edu.cmu.sphinx.linguist.acoustic.tiedstate

Class ScoreCachingSenone

java.lang.Object

edu.cmu.sphinx.linguist.acoustic.tiedstate.ScoreCachingSenone

All Implemented Interfaces:

Senone, java.io.Serializable

Direct Known Subclasses:

CompositeSenone, DiagGmm, GaussianMixture

public abstract class ScoreCachingSenone
extends java.lang.Object
implements Senone

Implements a Senone that contains a cache of the last scored data.

Subclasses should implement the abstract calculateScore(edu.cmu.sphinx.frontend.Data) method, which is called by the getScore(edu.cmu.sphinx.frontend.Data) method to calculate the score for each cache miss.

Note: this implementation is thread-safe and can be safely used across different threads without external synchronization.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
ScoreCachingSenone()

Method Summary

Modifier and Type	Method and Description
protected abstract float	calculateScore(Data feature) Calculates the score for this senone based upon the given feature.
float	getScore(Data feature) Gets the cached score for this senone based upon the given feature.

Methods inherited from class java.lang.Object

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

Methods inherited from interface edu.cmu.sphinx.linguist.acoustic.tiedstate.Senone

calculateComponentScore, dump, getID, getLogMixtureWeights, getMixtureComponents

Constructor Detail

ScoreCachingSenone

public ScoreCachingSenone()

Method Detail

getScore

public float getScore(Data feature)

Gets the cached score for this senone based upon the given feature. If the score was not cached, it is calculated using calculateScore(edu.cmu.sphinx.frontend.Data), cached, and then returned.

Specified by:

getScore in interface Senone

Parameters:

feature - the feature vector to score this senone against

Returns:

the score for this senone in LogMath log base

calculateScore

protected abstract float calculateScore(Data feature)

Calculates the score for this senone based upon the given feature.

Parameters:

feature - the feature vector to score this senone against

Returns:

the score for this senone in LogMath log base