AnotherFoxGuy/webrtc-audio-processing
1
0
Fork 0
Code Pull Requests Activity

Bump to WebRTC M120 release

Browse Source 
Some API deprecation -- ExperimentalAgc and ExperimentalNs are gone.
We're continuing to carry iSAC even though it's gone upstream, but maybe
we'll want to drop that soon.
This commit is contained in:
Arun Raghavan
2023-12-12 10:42:58 -05:00
parent 9a202fb8c2
commit c6abf6cd3f
479 changed files with 20900 additions and 11996 deletions
Download Patch File Download Diff File Expand all files Collapse all files

66
webrtc/modules/audio_processing/agc2/rnn_vad/pitch_search.cc
View File

@ -18,38 +18,52 @@
namespace webrtc {
namespace rnn_vad {
PitchEstimator::PitchEstimator()
: pitch_buf_decimated_(kBufSize12kHz),
pitch_buf_decimated_view_(pitch_buf_decimated_.data(), kBufSize12kHz),
auto_corr_(kNumInvertedLags12kHz),
auto_corr_view_(auto_corr_.data(), kNumInvertedLags12kHz) {
RTC_DCHECK_EQ(kBufSize12kHz, pitch_buf_decimated_.size());
RTC_DCHECK_EQ(kNumInvertedLags12kHz, auto_corr_view_.size());
}
PitchEstimator::PitchEstimator(const AvailableCpuFeatures& cpu_features)
: cpu_features_(cpu_features),
y_energy_24kHz_(kRefineNumLags24kHz, 0.f),
pitch_buffer_12kHz_(kBufSize12kHz),
auto_correlation_12kHz_(kNumLags12kHz) {}
PitchEstimator::~PitchEstimator() = default;
PitchInfo PitchEstimator::Estimate(
rtc::ArrayView<const float, kBufSize24kHz> pitch_buf) {
int PitchEstimator::Estimate(
rtc::ArrayView<const float, kBufSize24kHz> pitch_buffer) {
rtc::ArrayView<float, kBufSize12kHz> pitch_buffer_12kHz_view(
pitch_buffer_12kHz_.data(), kBufSize12kHz);
RTC_DCHECK_EQ(pitch_buffer_12kHz_.size(), pitch_buffer_12kHz_view.size());
rtc::ArrayView<float, kNumLags12kHz> auto_correlation_12kHz_view(
auto_correlation_12kHz_.data(), kNumLags12kHz);
RTC_DCHECK_EQ(auto_correlation_12kHz_.size(),
auto_correlation_12kHz_view.size());
// TODO(bugs.chromium.org/10480): Use `cpu_features_` to estimate pitch.
// Perform the initial pitch search at 12 kHz.
Decimate2x(pitch_buf, pitch_buf_decimated_view_);
auto_corr_calculator_.ComputeOnPitchBuffer(pitch_buf_decimated_view_,
auto_corr_view_);
std::array<size_t, 2> pitch_candidates_inv_lags = FindBestPitchPeriods(
auto_corr_view_, pitch_buf_decimated_view_, kMaxPitch12kHz);
// Refine the pitch period estimation.
Decimate2x(pitch_buffer, pitch_buffer_12kHz_view);
auto_corr_calculator_.ComputeOnPitchBuffer(pitch_buffer_12kHz_view,
auto_correlation_12kHz_view);
CandidatePitchPeriods pitch_periods = ComputePitchPeriod12kHz(
pitch_buffer_12kHz_view, auto_correlation_12kHz_view, cpu_features_);
// The refinement is done using the pitch buffer that contains 24 kHz samples.
// Therefore, adapt the inverted lags in |pitch_candidates_inv_lags| from 12
// Therefore, adapt the inverted lags in `pitch_candidates_inv_lags` from 12
// to 24 kHz.
pitch_candidates_inv_lags[0] *= 2;
pitch_candidates_inv_lags[1] *= 2;
size_t pitch_inv_lag_48kHz =
RefinePitchPeriod48kHz(pitch_buf, pitch_candidates_inv_lags);
// Look for stronger harmonics to find the final pitch period and its gain.
RTC_DCHECK_LT(pitch_inv_lag_48kHz, kMaxPitch48kHz);
last_pitch_48kHz_ = CheckLowerPitchPeriodsAndComputePitchGain(
pitch_buf, kMaxPitch48kHz - pitch_inv_lag_48kHz, last_pitch_48kHz_);
return last_pitch_48kHz_;
pitch_periods.best *= 2;
pitch_periods.second_best *= 2;
// Refine the initial pitch period estimation from 12 kHz to 48 kHz.
// Pre-compute frame energies at 24 kHz.
rtc::ArrayView<float, kRefineNumLags24kHz> y_energy_24kHz_view(
y_energy_24kHz_.data(), kRefineNumLags24kHz);
RTC_DCHECK_EQ(y_energy_24kHz_.size(), y_energy_24kHz_view.size());
ComputeSlidingFrameSquareEnergies24kHz(pitch_buffer, y_energy_24kHz_view,
cpu_features_);
// Estimation at 48 kHz.
const int pitch_lag_48kHz = ComputePitchPeriod48kHz(
pitch_buffer, y_energy_24kHz_view, pitch_periods, cpu_features_);
last_pitch_48kHz_ = ComputeExtendedPitchPeriod48kHz(
pitch_buffer, y_energy_24kHz_view,
/*initial_pitch_period_48kHz=*/kMaxPitch48kHz - pitch_lag_48kHz,
last_pitch_48kHz_, cpu_features_);
return last_pitch_48kHz_.period;
}
} // namespace rnn_vad