ArSpeechSynth Class Reference

#include <ArSpeech.h>

List of all members.

---

Detailed Description

Abstract interface to speech synthesis.

This class defines the abstract interface for speech synthesizers used with Aria. Implementations are provided in the separate ArSpeechSynth_Cepstral and ArSpeechSynth_Festival libraries. This class provides a common-denominator interface. Implementations (especially ArCepstral) may support more features, or behave differently; refer to their documentation for more information.

This class registers one parameter with the global ArConfig object in the "Speech Synthesis" section: The parameter is named "voice" and sets the voice used for speech synthesis, if multiple voices is supported and the new voice can be loaded.

Public Member Functions
virtual void	addToConfig (ArConfig *config)
	ArSpeechSynth ()
virtual int	getAudioSampleRate ()=0
virtual const char *	getCurrentVoiceName ()=0
ArRetFunctorC< bool, ArSpeechSynth > *	getInitCallback ()
ArFunctorC< ArSpeechSynth > *	getInterruptCallback ()
ArRetFunctor2C< bool, ArSpeechSynth, const char *, const char * > *	getSpeakCallback (void)
virtual std::list< std::string >	getVoiceNames ()=0
virtual bool	init ()
virtual void	interrupt ()=0
virtual void	lock ()
void	setAudioCallback (ArRetFunctor2< bool, ArTypes::Byte2 *, int > *cb)
virtual void	setAudioSampleRate (int rate)=0
virtual bool	setVoice (const char *name)=0
virtual bool	speak (const char *str, const char *voiceParams=NULL)
virtual bool	speak (const char *str, const char *voiceParams, ArRetFunctor2< bool, ArTypes::Byte2 *, int > *audioOutputCB, unsigned short sampleRate=0)=0
virtual bool	speakf (const char *fmt,...)=0
virtual void	unlock ()
Protected Attributes
ArRetFunctor2< bool, ArTypes::Byte2 *, int > *	myAudioPlaybackCB
	If set, send audio to this callback instead of playing it directly.
ArRetFunctorC< bool, ArSpeechSynth >	myInitCB
ArFunctorC< ArSpeechSynth >	myInterruptCB
ArRetFunctor2C< bool, ArSpeechSynth, const char *, const char * >	mySpeakCB

---

Constructor & Destructor Documentation

ArSpeechSynth::ArSpeechSynth (   )

Don't forget to call this from derived classes.

---

Member Function Documentation

void ArSpeechSynth::addToConfig (  ArConfig *  config  )  [virtual]

Use the given ArConfig object to read parameters such as voice, speaking rate, volume.

virtual const char* ArSpeechSynth::getCurrentVoiceName (   )  [pure virtual]

Get name of current voice, if set with setVoice (else returns NULL)

ArRetFunctorC< bool, ArSpeechSynth > * ArSpeechSynth::getInitCallback (   )

Returns: a functor for init() to use with ArSoundsQueue

ArFunctorC< ArSpeechSynth > * ArSpeechSynth::getInterruptCallback (   )

Returns: a functor for interrupt()

ArRetFunctor2C< bool, ArSpeechSynth, const char *, const char * > * ArSpeechSynth::getSpeakCallback (  void   )

Returns: a functor for speak() to use with ArSoundsQueue

virtual std::list<std::string> ArSpeechSynth::getVoiceNames (   )  [pure virtual]

Return a list of available voice names, if possible.

bool ArSpeechSynth::init (   )  [virtual]

Perform synthesizer initialization, if necessary. You must call this method. (Subclass implementations should call this method *after* initializing their speech engine.)

virtual void ArSpeechSynth::interrupt (   )  [pure virtual]

If any speech is currently ongoing, interrupt it.

virtual void ArSpeechSynth::lock (  void   )  [inline, virtual]

Lock, if neccesary

void ArSpeechSynth::setAudioCallback (  ArRetFunctor2< bool, ArTypes::Byte2 *, int > *  cb  )

Instead of playing synthesized audio using the synthesizer's internal audio playback, call the given callback when a chunk of audio has been synthesized. Audio is passed to the callback in the first parameter as signed 16-bit samples (PCM16). The sample rate is 16kHz but may be changed with setAudioSampleRate(). The second parameter is the number of samples. The return value from the callback is ignored.

virtual void ArSpeechSynth::setAudioSampleRate (  int  rate  )  [pure virtual]

Change audio sample rate (Hz). Normal rate is 16000 Hz. Suggested values are 8000, 16000, or 44400

virtual bool ArSpeechSynth::setVoice (  const char *  name  )  [pure virtual]

Set the current voice by name. Replaces fully any previous required voice criteria. See also: getVoiceNames

bool ArSpeechSynth::speak (  const char *  str, const char *  voiceParams = NULL )  [virtual]

Speaks the given text. Parameters: str  The text to speak. voiceParams  Voice selection criteria expression (implementation-specific)

virtual bool ArSpeechSynth::speak (  const char *  str, const char *  voiceParams, ArRetFunctor2< bool, ArTypes::Byte2 *, int > *  audioOutputCB, unsigned short  sampleRate = 0 )  [pure virtual]

Speaks the given text. Parameters: str  The text to speak. voiceParams  Voice selection criteria expression (implementation-specific) audioOutputCB  If not NULL, send synthesized audio data to this callback (may be called several times). Otherwise, play using default AudioCallback, or directly out the speakers sampleRate  if given, temporarily use this sample rate for this speech, then restore. If 0, use current sample rate.

virtual bool ArSpeechSynth::speakf (  const char *  fmt,   ... )  [pure virtual]

Speaks the given string, using current voice and output settings, taking varargs and a format string (like printf)

virtual void ArSpeechSynth::unlock (  void   )  [inline, virtual]

Unlock, if neccesary

---

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

• ArSpeech.h
• ArSpeech.cpp

---