A B C D E F G H I L M N O P Q R S T U V W Z

A

absCut(short[]) - Static method in class ocvolume.dsp.endPt

a method to use this class
calls: zeroCrossingBoolean, avgEnergy, zeroCrossing, chopping
called by: volume

add(point, double) - Method in class ocvolume.vq.centroid

add point to centroid's cell
calls: none
called by: codebook

addSound(short[], int) - Method in class ocvolume.audio.objSound

combining samples calls: none called by: volume

alpha - Variable in class ocvolume.hmm.markov

forward variable alpha

avgEnergy(short[]) - Static method in class ocvolume.dsp.endPt

5ms frame (80 samples / frame)
determent the noise location base on the energy
calls: none
called by: absCut

B

beta - Variable in class ocvolume.hmm.markov

backward variable beta

byteArrayComplete() - Method in class ocvolume.audio.micInput

return true if a word is stored in the buffer
calls: none
called by: volume

C

centroid - class ocvolume.vq.centroid.

last updated on June 15, 2002
description: centroid of a codebook calls: point
called by: codebook
input: k-dimensional points
output: distortion measure

centroid(double[]) - Constructor for class ocvolume.vq.centroid

constructor to create a centroid from input coordinates
calls: none
called by: codebook

centroids - Variable in class ocvolume.vq.codebook

centroids array

changeCo(double[]) - Method in class ocvolume.vq.point

replace coordinates with new ones
calls: none
called by: centroid, codebook

chopping(short[], int[]) - Static method in class ocvolume.dsp.endPt

cut the sample into smaller part
calls: none
called by: absCut

close() - Method in class ocvolume.database.database

close connection to database
calls: none
called by: volume, chgToDB

codebook - class ocvolume.vq.codebook.

last updated on June 15, 2002
description: codebook for Vector Quantization component
calls: centroid, point
called by: volume, train
input: speech signal
output: set of centroids, set of indices

codebook_size - Variable in class ocvolume.vq.codebook

codebook size - number of codewords (codevectors)
default is: 256

codebook(point[]) - Constructor for class ocvolume.vq.codebook

constructor to train a codebook with given training points and default codebook size (256)
calls: none
called by: trainCodebook

codebook(point[], int) - Constructor for class ocvolume.vq.codebook

constructor to train a codebook with given training points and codebook size
calls: none
called by: trainCodebook

codebook(ResultSet, int, int) - Constructor for class ocvolume.vq.codebook

constructor to load a codebook from a Result Set
calls: centroid
called by: volume

codebook(String) - Constructor for class ocvolume.vq.codebook

constructor to load a saved codebook from external file
calls: centroid
called by: volume

computeAlpha() - Method in class ocvolume.hmm.markov

calculate forward variable alpha
calls: none
called by: markov

computeBeta() - Method in class ocvolume.hmm.markov

calculate backward variable beta for later use with Re-Estimation method
calls: none
called by: markov

computeFFT(double[]) - Static method in class ocvolume.dsp.fft

performs Fast Fourier Transformation
calls: none
called by: featureExtraction

coordinates - Variable in class ocvolume.vq.point

k-dimensional coordinates array

createTable(String, String[], String[]) - Method in class ocvolume.database.database

creates a table in database
calls: none
called by: chgToDB

currentSeq - Variable in class ocvolume.hmm.markov

current observation sequence

D

database - class ocvolume.database.database.

last updated on June 15, 2002
description: class used to manipulate database
calls: none
called by: volume
input: data to be stored in database
output: data read from database

database(String) - Constructor for class ocvolume.database.database

constructor to load database
calls: none
called by: volume, chgToDB

delta - Variable in class ocvolume.hmm.markov

number of states the model is allowed to jump

dimension - Variable in class ocvolume.vq.point

k-dimensions

dimension - Variable in class ocvolume.vq.codebook

dimension

distortion - Variable in class ocvolume.vq.centroid

distortion measure - sum of all points' distances from the centroid

DOUBLE - Static variable in class ocvolume.database.database

data type DOUBLE

E

endPt - class ocvolume.dsp.endPt.

last modified: June 15, 2002
description: this is the detection of the noise and search for the spot where the acutal word is being sounded, and remove the noise from the sample.
calls: none
called by: volume, train
input: speech signal
output: modified speech signal

endPt() - Constructor for class ocvolume.dsp.endPt

equals(point, point) - Static method in class ocvolume.vq.point

check whether two points are identical
calls: none
called by: centroid, codebook

F

featureExtraction - class ocvolume.dsp.featureExtraction.

last updated on June 15, 2002
description: feature extraction class used to extract mel-frequency cepstral coefficients from input signal
calls: none
called by: volume, train
input: speech signal
output: mel-frequency cepstral coefficient

featureExtraction() - Constructor for class ocvolume.dsp.featureExtraction

fft - class ocvolume.dsp.fft.

last updated on June 15, 2002
description: FFT class for real signals.

