Advances in Audio and Speech Signal Processing: Technologies and Applications

Advances in Audio and
Speech Signal Processing:
Technologies and Applications

Chapter.I
Introduction.to.Audio.and.Speech.Signal.Processing................................................. 1
Hector Perez-Meana, National Polytechnic Institute, Mexico
Mariko Nakano-Miyatake, National Polytechnic Institute, Mexico
Section.I
Audio.and.Speech.Signal.Processing.Technology
 

Chapter.II
Digital.Filters.for.Digital.Audio.Effects..................................................................... 22
Gordana Jovanovic Dolecek, National Institute of Astrophysics, Mexico
Alfonso Fernandez-Vazquez, National Institute of Astrophysics, Mexico
 

Chapter.III
Spectral-Based.Analysis.and.Synthesis.of.Audio.Signals......................................... 56
Paulo A.A. Esquef, Nokia Institute of Technology, Brazil
Luiz W.P. Biscainho, Federal University of Rio de Janeiro, Brazil

Chapter.IV
DSP.Techniques.for.Sound.Enhancement.of.Old.Recordings................................. 93
Paulo A.A. Esquef, Nokia Institute of Technology, Brazil
Luiz W.P. Biscainho, Federal University of Rio de Janeiro, Brazil
Section.II
Speech.and.Audio.Watermarking.Methods
 

Chapter.V
Digital.Watermarking.Techniques.for.Audio.and.Speech.Signals......................... 132
Aparna Gurijala, Michigan State University, USA
John R. Deller, Jr., Michigan State University, USA
 

Chapter.VI
Audio.and.Speech.Watermarking.and.Quality.Evaluation................................... 161
Ronghui Tu, University of Ottawa, Canada
Jiying Zhao, University of Ottawa, Canada
Section.III
Adaptive.Filter.Algorithms
 

Chapter.VII
Adaptive.Filters:.Structures,.Algorithms,.and.Applications.................................. 190
Sergio L. Netto, Federal University of Rio de Janeiro, Brazil
Luiz W.P. Biscainho, Federal University of Rio de Janeiro, Brazil
 

Chapter.VIII
Adaptive.Digital.Filtering.and.Its.Algorithms.for.Acoustic.
Echo.Canceling...........................................................................................................225
Mohammad Reza Asharif, University of Okinawa, Japan
Rui Chen, University of Okinawa, Japan
 

Chapter.IX
Active.Noise.Canceling:.Structures.and.Adaption.Algorithms.............................. 286
Hector Perez-Meana, National Polytechnic Institute, Mexico
Mariko Nakano-Miyatake, National Polytechnic Institute, Mexico
 

Chapter.X
Differentially Fed Artificial Neural Networks for Speech Signal Prediction........ 309
Manjunath Ramachandra Iyer, Banglore University, India
Section.IV
Feature.Extraction.Algorithms.and.Speech.Speaker.Recognition
 

Chapter.XI
Introduction.to.Speech.Recognition.........................................................................325
Sergio Suárez-Guerra, National Polytechnic Institute, Mexico
Jose Luis Oropeza-Rodriguez, National Polytechnic Institute, Mexico
 

Chapter.XII
Advanced.Techniques.in.Speech.Recognition......................................................... 349
Jose Luis Oropeza-Rodriguez, National Polytechnic Institute, Mexico
Sergio Suárez-Guerra, National Polytechnic Institute, Mexico
 

Chapter.XIII
Speaker.Recognition.................................................................................................. 371
Shung-Yung Lung, National University of Taiwan, Taiwan
 

Chapter.XIV
Speech.Technologies.for.Language.Therapy........................................................... 408
Ingrid Kirschning, University de las Americas, Mexico
Ronald Cole, University of Colorado, USA
 

About.the.Authors......................................................................................................434
Index............................................................................................................................439

