1/*
2 * Copyright (c) 2012
3 *      MIPS Technologies, Inc., California.
4 *
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
7 * are met:
8 * 1. Redistributions of source code must retain the above copyright
9 *    notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 *    notice, this list of conditions and the following disclaimer in the
12 *    documentation and/or other materials provided with the distribution.
13 * 3. Neither the name of the MIPS Technologies, Inc., nor the names of its
14 *    contributors may be used to endorse or promote products derived from
15 *    this software without specific prior written permission.
16 *
17 * THIS SOFTWARE IS PROVIDED BY THE MIPS TECHNOLOGIES, INC. ``AS IS'' AND
18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20 * ARE DISCLAIMED.  IN NO EVENT SHALL THE MIPS TECHNOLOGIES, INC. BE LIABLE
21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27 * SUCH DAMAGE.
28 *
29 * Author:  Bojan Zivkovic   (bojan@mips.com)
30 *
31 * AAC encoder psychoacoustic model routines optimized
32 * for MIPS floating-point architecture
33 *
34 * This file is part of FFmpeg.
35 *
36 * FFmpeg is free software; you can redistribute it and/or
37 * modify it under the terms of the GNU Lesser General Public
38 * License as published by the Free Software Foundation; either
39 * version 2.1 of the License, or (at your option) any later version.
40 *
41 * FFmpeg is distributed in the hope that it will be useful,
42 * but WITHOUT ANY WARRANTY; without even the implied warranty of
43 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
44 * Lesser General Public License for more details.
45 *
46 * You should have received a copy of the GNU Lesser General Public
47 * License along with FFmpeg; if not, write to the Free Software
48 * Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
49 */
50
51/**
52 * @file
53 * Reference: libavcodec/aacpsy.c
54 */
55
56#ifndef AVCODEC_MIPS_AACPSY_MIPS_H
57#define AVCODEC_MIPS_AACPSY_MIPS_H
58
59#include "libavutil/mips/asmdefs.h"
60
61#if HAVE_INLINE_ASM && HAVE_MIPSFPU && ( PSY_LAME_FIR_LEN == 21 )
62#if !HAVE_MIPS32R6 && !HAVE_MIPS64R6
63static void calc_thr_3gpp_mips(const FFPsyWindowInfo *wi, const int num_bands,
64                               AacPsyChannel *pch, const uint8_t *band_sizes,
65                               const float *coefs, const int cutoff)
66{
67    int i, w, g;
68    int start = 0, wstart = 0;
69    for (w = 0; w < wi->num_windows*16; w += 16) {
70        wstart = 0;
71        for (g = 0; g < num_bands; g++) {
72            AacPsyBand *band = &pch->band[w+g];
73
74            float form_factor = 0.0f;
75            float Temp;
76            band->energy = 0.0f;
77            if (wstart < cutoff) {
78                for (i = 0; i < band_sizes[g]; i+=4) {
79                    float a, b, c, d;
80                    float ax, bx, cx, dx;
81                    float *cf = (float *)&coefs[start+i];
82
83                    __asm__ volatile (
84                        "lwc1   %[a],   0(%[cf])                \n\t"
85                        "lwc1   %[b],   4(%[cf])                \n\t"
86                        "lwc1   %[c],   8(%[cf])                \n\t"
87                        "lwc1   %[d],   12(%[cf])               \n\t"
88                        "abs.s  %[a],   %[a]                    \n\t"
89                        "abs.s  %[b],   %[b]                    \n\t"
90                        "abs.s  %[c],   %[c]                    \n\t"
91                        "abs.s  %[d],   %[d]                    \n\t"
92                        "sqrt.s %[ax],  %[a]                    \n\t"
93                        "sqrt.s %[bx],  %[b]                    \n\t"
94                        "sqrt.s %[cx],  %[c]                    \n\t"
95                        "sqrt.s %[dx],  %[d]                    \n\t"
96                        "madd.s %[e],   %[e],   %[a],   %[a]    \n\t"