FFT - Static variable in class ocvolume.dsp.featureExtraction

Fast Fourier Transformation

fft() - Constructor for class ocvolume.dsp.fft

fftSize - Static variable in class ocvolume.dsp.featureExtraction

FFT Size (Must be be a power of 2)

format - Static variable in class ocvolume.audio.micInput

audio format

frameLength - Static variable in class ocvolume.dsp.featureExtraction

Number of samples per frame

frames - Static variable in class ocvolume.dsp.featureExtraction

All the frames of the input signal

framing(double[]) - Static method in class ocvolume.dsp.featureExtraction

performs Frame Blocking to break down a speech signal into frames
calls: none
called by: featureExtraction

freqToMel(double) - Static method in class ocvolume.dsp.featureExtraction

convert frequency to mel-frequency
calls: none
called by: featureExtraction

G

getAllCo() - Method in class ocvolume.vq.point

get all the coordinates
calls: none
called by: codebook

getAvgEnergy() - Method in class ocvolume.audio.objFrame

return squared mean energy
calls: none
called by: volume

getCo(int) - Method in class ocvolume.vq.point

get coordinate at specific index
calls: none
called by: centroid, codebook

getDimension() - Method in class ocvolume.vq.point

get the k-dimensional space that the point is in
calls: none
called by: centroid, codebook

getDistortion() - Method in class ocvolume.vq.centroid

returns the distortion measure of the current cell
calls: none
called by: codebook

getDistortion(point[]) - Method in class ocvolume.vq.codebook

calculates the distortion
calls: none
called by: volume

getNumPts() - Method in class ocvolume.vq.centroid

returns the number of points in this cell
calls: none
called by: codebook

getPoint(int) - Method in class ocvolume.vq.centroid

get a point at specified index
calls: none
called by: codebook

getProbability(int[]) - Method in class ocvolume.hmm.markov

returns the probability calculated from the testing sequence
calls: none
called by: volume

getRowCount(String) - Method in class ocvolume.database.database

get number of rows in a table
calls: none
called by: volume

getSample() - Method in class ocvolume.audio.objFrame

return 100ms speech signal
calls: none
called by: volume

getSound() - Method in class ocvolume.audio.objSound

return speech signal calls: none called by: volume

getSoundAt(int) - Method in class ocvolume.audio.objSound

return speech signal with a given time(index) calls: none called by: volume

getSoundSize() - Method in class ocvolume.audio.objSound

return the signal calls: none called by: volume

getWord(short[]) - Method in class ocvolume.ocvolume

H

hammingWindow - Static variable in class ocvolume.dsp.featureExtraction

hamming window values

I

imag - Static variable in class ocvolume.dsp.fft

imaginary part

initialize() - Method in class ocvolume.vq.codebook

creates a codebook using LBG algorithm which includes K-means
calls: centroid
called by: codebook

insertRow(String, double[]) - Method in class ocvolume.database.database

inserts a row to the specified table
calls: none
called by: chgToDB

L

len_obSeq - Variable in class ocvolume.hmm.markov

length of observation sequence

log10(double) - Static method in class ocvolume.dsp.featureExtraction

calculates logarithm with base 10
calls: none
called by: featureExtraction

lowerFilterFreq - Static variable in class ocvolume.dsp.featureExtraction

lower limit of filter (or 64 Hz?)

M

magnitudeSpectrum(double[]) - Static method in class ocvolume.dsp.featureExtraction

computes the magnitude spectrum of the input frame
calls: none
called by: featureExtraction

main(String[]) - Static method in class ocvolume.train

create an instance of train

markov - class ocvolume.hmm.markov.

last updated on June 15, 2002
description: this class represents a left-to-right Hidden Markov Model and its essential methods for speech recognition.

markov(int, int) - Constructor for class ocvolume.hmm.markov

class constructor - used to create a left-to-right model with multiple observation sequences for training
calls: none
called by: trainHMM

markov(String) - Constructor for class ocvolume.hmm.markov

class constructor - used to create a model from a saved file
calls: none
called by: volume, trainHMM

micInput - class ocvolume.audio.micInput.

