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
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
|
// Copyright 2016 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.
#ifdef UNSAFE_BUFFERS_BUILD
// TODO(crbug.com/351564777): Remove this and convert code to safer constructs.
#pragma allow_unsafe_buffers
#endif
#include "third_party/blink/renderer/platform/audio/delay.h"
#include <xmmintrin.h>
#include <array>
namespace blink {
ALWAYS_INLINE static __m128i WrapIndexVector(__m128i v_write_index,
__m128i v_buffer_length) {
// Wrap the write_index if any index is past the end of the buffer.
// This implements
//
// if (write_index >= buffer_length)
// write_index -= buffer_length
// There's no mm_cmpge_epi32, so we need to use mm_cmplt_epi32. Thus, the
// above becomes
//
// if (!(write_index < buffer_length))
// write_index -= buffer_length
// If write_index < buffer_length, set cmp = 0xffffffff. Otherwise 0.
__m128i cmp = _mm_cmplt_epi32(v_write_index, v_buffer_length);
// Invert cmp and bitwise-and with buffer_length to get buffer_length or 0
// depending on whether write_index >= buffer_length or not. Subtract from
// write_index to wrap it.
return _mm_sub_epi32(v_write_index, _mm_andnot_si128(cmp, v_buffer_length));
}
ALWAYS_INLINE static __m128 WrapPositionVector(__m128 v_position,
__m128 v_buffer_length) {
// Wrap the read position if it exceed the buffer length.
// This implements
//
// if (position >= buffer_length)
// read_position -= buffer_length
// If position >= buffer length, set cmp = 0xffffffff. Otherwise 0.
__m128 cmp = _mm_cmpge_ps(v_position, v_buffer_length);
// Bitwise-and buffer_length with cmp to get buffer_length or 0 depending on
// whether read_position >= buffer length or not. Then subtract from the
// position to wrap it.
return _mm_sub_ps(v_position, _mm_and_ps(v_buffer_length, cmp));
}
std::tuple<unsigned, int> Delay::ProcessARateVector(
float* destination,
uint32_t frames_to_process) const {
const int buffer_length = buffer_.size();
const float* buffer = buffer_.Data();
const float sample_rate = sample_rate_;
const float* delay_times = delay_times_.Data();
int w_index = write_index_;
const __m128 v_sample_rate = _mm_set1_ps(sample_rate);
const __m128 v_all_zeros = _mm_setzero_ps();
// The buffer length as a float and as an int so we don't need to constant
// convert from one to the other.
const __m128 v_buffer_length_float = _mm_set1_ps(buffer_length);
const __m128i v_buffer_length_int = _mm_set1_epi32(buffer_length);
// How much to increment the write index each time through the loop.
const __m128i v_incr = _mm_set1_epi32(4);
// Temp arrays for storing the samples needed for interpolation
std::array<float, 4> sample1 __attribute((aligned(16)));
std::array<float, 4> sample2 __attribute((aligned(16)));
// Initialize the write index vector, and wrap the values if needed.
__m128i v_write_index =
_mm_set_epi32(w_index + 3, w_index + 2, w_index + 1, w_index + 0);
v_write_index = WrapIndexVector(v_write_index, v_buffer_length_int);
const int number_of_loops = frames_to_process / 4;
int k = 0;
for (int n = 0; n < number_of_loops; ++n, k += 4) {
// It's possible that `delay_time` contains negative values. Make sure
// they are greater than zero.
const __m128 v_delay_time = _mm_max_ps(_mm_loadu_ps(delay_times + k),
v_all_zeros);
const __m128 v_desired_delay_frames =
_mm_mul_ps(v_delay_time, v_sample_rate);
// read_position = write_index + buffer_length - desired_delay_frames. Wrap
// the position if needed.
__m128 v_read_position =
_mm_add_ps(_mm_cvtepi32_ps(v_write_index),
_mm_sub_ps(v_buffer_length_float, v_desired_delay_frames));
v_read_position =
WrapPositionVector(v_read_position, v_buffer_length_float);
// Get indices into the buffer for the samples we need for interpolation.
const __m128i v_read_index1 = WrapIndexVector(
_mm_cvttps_epi32(v_read_position), v_buffer_length_int);
const __m128i v_read_index2 = WrapIndexVector(
_mm_add_epi32(v_read_index1, _mm_set1_epi32(1)), v_buffer_length_int);
const __m128 interpolation_factor =
_mm_sub_ps(v_read_position, _mm_cvtepi32_ps(v_read_index1));
const uint32_t* read_index1 =
reinterpret_cast<const uint32_t*>(&v_read_index1);
const uint32_t* read_index2 =
reinterpret_cast<const uint32_t*>(&v_read_index2);
for (int m = 0; m < 4; ++m) {
sample1[m] = buffer[read_index1[m]];
sample2[m] = buffer[read_index2[m]];
}
const __m128 v_sample1 = _mm_load_ps(sample1.data());
const __m128 v_sample2 = _mm_load_ps(sample2.data());
v_write_index = _mm_add_epi32(v_write_index, v_incr);
v_write_index = WrapIndexVector(v_write_index, v_buffer_length_int);
const __m128 sample = _mm_add_ps(
v_sample1,
_mm_mul_ps(interpolation_factor, _mm_sub_ps(v_sample2, v_sample1)));
_mm_store_ps(destination + k, sample);
}
// Update |w_index|_ based on how many frames we processed here, wrapping
// around if needed.
w_index = write_index_ + k;
if (w_index >= buffer_length) {
w_index -= buffer_length;
}
return std::make_tuple(k, w_index);
}
void Delay::HandleNaN(float* delay_times,
uint32_t frames_to_process,
float max_time) {
unsigned k = 0;
const unsigned number_of_loops = frames_to_process / 4;
__m128 v_max_time = _mm_set1_ps(max_time);
// This is approximately 4 times faster than the scalar version.
for (unsigned loop = 0; loop < number_of_loops; ++loop, k += 4) {
__m128 x = _mm_loadu_ps(delay_times + k);
// 0xffffffff if x is NaN. Otherwise 0
__m128 cmp = _mm_cmpunord_ps(x, x);
// Use cmp as a mask to set a component of x to 0 if is NaN. Otherwise,
// preserve x.
x = _mm_andnot_ps(cmp, x);
// Now set cmp to be max_time if the value is 0xffffffff or 0.
cmp = _mm_and_ps(cmp, v_max_time);
// Merge i (bitwise or) x and cmp. This makes x = max_time if x was NaN and
// preserves x if not.
x = _mm_or_ps(x, cmp);
_mm_storeu_ps(delay_times + k, x);
}
// Handle any frames not done in the loop above.
for (; k < frames_to_process; ++k) {
if (std::isnan(delay_times[k])) {
delay_times[k] = max_time;
}
}
}
} // namespace blink
|