Comment to request full ebook

KEYWORDS

Index
A
absolute category rating (ACR) method
178
acoustic
echo cancellation systems 222
impulse responses 229, 275
noise control (ANC) 269
active
noise
cancellation (ANC) 286, 287
control 222
adaptive
digital filtering (ADF) 225
echo cancellation 2
filter 198
linear combiner 193
noise canceller 5
advanced audio coding (AAC) 148
affine-projection 190
air traffic control 136
allpass
filters 25, 51
reverberator 42
American Defense Department (DoD) 13
amplitude estimator 120
analog
-to-digital (A/D) 289
hole 144
analysis tool 120
ANC
algorithm 305
filter 291
systems 18
anti-aliasing filter 289
artificial
larynx transducer (ALT) 14
neural networks 179
neuronal networks 331
reverberation 24
spectrum 80
vibrato 84
ASR system 328
audio
and speech watermarking techniques
136
applications 94
de-clicking 96, 99
de-hissing 96
de-noising 83
equalization 45
morphing 84
segment 83
signal 59
Stirmark 156
watermarking algorithm 139
authentication 136
autocorrelation method 153
automatic
gain control (AGC) 191
speech recognition (ASR) 325, 350
systems 350
autoregressive (AR) 74, 96
based
audio de-clicker 109
interpolators 114, 115
linear prediction 78
model 100, 108
coefficients 100
estimators 101
order 113
parameters 337
separation method 100
synthesis filter 114
B
backward-extrapolated fragment 114
bandpass filters 74
bidirectional channel 2
binary phase shift keying (BPSK)
152, 171
Bit-stream watermarking 148
block
division 181
LMS (BLMS) 245
bounded-Q transform 74
broadband
audio watermarking algorithms 151
noise (hiss) 95
broadcast monitoring 136, 137, 166
C
capella 81
Center for Spoken Language Research
(CSLR) 412
cepstrales coefficients (CLPC) 345
chorusing 30
classifier techniques 375
codebook design 382
codebook exited linear predictive coding
(CELP) 10
/hybrid codecs 180
codecs 13
coefficient 99, 145, 227, 253
quantization 181
selection 181
collusion attack 144
comb
filter 22, 40
reverberator 42
computational
cost 165
model 179
consonants 409
Constant-Q transform (CQT) 74
content
authentication 137, 166
management 137
continuous density hidden Markov models
(CDHMMs) 349, 358
copy
attack 144
control 136, 137, 166
copyright
protection 136, 137, 165
correlation
function 262, 275
LMS (CLMS) 258
coustic
impulse response 245
coversignal 133
-to-watermark 150
cropping attack 135
cross-correlation
coefficients 99
function 263
vector 257
cryptographic attacks 168
CSLU toolkit 410
D
DANN 314
data payload 132, 134, 164
data time warping (DTW) 350
algorithm 350
dead-lock attack 144
demand only O(N) 122
desynchronized stegosignal 135
detection blindness 164
digital
-to-analog (D/A) 143
audio
technology 94
watermarking 185
filters 226
literacy 225, 325
record 133
signal processing (DSP) 2, 23, 191
chips 13
solutions 111
diphthongs 420
direct-sequence spread spectrum (DSSS)
169
discrete
-time Fourier transform (DTFT) 59, 379
cosine transform (DCT) 82, 139, 145
Fourier
series (DFS) 379
transform (DFT) 58, 117, 175
wavelet transform (DWT)
120, 181, 264, 394
disyllables words 363
dither
modulation (DM) 148
vectors 148
double-talk detector (DTD) 254
dragon dictate 349
dual port RAM (DPR) 251
DWT
coefficient 122
decomposition 181
domain-based watermarking algorithm
183
dynamic
programming (DP) 391
time warping (DTW) 331, 403
E
e-model 179
echo
hiding techniques 140
return loss (ERL) 228
return loss enhancement (ERLE) 252
speech signals 35
ECLMS algorithm 267
embedding patient information 138
emerging track 75
endpoint detection 379
energy
analysis 336
function of the high frequency (ERO
parameter) 365
environmental noise 180
Ephraim and Malah suppression rule
(EMSR) 120
equation error (EE) 190, 211
solution 217
error
measurement 192
microphone 289
esophageal speech 14
Eucldian distance 395
evolving track 75
expectation maximization (EM) 358
experimental design 371
F
fast
Fourier transform (FFT)
58, 178, 208, 330, 379
buffer 70
