chromium/components/speech/endpointer/energy_endpointer.h

// Copyright 2024 The Chromium Authors
// Use of this source code is governed by a BSD-style license that can be
// found in the LICENSE file.

// The EnergyEndpointer class finds likely speech onset and offset points.
//
// The implementation described here is about the simplest possible.
// It is based on timings of threshold crossings for overall signal
// RMS. It is suitable for light weight applications.
//
// As written, the basic idea is that one specifies intervals that
// must be occupied by super- and sub-threshold energy levels, and
// defers decisions re onset and offset times until these
// specifications have been met.  Three basic intervals are tested: an
// onset window, a speech-on window, and an offset window.  We require
// super-threshold to exceed some mimimum total durations in the onset
// and speech-on windows before declaring the speech onset time, and
// we specify a required sub-threshold residency in the offset window
// before declaring speech offset. As the various residency requirements are
// met, the EnergyEndpointer instance assumes various states, and can return the
// ID of these states to the client (see EpStatus below).
//
// The levels of the speech and background noise are continuously updated. It is
// important that the background noise level be estimated initially for
// robustness in noisy conditions. The first frames are assumed to be background
// noise and a fast update rate is used for the noise level. The duration for
// fast update is controlled by the fast_update_dur_ paramter.
//
// If used in noisy conditions, the endpointer should be started and run in the
// EnvironmentEstimation mode, for at least 200ms, before switching to
// UserInputMode.
// Audio feedback contamination can appear in the input audio, if not cut
// out or handled by echo cancellation. Audio feedback can trigger a false
// accept. The false accepts can be ignored by setting
// ep_contamination_rejection_period.

#ifndef COMPONENTS_SPEECH_ENDPOINTER_ENERGY_ENDPOINTER_H_
#define COMPONENTS_SPEECH_ENDPOINTER_ENERGY_ENDPOINTER_H_

#include <stdint.h>

#include <memory>
#include <vector>

#include "components/speech/endpointer/energy_endpointer_params.h"

namespace speech {

// Endpointer status codes
enum EpStatus { … };

class EnergyEndpointer { … };

}  // namespace speech

#endif  // COMPONENTS_SPEECH_ENDPOINTER_ENERGY_ENDPOINTER_H_