io.spokestack.spokestack

Class SpeechPipeline

java.lang.Object

io.spokestack.spokestack.SpeechPipeline

All Implemented Interfaces:

AutoCloseable

public final class SpeechPipeline
extends Object
implements AutoCloseable

Spokestack speech pipeline.

This is the primary client entry point to the Spokestack framework. It dynamically binds to configured components that implement the pipeline interfaces for reading audio frames and performing speech recognition tasks.

Clients initialize the pipeline using the SpeechPipeline.Builder class. The following is an example pipeline that reads the android microphone and performs voice activity detection (VAD).

 
  SpeechPipeline pipeline = new SpeechPipeline.Builder()
      .setInputClass("io.spokestack.spokestack.android.MicrophoneInput")
      .addStageClass("io.spokestack.spokestack.webrtc.VoiceActivityTrigger")
      .addStageClass("io.spokestack.spokestack.ActivationTimeout")
      .setProperty("sample-rate", 16000)
      .setProperty("frame-width", 20)
      .setProperty("buffer-width", 300)
      .addOnSpeechEventListener(this)
      .build();
  pipeline.start();
 
 

The pipeline may be stopped/restarted any number of times during its lifecycle. While stopped, the pipeline consumes as few resources as possible. The pipeline runs asynchronously on a dedicated thread, so it does not block the caller to perform I/O or speech processing.

When running, the pipeline communicates with the client via the event interface on the speech context. All calls to event handlers are made in the context of the pipeline's thread, so event handlers should not perform blocking operations, and should use message passing when communicating with UI components, etc.

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	SpeechPipeline.Builder speech pipeline builder.

Field Summary

Fields

Modifier and Type	Field and Description
static int	DEFAULT_BUFFER_WIDTH audio frame buffer width, in ms.
static int	DEFAULT_FRAME_WIDTH audio frame width, in ms.
static int	DEFAULT_SAMPLE_RATE audio sample rate default, in samples/sec.

Method Summary

Modifier and Type	Method and Description
void	activate() manually activate the speech pipeline.
void	addListener(OnSpeechEventListener listener) Add a new listener to receive events from the speech pipeline.
void	close() shuts down the pipeline and releases its resources.
void	deactivate() manually deactivate the speech pipeline.
SpeechConfig	getConfig()
SpeechContext	getContext()
boolean	isPaused()
boolean	isRunning()
void	pause() Pauses the speech pipeline, temporarily stopping passive listening.
void	removeListener(OnSpeechEventListener listener) Remove a pipeline listener, allowing it to be garbage collected.
void	resume() Resumes a paused speech pipeline, returning the pipeline to a passive listening state.
void	start() starts up the speech pipeline.
void	stop() stops the speech pipeline and releases all resources.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

DEFAULT_SAMPLE_RATE

public static final int DEFAULT_SAMPLE_RATE

audio sample rate default, in samples/sec.

See Also:

Constant Field Values

DEFAULT_FRAME_WIDTH

public static final int DEFAULT_FRAME_WIDTH

audio frame width, in ms.

See Also:

Constant Field Values

DEFAULT_BUFFER_WIDTH

public static final int DEFAULT_BUFFER_WIDTH

audio frame buffer width, in ms.

See Also:

Constant Field Values

Method Detail

close

public void close()

shuts down the pipeline and releases its resources.

Specified by:

close in interface AutoCloseable

getConfig

public SpeechConfig getConfig()

Returns:

current pipeline configuration.

getContext

public SpeechContext getContext()

Returns:

current pipeline configuration.

isRunning

public boolean isRunning()

Returns:

true if the pipeline has been started and is not paused, false otherwise.

isPaused

public boolean isPaused()

Returns:

true if the pipeline has been paused, false otherwise.

activate

public void activate()

manually activate the speech pipeline.

deactivate

public void deactivate()

manually deactivate the speech pipeline.

addListener

public void addListener(OnSpeechEventListener listener)

Add a new listener to receive events from the speech pipeline.

Parameters:

listener - The listener to add.

removeListener

public void removeListener(OnSpeechEventListener listener)

Remove a pipeline listener, allowing it to be garbage collected.

Parameters:

listener - The listener to remove.

start

public void start()
           throws Exception

starts up the speech pipeline.

Throws:

Exception - on configuration/startup error

pause

public void pause()

Pauses the speech pipeline, temporarily stopping passive listening.

Note that active listening (an active ASR stage) cannot be paused, so the pipeline is deactivated before it is paused. This may prevent the delivery of a SpeechContext.Event.RECOGNIZE event if an ASR request is currently in progress.

While paused, the pipeline will not respond to the wakeword, but in order to support a quick resume(), it will retain control of the microphone. No audio is explicitly read or analyzed. To fully release the pipeline's resources, see stop().

resume

public void resume()

Resumes a paused speech pipeline, returning the pipeline to a passive listening state.

stop

public void stop()

stops the speech pipeline and releases all resources.