| OLD | NEW |
| 1 /* | 1 /* |
| 2 * Copyright (c) 2012 The WebRTC project authors. All Rights Reserved. | 2 * Copyright (c) 2012 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 #include "webrtc/modules/audio_coding/neteq/preemptive_expand.h" | 11 #include "webrtc/modules/audio_coding/neteq/preemptive_expand.h" |
| 12 | 12 |
| 13 #include <algorithm> // min, max | 13 #include <algorithm> // min, max |
| 14 | 14 |
| 15 #include "webrtc/common_audio/signal_processing/include/signal_processing_librar
y.h" | 15 #include "webrtc/common_audio/signal_processing/include/signal_processing_librar
y.h" |
| 16 | 16 |
| 17 namespace webrtc { | 17 namespace webrtc { |
| 18 | 18 |
| 19 PreemptiveExpand::ReturnCodes PreemptiveExpand::Process( | 19 PreemptiveExpand::ReturnCodes PreemptiveExpand::Process( |
| 20 const int16_t* input, | 20 const int16_t* input, |
| 21 int input_length, | 21 size_t input_length, |
| 22 int old_data_length, | 22 size_t old_data_length, |
| 23 AudioMultiVector* output, | 23 AudioMultiVector* output, |
| 24 int16_t* length_change_samples) { | 24 size_t* length_change_samples) { |
| 25 old_data_length_per_channel_ = old_data_length; | 25 old_data_length_per_channel_ = old_data_length; |
| 26 // Input length must be (almost) 30 ms. | 26 // Input length must be (almost) 30 ms. |
| 27 // Also, the new part must be at least |overlap_samples_| elements. | 27 // Also, the new part must be at least |overlap_samples_| elements. |
| 28 static const int k15ms = 120; // 15 ms = 120 samples at 8 kHz sample rate. | 28 static const size_t k15ms = 120; // 15 ms = 120 samples at 8 kHz sample rate. |
| 29 if (num_channels_ == 0 || | 29 if (num_channels_ == 0 || |
| 30 input_length / num_channels_ < (2 * k15ms - 1) * fs_mult_ || | 30 input_length / num_channels_ < (2 * k15ms - 1) * fs_mult_ || |
| 31 old_data_length >= input_length / num_channels_ - overlap_samples_) { | 31 old_data_length >= input_length / num_channels_ - overlap_samples_) { |
| 32 // Length of input data too short to do preemptive expand. Simply move all | 32 // Length of input data too short to do preemptive expand. Simply move all |
| 33 // data from input to output. | 33 // data from input to output. |
| 34 output->PushBackInterleaved(input, input_length); | 34 output->PushBackInterleaved(input, input_length); |
| 35 return kError; | 35 return kError; |
| 36 } | 36 } |
| 37 const bool kFastMode = false; // Fast mode is not available for PE Expand. | 37 const bool kFastMode = false; // Fast mode is not available for PE Expand. |
| 38 return TimeStretch::Process(input, input_length, kFastMode, output, | 38 return TimeStretch::Process(input, input_length, kFastMode, output, |
| 39 length_change_samples); | 39 length_change_samples); |
| 40 } | 40 } |
| 41 | 41 |
| 42 void PreemptiveExpand::SetParametersForPassiveSpeech(size_t len, | 42 void PreemptiveExpand::SetParametersForPassiveSpeech(size_t len, |
| 43 int16_t* best_correlation, | 43 int16_t* best_correlation, |
| 44 int* peak_index) const { | 44 size_t* peak_index) const { |
| 45 // When the signal does not contain any active speech, the correlation does | 45 // When the signal does not contain any active speech, the correlation does |
| 46 // not matter. Simply set it to zero. | 46 // not matter. Simply set it to zero. |
| 47 *best_correlation = 0; | 47 *best_correlation = 0; |
| 48 | 48 |
| 49 // For low energy expansion, the new data can be less than 15 ms, | 49 // For low energy expansion, the new data can be less than 15 ms, |
