![1
CHAPTER
Introduction to Signal Processing
Theory
Isabela F. Apolinário and Paulo S.R. Diniz
Program of Electrical Engineering and the Department of Electronics and Computer Engineering, COPPE/Poli,
Universidade Federal do Rio de Janeiro, Brazil
1.01.1 Introduction
Signal processing is a key area of knowledge that finds applications in virtually all aspects of modern
life. Indeed the human beings are employing signal processing tools for centuries without realizing it
[1]. In present days the younger generation might not be able to understand how one can live without
carrying a mobile phone, traveling long distances without an almost self piloted airplane, exploring other
planets without human presence, and utilizing a medical facility without a wide range of diagnostic and
intervention equipments.
Signal processing consists of mapping or transforming information bearing signals into another form
of signals at the output, aiming at some application benefits. This mapping defines a continuous or analog
system if it involves functions representing the input and output signals. On the other hand, the system
is discrete or digital if its input and output signals are represented by sequences of numbers.
Signal processing theory is very rich and as a reward it has been finding applications in uncountable
areas, among which we can mention bioengineering, communications, control, surveillance, environ-
ment monitoring, oceanography, and astronomy, just to mention a few. In particular, the enormous
advance in digital integrated circuit technology, responsible for the computing and information revo-
lutions that we are so used today, enabled the widespread use of Digital Signal Processing Systems.
These systems allowed advances in fields such as: speech, audio, image, video, multirate processing,
besides in several applications of wireless communications and digital control.
This chapter is an overview of the classical signal processing tools whose details are encountered in
numerous textbooks such as: [1–15]. The topics treated in the following chapters entail the description
of many signal processing tools, usually not covered in the classical textbooks available in the market.
As a result, we believe that the readers can benefit from reading many of these chapters, if not all, in
order to deepen and widen their current knowledge as well as to start exploiting new ideas.
1.01.2 Continuous-time signals and systems
The processing of signals starts by accessing the type of information we want to deal with. Many signals
originating from nature are continuous in time and as such, the processing involved must include analog
signal acquisition systems followed by the implementation of a continuous-time system if the processing
Academic Press Library in Signal Processing. http://dx.doi.org/10.1016/B978-0-12-396502-8.00001-2
© 2014 Elsevier Ltd. All rights reserved.
3](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-31-320.jpg)
![1.01.3 Discrete-Time Signals and Systems 5
frequency variable. The Fourier transform is described by
X() =
∞
−∞
x(t)e−jt
dt ⇐⇒ x(t) =
1
2π
∞
−∞
X()ejt
d.
The representation of a time-domain periodic function of a continuous and real (time) variable is
performed by the Fourier series. In the frequency-domain, the representation consists of a non-periodic
function of a discrete and integer frequency variable, as following described:
X(k) =
1
T
T
0
x(t)e−j(2π/T )kt
dt ⇐⇒ x(t) =
∞
k=−∞
X(k)ej(2π/T )kt
.
1.01.3 Discrete-time signals and systems
A discrete-time signal is represented by a sequence of numbers and can be denoted as x(n) with n ∈ Z,
where Z is the set of integer numbers. The samples x(n) might represent numerically the amplitude of
a continuous-time signal sample at every T seconds or can be originated from a discrete information.
In many applications, the sampling period T represents the time interval between two acquired samples
of the continuous-time signal. However, in other situations it might represent the distance between two
sensorsoftwopixelsofanimage,ortheseparationbetweentwoantennas,justtomentionafewexamples.
Discrete-time systems map input sequences into output sequences. A general representation of this
mapping is
y(n) = H{x(n)},
where H{·} denotes the operation performed by the discrete-time system. Figure 1.2 depicts the input-
to-output mapping of sequences. According to the properties of H{·}, the discrete-time system might
be LTI. This way, it benefits from a number of analysis and design tools such as frequency-domain
representations. However, there are many applications employing nonlinear, time-varying, and even
non-causal systems [2,5,7,8,11,12,16].
If the system is linear and time-invariant, as will be described in Chapter 2, there are several tools
to describe the mapping features of the system. Typically, a general description of discrete-time sys-
tems through a difference equation can be solved and analyzed by employing the z transform. Also a
x(n)
y(n)
n
n
H{·}
FIGURE 1.2
Discrete-time signal representation.](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-33-320.jpg)
![6 CHAPTER 1 Introduction to Signal Processing Theory
discrete-time LTI system is fully described by its impulse response, whereas a nonlinear system can not
be fully characterized by its impulse response.
The z transform is the key tool to represent a time-domain non-periodic function of a discrete
and integer variable through a frequency-domain non-periodic function of a continuous and complex
frequency variable. That is,
X(z) =
∞
n=−∞
x(n)z−n
⇐⇒ x(n) =
1
2πj
C
X(z)zn−1
dz.
The discrete-time Fourier transform (DTFT) represents a time-domain non-periodic function of a
discrete and integer variable by a periodic function of a continuous frequency variable in the frequency
domain as follows:
X(ejω
) =
∞
n=−∞
x(n)e−jωn
⇐⇒ x(n) =
1
2π
∞
−∞
X(ejω
)ejωn
dω.
Finally, the discrete Fourier transform (DFT) is the right tool to represent a time-domain periodic
function of a discrete and integer variable through a periodic function of a discrete and integer frequency
variable in the frequency domain. The DFT is defined as
X(k) =
N−1
n=0
x(n)e−j(2π/N)kn
⇐⇒ x(n) =
1
N
N−1
k=0
X(k)ej(2π/N)kn
.
The DFT plays a key role in digital signal processing since it represents a finite-length sequence in
the time domain through a finite-length sequence in the frequency domain. Since both domains utilize
sequences, this feature makes the DFT a natural choice for time-frequency representation of information
in a digital computer. A little mental exercise allows us to infer that the DFT is the perfect representation
in the frequency domain of a finite-length sequence in the time domain, if we interpret the latter as a
period of a periodic infinite-length sequence. In addition, by employing appropriate zero-padding in
two finite-length sequences, the product of their DFTs represents the DFT of their linear convolution
(see Chapter 3), turning the DFT a valuable tool for LTI system implementation. There is a plethora of
applications for the DFT in signal processing and communications and some of them can be accessed
in the references [2–13].
Often, whenever a DSP system is LTI, it can be described through a difference equation as follows:
N
i=0
ai y(n − i) +
M
l=0
bl x(n − l) = 0.
The above description is suitable for implementation in digital computers. The difference equation
represents a causal LTI system if its auxiliary conditions correspond to its initial conditions and those
are zeros [2]. Assuming a0 = 1, the above equation can be rewritten as
y(n) = −
N
i=1
ai y(n − i) +
M
l=0
bl x(n − l).](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-34-320.jpg)
![1.01.4 Random Signals and Stochastic Processes 7
This class of system has infinite impulse response (i.e., y(n) = 0 when n → ∞) and, as such, it is
called IIR system or filter.
The nonrecursive system generates the output signal from past input samples, that is,
y(n) =
M
l=0
bl x(n − l).
In this case, the resulting system has finite impulse response and is known as FIR systems or filters.
1.01.4 Random signals and stochastic processes
In most practical applications, we have to deal with signals that can not be described by a function
of time, since their waveforms follow a random pattern [17,18]. Take, for example, the thermal noise
inherent to any material. Despite the lack of exact knowledge of the signal values, it is possible to
analyze and extract information the signals contain by employing the mathematical tools available to
deal with random signals. Chapter 4 will present a compact and yet clarifying review of random signals
and stochastic processes which are crucial to understand several concepts that will be discussed in the
more advanced chapters.