description: continuous record voice from microphone, with at least 600ms pause in between words or discrete recording with at least 100ms in the front and end being silent default is continuous recording
calls: none
called by: volume
input: digitalize speech signal
output: speech signal stored in array

micInput() - Constructor for class ocvolume.audio.micInput

Constructor

MIN_DISTORTION - Variable in class ocvolume.vq.codebook

minimum distortion

N

newWord() - Method in class ocvolume.audio.micInput

read the next element in the word buffer
calls: none
called by: volume

num_obSeq - Variable in class ocvolume.hmm.markov

number of observation sequence

num_states - Variable in class ocvolume.hmm.markov

number of state in the model example: number of urns

num_symbols - Variable in class ocvolume.hmm.markov

number of observation symbols per state example: how many different colour balls there are

numCepstra - Static variable in class ocvolume.dsp.featureExtraction

Number of MFCCs per frame

numMelFilters - Static variable in class ocvolume.dsp.featureExtraction

number of mel filters (SPHINX-III uses 40)

numPoints - Static variable in class ocvolume.dsp.fft

number of points

O

objFrame - class ocvolume.audio.objFrame.

last modified: June 13, 2002
description: store 100ms of samples (1600 samples), and calculate the averge square mean
calls: none
called by: micInput
input: 100ms signal
output: 100ms signal, squared mean energy

objFrame(short[]) - Constructor for class ocvolume.audio.objFrame

constructor to calculate squared mean energy
calls: none
called by: volume

objSound - class ocvolume.audio.objSound.

last modified: June 13, 2002
description: combine several 100ms frames
calls: none
called by: micInput
input: speech signal
output: speech signal

objSound(int) - Constructor for class ocvolume.audio.objSound

constructor to create a new short array calls: none called by: volume

obSeq - Variable in class ocvolume.hmm.markov

discrete set of observation symbols example: sequence of colour of balls

ocvolume - package ocvolume

ocvolume - class ocvolume.ocvolume.

speech recognitione engine last updated on June 18, 2002

ocvolume.audio - package ocvolume.audio

ocvolume.database - package ocvolume.database

ocvolume.dsp - package ocvolume.dsp

ocvolume.hmm - package ocvolume.hmm

ocvolume.vq - package ocvolume.vq

ocvolume.wav - package ocvolume.wav

ocvolume(String, String) - Constructor for class ocvolume.ocvolume

constructor to create a speech recognition engine using VQ for recognition

ocvolume(String, String, String) - Constructor for class ocvolume.ocvolume

constructor to create a speech recognition engine using HMM for recognition

output - Variable in class ocvolume.hmm.markov

discrete output probability example: probability of a specific output from a state

P

pi - Variable in class ocvolume.hmm.markov

initial state distribution example: which state is the starting state

point - class ocvolume.vq.point.

last updated on June 15, 2002
description: point class which stores coordinates in k-dimensional space
calls: none
called by: centroid, codebook
input: set of co-ordinates
output: none

point(double[]) - Constructor for class ocvolume.vq.point

constructor to create a point with k-dimensional coordinates array
calls: none
called by: centroid, codebook

preEmphasis(short[]) - Static method in class ocvolume.dsp.featureExtraction

perform pre-emphasis to equalize amplitude of high and low frequency
calls: none
called by: featureExtraction

preEmphasisAlpha - Static variable in class ocvolume.dsp.featureExtraction

Pre-Emphasis Alpha (Set to 0 if no pre-emphasis should be performed)

process(short[]) - Static method in class ocvolume.dsp.featureExtraction

takes a speech signal and returns the Mel-Frequency Cepstral Coefficient (MFCC)
calls: fft
called by: volume, train

pt - Variable in class ocvolume.vq.codebook

training points

pts - Variable in class ocvolume.vq.centroid

stores the points that belong to this centroid or cell

Q

q - Variable in class ocvolume.hmm.markov

best state sequence

quantize(point[]) - Method in class ocvolume.vq.codebook

quantize the input array of points in k-dimensional space
calls: none
called by: volume

