org.vosk

Class Recognizer

java.lang.Object

com.sun.jna.PointerType

org.vosk.Recognizer

All Implemented Interfaces:

com.sun.jna.NativeMapped, java.lang.AutoCloseable

public class Recognizer
extends com.sun.jna.PointerType
implements java.lang.AutoCloseable

Constructor Summary

Constructors

Constructor and Description
Recognizer(Model model, float sampleRate) Creates the recognizer object.
Recognizer(Model model, float sampleRate, SpeakerModel spkModel) Creates the recognizer object with speaker recognition.
Recognizer(Model model, float sampleRate, java.lang.String grammar) Creates the recognizer object with the phrase list.

Method Summary

Modifier and Type	Method and Description
boolean	acceptWaveForm(byte[] data, int len) Accept and process new chunk of voice data.
boolean	acceptWaveForm(float[] data, int len)
boolean	acceptWaveForm(short[] data, int len)
void	close() Releases recognizer object.
java.lang.String	getFinalResult() Returns speech recognition result.
java.lang.String	getPartialResult() Returns partial speech recognition.
java.lang.String	getResult() Returns speech recognition result
void	reset() Resets the recognizer.
void	setGrammar(java.lang.String grammar) Reconfigures recognizer to use grammar.
void	setMaxAlternatives(int maxAlternatives) Configures recognizer to output n-best results.
void	setPartialWords(boolean partial_words) Like above return words and confidences in partial results.
void	setSpeakerModel(SpeakerModel spkModel) Adds speaker model to already initialized recognizer.
void	setWords(boolean words) Enables words with times in the output

Methods inherited from class com.sun.jna.PointerType

equals, fromNative, getPointer, hashCode, nativeType, setPointer, toNative, toString

Methods inherited from class java.lang.Object

clone, finalize, getClass, notify, notifyAll, wait, wait, wait

Constructor Detail

Recognizer

public Recognizer(Model model,
                  float sampleRate)
           throws java.io.IOException

Creates the recognizer object. The recognizers process the speech and return text using shared model data

Parameters:

model - VoskModel containing static data for recognizer. Model can be shared across recognizers, even running in different threads.

sampleRate - The sample rate of the audio you are going to feed into the recognizer. Make sure this rate matches the audio content, it is a common issue causing accuracy problems.

Throws:

java.io.IOException - if the recognizer could not be created

Recognizer

public Recognizer(Model model,
                  float sampleRate,
                  SpeakerModel spkModel)

Creates the recognizer object with speaker recognition. With the speaker recognition mode the recognizer not just recognize text but also return speaker vectors one can use for speaker identification

Parameters:

model - VoskModel containing static data for recognizer. Model can be shared across recognizers, even running in different threads.

sampleRate - The sample rate of the audio you are going to feed into the recognizer. Make sure this rate matches the audio content, it is a common issue causing accuracy problems.

spkModel - speaker model for speaker identification

Recognizer

public Recognizer(Model model,
                  float sampleRate,
                  java.lang.String grammar)

Creates the recognizer object with the phrase list. Sometimes when you want to improve recognition accuracy and when you don't need to recognize large vocabulary you can specify a list of phrases to recognize. This will improve recognizer speed and accuracy but might return [unk] if user said something different. Only recognizers with lookahead models support this type of quick configuration. Precompiled HCLG graph models are not supported.

Parameters:

model - VoskModel containing static data for recognizer. Model can be shared across recognizers, even running in different threads.

sampleRate - The sample rate of the audio you are going to feed into the recognizer. Make sure this rate matches the audio content, it is a common issue causing accuracy problems.

grammar - The string with the list of phrases to recognize as JSON array of strings, for example "["one two three four five", "[unk]"]".

Method Detail

setMaxAlternatives

public void setMaxAlternatives(int maxAlternatives)

Configures recognizer to output n-best results.

   {
      "alternatives": [
          { "text": "one two three four five", "confidence": 0.97 },
          { "text": "one two three for five", "confidence": 0.03 },
      ]
   }
 

Parameters:

maxAlternatives - - maximum alternatives to return from recognition results

setWords

public void setWords(boolean words)

Enables words with times in the output

   "result" : [{
       "conf" : 1.000000,
       "end" : 1.110000,
       "start" : 0.870000,
       "word" : "what"
     }, {
       "conf" : 1.000000,
       "end" : 1.530000,
       "start" : 1.110000,
       "word" : "zero"
     }, {
       "conf" : 1.000000,
       "end" : 1.950000,
       "start" : 1.530000,
       "word" : "zero"
     }, {
       "conf" : 1.000000,
       "end" : 2.340000,
       "start" : 1.950000,
       "word" : "zero"
     }, {
       "conf" : 1.000000,
       "end" : 2.610000,
       "start" : 2.340000,
       "word" : "one"
     }],
 

Parameters:

words - - boolean value

setPartialWords

public void setPartialWords(boolean partial_words)

Like above return words and confidences in partial results.

Parameters:

partial_words - - boolean value

setSpeakerModel

public void setSpeakerModel(SpeakerModel spkModel)

Adds speaker model to already initialized recognizer. Can add speaker recognition model to already created recognizer. Helps to initialize speaker recognition for grammar-based recognizer.

Parameters:

spkModel - Speaker recognition model

acceptWaveForm

public boolean acceptWaveForm(byte[] data,
                              int len)

Accept and process new chunk of voice data.

Parameters:

data - - audio data in PCM 16-bit mono format

len - - length of the audio data

Returns:

1 if silence is occurred and you can retrieve a new utterance with result method 0 if decoding continues -1 if exception occurred

acceptWaveForm

public boolean acceptWaveForm(short[] data,
                              int len)

acceptWaveForm

public boolean acceptWaveForm(float[] data,
                              int len)

getResult

public java.lang.String getResult()

Returns speech recognition result

Returns:

the result in JSON format which contains decoded line, decoded words, times in seconds and confidences. You can parse this result with any json parser

  {
    "text" : "what zero zero zero one"
  }
 

If alternatives enabled it returns result with alternatives, see also #setMaxAlternatives(). If word times enabled returns word time, see also #setWordTimes().

getPartialResult

public java.lang.String getPartialResult()

Returns partial speech recognition.

Returns:

partial speech recognition text which is not yet finalized. result may change as recognizer process more data.

 {
    "partial" : "cyril one eight zero"
 }
 

getFinalResult

public java.lang.String getFinalResult()

Returns speech recognition result. Same as result, but doesn't wait for silence. You usually call it in the end of the stream to get final bits of audio. It flushes the feature pipeline, so all remaining audio chunks got processed.

Returns:

speech result in JSON format.

setGrammar

public void setGrammar(java.lang.String grammar)

Reconfigures recognizer to use grammar.

Parameters:

grammar - Set of phrases in JSON array of strings or "[]" to use default model graph.

See Also:

Recognizer(Model, float, String)

reset

public void reset()

Resets the recognizer. Resets current results so the recognition can continue from scratch.

close

public void close()

Releases recognizer object. Underlying model is also unreferenced and if needed, released.

Specified by:

close in interface java.lang.AutoCloseable