The theory starts by defining a random variable as a mapping of the result of a random process
experiment onto the set of numbers. A random variable X is a function that assigns a number x to every
outcome of a random process experiment. An example is given in Figure 1.3(a), in which each outcome
of a throw of a dice, numbers 1–6, corresponds to a determined value, x1–x6, respectively.
x1
X
x 1
x3
x2
x4 t
x
1
(
t
)
x
2
(
t
)
x
3
(
t
)
x
n
(
t
)
t
t
t
{X }
(b)
(a)
FIGURE 1.3
Examples of random variable and stochastic process. (a) Random variable. (b) Stochastic process.](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-35-320.jpg)
![1.01.6 FIR and IIR Filter Design 11
0 0.02 0.04 0.06 0.08 0.1
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
time(s)
Amplitude
x(t)
x(n)
(a)
0 0.02 0.04 0.06 0.08 0.1
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
time(s)
Amplitude
x(t)
xQ
(n)
(b)
0 0.02 0.04 0.06 0.08 0.1
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
time(s)
Amplitude
x(t)
x
ZOH
(t)
0 0.02 0.04 0.06 0.08 0.1
−1
−0.8
−0.6
−0.4
−0.2
0
0.2
0.4
0.6
0.8
1
time(s)
Amplitude
x(t)
x
rec
(t)
(d)
(c)
FIGURE 1.7
A real digital signal generation. (a) A/D converter: sampling. (b) A/D converter: quantization. (c) D/A converter:
zero-order hold. (d) D/A converter: lowpass filter.
the possibility of designing structurally induced linear-phase filters. On the other hand, an FIR filter
requires a higher number of multiplications, additions and storage elements when compared to their IIR
counterparts in order to satisfy a prescribed specification for the magnitude response.
There are many methods to design an FIR filter [19], ranging from the simple ones, such as the
window and frequency sampling methods, to more efficient and sophisticated, that rely on some
kind of optimization method [20–22]. The simple methods lead to solutions which are far from opti-
mum in most practical applications, in the sense that they either require higher order or cannot sat-
isfy prescribed specifications. A popular optimization solution is the minimax method based on the
Remez exchange algorithm, which leads to transfer functions with minimum order to satisfy prescribed
specifications.](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-39-320.jpg)
![12 CHAPTER 1 Introduction to Signal Processing Theory
0 500 1000 1500 2000
−120
−100
−80
−60
−40
−20
0
Frequency (Hz)
Magnitude
response
(dB)
Minimax method
WLS method
0 200 400 600 800 1200
−40
−30
−25
−20
−15
−10
−5
5
10
Frequency (Hz)
Magnitude
response
(dB)
Minimax method
WLS method
1 2
Αr
+ δp
− δp
f f
(b)
(a)
FIGURE 1.8
Typical magnitude responses of FIR filters: Minimax and WLS. (a) Magnitude response. (b) Specifications
for a lowpass filter using minimax design.
Nowadays, with the growing computational power, it is possible to design quite flexible FIR filters
by utilizing weighted least squares (WLS) solutions. These filters are particularly suitable for multirate
systems, since the resulting transfer functions can have decreasing energy in their stopband [2]. The WLS
method enables the design of filters satisfying prescribed specifications, such as the maximum deviation
in the passband and in part of the stopband, while minimizing the energy in a range of frequencies in the
stopband. As can be observed in Figure 1.8, for a given filter order, the minimax design leads to lower
maximum stopband attenuation, whereas the WLS solution leads to lower stopband energy. Note there
are solutions in between where some maximum values of stopband ripples can be minimized while
the energy of the remaining ones is also minimized. It is possible to observe that the WLS solution
does not satisfy the prescribed specifications given that it has the same filter order as the minimax
solution.
The typical transfer function of an IIR filter is given by
H(z) =
Y(z)
X(z)
=
M
l=0 bl z−l
1 +
N
i=1 ai z−i
,
which can be rewritten in a form explicitly showing the poles positions as
H(z) = H0
M
l=0 (1 − z−1zl)
N
i=0 (1 − z−1 pi )
= H0zN−M
M
l=0 (z − zl)
N
i=0 (z − pi )
.
IIR filters are usually designed by using well-established analog filter approximations, such as But-
terworth, Chebyshev, and elliptic methods (Figure 1.9) The prototype analog transfer function is then](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-40-320.jpg)
![1.01.7 Digital Filter Structures and Implementations 13
0 500 1000 1500 2000
−100
−90
−80
−70
−60
−50
−40
−30
−20
−10
0
Frequency (Hz)
Magnitude
response
(dB)
Elliptic approximation
Butherworth approx.
Chebyshev Type I approx.
FIGURE 1.9
Typical magnitude response of an IIR filter.
transformed into a digital transfer function by using an adequate transformation method. The most
widely used transformation methods are bilinear and the impulse invariance methods [2].
1.01.7 Digital filter structures and implementations
From a closer examination of the FIR and IIR transfer functions, one can infer that they can be realized
with three basic operators: adder, multiplier and delay (represented by z−1).
For example, Figure 1.10 shows a linear-phase FIR filter realization of odd order. Linear-phase filters
have symmetric or anti-symmetric impulse response and the symmetry should be exploited in order to
minimize the number of multipliers in the filter implementation, as illustrates Figure 1.10.
Figure 1.11 depicts a possible direct-form realization of an IIR transfer function. This realization
requires the minimum number of multiplications, adders and delays to implement the desired IIR transfer
function.
There are many alternative structures to implement the FIR and IIR filters. For example, IIR transfer
functions can be implemented as a product or as a summation of lower order IIR structures by starting
with their description as:
H(z) =
m
k=1
γ0k + γ1kz−1 + γ2kz−2
1 + m1kz−1 + m2kz−2
= h
p
0 +
m
k=1
γ
p
1k z + γ
p
2k
z2 + m1kz + m2k
,
respectively. The right choice for a filter realization must take into consideration some issues, such as:
modularity, quantization effects, design simplicity, power consumption, etc. [2]. Chapter 6 will address](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-41-320.jpg)
![14 CHAPTER 1 Introduction to Signal Processing Theory
z−1
z−1
z−1
z−1
z−1
× × ×
+
+ + +
x(n)
y(n)
b0 b1 b2
FIGURE 1.10
Odd-order linear-phase FIR filter structure with M = 5.
advanced methods for FIR and IIR filter realization and implementation, where elegant and creative
solutions are introduced in a clear manner. This chapter will also discuss strategies to implement digital
filters efficiently.
1.01.8 Multirate signal processing
In digital signal processing it is easy to change the sampling rate of the underlying sequences by a ratio
of integer values. This feature is highly desirable whenever we want to merge information of systems
employing distinct sampling rates. In these cases, those rates should be made compatible [6,9,10].
Systems that internally employ multiple sampling rates are collectively referred to as multirate
systems. In most cases, these systems are time-varying or, at most, periodically time-invariant.
The basic operations of the multirate systems are the decimation and interpolation. The right compo-
sition of these operations allows arbitrary rational sampling rate changes. If this sampling rate change
is arbitrary and non-rational, the only solution is to recover the bandlimited continuous-time signal x(t)
from its samples x(m), and then re-sample it with a different sampling rate, thus generating a distinct
discrete-time signal x(n).