coefficients 178
kernel 264
transversal filter (FTF) 208
feature
extraction 374
selection 371
fidelity 134
filter
banks 378
coefficients 340
fingerprinting 138, 166
finite impulse response (FIR)
38, 191, 366
adaptive filters 4
flanging 30, 43
formant
amplitudes 373
frequencies 373
frequency transitions 373
forward-extrapolated signal 114
Fourier transform 379
fragile speech watermarking algorithm
155
frame 378
frequency
bin adaptive filtering (FBAF) 245, 247
domain adaptive filtering (FDAF) 245
domain ECLMS (FECLMS)
algorithm 262, 267
domain synthesis 80
hopping spread spectrum (FHSS) 169
fundamental tone (T0) 345
fuzzy C-means algorithm (FCM) 387
FxRLS algorithms 286, 291
G
Gaussian
densities 356
estimators 314
HMM 422
mixture model (GMM)
179, 311, 312, 378
mixtures 356
genetic algorithm(GA) 382
geometric attacks 144, 168
group vector quantization (GVQ) 380
H
Hanning window 71, 72
harmonic
and individual lines plus noise (HILN)
87
signal 84
head-related
impulse response 218
transfer functions 218
hidden Markov models (HMMs) 78, 312,
326, 350, 396, 408
-neural networks (HMM-NN) 327
hidden wavelet Markov model (HWMM)
378
human
auditory system (HAS) 18, 138, 162
-based audio encoding operations 139
-based perceptual models 141, 151
-based psychoacoustic auditory model
141
visual system (HVS) 162
hybrid 379
methods 331
transformer 2
I
IBM's SpeechView 417
imperceptibility 164
implicit 379
individualized instruction 415
infinite impulse response (IIR)
37, 191, 227
filter 36
information embedding rate 143
input signal power spectral 344
instructor 418
integer 395
intelligent learning systems 412
international telecommunication union
(ITU) 179
interpolation 83
interpolators 111
Intonation 409
inusoidal modeling parameters 78
inverse
discrete wavelet transform (IDWT)
121, 266
reconstruction 183
Fourier transform 174, 344
inversion 144
isolated peaks 71
iteration 183
J
Jean Piaget School 422
K
Karhunen-Loeve
expansion 384
transform (KLT) 377, 384
kernels 146
keys 144
L
Lagrange multipliers 386
language therapy 409, 411
learning vector quantization (LVQ) 379
algorithm 380
least mean square (LMS)
190, 191, 203, 234
-based algorithm 191
-Newton algorithm 205
-type family 204
algorithm 201, 208
convergence speed 201
least significant bit (LSB) 140
-based watermarking schemes 147
least square
error (LSE) 154
least squares
autoregressive (LSAR) 112
interpolator 113
spectral optimization method 74
Linde-Buzo-Grey (LBG)
algorithm 379, 380, 382
codebooks 381
linear prediction (LP) 140, 338
coding (LPC) 65, 198, 227, 330, 350
method 337
linguistic message 373
local maxima 73
log-to-digital (A/D) 143
logarithm 344
logarithmic spectrum 344
long-playing (LP) 94
cepstral coefficients 389
model 152, 153
parameters 153
long-term 153
parametric code 153
loudspeaker enclosure microphone system
(LEMS) 253
low-bit audio coding 83
low-order AR models 80
lowpass
filtering 26
reverberation filters 25
M
Markov model 313
MATLAB 25
matrix quantization (MQ) 379
maximum a posteriori (MAP) 101, 359
mean opinion score (MOS) 178
value 180
mean square consistency 194
mean squared error (MSE) 142, 227, 253
measure analysis 371
minimum-distance decoder 148
misadjustment 194
model
-based detectors 111
language 366
parameters 101
modified discrete cosine transform
(MDCT) 149
modulated complex lapped transform
(MCLT) 170
modulation 409
monosyllabic words 363
morph 84
mother-wavelet 121
moving-average autoregressive (ARMA)
74
MPEG
-1 psychoacoustic model 174
audio compression 140
compression 139
MQ codebooks 379
multi-band excitation (MBE) 13
multi-resolution singular value decomposition
(MSVD) 386
multimedia
data 137
design 415
multiple
echo filtering 35
echo filters 25
transcoding 180
multipulse excited (MPE) 10
mutually independent 195
N
natural
reverberation 53
sound recognition 17
Newton
-like algorithm 204
-type algorithm 202
algorithm 195
Neyman-Pearson paradigm 135
