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 1.8 2001/12/10 04:53:10 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 #define MODULE_NAME downmix3dn
25 #include "modules_inner.h"
27 /*****************************************************************************
29 *****************************************************************************/
34 #include "ac3_downmix.h"
36 const float sqrt2_3dn __asm__ ("sqrt2_3dn") = 0.7071068;
38 void _M( downmix_3f_2r_to_2ch ) (float * samples, dm_par_t * dm_par)
40 __asm__ __volatile__ (
43 "movl $128, %%ebx\n" /* loop counter */
45 "movd (%%ecx), %%mm5\n" /* unit */
46 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
48 "movd 4(%%ecx), %%mm6\n" /* clev */
49 "punpckldq %%mm6, %%mm6\n" /* clev | clev */
51 "movd 8(%%ecx), %%mm7\n" /* slev */
52 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
56 "movq (%%eax), %%mm0\n" /* left */
57 "movq 2048(%%eax), %%mm1\n" /* right */
58 "movq 1024(%%eax), %%mm2\n" /* center */
59 "movq 3072(%%eax), %%mm3\n" /* leftsur */
60 "movq 4096(%%eax), %%mm4\n" /* rightsur */
61 "pfmul %%mm5, %%mm0\n"
62 "pfmul %%mm5, %%mm1\n"
63 "pfmul %%mm6, %%mm2\n"
64 "pfadd %%mm2, %%mm0\n"
65 "pfadd %%mm2, %%mm1\n"
66 "pfmul %%mm7, %%mm3\n"
67 "pfmul %%mm7, %%mm4\n"
68 "pfadd %%mm3, %%mm0\n"
69 "pfadd %%mm4, %%mm1\n"
71 "movq %%mm0, (%%eax)\n"
72 "movq %%mm1, 1024(%%eax)\n"
81 : "a" (samples), "c" (dm_par));
84 void _M( downmix_2f_2r_to_2ch ) (float *samples, dm_par_t * dm_par)
86 __asm__ __volatile__ (
89 "movl $128, %%ebx\n" /* loop counter */
91 "movd (%%ecx), %%mm5\n" /* unit */
92 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
94 "movd 8(%%ecx), %%mm7\n" /* slev */
95 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
99 "movq (%%eax), %%mm0\n" /* left */
100 "movq 1024(%%eax), %%mm1\n" /* right */
101 "movq 2048(%%eax), %%mm3\n" /* leftsur */
102 "movq 3072(%%eax), %%mm4\n" /* rightsur */
103 "pfmul %%mm5, %%mm0\n"
104 "pfmul %%mm5, %%mm1\n"
105 "pfmul %%mm7, %%mm3\n"
106 "pfmul %%mm7, %%mm4\n"
107 "pfadd %%mm3, %%mm0\n"
108 "pfadd %%mm4, %%mm1\n"
110 "movq %%mm0, (%%eax)\n"
111 "movq %%mm1, 1024(%%eax)\n"
120 : "a" (samples), "c" (dm_par));
123 void _M( downmix_3f_1r_to_2ch ) (float *samples, dm_par_t * dm_par)
125 __asm__ __volatile__ (
128 "movl $128, %%ebx\n" /* loop counter */
130 "movd (%%ecx), %%mm5\n" /* unit */
131 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
133 "movd 4(%%ecx), %%mm6\n" /* clev */
134 "punpckldq %%mm6, %%mm6\n" /* clev | clev */
136 "movd 8(%%ecx), %%mm7\n" /* slev */
137 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
141 "movq (%%eax), %%mm0\n" /* left */
142 "movq 2048(%%eax), %%mm1\n" /* right */
143 "movq 1024(%%eax), %%mm2\n" /* center */
144 "movq 3072(%%eax), %%mm3\n" /* sur */
145 "pfmul %%mm5, %%mm0\n"