If we assume that a discrete-time signal signal x(m) was generated from an continuous-time signal
y(t) with sampling period T1, so that x(m) = y(mT1), with m = . . . , 0, 1, 2 . . ., the sampling theorem
requires that y(t) should be bandlimited to the range [− π
T1
, π
T1
]. The sampled continuous-time signal is
given by:
y
(t) =
∞
m=−∞
x(m)δ(t − mT1).](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-42-320.jpg)
![1.01.8 Multirate Signal Processing 15
z−1
× +
× +
z−1
× +
z−1
× +
×
×
×
x(n) b0
b1
b2
− a1
− a2
− aN
bM
y(n)
FIGURE 1.11
Direct-Form IIR filter structure, for M = N.
The spectrum of the sampled signal is periodic with period 2π
T1
. The original continuous-time signal
y(t) can be recovered from y(t) through an ideal lowpass filter. This interpolation filter, whose impulse
response is denoted as h(t), has ideal frequency response H(ejω) as follows:
H(ejω
) =
1, ω ∈ [− π
T1
, π
T1
]
0, otherwise,
so that
y(t) = y
(t) ∗ h(t) =
1
T1
∞
m=−∞
x(m)sinc
π
T1
(t − mT1).
By re-sampling y(t) with period T2, we obtain the discrete-time signal in the new desired sampling
rate as follows:
x(n) =
1
T1
∞
m=−∞
x(m)sinc
π
T1
(nT2 − mT1),
where x(n) = y(nT2), with n = . . . , 0, 1, 2 . . ..
In this general equation governing sampling rate changes, there is no restriction on the values of T1
and T2. However, if T2 T1 and aliasing is to be avoided, the interpolation filter must have a zero gain
for ω /
∈ [− π
T2
, π
T2
].](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-43-320.jpg)
![16 CHAPTER 1 Introduction to Signal Processing Theory
In the case the sampling rate change corresponds to a ratio of integer numbers, all we need is simple
decimation and interpolation operations. The decimation operation, or down-sampling, consists of
retaining every Mth sample of a given sequence x(m). Since we are disposing samples from the original
sequence, either the signal is sufficiently limited in band or the signal has to be filtered by a lowpass
filter so that the decimated signal retains the useful information of the original signal. Decimation is a
time-varying operation, since, if x(m) is shifted by m0, the output signal will, in general, differ from
the unshifted output shifted by m0. Indeed, decimation is a periodically time-invariant operation, which
consist of an extra degree of freedom with respect to the traditional time-invariant systems.
The interpolation, or up-sampling, of a discrete-time signal x(m) by a factor of L consists of inserting
L − 1 zeros between each of its samples. In the frequency domain, the spectrum of the up-sampled
signal is periodically repeated. Given that the spectrum of the original discrete-time signal is periodic
with period 2π, the interpolated signal will have period 2π
L . Therefore, in order to obtain a smooth
interpolated version of x(m), the spectrum of the interpolated signal must have the same shape of the
spectrum of x(m). This can be obtained by filtering out the repetitions of the spectra beyond [−π
L , π
L ].
Thus, the up-sampling operation is generally followed by a lowpass filter. The interpolation is only
periodically time-invariant and does not entail any loss of information.
Figure 1.12 illustrates discrete-time signal decimation and interpolation operations. As can be
observed in Figure 1.12(a), the decimation factor of two widens the spectrum of the sampled sig-
nal in comparison to the new sampling rate. Figure 1.12(b) depicts the effect of increasing the sampling
rate of a given sequence by two, where it can be seen that the frequency contents of the original signal
repeats twice as as often and appears narrower in the new sampling rate.
Chapter 7 will further discuss the theory and practice of multirate signal processing. This chapter will
show how sophisticated real-life signal processing problems can be addressed starting from the basic
theory. In particular, the authors address the design of flexible communications systems incorporating
several advanced signal processing tools [4,23–25].
1.01.9 Filter banks and wavelets
In many applications it is desirable to decompose a wideband discrete-time signal into several non-
overlapping narrowband subsignals in order to be transmitted, quantized, or stored. In this case, each
narrow-band channel can have its sampling rate reduced, since the subband signals have lower bandwidth
than their originator signal. The signal processing tool that performs these tasks is the analysis filter
bank. The analysis filters, represented by the transfer functions Hi (z), for i = 0, 1, . . . , M − 1, consist
of a lowpass filter H0(z), bandpass filters Hi (z), for i = 1, 2, . . . , M −2, and a highpass filter HM−1(z).
Ideally, these filters have non-overlapping passbands, whereas their spectrum combination should cover
the overall spectrum of the input signal. The outputs of the analysis filters, denoted as xi (n), for i =
0, 1, . . . , M−1, have together M times the number of samples of the original signal x(n). However, since
each subband signal occupies a spectrum band M times narrower than the input signal, they can be deci-
mated so the number of samples in the subbands does not increase. Indeed, if the input signal is uniformly
split into subbands, we can decimate each xi (n) by a factor of L smaller or equal to M without generating
unrecoverable aliasing effects. Figure 1.13 shows the general configuration of a maximally (or critically)
decimated analysis filter, where the decimation factor is equal to the number of subbands, i.e., L = M.](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-44-320.jpg)
![1.01.9 Filter Banks and Wavelets 17
x̄(n)
m
x(m )
x̄(n)
M
0 1 2 12
X (ejω)
2π
−2π p
− p
X (ejω)
2π
−2π 2ωp
−2ωp
n
0 1 2 6
3 4 5
x(m )
(a) Signal decimation with M = 2.
2π
2
− 4π
2
3
x(m )
5 4
n
x(m )
x̄(n)
L
0 1 2 12
X (ejω)
2π
−2π ωp
−ωp
X (ej )
− 2π
2
ωp
2
−
ωp
2
x̄(n)
m
0 1 2 6
4π
2
(b) Signal interpolation with L = 2.
FIGURE 1.12
Signal decimation and interpolation. (a) Signal decimation with M = 2. (b) Signal interpolation with L = 2.
Whenever the input signal is split in subbands and decimated, if L ≤ M, it is always possible to
recover the input signal by properly designing the analysis filters in conjunction with the synthesis filters
Gi (z), for i = 0, 1, . . . , M −1. The synthesis filters are placed after interpolators and they have the task
of smoothing the up-sampled signals. Figure 1.14 illustrates the general configuration of the synthesis
filter bank. A key feature of the synthesis filter bank is to cancel out or reduce the aliasing effects.
If the signals in the subbands are not modified, the filter bank output y(n) can be a delayed copy
signal of x(n). The cascade connection of the analysis and synthesis filter banks satisfying this condition
is called a perfect reconstruction filter bank.
The cascade of an M-channel synthesis filter bank with an M-channel analysis filter bank gives
rise to a transmultiplex system. The transmultiplex model is widely used in communications to model
multiuser, multicarrier, and multiple access systems [4,23,24]. Chapter 7 will utilize multirate signal
processing and transmultiplexers to discuss a design procedure to be applied in cognitive radio.
Chapter 8 will present several methods for designing the analysis filters Hi (z) and the synthesis filters
Gi (z), such that perfect reconstruction, as well as other features, are met. This chapter will also discuss](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-45-320.jpg)
![This ebook is for the use of anyone anywhere in the United States
and most other parts of the world at no cost and with almost no
restrictions whatsoever. You may copy it, give it away or re-use it
under the terms of the Project Gutenberg License included with this
ebook or online at www.gutenberg.org. If you are not located in the
United States, you will have to check the laws of the country where
you are located before using this eBook.
