edu.cmu.sphinx.frontend.util

Class ConcatFileDataSource

java.lang.Object

edu.cmu.sphinx.util.props.ConfigurableAdapter

edu.cmu.sphinx.frontend.BaseDataProcessor

edu.cmu.sphinx.frontend.util.StreamDataSource

edu.cmu.sphinx.frontend.util.ConcatFileDataSource

All Implemented Interfaces:

DataProcessor, Configurable, ReferenceSource

public class ConcatFileDataSource
extends StreamDataSource
implements ReferenceSource

Concatenates a list raw headerless audio files as one continuous audio stream. A DataStartSignal will be placed before the start of the first file, and a DataEndSignal after the last file. No DataStartSignal or DataEndSignal will be placed between them. Optionally, silence can be added in-between the audio files by setting the property:

edu.cmu.sphinx.frontend.util.ConcatFileDataSource.silenceFile

to a audio file for silence. By default, no silence is added. Moreover, one can also specify how many files to skip for every file read.

You can also specify the name of a transcript file to write the transcription to. The transcription will be written in HUB-4 style. A sample HUB-4 transcript looks like:

 bn99en_1 1 peter_jennings 0.806084 7.079850 <o,f4,male> Tonight this
 Thursday big pressure on the Clinton administration to do something about
 the latest killing in Yugoslavia
 bn99en_1 1 peter_jennings 7.079850 14.007608 <o,fx,male> Airline passengers
 and outrageous behavior at thirty thousand feet What can an airline do
 ...
 bn99en_1 1 inter_segment_gap 23.097000 28.647000 <o,fx,>
 ...
 

The format of each line is:

 test_set_name category speaker_name start_time_in_seconds
 end_time_in_seconds <category,hub4_focus_conditions,speaker_sex> transcript
 

In our example above,

 test_set_name is "bn99en_1"
 category is "1"
 speaker_name is "peter_jennings"
 start_time_in_seconds is "0.806084"
 end_time_in_seconds is "7.079850"
 category is "o" for "Overall"
 hub4_focus_conditions is:
     "f0" for "Baseline//Broadcast//Speech"
     "f1" for "Spontaneous//Broadcast//Speech"
     "f2" for "Speech Over//Telephone//Channels"
     "f3" for "Speech in the//Presence of//Background Music"
     "f4" for "Speech Under//Degraded//Acoustic Conditions"
     "f5" for "Speech from//Non-Native//Speakers"
     "fx" for "All other speech"
 speaker_sex is "male"
 transcript is "Tonight this Thursday big pressure on the Clinton
 administration to do something about the latest killing in Yugoslavia
 

The ConcatFileDataSource will produce such a transcript if the name of the file to write to is supplied in the constructor. This transcript file will be used in detected gap insertion errors, because it accurately describes the "correct" sequence of speech and silences in the concatenated version of the audio files.

Field Summary

Fields

Modifier and Type	Field and Description
static java.lang.String	PROP_ADD_RANDOM_SILENCE The property that specifies whether to add random silence.
static java.lang.String	PROP_BATCH_FILE The property for the file containing a list of audio files to read from.
static java.lang.String	PROP_MAX_SILENCE The property that specifies the maximum number of times the silence file is added between files.
static java.lang.String	PROP_SILENCE_FILE The property that specifies the silence audio file, if any.
static java.lang.String	PROP_SKIP The property that specifies the number of files to skip for every file read.
static java.lang.String	PROP_START_FILE The property that specifies which file to start at.
static java.lang.String	PROP_TOTAL_FILES The property that specifies the total number of files to read.
static java.lang.String	PROP_TRANSCRIPT_FILE The property that specifies the name of the transcript file.

Fields inherited from class edu.cmu.sphinx.frontend.util.StreamDataSource

bitsPerSample, PROP_BIG_ENDIAN_DATA, PROP_BITS_PER_SAMPLE, PROP_BYTES_PER_READ, PROP_SAMPLE_RATE, PROP_SIGNED_DATA, sampleRate

Fields inherited from class edu.cmu.sphinx.util.props.ConfigurableAdapter

logger

Constructor Summary

Constructors

Constructor and Description
ConcatFileDataSource()
ConcatFileDataSource(int sampleRate, int bytesPerRead, int bitsPerSample, boolean bigEndian, boolean signedData, boolean addRandomSilence, int maxSilence, int skip, java.lang.String silenceFileName, int startFile, int totalFiles, java.lang.String transcriptFile, java.lang.String batchFile)

Method Summary

Modifier and Type	Method and Description
java.util.List<java.lang.String>	getReferences() Returns a list of all reference text.
java.lang.String	getTranscriptFile() Returns the name of the transcript file.
void	initialize() Initializes a ConcatFileDataSource.
void	newProperties(PropertySheet ps) This method is called when this configurable component needs to be reconfigured.

