfmul %%mm7, %%mm4\n"
102 "pfadd %%mm3, %%mm0\n"
103 "pfadd %%mm4, %%mm1\n"
105 "movq %%mm0, (%%eax)\n"
106 "movq %%mm1, 1024(%%eax)\n"
115 : "a" (samples), "c" (dm_par));
118 void E_( downmix_3f_1r_to_2ch ) (float *samples, dm_par_t * dm_par)
120 __asm__ __volatile__ (
123 "movl $128, %%ebx\n" /* loop counter */
125 "movd (%%ecx), %%mm5\n" /* unit */
126 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
128 "movd 4(%%ecx), %%mm6\n" /* clev */
129 "punpckldq %%mm6, %%mm6\n" /* clev | clev */
131 "movd 8(%%ecx), %%mm7\n" /* slev */
132 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
136 "movq (%%eax), %%mm0\n" /* left */
137 "movq 2048(%%eax), %%mm1\n" /* right */
138 "movq 1024(%%eax), %%mm2\n" /* center */
139 "movq 3072(%%eax), %%mm3\n" /* sur */
140 "pfmul %%mm5, %%mm0\n"
141 "pfmul %%mm5, %%mm1\n"
142 "pfmul %%mm6, %%mm2\n"
143 "pfadd %%mm2, %%mm0\n"
144 "pfmul %%mm7, %%mm3\n"
145 "pfadd %%mm2, %%mm1\n"
146 "pfsub %%mm3, %%mm0\n"
147 "pfadd %%mm3, %%mm1\n"
149 "movq %%mm0, (%%eax)\n"
150 "movq %%mm1, 1024(%%eax)\n"
159 : "a" (samples), "c" (dm_par));
162 void E_( downmix_2f_1r_to_2ch ) (float *samples, dm_par_t * dm_par)
164 __asm__ __volatile__ (
167 "movl $128, %%ebx\n" /* loop counter */
169 "movd (%%ecx), %%mm5\n" /* unit */
170 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
172 "movd 8(%%ecx), %%mm7\n" /* slev */
173 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
177 "movq (%%eax), %%mm0\n" /* left */
178 "movq 1024(%%eax), %%mm1\n" /* right */
179 "movq 2048(%%eax), %%mm3\n" /* sur */
180 "pfmul %%mm5, %%mm0\n"
181 "pfmul %%mm5, %%mm1\n"
182 "pfmul %%mm7, %%mm3\n"
183 "pfsub %%mm3, %%mm0\n"
184 "pfadd %%mm3, %%mm1\n"
186 "movq %%mm0, (%%eax)\n"
187 "movq %%mm1, 1024(%%eax)\n"
196 : "a" (samples), "c" (dm_par));
199 void E_( downmix_3f_0r_to_2ch ) (float *samples, dm_par_t * dm_par)
201 __asm__ __volatile__ (
204 "movl $128, %%ebx\n" /* loop counter */
206 "movd (%%ecx), %%mm5\n" /* unit */
207 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
209 "movd 4(%%ecx), %%mm6\n" /* clev */
210 "punpckldq %%mm6, %%mm6\n" /* clev | clev */
214 "movq (%%eax), %%mm0\n" /*left */
215 "movq 2048(%%eax), %%mm1\n" /* right */
216 "movq 1024(%%eax), %%mm2\n" /* center */
217 "pfmul %%mm5, %%mm0\n"
218 "pfmul %%mm5, %%mm1\n"
219 "pfmul %%mm6, %%mm2\n"
220 "pfadd %%mm2, %%mm0\n"
221 "pfadd %%mm2, %%mm1\n"
223 "movq %%mm0, (%%eax)\n"
224 "movq %%mm1, 1024(%%eax)\n"
233 : "a" (samples), "c" (dm_par));
236 void E_( stream_sample_1ch_to_s16 ) (s16 *s16_samples, float *left)
238 __asm__ __volatile__ (
243 "movl $sqrt2_3dn, %%edx\n"
244 "movd (%%edx), %%mm7\n"
245 "punpckldq %%mm7, %%mm7\n" /* sqrt2 | sqrt2 */
250 "movq (%%ecx), %%mm0\n" /* c1 | c0 */
251 "pfmul %%mm7, %%mm0\n"
253 "pf2id %%mm0, %%mm0\n" /* c1 c0 --> mm0, int_32 */
255 "packssdw %%mm0, %%mm0\n" /* c1 c1 c0 c0 --> mm0, int_16 */
257 "movq %%mm0, (%%eax)\n"
267 : "=a" (s16_samples), "=c" (left)
268 : "a" (s16_samples), "c" (left));
271 void E_( stream_sample_2ch_to_s16 ) (s16 *s16_samples, float *left, float *right)
274 __asm__ __volatile__ (
281 "movq (%%ecx), %%mm0\n" /* l1 | l0 */
282 "movq (%%edx), %%mm1\n" /* r1 | r0 */
283 "movq %%mm0, %%mm2\n" /* l1 | l0 */
284 "punpckldq %%mm1, %%mm0\n" /* r0 | l0 */
285 "punpckhdq %%mm1, %%mm2\n" /* r1 | l1 */
287 "pf2id %%mm0, %%mm0\n" /* r0 l0 --> mm0, int_32 */
288 "pf2id %%mm2, %%mm2\n" /* r0 l0 --> mm0, int_32 */
290 "packssdw %%mm2, %%mm0\n" /* r1 l1 r0 l0 --> mm0, int_16 */
292 "movq %%mm0, (%%eax)\n"
293 "movq %%mm2, 8(%%eax)\n"
303 : "=a" (s16_samples), "=c" (left), "=d" (right)
304 : "a" (s16_samples), "c" (left), "d" (right));