Title: Misinforming a Nation
Author: Willard Huntington Wright
Release date: December 20, 2019 [eBook #60985]
Most recently updated: October 17, 2024
Language: English
Credits: Produced by WebRover, MWS and the Online Distributed
Proofreading Team at http://www.pgdp.net (This file was
produced from images generously made available by The
Internet Archive)
*** START OF THE PROJECT GUTENBERG EBOOK MISINFORMING A
NATION ***](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-55-320.jpg)
![to find such depth of feeling compatible with the power of minutely
and publicly recording what is felt.” Loti, like de Musset, lacked that
prudish restraint which is so admirable a virtue in English writers.
Daudet, in a short and very inadequate biography, is written down
as an imitator of Dickens; and in Anatole France’s biography, which
is shorter than Marryat’s or Mrs. Oliphant’s, no adequate indication
of his genius is given.
Zola is treated with greater unfairness than perhaps any other
French author. Zola has always been disliked in England, and his
English publisher was jailed by the guardians of British morals. But it
is somewhat astonishing to find to what lengths this insular
prejudice has gone in the Encyclopædia Britannica. Zola’s biography,
which is shorter than Mrs. Humphry Ward’s, is written by a former
Accountant General of the English army, and contains adverse
comment because he did not idealize “the nobler elements in human
nature,” although, it is said, “his later books show improvement.”
Such scant treatment of Zola reveals the unfairness of extreme
prejudice, for no matter what the nationality, religion, or taste of the
critic, he must, in all fairness, admit that Zola is a more important
and influential figure in modern letters than Mrs. Humphry Ward.
In the biography of George Sand we learn that “as a thinker,
George Eliot is vastly [sic] superior; her knowledge is more
profound, and her psychological analysis subtler and more scientific.”
Almost nothing is said of Constant’s writings; and in the mere half-
column sketch of Huysmans there are only a few biographical facts
with a list of his books. Of Stendhal there is practically no criticism;
and Coppée “exhibits all the defects of his qualities.” René Bazin
draws only seventeen lines—a bare record of facts; and Édouard Rod
is given a third of a column with no criticism.
Despite the praise given Victor Hugo, his biography, from a critical
standpoint, is practically worthless. In it there is no sense of critical
proportion: it is a mere panegyric which definitely states that Hugo
was greater than Balzac. This astonishing and incompetent praise is
accounted for when we discover that it was written by Swinburne](https://image.slidesharecdn.com/31150-250313074641-6ddace0a/85/Signal-processing-theory-and-machine-learning-1st-Edition-Diniz-83-320.jpg)
Signal processing theory and machine learning 1st Edition Diniz
- 1. Instant Ebook Access, One Click Away – Begin at ebookgate.com Signal processing theory and machine learning 1st Edition Diniz https://ebookgate.com/product/signal-processing-theory-and- machine-learning-1st-edition-diniz/ OR CLICK BUTTON DOWLOAD EBOOK Get Instant Ebook Downloads – Browse at https://ebookgate.com Click here to visit ebookgate.com and download ebook now
- 2. Instant digital products (PDF, ePub, MOBI) available Download now and explore formats that suit you... Digital Signal Processing System Analysis and Design 1st Edition Paulo S. R. Diniz https://ebookgate.com/product/digital-signal-processing-system- analysis-and-design-1st-edition-paulo-s-r-diniz/ ebookgate.com Machine Learning in Bio Signal Analysis and Diagnostic Imaging 1st Edition Nilanjan Dey https://ebookgate.com/product/machine-learning-in-bio-signal-analysis- and-diagnostic-imaging-1st-edition-nilanjan-dey/ ebookgate.com Bayesian Signal Processing Classical Modern and Particle Filtering Methods Adaptive and Learning Systems for Signal Processing Communications and Control Series 1st Edition James V. Candy https://ebookgate.com/product/bayesian-signal-processing-classical- modern-and-particle-filtering-methods-adaptive-and-learning-systems- for-signal-processing-communications-and-control-series-1st-edition- james-v-candy/ ebookgate.com Information Fusion in Signal and Image Processing Digital Signal and Image Processing 1st Edition Isabelle Bloch https://ebookgate.com/product/information-fusion-in-signal-and-image- processing-digital-signal-and-image-processing-1st-edition-isabelle- bloch/ ebookgate.com
- 3. Probability random variables and random processes theory and signal processing applications 1st Edition Shynk https://ebookgate.com/product/probability-random-variables-and-random- processes-theory-and-signal-processing-applications-1st-edition-shynk/ ebookgate.com The Digital Signal Processing Handbook Video Speech and Audio Signal Processing 2nd Edition Vijay K. Madisetti https://ebookgate.com/product/the-digital-signal-processing-handbook- video-speech-and-audio-signal-processing-2nd-edition-vijay-k- madisetti/ ebookgate.com Introduction to Machine Learning Second Edition Adaptive Computation and Machine Learning Ethem Alpaydin https://ebookgate.com/product/introduction-to-machine-learning-second- edition-adaptive-computation-and-machine-learning-ethem-alpaydin/ ebookgate.com Introduction to IoT with Machine Learning and Image Processing using Raspberry Pi 1st Edition Shrirang Ambaji Kulkarni https://ebookgate.com/product/introduction-to-iot-with-machine- learning-and-image-processing-using-raspberry-pi-1st-edition-shrirang- ambaji-kulkarni/ ebookgate.com DSP for MATLAB and LabVIEW I Fundamentals of Discrete Signal Processing Synthesis Lectures on Signal Processing Isen https://ebookgate.com/product/dsp-for-matlab-and-labview-i- fundamentals-of-discrete-signal-processing-synthesis-lectures-on- signal-processing-isen/ ebookgate.com
- 6. Academic Press is an imprint of Elsevier The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, UK 225 Wyman Street, Waltham, MA 02451, USA First edition 2014 Copyright © 2014 Elsevier Ltd. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted in any form or by any means electronic, mechanical, photocopying, recording or otherwise without the prior written permission of the publisher. Permissions may be sought directly from Elsevier’s Science & Technology Rights Department in Oxford, UK: phone (+44) (0) 1865 843830; fax (+44) (0) 1865 853333; email: permissions@elsevier.com. Alternatively you can submit your request online by visiting the Elsevier web site at http://elsevier.com/locate/permissions, and selecting Obtaining permission to use Elsevier material. Notice No responsibility is assumed by the publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, indepen- dent verification of diagnoses and drug dosages should be made. Library of Congress Cataloging in Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-396502-8 For information on all Elsevier publications visit our website at www.store.elsevier.com Printed and bound in Poland. 14 15 16 17 10 9 8 7 6 5 4 3 2 1
- 7. xxvii Signal Processing at Your Fingertips! Let us flash back to the 1970s when the editors-in-chief of this e-reference were graduate students. One of the time-honored traditions then was to visit the libraries several times a week to keep track of the latest research findings. After your advisor and teachers, the librarians were your best friends. We visited the engineering and mathematics libraries of our Universities every Friday afternoon and poured over the IEEE Transactions, Annals of Statistics, the Journal of Royal Statistical Society, Biometrika, and other journals so that we could keep track of the recent results published in these journals. Another ritual that was part of these outings was to take sufficient number of coins so that papers of interest could be xeroxed. As there was no Internet, one would often request copies of reprints from authors by mailing postcards and most authors would oblige. Our generation maintained thick folders of hard- copies of papers. Prof. Azriel Rosenfeld (one of RC’s mentors) maintained a library of over 30,000 papers going back to the early 1950s! Another fact to recall is that in the absence of Internet, research results were not so widely dis- seminated then and even if they were, there was a delay between when the results were published in technologically advanced western countries and when these results were known to scientists in third world countries. For example, till the late 1990s, scientists in US and most countries in Europe had a lead time of at least a year to 18months since it took that much time for papers to appear in journals after submission. Add to this the time it took for the Transactions to go by surface mails to various libraries in the world. Scientists who lived and worked in the more prosperous