WakeWordDetector< T, N > Class Template Reference

Template-based wake word detector for microcontrollers using dominant frequency patterns. More...

#include <WakeWordDetector.h>

Inheritance diagram for WakeWordDetector< T, N >:

Classes
struct	Template

Public Types
using	WakeWordCallback = void(*)(const char *name)

Public Member Functions
	WakeWordDetector (AudioFFTBase &fft)
virtual void	addNotifyAudioChange (AudioInfoSupport &bi)
	Adds target to be notified about audio changes.
void	addTemplate (const Vector< FrequencyFrame< N > > &frames, float threshold_percent, const char *name)
virtual AudioInfo	audioInfo () override
	provides the actual input AudioInfo
virtual AudioInfo	audioInfoOut ()
virtual int	availableForWrite () override
virtual bool	begin ()
virtual bool	begin (AudioInfo info)
virtual void	clearNotifyAudioChange ()
	Deletes all change notify subscriptions.
virtual void	end ()
virtual void	flush ()
virtual bool	isDeletable ()
	If true we need to release the related memory in the destructor.
bool	isNotifyActive ()
	Checks if the automatic AudioInfo update is active.
bool	isRecording () const
virtual	operator bool ()
virtual bool	removeNotifyAudioChange (AudioInfoSupport &bi)
	Removes a target in order not to be notified about audio changes.
virtual void	setAudioInfo (AudioInfo newInfo) override
	Defines the input AudioInfo.
void	setNotifyActive (bool flag)
	Deactivate/Reactivate automatic AudioInfo updates: (default is active)
void	setWakeWordCallback (WakeWordCallback cb)
void	startRecording ()
Vector< FrequencyFrame< N > >	stopRecording ()
size_t	write (const uint8_t *buf, size_t size) override
virtual size_t	write (uint8_t ch) override
virtual void	writeSilence (size_t len)

Static Public Member Functions
static void	fftResult (AudioFFTBase &fft)

Protected Member Functions
float	matchTemplate (Template &tmpl)
void	notifyAudioChange (AudioInfo info)

Protected Attributes
Vector< T >	_buffer
	Buffer for incoming PCM samples.
WakeWordCallback	_callback = nullptr
size_t	_frame_pos
	Current position in frame buffer.
bool	_is_recording = false
	True if currently recording a template.
size_t	_max_template_len = 0
	Length of the longest template.
Vector< FrequencyFrame< N > >	_recent_frames
	Recent frames for comparison.
Vector< Template >	_templates
	List of wake word templates.
int	_timeout = 10
AudioInfo	cfg
bool	is_active = false
bool	is_notify_active = true
Vector< AudioInfoSupport * >	notify_vector
AudioFFTBase *	p_fft = nullptr
SingleBuffer< uint8_t >	tmp {MAX_SINGLE_CHARS}
int	tmpPos = 0

Detailed Description

template<typename T = int16_t, size_t N = 3>
class audio_tools::WakeWordDetector< T, N >

Template-based wake word detector for microcontrollers using dominant frequency patterns.

This class detects wake words by comparing the sequence of the top N dominant frequencies in each audio frame to stored templates for each wake word. When the percentage of matching frames exceeds a configurable threshold, the corresponding wake word is considered detected.

Template Parameters

N	Number of dominant frequencies to track per frame (default: 3)

Usage:

• Record each wake word and extract the top N frequencies for each frame to build templates.
• Instantiate WakeWordDetector<N> and add templates for each wake word.
• Register a callback to handle detection events using setWakeWordCallback().

Example:

audio_tools::WakeWordDetector<3> detector(fft);
detector.addTemplate(my_template_frames, 80.0f, "hello");
detector.setWakeWordCallback([](const char* name) { Serial.println(name); });
// ...

Member Typedef Documentation

WakeWordCallback

template<typename T = int16_t, size_t N = 3>

using WakeWordCallback = void (*)(const char* name)

Constructor & Destructor Documentation

WakeWordDetector()

template<typename T = int16_t, size_t N = 3>

WakeWordDetector ( AudioFFTBase &  fft )

inline

Member Function Documentation

addNotifyAudioChange()

virtual void addNotifyAudioChange ( AudioInfoSupport &  bi )

inlinevirtualinherited

Adds target to be notified about audio changes.