| 50 // but we must ensure that best_correlation is not larger than the length of | 50 // but we must ensure that best_correlation is not larger than the length of |
| 51 // the new data. | 51 // the new data. |
| 52 // but we must ensure that best_correlation is not larger than the new data. | 52 // but we must ensure that best_correlation is not larger than the new data. |
| 53 *peak_index = std::min(*peak_index, | 53 *peak_index = std::min(*peak_index, |
| 54 static_cast<int>(len - old_data_length_per_channel_)); | 54 len - old_data_length_per_channel_); |
| 55 } | 55 } |
| 56 | 56 |
| 57 PreemptiveExpand::ReturnCodes PreemptiveExpand::CheckCriteriaAndStretch( | 57 PreemptiveExpand::ReturnCodes PreemptiveExpand::CheckCriteriaAndStretch( |
| 58 const int16_t* input, | 58 const int16_t* input, |
| 59 size_t input_length, | 59 size_t input_length, |
| 60 size_t peak_index, | 60 size_t peak_index, |
| 61 int16_t best_correlation, | 61 int16_t best_correlation, |
| 62 bool active_speech, | 62 bool active_speech, |
| 63 bool /*fast_mode*/, | 63 bool /*fast_mode*/, |
| 64 AudioMultiVector* output) const { | 64 AudioMultiVector* output) const { |
| 65 // Pre-calculate common multiplication with |fs_mult_|. | 65 // Pre-calculate common multiplication with |fs_mult_|. |
| 66 // 120 corresponds to 15 ms. | 66 // 120 corresponds to 15 ms. |
| 67 int fs_mult_120 = fs_mult_ * 120; | 67 size_t fs_mult_120 = static_cast<size_t>(fs_mult_ * 120); |
| 68 assert(old_data_length_per_channel_ >= 0); // Make sure it's been set. | |
| 69 // Check for strong correlation (>0.9 in Q14) and at least 15 ms new data, | 68 // Check for strong correlation (>0.9 in Q14) and at least 15 ms new data, |
| 70 // or passive speech. | 69 // or passive speech. |
| 71 if (((best_correlation > kCorrelationThreshold) && | 70 if (((best_correlation > kCorrelationThreshold) && |
| 72 (old_data_length_per_channel_ <= fs_mult_120)) || | 71 (old_data_length_per_channel_ <= fs_mult_120)) || |
| 73 !active_speech) { | 72 !active_speech) { |
| 74 // Do accelerate operation by overlap add. | 73 // Do accelerate operation by overlap add. |
| 75 | 74 |
| 76 // Set length of the first part, not to be modified. | 75 // Set length of the first part, not to be modified. |
| 77 size_t unmodified_length = std::max(old_data_length_per_channel_, | 76 size_t unmodified_length = std::max(old_data_length_per_channel_, |
| 78 fs_mult_120); | 77 fs_mult_120); |
| (...skipping 21 matching lines...) Expand all Loading... |
| 100 // Accelerate not allowed. Simply move all data from decoded to outData. | 99 // Accelerate not allowed. Simply move all data from decoded to outData. |
| 101 output->PushBackInterleaved(input, input_length); | 100 output->PushBackInterleaved(input, input_length); |
| 102 return kNoStretch; | 101 return kNoStretch; |
| 103 } | 102 } |
| 104 } | 103 } |
| 105 | 104 |
| 106 PreemptiveExpand* PreemptiveExpandFactory::Create( | 105 PreemptiveExpand* PreemptiveExpandFactory::Create( |
| 107 int sample_rate_hz, | 106 int sample_rate_hz, |
| 108 size_t num_channels, | 107 size_t num_channels, |
| 109 const BackgroundNoise& background_noise, | 108 const BackgroundNoise& background_noise, |
| 110 int overlap_samples) const { | 109 size_t overlap_samples) const { |
| 111 return new PreemptiveExpand( | 110 return new PreemptiveExpand( |
| 112 sample_rate_hz, num_channels, background_noise, overlap_samples); | 111 sample_rate_hz, num_channels, background_noise, overlap_samples); |
| 113 } | 112 } |
| 114 | 113 |
| 115 } // namespace webrtc | 114 } // namespace webrtc |
| OLD | NEW |