countries were aware of the progress in their fields by visiting each other or attending conferences. Let us race back to 21st century! We live and experience a world which is fast changing with rates unseen before in the human history. The era of Information and Knowledge societies had an impact on all aspects of our social as well as personal lives. In many ways, it has changed the way we experience and understand the world around us; that is, the way we learn. Such a change is much more obvious to the younger generation, which carries much less momentum from the past, compared to us, the older generation. A generation which has grew up in the Internet age, the age of Images and Video games, the age of IPAD and Kindle, the age of the fast exchange of information. These new technologies comprise a part of their “real” world, and Education and Learning can no more ignore this reality. Although many questions are still open for discussions among sociologists, one thing is certain. Electronic publishing and dissemination, embodying new technologies, is here to stay. This is the only way that effective pedagogic tools can be developed and used to assist the learning process from now on. Many kids in the early school or even preschool years have their own IPADs to access information in the Internet. When they grow up to study engineering, science, or medicine or law, we doubt if they ever will visit a library as they would by then expect all information to be available at their fingertips, literally! Another consequence of this development is the leveling of the playing field. Many institutions in lesser developed countries could not afford to buy the IEEE Transactions and other journals of repute. Even if they did, given the time between submission and publication of papers in journals and the time it took for the Transactions to be sent over surface mails, scientists and engineers in lesser developed countries were behind by twoyears or so.Also, most libraries did not acquire the proceedings of confer- ences and so there was a huge gap in the awareness of what was going on in technologically advanced Introduction
- 8. xxviii Introduction countries. The lucky few who could visit US and some countries in Europe were able to keep up with the progress in these countries. This has changed. Anyone with an Internet connection can request or download papers from the sites of scientists. Thus there is a leveling of the playing field which will lead to more scientist and engineers being groomed all over the world. The aim of Online Reference for Signal Processing project is to implement such a vision. We all know that asking any of our students to search for information, the first step for him/her will be to click on the web and possibly in the Wikipedia. This was the inspiration for our project. To develop a site, related to the Signal Processing, where a selected set of reviewed articles will become available at a first “click.” However, these articles are fully refereed and written by experts in the respected topic. Moreover, the authors will have the “luxury” to update their articles regularly, so that to keep up with the advances that take place as time evolves. This will have a double benefit. Such articles, besides the more classical material, will also convey the most recent results providing the students/researchers with up-to-date information. In addition, the authors will have the chance of making their article a more “permanent” source of reference, that keeps up its freshness in spite of the passing time. The other major advantage is that authors have the chance to provide, alongside their chapters, any multimedia tool in order to clarify concepts as well as to demonstrate more vividly the performance of various methods, in addition to the static figures and tables. Such tools can be updated at the author’s will, building upon previous experience and comments. We do hope that, in future editions, this aspect of this project will be further enriched and strengthened. In the previously stated context, the Online Reference in Signal Processing provides a revolutionary way of accessing, updating and interacting with online content. In particular, the Online Reference will be a living, highly structured, and searchable peer-reviewed electronic reference in signal/image/video Processing and related applications, using existing books and newly commissioned content, which gives tutorial overviews of the latest technologies and research, key equations, algorithms, applications, standards, code, core principles, and links to key Elsevier journal articles and abstracts of non-Elsevier journals. The audience of the Online Reference in Signal Processing is intended to include practicing engi- neers in signal/image processing and applications, researchers, PhD students, post Docs, consultants, and policy makers in governments. In particular, the readers can be benefited in the following needs: • To learn about new areas outside their own expertise. • To understand how their area of research is connected to other areas outside their expertise. • To learn how different areas are interconnected and impact on each other: the need for a “helicopter” perspective that shows the “wood for the trees.” • To keep up-to-date with new technologies as they develop: what they are about, what is their potential, what are the research issues that need to be resolved, and how can they be used. • To find the best and most appropriate journal papers and keeping up-to-date with the newest, best papers as they are written. • To link principles to the new technologies. The Signal Processing topics have been divided into a number of subtopics, which have also dic- tated the way the different articles have been compiled together. Each one of the subtopics has been coordinated by an AE (Associate Editor). In particular:
- 9. xxix Introduction 1. Signal Processing Theory (Prof. P. Diniz) 2. Machine Learning (Prof. J. Suykens) 3. DSP for Communications (Prof. N. Sidiropulos) 4. Radar Signal Processing (Prof. F. Gini) 5. Statistical SP (Prof. A. Zoubir) 6. Array Signal Processing (Prof. M. Viberg) 7. Image Enhancement and Restoration (Prof. H. J. Trussell) 8. Image Analysis and Recognition (Prof. Anuj Srivastava) 9. Video Processing (other than compression), Tracking, Super Resolution, Motion Estimation, etc. (Prof. A. R. Chowdhury) 10. Hardware and Software for Signal Processing Applications (Prof. Ankur Srivastava) 11. Speech Processing/Audio Processing (Prof. P. Naylor) 12. Still Image Compression 13. Video Compression We would like to thank all the Associate Editors for all the time and effort in inviting authors as well as coordinating the reviewing process. The Associate Editors have also provided succinct summaries of their areas. The articles included in the current editions comprise the first phase of the project. In the second phase, besides the updates of the current articles, more articles will be included to further enrich the existing number of topics. Also, we envisage that, in the future editions, besides the scientific articles we are going to be able to include articles of historical value. Signal Processing has now reached an age that its history has to be traced back and written. Last but not least, we would like to thank all the authors for their effort to contribute in this new and exciting project. We earnestly hope that in the area of Signal Processing, this reference will help level the playing field by highlighting the research progress made in a timely and accessible manner to anyone who has access to the Internet. With this effort the next breakthrough advances may be coming from all around the world. The companion site for this work: http://booksite.elsevier.com/9780124166165 includes multimedia files (Video/Audio) and MATLAB codes for selected chapters. Rama Chellappa Sergios Theodoridis
