OLD | NEW |
1 /* | 1 /* |
2 * Copyright (c) 2014 The WebRTC project authors. All Rights Reserved. | 2 * Copyright (c) 2014 The WebRTC project authors. All Rights Reserved. |
3 * | 3 * |
4 * Use of this source code is governed by a BSD-style license | 4 * Use of this source code is governed by a BSD-style license |
5 * that can be found in the LICENSE file in the root of the source | 5 * that can be found in the LICENSE file in the root of the source |
6 * tree. An additional intellectual property rights grant can be found | 6 * tree. An additional intellectual property rights grant can be found |
7 * in the file PATENTS. All contributing project authors may | 7 * in the file PATENTS. All contributing project authors may |
8 * be found in the AUTHORS file in the root of the source tree. | 8 * be found in the AUTHORS file in the root of the source tree. |
9 */ | 9 */ |
10 | 10 |
11 // | 11 // |
12 // Command line tool for speech intelligibility enhancement. Provides for | 12 // Command line tool for speech intelligibility enhancement. Provides for |
13 // running and testing intelligibility_enhancer as an independent process. | 13 // running and testing intelligibility_enhancer as an independent process. |
14 // Use --help for options. | 14 // Use --help for options. |
15 // | 15 // |
16 | 16 |
17 #include <stdint.h> | |
18 #include <stdlib.h> | |
19 #include <sys/stat.h> | 17 #include <sys/stat.h> |
20 #include <sys/types.h> | |
21 #include <string> | |
22 | 18 |
23 #include "gflags/gflags.h" | 19 #include "gflags/gflags.h" |
24 #include "testing/gtest/include/gtest/gtest.h" | 20 #include "testing/gtest/include/gtest/gtest.h" |
25 #include "webrtc/base/checks.h" | |
26 #include "webrtc/base/criticalsection.h" | 21 #include "webrtc/base/criticalsection.h" |
27 #include "webrtc/common_audio/include/audio_util.h" | 22 #include "webrtc/common_audio/include/audio_util.h" |
28 #include "webrtc/common_audio/real_fourier.h" | |
29 #include "webrtc/common_audio/wav_file.h" | 23 #include "webrtc/common_audio/wav_file.h" |
30 #include "webrtc/modules/audio_processing/audio_buffer.h" | 24 #include "webrtc/modules/audio_processing/audio_buffer.h" |
31 #include "webrtc/modules/audio_processing/include/audio_processing.h" | |
32 #include "webrtc/modules/audio_processing/intelligibility/intelligibility_enhanc
er.h" | 25 #include "webrtc/modules/audio_processing/intelligibility/intelligibility_enhanc
er.h" |
33 #include "webrtc/modules/audio_processing/intelligibility/intelligibility_utils.
h" | |
34 #include "webrtc/modules/audio_processing/noise_suppression_impl.h" | 26 #include "webrtc/modules/audio_processing/noise_suppression_impl.h" |
35 #include "webrtc/system_wrappers/include/critical_section_wrapper.h" | |
36 #include "webrtc/test/testsupport/fileutils.h" | |
37 | 27 |
38 using std::complex; | 28 using std::complex; |
39 | 29 |
40 namespace webrtc { | 30 namespace webrtc { |
41 namespace { | 31 namespace { |
42 | 32 |
43 DEFINE_double(clear_alpha, 0.9, "Power decay factor for clear data."); | 33 DEFINE_double(clear_alpha, 0.9, "Power decay factor for clear data."); |
44 DEFINE_int32(sample_rate, | 34 DEFINE_int32(sample_rate, |
45 16000, | 35 16000, |
46 "Audio sample rate used in the input and output files."); | 36 "Audio sample rate used in the input and output files."); |
47 DEFINE_int32(ana_rate, | 37 DEFINE_int32(ana_rate, |
48 60, | 38 60, |
