blob: 6df0ead2716c951beabc537457e46c619fd37ff4 [file] [log] [blame]
* Copyright (c) 2018 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 <memory>
#include <vector>
#include "api/array_view.h"
#include "common_audio/resampler/include/push_resampler.h"
#include "modules/audio_processing/agc2/cpu_features.h"
#include "modules/audio_processing/include/audio_frame_view.h"
namespace webrtc {
// Wraps a single-channel Voice Activity Detector (VAD) which is used to analyze
// the first channel of the input audio frames. Takes care of resampling the
// input frames to match the sample rate of the wrapped VAD and periodically
// resets the VAD.
class VoiceActivityDetectorWrapper {
// Single channel VAD interface.
class MonoVad {
virtual ~MonoVad() = default;
// Returns the sample rate (Hz) required for the input frames analyzed by
// `ComputeProbability`.
virtual int SampleRateHz() const = 0;
// Resets the internal state.
virtual void Reset() = 0;
// Analyzes an audio frame and returns the speech probability.
virtual float Analyze(rtc::ArrayView<const float> frame) = 0;
// Ctor. `vad_reset_period_ms` indicates the period in milliseconds to call
// `MonoVad::Reset()`; it must be equal to or greater than the duration of two
// frames. Uses `cpu_features` to instantiate the default VAD.
VoiceActivityDetectorWrapper(int vad_reset_period_ms,
const AvailableCpuFeatures& cpu_features,
int sample_rate_hz);
// Ctor. Uses a custom `vad`.
VoiceActivityDetectorWrapper(int vad_reset_period_ms,
std::unique_ptr<MonoVad> vad,
int sample_rate_hz);
VoiceActivityDetectorWrapper(const VoiceActivityDetectorWrapper&) = delete;
VoiceActivityDetectorWrapper& operator=(const VoiceActivityDetectorWrapper&) =
// Initializes the VAD wrapper.
void Initialize(int sample_rate_hz);
// Analyzes the first channel of `frame` and returns the speech probability.
// `frame` must be a 10 ms frame with the sample rate specified in the last
// `Initialize()` call.
float Analyze(AudioFrameView<const float> frame);
const int vad_reset_period_frames_;
int frame_size_;
int time_to_vad_reset_;
PushResampler<float> resampler_;
std::unique_ptr<MonoVad> vad_;
std::vector<float> resampled_buffer_;
} // namespace webrtc