- 10. xxxi RamaChellappareceivedtheB.E.(Hons.)degreeinElectronicsandCommunication Engineering from the University of Madras, India in 1975 and the M.E. (with Distinction) degree from the Indian Institute of Science, Bangalore, India in 1977. He received the M.S.E.E. and Ph.D. Degrees in Electrical Engineering from Purdue University, West Lafayette, IN, in 1978 and 1981, respectively. During 1981– 1991, he was a faculty member in the department of EE-Systems at University of Southern California (USC). Since 1991, he has been a Professor of Electrical and Computer Engineering (ECE) and an affiliate Professor of Computer Science at University of Maryland (UMD), College Park. He is also affiliated with the Center for Automation Research, the Institute for Advanced Computer Studies (Permanent Member) and is serving as the Chair of the ECE department. In 2005, he was named a Minta Martin Professor of Engineering. His current research interests are face recognition, clustering and video summariza- tion, 3D modeling from video, image and video-based recognition of objects, events and activities, dictionary-based inference, compressive sensing, domain adaptation and hyper spectral processing. Prof. Chellappa received an NSF Presidential Young Investigator Award, four IBM Faculty Development Awards, an Excellence in Teaching Award from the School of Engineering at USC, and two paper awards from the International Association of Pattern Recognition (IAPR). He is a recipient of the K.S. Fu Prize from IAPR. He received the Society, Technical Achievement, and Meritorious Service Awards from the IEEE Signal Processing Society. He also received the Technical Achievement and Meritorious Service Awards from the IEEE Computer Society. At UMD, he was elected as a Distinguished Faculty Research Fellow, as a Distinguished Scholar-Teacher, received an Outstanding Innovator Award from the Office of Technology Commercialization, and an Outstanding GEMSTONE Mentor Award from the Honors College. He received the Outstanding Faculty Research Award and the Poole and Kent Teaching Award for Senior Faculty from the College of Engineering. In 2010, he was recognized as an Outstanding ECE by Purdue University. He is a Fellow of IEEE, IAPR, OSA, and AAAS. He holds four patents. Prof. Chellappa served as the Editor-in-Chief of IEEE Transactions on PatternAnalysis and Machine Intelligence. He has served as a General and Technical Program Chair for several IEEE international and national conferences and workshops. He is a Golden Core Member of the IEEE Computer Society and served as a Distinguished Lecturer of the IEEE Signal Processing Society. Recently, he completed a two-year term as the President of the IEEE Biometrics Council. About the Editors
- 11. xxxii About the Editors Sergios Theodoridis is currently Professor of Signal Processing and Communications in the Department of Informatics and Telecommunications of the University of Athens. His research interests lie in the areas of Adaptive Algorithms and Communications, Machine Learning and Pattern Recognition, Signal Processing for Audio Processing and Retrieval. He is the co-editor of the book “Efficient Algorithms for Signal Processing and System Identification,” Prentice Hall 1993, the co-author of the best selling book “Pattern Recognition,” Academic Press, 4th ed. 2008, the co-author of the book “Introduction to Pattern Recognition: A MATLAB Approach,” Academic Press, 2009, and the co-author of three books in Greek, two of them for the Greek Open University. He is Editor-in-Chief for the Signal Processing Book Series, Academic Press and for the E-Reference Signal Processing, Elsevier. He is the co-author of six papers that have received best paper awards including the 2009 IEEE Computational Intelligence Society Transactions on Neural Networks Outstanding paper Award. He has served as an IEEE Signal Processing Society Distinguished Lecturer. He was Otto Monstead Guest Professor, Technical University of Denmark, 2012, and holder of the Excellence Chair, Department of Signal Processing and Communications, University Carlos III, Madrid, Spain, 2011. He was the General Chairman of EUSIPCO-98, the Technical Program co-Chair for ISCAS-2006 and ISCAS-2013, and co-Chairman and co-Founder of CIP-2008 and co-Chairman of CIP-2010. He has served as President of the European Association for Signal Processing (EURASIP) and as member of the Board of Governors for the IEEE CAS Society. He currently serves as member of the Board of Governors (Member-at-Large) of the IEEE SP Society. He has served as a member of the Greek National Council for Research and Technology and he was Chairman of the SP advisory committee for the Edinburgh Research Partnership (ERP). He has served as Vice Chairman of the Greek Pedagogical Institute and he was for 4years member of the Board of Directors of COSMOTE (the Greek mobile phone operating company). He is Fellow of IET, a Corresponding Fellow of the Royal Society of Edinburgh (RSE), a Fellow of EURASIP, and a Fellow of IEEE.
- 12. xxxiii Paulo S. R. Diniz was born in Niterói, Brazil. He received the Electronics Engineering degree (Cum Laude) from the Federal University of Rio de Janeiro (UFRJ) in 1978, the M.Sc. degree from COPPE/UFRJ in 1981, and the Ph.D. from Concordia University, Montreal, P.Q., Canada, in 1984, all in electrical engineering. Since 1979 he has been with the Department of Electronic Engineering (the undergraduate department) UFRJ. He has also been with the Program of Electrical Engineering (the graduate studies department), COPPE/UFRJ, since 1984, where he is presently a Professor. He served as Undergraduate Course Coordinator and as Chairman of the Graduate Department. He has also received the Rio de Janeiro State Scientist award, from the Governor of Rio de Janeiro state. From January 1991 to July 1992, he was a visiting Research Associate in the Department of Electri- cal and Computer Engineering of University of Victoria, Victoria, B.C., Canada. He also holds a Do- cent position at Helsinki University of Technology (now Aalto University). From January 2002 to June 2002, he was a Melchor Chair Professor in the Department of Electrical Engineering of University of Notre Dame, Notre Dame, IN, USA. His teaching and research interests are in analog and digital signal processing, adaptive signal processing, digital communications, wireless communications, multirate systems, stochastic processes, and electronic circuits. He has published several refereed papers in some of these areas and wrote the books ADAPTIVE FILTERING: Algorithms and Practical Implementa- tion, Springer, NY, Fourth Edition 2013, DIGITAL SIGNAL PROCESSING: System Analysis and Design, Cambridge University Press, Cambridge, UK, Second Edition 2010 (with E. A. B. da Silva and S. L. Netto), and Block Transceivers: OFDM and Beyond, Morgan & Claypool, Fort Collins, CO, 2012, (with W. A. Martins and M. V. S. Lim). He has served as General co-Chair of ISCAS2011 and Technical Program Chair of the 1995 MWSCAS both held in Rio de Janeiro, Brazil. He was also the Technical Program co-Chair of SPAWC2008. He has been on the technical committee of several inter- national conferences including ICASSP, ISCAS, ICECS, EUSIPCO, and MWSCAS. He has served as the Vice President for region 9 of the IEEE Circuits and Systems Society and as Chairman of the DSP technical committee of the same Society. He is also a Fellow of IEEE (for fundamental contributions to the design and implementation of fixed and adaptive filters and Electrical Engineering Education). He has served as Associate Editor for the following Journals: IEEE Transactions on Circuits and Systems II: Analog and Digital Signal Processing from 1996 to 1999, IEEE Transactions on Signal Processing from 1999 to 2002, and the Circuits, Systems and Signal Processing Journal from 1998 to 2002. He was a Distinguished Lecturer of the IEEE Circuits and Systems Society for the year 2000–2001. In 2004 he served as Distinguished Lecturer of the IEEE Signal Processing Society and received the 2004 Education Award of the IEEE Circuits and Systems Society. Section Editors Section 1
- 13. xxxiv Section Editors Johan A.K. Suykens received the master degree in ElectroMechanical Engineering and the Ph.D. degree inApplied Sciences from the Katholieke Universiteit Leuven, in 1989 and 1995, respectively. In 1996 he has been a Visiting Postdoctoral Researcher at the University of California, Berkeley. He has been a Postdoctoral Researcher with the Fund for Scientific Research FWO Flanders and is currently a Professor (Hoogleraar) with KU Leuven. He is author of the books “Artificial Neural Networks for Modeling and Control of Non-linear Systems” (Kluwer Academic Publishers) and “Least Squares Support Vector Machines” (World Scientific), co-author of the book “Cellular Neural Networks, Multi-Scroll Chaos and Synchronization” (World Scientific), and editor of the books “Nonlinear Modeling: Advanced Black-Box Techniques” (Kluwer Academic Publishers) and “Advances in Learning Theory: Methods, Models and Applications” (IOS Press). He is a Senior IEEE member and has served as associate edi- tor for the IEEE Transactions on Circuits and Systems (1997–1999 and 2004–2007) and for the IEEE Transactions on Neural Networks (1998–2009). He received an IEEE Signal Processing Society 1999 Best Paper (Senior) Award and several Best Paper Awards at International Conferences. He is a recipi- ent of the International Neural Networks Society INNS 2000 Young Investigator Award for significant contributions in the field of neural networks. He has served as a Director and Organizer of the NATO Advanced Study Institute on Learning Theory and Practice (Leuven 2002), as a program co-chair for the International Joint Conference on Neural Networks 2004 and the International Symposium on Non-linear Theory and its Applications 2005, as an organizer of the International Symposium on Synchronization in Complex Networks 2007 and a co-organizer of the NIPS 2010 workshop on Tensors, Kernels and Machine Learning. He has been awarded an ERC Advanced Grant 2011. Section 2