Reimplemented in CodecNOP, EncodedAudioOutput, EncodedAudioStream, AACDecoderFDK, DecoderBasic, CodecChain, MP3DecoderHelix, MP3DecoderMAD, OggContainerDecoder, RTSPClient< TcpClient, UdpSocket >, Pipeline, and Pipeline::ModifyingStreamAdapter.

addTemplate()

template<typename T = int16_t, size_t N = 3>

void addTemplate ( const Vector< FrequencyFrame< N > > &  frames, float  threshold_percent, const char *  name  )

inline

audioInfo()

virtual AudioInfo audioInfo ( )

inlineoverridevirtualinherited

provides the actual input AudioInfo

Implements AudioInfoSupport.

Reimplemented in AdapterPrintToAudioOutput, EncodedAudioOutput, and AdapterAudioStreamToAudioOutput.

audioInfoOut()

virtual AudioInfo audioInfoOut ( )

inlinevirtualinherited

provides the actual output AudioInfo: this is usually the same as audioInfo() unless we use a transforming stream

Reimplemented in PureDataStream, PWMAudioOutput, ChannelFormatConverterStreamT< T >, ChannelFormatConverterStream, NumberFormatConverterStreamT< TFrom, TTo >, NumberFormatConverterStream, FormatConverterStream, Pipeline, ResampleStream, and ResampleStreamT< TInterpolator >.

availableForWrite()

virtual int availableForWrite ( )

inlineoverridevirtualinherited

Reimplemented from Print.

Reimplemented in AudioESP32ULP, RTSPOutput< Platform >, EncodedAudioOutput, AudioServerEx, AudioSyncWriter, AdapterAudioStreamToAudioOutput, CsvOutput< T >, MemoryOutput, and PWMAudioOutput.

begin() [1/2]

virtual bool begin ( )

inlinevirtualinherited

Reimplemented in OSCContainerEncoder::OSCOutput, AudioOutputWithCallback, AudioServerEx, RTSPOutput< Platform >, PitchShiftOutput< T, BufferT >, PWMAudioOutput, RTTTLOutput< T >, EncodedAudioOutput, OggContainerOutput, FFTEffect, FFTPitchShift, R2ROutput, AdapterAudioStreamToAudioOutput, ChannelsSelectOutput, MetaDataFilter, CsvOutput< T >, HexDumpOutput, MemoryOutput, and RTTTLOutput< T >.

begin() [2/2]

virtual bool begin ( AudioInfo  info )

inlinevirtualinherited

Reimplemented in AudioESP32ULP, RTSPOutput< Platform >, RTTTLOutput< T >, ChannelsSelectOutput, CsvOutput< T >, and EncodedAudioOutput.

clearNotifyAudioChange()

virtual void clearNotifyAudioChange ( )

inlinevirtualinherited

Deletes all change notify subscriptions.

Reimplemented in RTSPClient< TcpClient, UdpSocket >.

end()

virtual void end ( )

inlinevirtualinherited

Reimplemented in AudioESP32ULP, AudioServerEx, AudioSyncWriter, RTSPOutput< Platform >, PitchShiftOutput< T, BufferT >, MetaDataOutput, EncodedAudioOutput, OggContainerOutput, AdapterAudioStreamToAudioOutput, MetaDataFilter, and PWMAudioOutput.

fftResult()

template<typename T = int16_t, size_t N = 3>

static void fftResult ( AudioFFTBase &  fft )

inlinestatic

flush()

virtual void flush ( )

inlinevirtualinherited

Reimplemented from Print.

Reimplemented in MultiOutput, and HexDumpOutput.

isDeletable()

virtual bool isDeletable ( )

inlinevirtualinherited

If true we need to release the related memory in the destructor.