97                        "madd.s %[e],   %[e],   %[b],   %[b]    \n\t"
98                        "madd.s %[e],   %[e],   %[c],   %[c]    \n\t"
99                        "madd.s %[e],   %[e],   %[d],   %[d]    \n\t"
100                        "add.s  %[f],   %[f],   %[ax]           \n\t"
101                        "add.s  %[f],   %[f],   %[bx]           \n\t"
102                        "add.s  %[f],   %[f],   %[cx]           \n\t"
103                        "add.s  %[f],   %[f],   %[dx]           \n\t"
104
105                        : [a]"=&f"(a), [b]"=&f"(b),
106                          [c]"=&f"(c), [d]"=&f"(d),
107                          [e]"+f"(band->energy), [f]"+f"(form_factor),
108                          [ax]"=&f"(ax), [bx]"=&f"(bx),
109                          [cx]"=&f"(cx), [dx]"=&f"(dx)
110                        : [cf]"r"(cf)
111                        : "memory"
112                    );
113                }
114            }
115
116            Temp = sqrtf((float)band_sizes[g] / band->energy);
117            band->thr      = band->energy * 0.001258925f;
118            band->nz_lines = form_factor * sqrtf(Temp);
119            start += band_sizes[g];
120            wstart += band_sizes[g];
121        }
122    }
123}
124
125static void psy_hp_filter_mips(const float *firbuf, float *hpfsmpl, const float * psy_fir_coeffs)
126{
127    float sum1, sum2, sum3, sum4;
128    float *fb = (float*)firbuf;
129    float *fb_end = fb + AAC_BLOCK_SIZE_LONG;
130    float *hp = hpfsmpl;
131
132    float coeff0 = psy_fir_coeffs[1];
133    float coeff1 = psy_fir_coeffs[3];
134    float coeff2 = psy_fir_coeffs[5];
135    float coeff3 = psy_fir_coeffs[7];
136    float coeff4 = psy_fir_coeffs[9];
137
138    float f1 = 32768.0;
139    __asm__ volatile (
140        ".set push                                          \n\t"
141        ".set noreorder                                     \n\t"
142
143        "1:                                                 \n\t"
144        "lwc1   $f0,        40(%[fb])                       \n\t"
145        "lwc1   $f1,        4(%[fb])                        \n\t"
146        "lwc1   $f2,        80(%[fb])                       \n\t"
147        "lwc1   $f3,        44(%[fb])                       \n\t"
148        "lwc1   $f4,        8(%[fb])                        \n\t"
149        "madd.s %[sum1],    $f0,        $f1,    %[coeff0]   \n\t"
150        "lwc1   $f5,        84(%[fb])                       \n\t"
151        "lwc1   $f6,        48(%[fb])                       \n\t"
152        "madd.s %[sum2],    $f3,        $f4,    %[coeff0]   \n\t"
153        "lwc1   $f7,        12(%[fb])                       \n\t"
154        "madd.s %[sum1],    %[sum1],    $f2,    %[coeff0]   \n\t"
155        "lwc1   $f8,        88(%[fb])                       \n\t"
156        "lwc1   $f9,        52(%[fb])                       \n\t"
157        "madd.s %[sum2],    %[sum2],    $f5,    %[coeff0]   \n\t"
158        "madd.s %[sum3],    $f6,        $f7,    %[coeff0]   \n\t"
159        "lwc1   $f10,       16(%[fb])                       \n\t"
160        "lwc1   $f11,       92(%[fb])                       \n\t"
161        "madd.s %[sum1],    %[sum1],    $f7,    %[coeff1]   \n\t"
162        "lwc1   $f1,        72(%[fb])                       \n\t"
163        "madd.s %[sum3],    %[sum3],    $f8,    %[coeff0]   \n\t"
164        "madd.s %[sum4],    $f9,        $f10,   %[coeff0]   \n\t"
165        "madd.s %[sum2],    %[sum2],    $f10,   %[coeff1]   \n\t"
166        "madd.s %[sum1],    %[sum1],    $f1,    %[coeff1]   \n\t"
167        "lwc1   $f4,        76(%[fb])                       \n\t"
168        "lwc1   $f8,        20(%[fb])                       \n\t"
169        "madd.s %[sum4],    %[sum4],    $f11,   %[coeff0]   \n\t"
170        "lwc1   $f11,       24(%[fb])                       \n\t"
171        "madd.s %[sum2],    %[sum2],    $f4,    %[coeff1]   \n\t"
