vad_audio_proc.cc | Explore in Territory

/*
 *  Copyright (c) 2012 The WebRTC project authors. All Rights Reserved.
 *
 *  Use of this source code is governed by a BSD-style license
 *  that can be found in the LICENSE file in the root of the source
 *  tree. An additional intellectual property rights grant can be found
 *  in the file PATENTS.  All contributing project authors may
 *  be found in the AUTHORS file in the root of the source tree.
 */

#include "modules/audio_processing/vad/vad_audio_proc.h"

#include <math.h>
#include <stdio.h>
#include <string.h>

#include "common_audio/third_party/ooura/fft_size_256/fft4g.h"
#include "modules/audio_processing/vad/pitch_internal.h"
#include "modules/audio_processing/vad/pole_zero_filter.h"
#include "modules/audio_processing/vad/vad_audio_proc_internal.h"
#include "rtc_base/checks.h"
extern "C" {
#include "modules/audio_coding/codecs/isac/main/source/filter_functions.h"
#include "modules/audio_coding/codecs/isac/main/source/isac_vad.h"
#include "modules/audio_coding/codecs/isac/main/source/pitch_estimator.h"
#include "modules/audio_coding/codecs/isac/main/source/structs.h"
}

namespace webrtc {

// The following structures are declared anonymous in iSAC's structs.h. To
// forward declare them, we use this derived class trick.
struct VadAudioProc::PitchAnalysisStruct : public ::PitchAnalysisStruct { … };
struct VadAudioProc::PreFiltBankstr : public ::PreFiltBankstr { … };

static constexpr float kFrequencyResolution = …;
static constexpr int kSilenceRms = …;

// TODO(turajs): Make a Create or Init for VadAudioProc.
VadAudioProc::VadAudioProc()
    : … { … }

VadAudioProc::~VadAudioProc() { … }

void VadAudioProc::ResetBuffer() { … }

int VadAudioProc::ExtractFeatures(const int16_t* frame,
                                  size_t length,
                                  AudioFeatures* features) { … }

// Computes |kLpcOrder + 1| correlation coefficients.
void VadAudioProc::SubframeCorrelation(double* corr,
                                       size_t length_corr,
                                       size_t subframe_index) { … }

// Compute `kNum10msSubframes` sets of LPC coefficients, one per 10 ms input.
// The analysis window is 15 ms long and it is centered on the first half of
// each 10ms sub-frame. This is equivalent to computing LPC coefficients for the
// first half of each 10 ms subframe.
void VadAudioProc::GetLpcPolynomials(double* lpc, size_t length_lpc) { … }

// Fit a second order curve to these 3 points and find the location of the
// extremum. The points are inverted before curve fitting.
static float QuadraticInterpolation(float prev_val,
                                    float curr_val,
                                    float next_val) { … }

// 1 / A(z), where A(z) is defined by `lpc` is a model of the spectral envelope
// of the input signal. The local maximum of the spectral envelope corresponds
// with the local minimum of A(z). It saves complexity, as we save one
// inversion. Furthermore, we find the first local maximum of magnitude squared,
// to save on one square root.
void VadAudioProc::FindFirstSpectralPeaks(double* f_peak,
                                          size_t length_f_peak) { … }

// Using iSAC functions to estimate pitch gains & lags.
void VadAudioProc::PitchAnalysis(double* log_pitch_gains,
                                 double* pitch_lags_hz,
                                 size_t length) { … }

void VadAudioProc::Rms(double* rms, size_t length_rms) { … }

}  // namespace webrtc
chromium/third_party/webrtc/modules/audio_processing/vad/vad_audio_proc.cc
