Interrupt-driven ESpeak-NG instance. Generates a full frame of samples each cycle and uses the RawBuffer to safely hand data from the app to the decompressor. More...

#include <BackgroundAudioSpeech.h>

Public Member Functions
	BackgroundAudioSpeechClass (AudioOutputBase &d)
	Construct an output device using the specified physical audio output.

bool	setDevice (AudioOutputBase *d)
	Set an output device before `begin`

void	setGain (float scale)
	Set the gain multiplier (volume) for the stream. Takes effect immediately.

void	setVoice (BackgroundAudioVoice &v)
	Sets the voice parameters (language customization)

void	setRate (int rate)
	Set the speaking rate in ~wpm, after calling begin()

void	setPitch (int pitch)
	Adjust the pitch, 0...99, with 50 default. After begin()

void	setWordGap (int gap)
	Adjust the interword gap after begin()

bool	begin ()
	Starts the background speaker. Will initialize the output device and start sending silence immediately.

void	end ()
	Stops the process and the calls the output device's end to shut it down, too.

bool	playing ()
	Determines if the speaker has been started.

size_t	write (const void *data, size_t len)
	Writes a block of raw data to the decoder's buffer.

size_t	speak (const char *string)
	Speaks a C-String.

size_t	speak (const String &string)
	Speaks an Arduino String.

size_t	availableForWrite ()
	Gets number of bytes available to write to raw buffer.

size_t	available ()
	Gets number of bytes already in the raw buffer.

bool	done ()
	Determine if no more speech is present in the buffer.

uint32_t	frames ()
	Get number of "frames" processed by speaker.

uint32_t	shifts ()
	Get the number of input data shifts processed by decoder since `begin`

uint32_t	underflows ()
	Get the number of times the speaker has underflowed waiting on raw data since `begin`

uint32_t	errors ()
	Get the number of decoder errors since `begin`

uint32_t	dumps ()
	Get the number of full buffer dumps (catastrophic data error) since `begin`

void	pause ()
	Pause the decoder. Won't process raw input data and will transmit silence.

bool	paused ()
	Determine if the playback is paused.

void	unpause ()
	Unpause previously paused playback. Will start processing input data again.

void	flush ()
	Flushes any existing raw data, resets the processor to start a new speaking.

Detailed Description

template<class DataBuffer>
class BackgroundAudioSpeechClass< DataBuffer >

Interrupt-driven ESpeak-NG instance. Generates a full frame of samples each cycle and uses the RawBuffer to safely hand data from the app to the decompressor.

Template Parameters

DataBuffer The type of raw data buffer, either normal RawDataBuffer or ROMDataBuffer as appropriate.

Constructor & Destructor Documentation

◆ BackgroundAudioSpeechClass()

template<class DataBuffer >

BackgroundAudioSpeechClass< DataBuffer >::BackgroundAudioSpeechClass ( AudioOutputBase & d )

inline

Construct an output device using the specified physical audio output.

Parameters

[in] d AudioOutputBase device to send data to (or AudioMixerInput)

Member Function Documentation

◆ available()

template<class DataBuffer >

size_t BackgroundAudioSpeechClass< DataBuffer >::available ( )

inline

Gets number of bytes already in the raw buffer.

Returns: Bytes of raw data in the buffer

◆ availableForWrite()

template<class DataBuffer >

size_t BackgroundAudioSpeechClass< DataBuffer >::availableForWrite ( )

inline

Gets number of bytes available to write to raw buffer.

For ROM buffers, this is always 0 after a write has happened. Because ROM buffers don't actually allocate any RAM for buffering and simply store the pointer passed in from write there is no real space available for an app to write to. An app can simply flush and write a new complete buffer if needed ignoring availableForWrite in the ROM case.

Returns: Bytes that can be written

◆ begin()

template<class DataBuffer >

bool BackgroundAudioSpeechClass< DataBuffer >::begin ( )

inline

Starts the background speaker. Will initialize the output device and start sending silence immediately.

Returns: True on success, false if already started.

◆ done()

template<class DataBuffer >

bool BackgroundAudioSpeechClass< DataBuffer >::done ( )

inline

Determine if no more speech is present in the buffer.

Returns: True if no raw strings still left to process

◆ dumps()

template<class DataBuffer >

uint32_t BackgroundAudioSpeechClass< DataBuffer >::dumps ( )

inline

Get the number of full buffer dumps (catastrophic data error) since begin

Returns: Number of complete buffer throw-aways

◆ errors()

template<class DataBuffer >

uint32_t BackgroundAudioSpeechClass< DataBuffer >::errors ( )

inline

Get the number of decoder errors since begin

Returns: Number of decode errors encountered

◆ flush()

template<class DataBuffer >

void BackgroundAudioSpeechClass< DataBuffer >::flush ( )

inline

Flushes any existing raw data, resets the processor to start a new speaking.

