osmscout::PiperTTSEngine Class Reference

Piper (libpiper) based text-to-speech engine. More...

#include <PiperTTSEngine.h>

Inheritance diagram for osmscout::PiperTTSEngine:

Public Member Functions
Voice	getVoice () override
void	initVoice (const Voice &voice) override
	Initialize the engine for the given voice.
	PiperTTSEngine (const QString &espeakDataDir)
void	playMessage (QString message) override
	Prepare the message (when it is not cached yet) and play it right away.
void	prepareMessage (QString message) override
	Synthesize the message to an audio file (cached in a temporary directory) without playing it.
	~PiperTTSEngine () override
Public Member Functions inherited from osmscout::TTSEngine
TTSEngine &	operator= (const TTSEngine &)=delete
TTSEngine &	operator= (TTSEngine &&)=delete
	TTSEngine ()
	TTSEngine (const TTSEngine &)=delete
	TTSEngine (TTSEngine &&)=delete
	~TTSEngine () override

Protected Member Functions
void	cleanCache () override
	Remove the oldest synthesized WAV files when the cache directory exceeds the size limit.

Additional Inherited Members
Public Slots inherited from osmscout::TTSEngine
Signals inherited from osmscout::TTSEngine
void	playAudioFilesRequest (const QList< QUrl > &audioFiles)
Protected Attributes inherited from osmscout::TTSEngine
QThread *	thread
	engine background thread (owns itself, deleted on finish)

Detailed Description

Piper (libpiper) based text-to-speech engine.

Synthesis is performed with the (synchronous) Piper API on the engine background thread. Each message is rendered into a WAV file placed in the per-application cache directory (e.g. ~/.cache/... on Unix) and cached, so repeated messages are synthesized only once. Playback is delegated to the VoiceCorePlayer.

Support for Piper is optional; this class is only compiled and available when the library was built with libpiper (see the CMake / Meson build files). When libpiper is missing, NavigationModule simply keeps its TTS engine unset.

Constructor & Destructor Documentation

PiperTTSEngine()

osmscout::PiperTTSEngine::PiperTTSEngine ( const QString &  espeakDataDir )

explicit

Parameters

player	audio player used to play synthesized messages.
espeakDataDir	path to the espeak-ng data directory required by Piper.

~PiperTTSEngine()

osmscout::PiperTTSEngine::~PiperTTSEngine ( )

override

Member Function Documentation

cleanCache()

void osmscout::PiperTTSEngine::cleanCache ( )

overrideprotectedvirtual

Remove the oldest synthesized WAV files when the cache directory exceeds the size limit.

Called on the engine thread when the engine thread starts.

Reimplemented from osmscout::TTSEngine.

References osmscout::Log::Debug(), osmscout::log, and osmscout::Log::Warn().

getVoice()

Voice osmscout::PiperTTSEngine::getVoice ( )

overridevirtual

Implements osmscout::TTSEngine.

initVoice()

void osmscout::PiperTTSEngine::initVoice ( const Voice &  voice )

overridevirtual

Initialize the engine for the given voice.

Runs asynchronously on the engine thread.

Implements osmscout::TTSEngine.

References osmscout::AppendFileToDir(), osmscout::Log::Debug(), osmscout::Log::Error(), osmscout::ExistsInFilesystem(), osmscout::Voice::getDir(), osmscout::Voice::getModelFile(), osmscout::Voice::getName(), osmscout::Voice::isPiper(), osmscout::Voice::isValid(), osmscout::log, and osmscout::Log::Warn().

playMessage()

void osmscout::PiperTTSEngine::playMessage ( QString  message )

overridevirtual

Prepare the message (when it is not cached yet) and play it right away.

Runs asynchronously on the engine thread.

Implements osmscout::TTSEngine.

References osmscout::TTSEngine::playAudioFilesRequest().

prepareMessage()

void osmscout::PiperTTSEngine::prepareMessage ( QString  message )

overridevirtual

Synthesize the message to an audio file (cached in a temporary directory) without playing it.

Runs asynchronously on the engine thread.

Implements osmscout::TTSEngine.

---

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

• PiperTTSEngine.h
• PiperTTSEngine.cpp