172        "madd.s %[sum1],    %[sum1],    $f8,    %[coeff2]   \n\t"
173        "madd.s %[sum3],    %[sum3],    $f8,    %[coeff1]   \n\t"
174        "madd.s %[sum4],    %[sum4],    $f11,   %[coeff1]   \n\t"
175        "lwc1   $f7,        64(%[fb])                       \n\t"
176        "madd.s %[sum2],    %[sum2],    $f11,   %[coeff2]   \n\t"
177        "lwc1   $f10,       68(%[fb])                       \n\t"
178        "madd.s %[sum3],    %[sum3],    $f2,    %[coeff1]   \n\t"
179        "madd.s %[sum4],    %[sum4],    $f5,    %[coeff1]   \n\t"
180        "madd.s %[sum1],    %[sum1],    $f7,    %[coeff2]   \n\t"
181        "madd.s %[sum2],    %[sum2],    $f10,   %[coeff2]   \n\t"
182        "lwc1   $f2,        28(%[fb])                       \n\t"
183        "lwc1   $f5,        32(%[fb])                       \n\t"
184        "lwc1   $f8,        56(%[fb])                       \n\t"
185        "lwc1   $f11,       60(%[fb])                       \n\t"
186        "madd.s %[sum3],    %[sum3],    $f2,    %[coeff2]   \n\t"
187        "madd.s %[sum4],    %[sum4],    $f5,    %[coeff2]   \n\t"
188        "madd.s %[sum1],    %[sum1],    $f2,    %[coeff3]   \n\t"
189        "madd.s %[sum2],    %[sum2],    $f5,    %[coeff3]   \n\t"
190        "madd.s %[sum3],    %[sum3],    $f1,    %[coeff2]   \n\t"
191        "madd.s %[sum4],    %[sum4],    $f4,    %[coeff2]   \n\t"
192        "madd.s %[sum1],    %[sum1],    $f8,    %[coeff3]   \n\t"
193        "madd.s %[sum2],    %[sum2],    $f11,   %[coeff3]   \n\t"
194        "lwc1   $f1,        36(%[fb])                       \n\t"
195        PTR_ADDIU "%[fb],   %[fb],      16                  \n\t"
196        "madd.s %[sum4],    %[sum4],    $f0,    %[coeff3]   \n\t"
197        "madd.s %[sum3],    %[sum3],    $f1,    %[coeff3]   \n\t"
198        "madd.s %[sum1],    %[sum1],    $f1,    %[coeff4]   \n\t"
199        "madd.s %[sum2],    %[sum2],    $f0,    %[coeff4]   \n\t"
200        "madd.s %[sum4],    %[sum4],    $f10,   %[coeff3]   \n\t"
201        "madd.s %[sum3],    %[sum3],    $f7,    %[coeff3]   \n\t"
202        "madd.s %[sum1],    %[sum1],    $f6,    %[coeff4]   \n\t"
203        "madd.s %[sum2],    %[sum2],    $f9,    %[coeff4]   \n\t"
204        "madd.s %[sum4],    %[sum4],    $f6,    %[coeff4]   \n\t"
205        "madd.s %[sum3],    %[sum3],    $f3,    %[coeff4]   \n\t"
206        "mul.s  %[sum1],    %[sum1],    %[f1]               \n\t"
207        "mul.s  %[sum2],    %[sum2],    %[f1]               \n\t"
208        "madd.s %[sum4],    %[sum4],    $f11,   %[coeff4]   \n\t"
209        "madd.s %[sum3],    %[sum3],    $f8,    %[coeff4]   \n\t"
210        "swc1   %[sum1],    0(%[hp])                        \n\t"
211        "swc1   %[sum2],    4(%[hp])                        \n\t"
212        "mul.s  %[sum4],    %[sum4],    %[f1]               \n\t"
213        "mul.s  %[sum3],    %[sum3],    %[f1]               \n\t"
214        "swc1   %[sum4],    12(%[hp])                       \n\t"
215        "swc1   %[sum3],    8(%[hp])                        \n\t"
216        "bne    %[fb],      %[fb_end],  1b                  \n\t"
217        PTR_ADDIU "%[hp],   %[hp],      16                  \n\t"
218
219        ".set pop                                           \n\t"
220
221        : [sum1]"=&f"(sum1), [sum2]"=&f"(sum2),
222          [sum3]"=&f"(sum3), [sum4]"=&f"(sum4),
223          [fb]"+r"(fb), [hp]"+r"(hp)
224        : [coeff0]"f"(coeff0), [coeff1]"f"(coeff1),
225          [coeff2]"f"(coeff2), [coeff3]"f"(coeff3),
226          [coeff4]"f"(coeff4), [fb_end]"r"(fb_end), [f1]"f"(f1)
227        : "$f0", "$f1", "$f2", "$f3", "$f4", "$f5", "$f6",
228          "$f7", "$f8", "$f9", "$f10", "$f11",
229          "memory"
230    );
231}
232
233#define calc_thr_3gpp calc_thr_3gpp_mips
234#define psy_hp_filter psy_hp_filter_mips
235
236#endif /* !HAVE_MIPS32R6 && !HAVE_MIPS64R6 */
237#endif /* HAVE_INLINE_ASM && HAVE_MIPSFPU */
238#endif /* AVCODEC_MIPS_AACPSY_MIPS_H */
239