This is only needed to abort a currently speaking phrase. Multiple strings can be sent in one after the other with write or speak

◆ frames()

template<class DataBuffer >

uint32_t BackgroundAudioSpeechClass< DataBuffer >::frames ( )

inline

Get number of "frames" processed by speaker.

Returns: Number of frames, where frames are framelen stereo samples in size

◆ paused()

template<class DataBuffer >

bool BackgroundAudioSpeechClass< DataBuffer >::paused ( )

inline

Determine if the playback is paused.

Returns: True of WAV playback has been paused

◆ playing()

template<class DataBuffer >

bool BackgroundAudioSpeechClass< DataBuffer >::playing ( )

inline

Determines if the speaker has been started.

Returns: True if running

◆ setDevice()

template<class DataBuffer >

bool BackgroundAudioSpeechClass< DataBuffer >::setDevice ( AudioOutputBase * d )

inline

Set an output device before begin

Parameters

[in] d Pointer to AudioOutputDevice to send decoded data to

Returns: True on success

◆ setGain()

template<class DataBuffer >

void BackgroundAudioSpeechClass< DataBuffer >::setGain ( float scale )

inline

Set the gain multiplier (volume) for the stream. Takes effect immediately.

Parameters

[in] scale Floating point value from 0.0....16.0 to multiply all audio data by

◆ setPitch()

template<class DataBuffer >

void BackgroundAudioSpeechClass< DataBuffer >::setPitch ( int pitch )

inline

Adjust the pitch, 0...99, with 50 default. After begin()

Parameters

[in] pitch Pitch setting

◆ setRate()

template<class DataBuffer >

void BackgroundAudioSpeechClass< DataBuffer >::setRate ( int rate )

inline

Set the speaking rate in ~wpm, after calling begin()

Parameters

[in] rate WPM rate

◆ setVoice()

template<class DataBuffer >

void BackgroundAudioSpeechClass< DataBuffer >::setVoice ( BackgroundAudioVoice & v )

inline

Sets the voice parameters (language customization)

Parameters

[in] v Voice (voice_xxx) included from in libespeak-ng/dict folder

◆ setWordGap()

template<class DataBuffer >

void BackgroundAudioSpeechClass< DataBuffer >::setWordGap ( int gap )

inline

Adjust the interword gap after begin()

Parameters

[in] gap Gap in ~10ms units at normal rate

◆ shifts()

template<class DataBuffer >

uint32_t BackgroundAudioSpeechClass< DataBuffer >::shifts ( )

inline

Get the number of input data shifts processed by decoder since begin

Returns: Number of times data has been shifted in the raw input buffer

◆ speak() [1/2]

template<class DataBuffer >

size_t BackgroundAudioSpeechClass< DataBuffer >::speak ( const char * string )

inline

Speaks a C-String.

Parameters

[in] string The string to speak. Data is copied so this string can disappear afterwards

Returns: Number of bytes actually written to the buffer, or 0 on error (out of space)

◆ speak() [2/2]

template<class DataBuffer >

size_t BackgroundAudioSpeechClass< DataBuffer >::speak ( const String & string )

inline

Speaks an Arduino String.

Parameters

[in] string The string to speak. Data is copied so this string can disappear afterwards

Returns: Number of bytes actually written to the buffer, or 0 on error (out of space)

◆ underflows()

template<class DataBuffer >

uint32_t BackgroundAudioSpeechClass< DataBuffer >::underflows ( )

inline

Get the number of times the speaker has underflowed waiting on raw data since begin

Returns: Number of frames of underflow data have occurred

◆ write()

template<class DataBuffer >

size_t BackgroundAudioSpeechClass< DataBuffer >::write	(	const void *	data,
		size_t	len
	)

inline

Writes a block of raw data to the decoder's buffer.

Copies up to len bytes into the raw buffer for the object. Can be called before begin, and can write less fewer than the number of bytes requested if there is not enough space. Will not block.

For ROM buffers this does not actually copy data, only the pointer. Therefore, for ROM buffers you cannot concatenate data by calling multiple writes because the second write will throw out all the data from the first one. Use flush and write when ready for a completely new buffer.

Parameters

[in]	data	Uncompressed input data
[in]	len	Number of bytes to write

Returns: Number of bytes written

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

BackgroundAudioSpeech.h

Public Member Functions

Detailed Description

Constructor & Destructor Documentation

◆ BackgroundAudioSpeechClass()

Member Function Documentation

◆ available()

◆ availableForWrite()

◆ begin()

◆ done()

◆ dumps()

◆ errors()

◆ flush()

◆ frames()

◆ paused()

◆ playing()

◆ setDevice()

◆ setGain()

◆ setPitch()

◆ setRate()

◆ setVoice()

◆ setWordGap()

◆ shifts()

◆ speak() [1/2]

◆ speak() [2/2]

◆ underflows()

◆ write()