Reimplemented in AdapterPrintToAudioOutput, and AdapterAudioStreamToAudioOutput.

isNotifyActive()

bool isNotifyActive ( )

inlineinherited

Checks if the automatic AudioInfo update is active.

isRecording()

template<typename T = int16_t, size_t N = 3>

bool isRecording ( ) const

inline

matchTemplate()

template<typename T = int16_t, size_t N = 3>

float matchTemplate ( Template &  tmpl )

inlineprotected

notifyAudioChange()

void notifyAudioChange ( AudioInfo  info )

inlineprotectedinherited

operator bool()

virtual operator bool ( )

inlinevirtualinherited

Reimplemented in RTSPOutput< Platform >, EncodedAudioOutput, and AdapterAudioStreamToAudioOutput.

removeNotifyAudioChange()

virtual bool removeNotifyAudioChange ( AudioInfoSupport &  bi )

inlinevirtualinherited

Removes a target in order not to be notified about audio changes.

Reimplemented in RTSPClient< TcpClient, UdpSocket >.

setAudioInfo()

virtual void setAudioInfo ( AudioInfo  info )

inlineoverridevirtualinherited

Defines the input AudioInfo.

Implements AudioInfoSupport.

Reimplemented in ChannelsSelectOutput, AdapterPrintToAudioOutput, MultiOutput, AdapterAudioStreamToAudioOutput, CsvOutput< T >, PWMAudioOutput, and EncodedAudioOutput.

setNotifyActive()

void setNotifyActive ( bool  flag )

inlineinherited

Deactivate/Reactivate automatic AudioInfo updates: (default is active)

setWakeWordCallback()

template<typename T = int16_t, size_t N = 3>

void setWakeWordCallback ( WakeWordCallback  cb )

inline

startRecording()

template<typename T = int16_t, size_t N = 3>

void startRecording ( )

inline

stopRecording()

template<typename T = int16_t, size_t N = 3>

Vector< FrequencyFrame< N > > stopRecording ( )

inline

write() [1/2]

template<typename T = int16_t, size_t N = 3>

size_t write ( const uint8_t *  buf, size_t  size  )

inlineoverridevirtual

Implements AudioOutput.

write() [2/2]

virtual size_t write ( uint8_t  ch )

inlineoverridevirtualinherited

Reimplemented in MetaDataOutput, and MultiOutput.

writeSilence()

virtual void writeSilence ( size_t  len )

inlinevirtualinherited

Writes n 0 values (= silence)

Parameters

len

Member Data Documentation

_buffer

template<typename T = int16_t, size_t N = 3>

Vector<T> _buffer

protected

Buffer for incoming PCM samples.

_callback

template<typename T = int16_t, size_t N = 3>

WakeWordCallback _callback = nullptr

protected

_frame_pos

template<typename T = int16_t, size_t N = 3>

size_t _frame_pos

protected

Current position in frame buffer.

_is_recording

template<typename T = int16_t, size_t N = 3>

bool _is_recording = false

protected

True if currently recording a template.

_max_template_len

template<typename T = int16_t, size_t N = 3>

size_t _max_template_len = 0

protected

Length of the longest template.

_recent_frames

template<typename T = int16_t, size_t N = 3>

Vector<FrequencyFrame<N> > _recent_frames

protected

Recent frames for comparison.

_templates

template<typename T = int16_t, size_t N = 3>

Vector<Template> _templates

protected

List of wake word templates.

_timeout

int _timeout = 10

protectedinherited

cfg

AudioInfo cfg

protectedinherited

is_active

bool is_active = false

protectedinherited

is_notify_active

bool is_notify_active = true

protectedinherited

notify_vector

Vector<AudioInfoSupport*> notify_vector

protectedinherited

p_fft

template<typename T = int16_t, size_t N = 3>

AudioFFTBase* p_fft = nullptr

protected

tmp

SingleBuffer<uint8_t> tmp {MAX_SINGLE_CHARS}

protectedinherited

tmpPos

int tmpPos = 0

protectedinherited

---

The documentation for this class was generated from the following file:

• src/AudioTools/Sandbox/STT/WakeWordDetector.h