146 "pfmul %%mm5, %%mm1\n"
147 "pfmul %%mm6, %%mm2\n"
148 "pfadd %%mm2, %%mm0\n"
149 "pfmul %%mm7, %%mm3\n"
150 "pfadd %%mm2, %%mm1\n"
151 "pfsub %%mm3, %%mm0\n"
152 "pfadd %%mm3, %%mm1\n"
154 "movq %%mm0, (%%eax)\n"
155 "movq %%mm1, 1024(%%eax)\n"
164 : "a" (samples), "c" (dm_par));
167 void _M( downmix_2f_1r_to_2ch ) (float *samples, dm_par_t * dm_par)
169 __asm__ __volatile__ (
172 "movl $128, %%ebx\n" /* loop counter */
174 "movd (%%ecx), %%mm5\n" /* unit */
175 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
177 "movd 8(%%ecx), %%mm7\n" /* slev */
178 "punpckldq %%mm7, %%mm7\n" /* slev | slev */
182 "movq (%%eax), %%mm0\n" /* left */
183 "movq 1024(%%eax), %%mm1\n" /* right */
184 "movq 2048(%%eax), %%mm3\n" /* sur */
185 "pfmul %%mm5, %%mm0\n"
186 "pfmul %%mm5, %%mm1\n"
187 "pfmul %%mm7, %%mm3\n"
188 "pfsub %%mm3, %%mm0\n"
189 "pfadd %%mm3, %%mm1\n"
191 "movq %%mm0, (%%eax)\n"
192 "movq %%mm1, 1024(%%eax)\n"
201 : "a" (samples), "c" (dm_par));
204 void _M( downmix_3f_0r_to_2ch ) (float *samples, dm_par_t * dm_par)
206 __asm__ __volatile__ (
209 "movl $128, %%ebx\n" /* loop counter */
211 "movd (%%ecx), %%mm5\n" /* unit */
212 "punpckldq %%mm5, %%mm5\n" /* unit | unit */
214 "movd 4(%%ecx), %%mm6\n" /* clev */
215 "punpckldq %%mm6, %%mm6\n" /* clev | clev */
219 "movq (%%eax), %%mm0\n" /*left */
220 "movq 2048(%%eax), %%mm1\n" /* right */
221 "movq 1024(%%eax), %%mm2\n" /* center */
222 "pfmul %%mm5, %%mm0\n"
223 "pfmul %%mm5, %%mm1\n"
224 "pfmul %%mm6, %%mm2\n"
225 "pfadd %%mm2, %%mm0\n"
226 "pfadd %%mm2, %%mm1\n"
228 "movq %%mm0, (%%eax)\n"
229 "movq %%mm1, 1024(%%eax)\n"
238 : "a" (samples), "c" (dm_par));
241 void _M( stream_sample_1ch_to_s16 ) (s16 *s16_samples, float *left)
243 __asm__ __volatile__ (
248 "movl $sqrt2_3dn, %%edx\n"
249 "movd (%%edx), %%mm7\n"
250 "punpckldq %%mm7, %%mm7\n" /* sqrt2 | sqrt2 */
255 "movq (%%ecx), %%mm0\n" /* c1 | c0 */
256 "pfmul %%mm7, %%mm0\n"
258 "pf2id %%mm0, %%mm0\n" /* c1 c0 --> mm0, int_32 */
260 "packssdw %%mm0, %%mm0\n" /* c1 c1 c0 c0 --> mm0, int_16 */
262 "movq %%mm0, (%%eax)\n"
272 : "=a" (s16_samples), "=c" (left)
273 : "a" (s16_samples), "c" (left));
276 void _M( stream_sample_2ch_to_s16 ) (s16 *s16_samples, float *left, float *right)
279 __asm__ __volatile__ (
286 "movq (%%ecx), %%mm0\n" /* l1 | l0 */
287 "movq (%%edx), %%mm1\n" /* r1 | r0 */
288 "movq %%mm0, %%mm2\n" /* l1 | l0 */
289 "punpckldq %%mm1, %%mm0\n" /* r0 | l0 */
290 "punpckhdq %%mm1, %%mm2\n" /* r1 | l1 */
292 "pf2id %%mm0, %%mm0\n" /* r0 l0 --> mm0, int_32 */
293 "pf2id %%mm2, %%mm2\n" /* r0 l0 --> mm0, int_32 */
295 "packssdw %%mm2, %%mm0\n" /* r1 l1 r0 l0 --> mm0, int_16 */
297 "movq %%mm0, (%%eax)\n"
298 "movq %%mm2, 8(%%eax)\n"
308 : "=a" (s16_samples), "=c" (left), "=d" (right)
309 : "a" (s16_samples), "c" (left), "d" (right));