- 14. xxxv CHAPTER 1 Isabela Ferrão Apolinário was born in Brasília, Brazil. She is currently studying Electronics Engineering in the Federal University of Rio de Janeiro (UFRJ) and will soon join an M.Sc. course in the same university. Since 2010 she has been a member of the Audio Processing Group (GPA) from the Signal Processing Laboratory (LPS) and has been studying audio pro- cessing. She has worked as an undergraduate teaching assistance in Calculus II and Digital Signal Processing. She is currently a student member of the IEEE Circuits and Systems So- ciety. Her main interest is in digital audio signal processing, more specifically in Psychoacoustics and Audio 3D. CHAPTER 2 José Antonio Apolinário Junior graduated from the Military Academy of Agulhas Negras (AMAN), Brazil, in 1981 and received the B.Sc. degree from the Military Institute of Engineering (IME), Brazil, in 1988, the M.Sc. degree from the University of Brasília (UnB), Brazil, in 1993, and the D.Sc. degree from the Federal University of Rio de Janeiro (COPPE/UFRJ), Rio de Janeiro, Brazil, in 1998, all in electrical engineering. He is currently an Adjoint Professor with the Department of Electrical Engineering, IME, where he has already served as the Head of Department and as the Vice-Rector for Study and Research. He was a Visiting Professor at the em Escuela Politécnica del Ejército (ESPE), Quito, Ecuador, from 1999 to 2000 and a Visiting Researcher and twice a Visiting Professor at Helsinki University of Technology (HUT), Finland, in 1997, 2004, and 2006, respectively. His research interests comprise many aspects of linear and nonlinear digital signal processing, including adaptive filtering, speech, and array processing. He has organized and has been the first Chair of the Rio de Janeiro Chapter of the IEEE Communications Society. He has recently edited the book “QRDRLS Adaptive Filtering” (Springer, 2009) and served as the Finance Chair of IEEE ISCAS 2011 (Rio de Janeiro, Brazil, May 2011). He is a senior member of the IEEE. Carla Liberal Pagliari received the Ph.D. degree in electronic systems engi- neer-ing from the University of Essex, UK, in 2000. Since 1993 she has been with the Department of Electrical Engineering at the Military Institute of Engineering (IME), Rio de Janeiro, Brazil. She took part in the team that worked toward the development of the Brazilian Digital Television System. Her research interests include image processing, digital television, image and video coding, stereoscopic and multiview systems, and computer vision. She is a senior member of the IEEE and served as the Local Arrangements Chair Authors Biography
- 15. xxxvi Authors Biography of IEEE ISCAS 2011 (Rio de Janeiro, Brazil, May 2011). She is currently an Associate Editor of the journal Multidimensional Systems and Signal Processing. CHAPTER 3 Leonardo Gomes Baltar received his B.Sc. and M.Sc. degrees in Electrical Engineering from the Federal University of Rio de Janeiro (UFRJ), Brazil, in 2004 and 2006, respectively. Since 2006, he is with the Chair of Circuit Theory and Signal Processing at the Technical University of Munich (TUM), Germany, as a research and teaching assistant pursuing a Ph.D. degree. His activities are mainly in the area of digital signal processing techniques applied to wired and wireless communications systems including multirate systems, filter banks, block transforms, digital filters design, and efficient processing structures. He has been involved in the European FP7 project PHYDYAS and in projects with industrial partners. Josef A. Nossek received his Dipl.-Ing. and Dr. techn. degrees from the Technical University of Vienna, Austria, in 1974 and 1980, respectively. He was for 15years with Siemens AG, Munich, in the field of communications, including the position of Head of Radio Systems Design for digital commu- nications. In 1989 he joined the Technical University of Munich (TUM), Germany, where he is full Professor and Head of the Chair for Circuit Theory and Signal Processing. He is a Fellow of IEEE since 1993 for his contribu- tions to the design of discrete-time networks and technical leadership in the development of radio communication systems. He was the Editor-in-Chief of the IEEE Transactions on Circuits and Systems (1993–1995). He was President-Elect, President, and Past President of the IEEE Circuits and Systems Society (2001/02/03). In 1998 he received the Innovations Award of the Vodafone Foundation for excellent research in mobile communica- tions. In 2008 he received the Education Award of the IEEE Circuits and Systems Society. In 2011 he received the IEEE Guillemin-Cauer Best Paper Award. Since 2009 he is member of acatech (National Academy of Science and Engineering). He is author or co-author of numerous publications and has given a number of invited plenary lectures. He was the President of VDE (2007–2008), the German Association of Electrical, Electronics and Information Engineering, and Vice-President of VDE (2009–2010). CHAPTER 4 Luiz W. P. Biscainho was born in Rio de Janeiro, Brazil, in 1962. He received the Electronics Engineering degree (magna cum laude) from the EE (now Poli) at Universidade Federal do Rio de Janeiro (UFRJ), Brazil, in 1985, and the M.Sc. and D.Sc. degrees in Electrical Engineering from the COPPE at UFRJ in 1990 and 2000, respectively. Having worked in the telecommu- nication industry between 1985 and 1993, Dr. Biscainho is now Associate
- 16. xxxvii Authors Biography Professor at the Department of Electronics and Computer Engineering (DEL) of Poli and the Electrical Engineering Program (PEE) of COPPE (serving as Academic Coordinator in 2010), at UFRJ. His research area is digital signal processing, particularly audio processing and adap- tive systems. He is currently a member of the IEEE (Institute of Electrical and Electronics Engineers), the AES (Audio Engineering Society), the SBrT (Brazilian Telecommunications Society), and the SBC (Brazilian Computer Society). CHAPTER 5 Håkan Johansson received the Master of Science degree in computer science and the Licentiate, Doctoral, and Docent degrees in Electronics Systems from Linkoping University, Sweden, in 1995, 1997, 1998, and 2001, respec- tively. During 1998 and 1999 he held a postdoctoral position at Signal Processing Laboratory, Tampere University of Technology, Finland. He is currently Professor in Electronics Systems at the Department of Electrical Engineering of Linkoping University. His research encompasses design and implementation of efficient and flexible signal processing (SP) systems, mainly for communication applications. During the past decade, he has developed many different SP algorithms for various purposes, including filtering, sampling rate conversion, signal reconstruction, and parameter estimation. He has developed new estimation and compensation algorithms for errors in analog circuits such as compensation of mismatch errors in time-interleaved analog-to-digital converters and mixers. He is one of the founders of the company Signal Processing Devices Sweden AB that sells this type of advanced signal processing. He is the author or co-author of four books and some 160 international journal and conference papers. He is the co-author of three papers that have received best paper awards and he has authored one invited paper in IEEE Transactions and four invited chapters. He has served as Associate Editor for IEEE Trans. on Circuits and Systems I and II, IEEE Trans. Signal Processing, and IEEE Signal Processing Letters, and he is currently an Area Editor of the Elsevier Digital Signal Processing journal and a member of the IEEE Int. Symp. Circuits. Syst. DSP track committee. CHAPTER 6 Lars Wanhammar was born in Vansbro, Sweden, on August 19, 1944. He has received the following degrees: Teknisk magister (teknisk fysik) in 1970, civilin- genjör in 1980, teknisk doktor in 1981, and docent in 1986, in electrical engi- neering from Linköping University, Sweden. During 1964–1970 he worked at Televerket (Royal Swedish Telephone Board), Division of Communication and during 1970–1971 as a Lecturer at the technical college in Norrköping. Since 1971 he has been working at Linköping University, Department of Electrical Engineering, Division of Applied Electronics, as Assistant, Research Assistant, and from 1982 as Associate Professor (universitetslektor) and from 1997 as full Professor and Head of the Division of Electronics Systems, at the Department of Electrical Engineering, Linköping University, Sweden. From 2011 he is currently Professor Emeritus.