R

readWave(String) - Static method in class ocvolume.wav.waveIO

loading wave file
calls: none
called by: train

real - Static variable in class ocvolume.dsp.fft

real part

remove(point, double) - Method in class ocvolume.vq.centroid

removes a given point from the centroid's cell
calls: none
called by: codebook

removeOldWord() - Method in class ocvolume.audio.micInput

remove the first element in the word buffer
calls: none
called by: volume

retrieve(String) - Method in class ocvolume.database.database

retrieve data from entire table
calls: none
called by: volume

returnShort(int) - Method in class ocvolume.audio.micInput

return the signed short with given index
calls: none
called by: volume

returnShortArray() - Method in class ocvolume.audio.micInput

return the signed short array
calls: none
called by: volume

returnShortSize() - Method in class ocvolume.audio.micInput

return the length of the array of the short array
calls: none
called by: volume

run() - Method in class ocvolume.train

run() - Method in class ocvolume.audio.micInput

recording from the microphone
calls: none
called by: volume

S

SAMPLE_BIG_ENDIAN - Static variable in class ocvolume.audio.micInput

use little-endian
00000100 00000001
first byte 00000001
second byte 00000100

SAMPLE_BITS - Static variable in class ocvolume.audio.micInput

sample size - 16 bits

SAMPLE_BUFFER_SIZE - Static variable in class ocvolume.audio.micInput

buffersize 2 seconds

SAMPLE_CHANNELS - Static variable in class ocvolume.audio.micInput

sample channels

SAMPLE_ENCODING - Static variable in class ocvolume.audio.micInput

encoding method "pulse-code modulation" predefined encoding

SAMPLE_FRAME_RATE - Static variable in class ocvolume.audio.micInput

sample frame rate

SAMPLE_FRAME_SIZE - Static variable in class ocvolume.audio.micInput

sample frame size

SAMPLE_RATE - Static variable in class ocvolume.audio.micInput

capture amound of sample per second, can be changed in the constructor 16 kHz

samplingRate - Static variable in class ocvolume.dsp.featureExtraction

sample rate in Hz

save(String) - Method in class ocvolume.hmm.markov

save HMM model to file
calls: none
called by: trainHMM

saveToDB(String) - Method in class ocvolume.vq.codebook

save codebook to database
calls: database
called by: train

saveToFile(String) - Method in class ocvolume.vq.codebook

save codebook to txt file
calls: none
called by: train

scaleFactor - Variable in class ocvolume.hmm.markov

Scale Coefficient

setCo(int, double) - Method in class ocvolume.vq.point

set coordinate at specific index
calls: none
called by: centroid, codebook

setContinuous() - Method in class ocvolume.audio.micInput

set the recording method to continuous
calls: none
called by: volume

setDiscrete() - Method in class ocvolume.audio.micInput

set the recording method to discrete
calls: none
called by: volume

setNumObSeq(int) - Method in class ocvolume.hmm.markov

set the number of training sequences
calls: none
called by: trainHMM

setObSeq(int[]) - Method in class ocvolume.hmm.markov

set observation sequence
calls: none
called by: trainHMM

setTrainSeq(int[][]) - Method in class ocvolume.hmm.markov

set training sequences for re-estimation step
calls: none
called by: trainHMM

setTrainSeq(int, int[]) - Method in class ocvolume.hmm.markov

set a training sequence for re-estimation step
calls: none
called by: trainHMM

shiftInterval - Static variable in class ocvolume.dsp.featureExtraction

Number of overlapping samples (usually 50% of frame length)

SPLIT - Variable in class ocvolume.vq.codebook

split factor (should be in the range of 0.01 <= SPLIT <= 0.05)

split() - Method in class ocvolume.vq.codebook

splitting algorithm to increase number of centroids by multiple of 2
calls: centroid
called by: codebook

stopRecord() - Method in class ocvolume.audio.micInput

Stop the mic input
calls: none
called by: volume

T

total_pts - Variable in class ocvolume.vq.centroid

total number of points that belong to this centroid or cell

train - class ocvolume.train.

training class for VQ recognition last updated on June 15, 2002

train() - Constructor for class ocvolume.train

creates new form volume

train() - Method in class ocvolume.hmm.markov

train the hmm model until no more improvement
calls: none
called by: trainHMM

transition - Variable in class ocvolume.hmm.markov

state transition probability example: probability from one state to another state

U

update() - Method in class ocvolume.vq.centroid

update centroid by taking average of all points in the cell
calls: none
called by: codebook

upperFilterFreq - Static variable in class ocvolume.dsp.featureExtraction

upper limit of filter (or half of sampling freq.?)

V

vBufferWord - Variable in class ocvolume.audio.micInput

word buffer

viterbi(int[]) - Method in class ocvolume.hmm.markov

viterbi algorithm used to get best state sequence and probability
calls: none
called by: volume

W

waveIO - class ocvolume.wav.waveIO.

last modified: June 15, 2002
description: input and output to a wave file
calls: none
called: tarin
input: speech signal, path
output: none

waveIO() - Constructor for class ocvolume.wav.waveIO

wordBufferSize() - Method in class ocvolume.audio.micInput

return how many words are in the buffer
calls: none
called by: volume

writeWave(short[], String) - Static method in class ocvolume.wav.waveIO

write to wave file
calls: none
called by: train

Z

zeroCrossing(short[], int[], boolean[]) - Static method in class ocvolume.dsp.endPt

search for the end points based on zero-crossing and result from avgEnergy
calls: none
called by: absCut

zeroCrossingBoolean(short[]) - Static method in class ocvolume.dsp.endPt

Simply mark off the location of Zero-Crossing
calls: none
called by: absCut

A B C D E F G H I L M N O P Q R S T U V W Z