non-decimated filterbank 58
non-linear network 314
non-parametric 74
signal 120
normalization 257
normalized
digital frequencies 28
least mean square (NLMS) 235
version 190
Nyquist frequency 84
O
objective
function 192
testing method 179
optimization method 192
optimum Wiener solution 237
orthogonal 194
orthogonalized ANC 305
output-error (OE) 190
algorithm 212
P
packet losses 180
parabolic interpolation 69, 72
parameter
-based algorithms 179
estimation 71
parametric
audio coding 87
representation 83
watermarking 153
parent node 395
PARSHL program 75
peaking filters 45
peak parameters 67
percentage of correctly extracted watermark
bit (PCEW) 181
perceptual
evaluation of speech quality (PESQ) 179
algorithm 180
linear prediction analysis (PLP) 331
model 179
personal identification number (PIN) 319
phase vocoder 58
phonetic-based system 328
pitch 373
-based extended AR model 113
dynamics 373
synchronous overlap and add (PSOLA)
algorithm 155
polynomial filtering 102
post-masking 174
posteriori 120, 359
power spectral density 28
pre-echo phenomenon 63
pre-masking 174
preamplifier 289
Presentation 419
primitives 375
priori 106, 120, 200
probability density function (pdf) 13
protocol attacks 144, 168
pseudo
-noise (PN) 139
-random noise (PN) 181
psychoacoustical mask 174
psychoacoustic auditory model 139, 141
psychoacoustics 57
Q
quantization
index modulation (QIM) 145, 147
scale (QS) 181
quantized-data algorithms 204
R
re-synthesized signal 84
real life audio signals 62
recording/reproduction techniques 94
recursive least squares (RLS) 190, 205
algorithm 191
regular pulse excited (RPE) 10
code 11
render inharmonic 84
reverb 30
reverberation
filters 25, 42
time 24
rhythm 409
robust 134
bit-stream watermarking 145
S
SAR algorithm 275
SBD-ANC scheme 305
scanners 1, 22, 56, 93, 132, 225, 309
, 371
security 165
segmental CWR 142
segment durations 373
set-membership variation 190
shelving filters 45
short-term total energy function (STTEF)
365
short-time
fourier transform (STFT) 71
sinusoid 120
spectral attenuation (STSA) 116
-based de-hisser 119
signal
-based algorithms 179
-to-noise ratio (S/N) 107, 252
quality 140
segmentation 59
simple hyperstable algorithm for recursive
filters (SHARF) 213
single echo filter 25
sinusoidal
modeling 56, 81, 111
parameters 80
sinusoids
+noise synthesis scheme 81
smart acoustic room (SAR)
225, 227, 269
social agents 415
sound source separation 83
speaker
classifiers 371
model 378
recognition system 311
spectrograms 373
speech
/audio signals 136
quality evaluation 185
signal 35, 227, 248
preprocessing 374
spread spectrum (SS) 145
signaling 152
watermarking 145
statistical models 331
stegosignal 133
fidelity 139
stirmark benchmark for audio (SBMA)
168
stochastic
framework 359
model simplifications 360
sub-band
adaptive 205
methods 111
syllabic
-based system 328
units 329
synthesis
filters 10, 13
methods 87
systemic processes 134
T
temporal domain based algorithm 184
three-syllabic words 363
time
-scale modifications 83
domain adaptive filter (TDAF) 245
Toeplitz matrix 264
transaction tracking 138
transform
-domain 205
LMS 205
encryption coding (TEC) 149
transmission error 180
triangular window technique 69
two
-pass split window (TPSW) 66, 102
-based estimate 105
filtering procedure 103
-syllabic words 363
one-directional channels 2
V
vanishing track 75
variable delay 180
vector
codebooks 311
quantization (VQ) 379, 382
speaker model 379
quantizer codebook 12
virtual humans 414
VLSI technology 2
voice over internet protocol (VoIP) 179
vowels 409
VQ
-based classifier 380
model 379
W
watermark 143
algorithms 132
applications 132, 138
detector 166
embedding 182
literature 133
robustness 132
sequence 143
signal 144
fidelity 132
techniques 136
waveform
codecs 180
substitution schemes 111
wavelet
Markov model (WMM) 396
neural network (WNN) 378
transform (WT) 386
WECLMS algorithm 267
weighted least-squares (WLS) 205
white noise 120