Methods inherited from class edu.cmu.sphinx.frontend.util.StreamDataSource

getData, setInputStream, setInputStream

Methods inherited from class edu.cmu.sphinx.frontend.BaseDataProcessor

getPredecessor, setPredecessor

Methods inherited from class edu.cmu.sphinx.util.props.ConfigurableAdapter

getName, initLogger, toString

Methods inherited from class java.lang.Object

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

Field Detail

PROP_START_FILE

@S4Integer(defaultValue=1)
public static final java.lang.String PROP_START_FILE

The property that specifies which file to start at.

See Also:

Constant Field Values

PROP_SKIP

@S4Integer(defaultValue=0)
public static final java.lang.String PROP_SKIP

The property that specifies the number of files to skip for every file read.

See Also:

Constant Field Values

PROP_TOTAL_FILES

@S4Integer(defaultValue=-1)
public static final java.lang.String PROP_TOTAL_FILES

The property that specifies the total number of files to read. The default value should be no limit.

See Also:

Constant Field Values

PROP_SILENCE_FILE

@S4String
public static final java.lang.String PROP_SILENCE_FILE

The property that specifies the silence audio file, if any. If this property is null, then no silences are added in between files.

See Also:

Constant Field Values

PROP_ADD_RANDOM_SILENCE

@S4Boolean(defaultValue=false)
public static final java.lang.String PROP_ADD_RANDOM_SILENCE

The property that specifies whether to add random silence.

See Also:

Constant Field Values

PROP_MAX_SILENCE

@S4Integer(defaultValue=3)
public static final java.lang.String PROP_MAX_SILENCE

The property that specifies the maximum number of times the silence file is added between files. If PROP_ADD_RANDOM_SILENCE is set to true, the number of times the silence file is added is between 1 and this value. If PROP_ADD_RANDOM_SILENCE is set to false, this value will be the number of times the silence file is added. So if PROP_MAX_SILENCE is set to 3, then the silence file will be added three times between files.

See Also:

Constant Field Values

PROP_TRANSCRIPT_FILE

@S4String
public static final java.lang.String PROP_TRANSCRIPT_FILE

The property that specifies the name of the transcript file. If this property is set, a transcript file will be created. No transcript file will be created if this property is not set.

See Also:

Constant Field Values

PROP_BATCH_FILE

@S4String
public static final java.lang.String PROP_BATCH_FILE

The property for the file containing a list of audio files to read from.

See Also:

Constant Field Values

Constructor Detail

ConcatFileDataSource

public ConcatFileDataSource(int sampleRate,
                            int bytesPerRead,
                            int bitsPerSample,
                            boolean bigEndian,
                            boolean signedData,
                            boolean addRandomSilence,
                            int maxSilence,
                            int skip,
                            java.lang.String silenceFileName,
                            int startFile,
                            int totalFiles,
                            java.lang.String transcriptFile,
                            java.lang.String batchFile)

ConcatFileDataSource

public ConcatFileDataSource()

Method Detail

newProperties

public void newProperties(PropertySheet ps)
                   throws PropertyException

Description copied from interface: Configurable

This method is called when this configurable component needs to be reconfigured.

Specified by:

newProperties in interface Configurable

Overrides:

newProperties in class StreamDataSource

Parameters:

ps - a property sheet holding the new data

Throws:

PropertyException - if there is a problem with the properties.

initialize

public void initialize()

Initializes a ConcatFileDataSource.

Specified by:

initialize in interface DataProcessor

Overrides:

initialize in class StreamDataSource

getReferences

public java.util.List<java.lang.String> getReferences()

Returns a list of all reference text. Implements the getReferences() method of ReferenceSource.

Specified by:

getReferences in interface ReferenceSource

Returns:

a list of all reference text

getTranscriptFile

public java.lang.String getTranscriptFile()

Returns the name of the transcript file.

Returns:

the name of the transcript file