49 "Analysis rate; gains recalculated every N blocks."); | 39 "Analysis rate; gains recalculated every N blocks."); |
50 DEFINE_double(gain_limit, 1000.0, "Maximum gain change in one block."); | 40 DEFINE_double(gain_limit, 1000.0, "Maximum gain change in one block."); |
51 | 41 |
52 DEFINE_string(clear_file, "speech.wav", "Input file with clear speech."); | 42 DEFINE_string(clear_file, "speech.wav", "Input file with clear speech."); |
53 DEFINE_string(noise_file, "noise.wav", "Input file with noise data."); | 43 DEFINE_string(noise_file, "noise.wav", "Input file with noise data."); |
54 DEFINE_string(out_file, | 44 DEFINE_string(out_file, "proc_enhanced.wav", "Enhanced output file."); |
55 "proc_enhanced.wav", | |
56 "Enhanced output. Use '-' to " | |
57 "play through aplay immediately."); | |
58 | 45 |
59 const size_t kNumChannels = 1; | 46 const size_t kNumChannels = 1; |
60 | 47 |
61 // void function for gtest | 48 // void function for gtest |
62 void void_main(int argc, char* argv[]) { | 49 void void_main(int argc, char* argv[]) { |
63 google::SetUsageMessage( | 50 google::SetUsageMessage( |
64 "\n\nInput files must be little-endian 16-bit signed raw PCM.\n"); | 51 "\n\nInput files must be little-endian 16-bit signed raw PCM.\n"); |
65 google::ParseCommandLineFlags(&argc, &argv, true); | 52 google::ParseCommandLineFlags(&argc, &argv, true); |
66 | 53 |
67 size_t samples; // Number of samples in input PCM file | 54 size_t samples; // Number of samples in input PCM file |
(...skipping 53 matching lines...) Expand 10 before | Expand all | Expand 10 after Loading... |
121 ns.AnalyzeCaptureAudio(&capture_audio); | 108 ns.AnalyzeCaptureAudio(&capture_audio); |
122 ns.ProcessCaptureAudio(&capture_audio); | 109 ns.ProcessCaptureAudio(&capture_audio); |
123 enh.SetCaptureNoiseEstimate(ns.NoiseEstimate()); | 110 enh.SetCaptureNoiseEstimate(ns.NoiseEstimate()); |
124 enh.ProcessRenderAudio(&clear_cursor, FLAGS_sample_rate, kNumChannels); | 111 enh.ProcessRenderAudio(&clear_cursor, FLAGS_sample_rate, kNumChannels); |
125 clear_cursor += fragment_size; | 112 clear_cursor += fragment_size; |
126 noise_cursor += fragment_size; | 113 noise_cursor += fragment_size; |
127 } | 114 } |
128 | 115 |
129 FloatToFloatS16(&in_fpcm[0], samples, &in_fpcm[0]); | 116 FloatToFloatS16(&in_fpcm[0], samples, &in_fpcm[0]); |
130 | 117 |
131 if (FLAGS_out_file.compare("-") == 0) { | 118 WavWriter out_file(FLAGS_out_file, FLAGS_sample_rate, kNumChannels); |
132 const std::string temp_out_filename = | 119 out_file.WriteSamples(&in_fpcm[0], samples); |
133 test::TempFilename(test::WorkingDir(), "temp_wav_file"); | |
134 { | |
135 WavWriter out_file(temp_out_filename, FLAGS_sample_rate, kNumChannels); | |
136 out_file.WriteSamples(&in_fpcm[0], samples); | |
137 } | |
138 system(("aplay " + temp_out_filename).c_str()); | |
139 system(("rm " + temp_out_filename).c_str()); | |
140 } else { | |
141 WavWriter out_file(FLAGS_out_file, FLAGS_sample_rate, kNumChannels); | |
142 out_file.WriteSamples(&in_fpcm[0], samples); | |
143 } | |
144 } | 120 } |
145 | 121 |
146 } // namespace | 122 } // namespace |
147 } // namespace webrtc | 123 } // namespace webrtc |
148 | 124 |
149 int main(int argc, char* argv[]) { | 125 int main(int argc, char* argv[]) { |
150 webrtc::void_main(argc, argv); | 126 webrtc::void_main(argc, argv); |
151 return 0; | 127 return 0; |
152 } | 128 } |
OLD | NEW |