1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
|
/*
* Copyright (c) 2006 Luca Barbato <lu_zero@gentoo.org>
*
* This file is part of FFmpeg.
*
* FFmpeg is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* FFmpeg is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with FFmpeg; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*/
#include "libavcodec/fmtconvert.h"
#include "libavutil/ppc/util_altivec.h"
#include "libavutil/mem.h"
#include "dsputil_altivec.h"
static void int32_to_float_fmul_scalar_altivec(float *dst, const int *src,
float mul, int len)
{
union {
vector float v;
float s[4];
} mul_u;
int i;
vector float src1, src2, dst1, dst2, mul_v, zero;
zero = (vector float)vec_splat_u32(0);
mul_u.s[0] = mul;
mul_v = vec_splat(mul_u.v, 0);
for (i = 0; i < len; i += 8) {
src1 = vec_ctf(vec_ld(0, src+i), 0);
src2 = vec_ctf(vec_ld(16, src+i), 0);
dst1 = vec_madd(src1, mul_v, zero);
dst2 = vec_madd(src2, mul_v, zero);
vec_st(dst1, 0, dst+i);
vec_st(dst2, 16, dst+i);
}
}
static vector signed short float_to_int16_one_altivec(const float *src)
{
vector float s0 = vec_ld(0, src);
vector float s1 = vec_ld(16, src);
vector signed int t0 = vec_cts(s0, 0);
vector signed int t1 = vec_cts(s1, 0);
return vec_packs(t0,t1);
}
static void float_to_int16_altivec(int16_t *dst, const float *src, long len)
{
int i;
vector signed short d0, d1, d;
vector unsigned char align;
if (((long)dst) & 15) { //FIXME
for (i = 0; i < len - 7; i += 8) {
d0 = vec_ld(0, dst+i);
d = float_to_int16_one_altivec(src + i);
d1 = vec_ld(15, dst+i);
d1 = vec_perm(d1, d0, vec_lvsl(0, dst + i));
align = vec_lvsr(0, dst + i);
d0 = vec_perm(d1, d, align);
d1 = vec_perm(d, d1, align);
vec_st(d0, 0, dst + i);
vec_st(d1, 15, dst + i);
}
} else {
for (i = 0; i < len - 7; i += 8) {
d = float_to_int16_one_altivec(src + i);
vec_st(d, 0, dst + i);
}
}
}
static void float_to_int16_stride_altivec(int16_t *dst, const float *src,
long len, int stride)
{
int i;
vector signed short d, s;
for (i = 0; i < len - 7; i += 8) {
d = float_to_int16_one_altivec(src + i);
#define ASSIGN_S_VEC_SPLAT_D(j) \
s = vec_splat(d, j); \
vec_ste(s, 0, dst); \
dst += stride
ASSIGN_S_VEC_SPLAT_D(0);
ASSIGN_S_VEC_SPLAT_D(1);
ASSIGN_S_VEC_SPLAT_D(2);
ASSIGN_S_VEC_SPLAT_D(3);
ASSIGN_S_VEC_SPLAT_D(4);
ASSIGN_S_VEC_SPLAT_D(5);
ASSIGN_S_VEC_SPLAT_D(6);
ASSIGN_S_VEC_SPLAT_D(7);
#undef ASSIGN_S_VEC_SPLAT_D
}
}
static void float_to_int16_interleave_altivec(int16_t *dst, const float **src,
long len, int channels)
{
int i;
vector signed short d0, d1, d2, c0, c1, t0, t1;
vector unsigned char align;
if (channels == 1)
float_to_int16_altivec(dst, src[0], len);
else {
if (channels == 2) {
if (((long)dst) & 15) {
for (i = 0; i < len - 7; i += 8) {
d0 = vec_ld(0, dst + i);
t0 = float_to_int16_one_altivec(src[0] + i);
d1 = vec_ld(31, dst + i);
t1 = float_to_int16_one_altivec(src[1] + i);
c0 = vec_mergeh(t0, t1);
c1 = vec_mergel(t0, t1);
d2 = vec_perm(d1, d0, vec_lvsl(0, dst + i));
align = vec_lvsr(0, dst + i);
d0 = vec_perm(d2, c0, align);
d1 = vec_perm(c0, c1, align);
vec_st(d0, 0, dst + i);
d0 = vec_perm(c1, d2, align);
vec_st(d1, 15, dst + i);
vec_st(d0, 31, dst + i);
dst += 8;
}
} else {
for (i = 0; i < len - 7; i += 8) {
t0 = float_to_int16_one_altivec(src[0] + i);
t1 = float_to_int16_one_altivec(src[1] + i);
d0 = vec_mergeh(t0, t1);
d1 = vec_mergel(t0, t1);
vec_st(d0, 0, dst + i);
vec_st(d1, 16, dst + i);
dst += 8;
}
}
} else {
for (i = 0; i < channels; i++)
float_to_int16_stride_altivec(dst + i, src[i], len, channels);
}
}
}
void ff_fmt_convert_init_altivec(FmtConvertContext *c, AVCodecContext *avctx)
{
c->int32_to_float_fmul_scalar = int32_to_float_fmul_scalar_altivec;
if (!(avctx->flags & CODEC_FLAG_BITEXACT)) {
c->float_to_int16 = float_to_int16_altivec;
c->float_to_int16_interleave = float_to_int16_interleave_altivec;
}
}
|