- 17. xxxviii Authors Biography In addition, from 1995 to 2004 he worked as Adjunct Professor at the Norwegian Institute of Technology (NTNU) at the departments of Physical Electronics and Telecommunications. His research interests are primary theory and design of digital signal processing and telecom- munication systems, particularly analog and digital filters, and discrete transforms as well as com- putational properties of DSP algorithms, bit-serial, digit-serial and distributed arithmetic, CAD tools, and globally asynchronous locally synchronous techniques for ULSI. He is the author or co-author of five books in analog and digital filters, one in parallel process- ing in industrial real-time applications, and one in DSP integrated circuits. Ya Jun Yu received the B.Sc. and M.Eng. degrees in biomedical engineering from Zhejiang University, Hangzhou, China, in 1994 and 1997, respectively, and the Ph.D. degree in electrical and computer engineering from the National University of Singapore, Singapore, in 2004. From 1997 to 1998, she was a Teaching Assistant with Zhejiang University. She joined the Department of Electrical and Computer Engineering, National University of Singapore as a Post Master Fellow in 1998 and remained in the same department as a Research Engineer until 2004. She joined the Temasek Laboratories at Nanyang Technological University as a Research Fellow in 2004. Since 2005, she has been with the School of Electrical and Electronic Engineering, Nanyang Technological University, Singapore, where she is currently an Assistant Professor. Her research interests include digital signal processing and VLSI circuits and systems design. He has served as an associate editor for Circuits Systems and Signal Processing and IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II since 2009 and 2010, respectively. CHAPTER 7 Fred Harris holds the Signal Processing Chair of the Communication Systems and Signal Processing Institute at San Diego State University where since 1967 he has taught courses in Digital Signal Processing and Communication Systems. He holds 20 patents on digital receiver and DSP technology and lectures throughout the world on DSP applications. He consults for organi- zations requiring high-performance, cost-effective DSP solutions. He is an adjunct member of the IDA-Princeton Center for Communications Research. Fred Harris has written over 200 journal and conference papers, the most well known being his 1978 paper “On the use of Windows for Harmonic Analysis with the Discrete Fourier Transform”. He is the author of the book Multirate Signal Processing for Communication Systems and he has contributed to a number of other books on DSP applications including the “Source Coding” chapter in Bernard Sklar’s 1988 book, Digital Communications and the “Multirate FIR Filters for Interpolation and Resampling” and the “Time Domain Signal Processing with the DFT” chapters in Doug Elliot’s 1987 book Handbook of Digital Signal Processing, and “A most Efficient Digital Filter: The Two-Path Recursive All- Pass Filter” Chapter and the “Ultra Low Phase Noise DSP Oscillator” Chapter in Rick Lyon’s 2012 book Streamlining Digital Signal Processing. He is also co-author of the book Software Radio Sampling Rate Selection, Design and Synchronization.
- 18. xxxix Authors Biography In 1990 and 1991 he was the Technical and then the General Chair of the Asilomar Conference on Signals, Systems, and Computers and was Technical Chair of the 2003 Software Defined Radio Conference and of the 2006 Wireless Personal Multimedia Conference. He became a Fellow of the IEEE in 2003, cited for contributions of DSP to communications systems. In 2006 he received the Software Defined Radio Forum’s “Industry Achievement Award”. His paper at the 2006 SDR conference was selected for the best paper award as was his paper at the 2010 Autotestcon conference and again his paper at the 2011 Wireless Personal Mobile Communications Conference and once again the 2011 SDR conference. He is the former Editor- in-Chief of the Elsevier DSP Journal. The spelling of my name with all lower case letters is a source of distress for typists and spell checkers. A child at heart, I collect toy trains and old slide-rules. Elettra Venosa received the “Laurea” (BS/MS) degree, summa cum laude, in Electrical Engineering in January 2007 from Seconda Università degli Studi di Napoli, Italy. From January 2007 to November 2007 she was a researcher at the Italian National Inter-University Consortium for Telecommunications. In January 2011, she received the Ph.D. in Telecommunication/DSP from Seconda Università degli Studi di Napoli. From June 2008 to September 2008 she worked as a project manager for Kiranet –ICT Research Center – to develop an advanced radio identification system for avionics, in collabo- ration with the Italian Center for Aerospace Research (CIRA). From April 2009 to September 2009 she worked as associate researcher at Communications and Signal Processing Laboratory (CSPL) in the Department of Electrical and Computer Engineering at Drexel University, Philadelphia, where she worked on Sparse Sampling Techniques for Software Defined Radio Receivers. From January 2011 to July 2012 she worked as principal system engineer in IQ-Analog, CA, developing algorithms for digital correction in TI-ADCs. From August 2012 to December 2013 she was associate researcher in Qualcomm, CA. Her focus was to improve the current commercial modem architectures. Currently, she is work- ing, as a postdoctoral researcher, on Multirate Signal Processing for Software Defined Radios in the Department of Electrical Engineering at San Diego State University, San Diego, CA, USA where she also teaches graduate and undergraduate courses. She is also working as a soft- ware defined radio engineer in Space Micro, CA. She is the author of more than 40 scientific publications on SDR and of the book “Software Radio: Sampling Rate Selection Design and Synchronization”. Xiaofei Chen received the Bachelor’s degree in Electrical Engineering in June 2006 from Xi’an University of Posts & Telecommunications, China. In December 2008, he received the Master’s degree at Electrical & Computer Engineering department, San Diego State University, USA. In March 2009, he joined the Joint Doctoral Program between San Diego State University and University of California, San Diego. His current research interests are in the area of multirate signal processing and software defined radio.
- 19. xl Authors Biography CHAPTER 8 Trac D. Tran (S’94–M’98–SM’08) received the B.S. and M.S. degrees from the Massachusetts Institute of Technology, Cambridge, in 1993 and 1994, respectively, and the Ph.D. degree from the University of Wisconsin, Madison, in 1998, all in electrical engineering. In July 1998, he joined the Department of Electrical and Computer Engineering, The Johns Hopkins University, Baltimore, MD, where he cur- rently holds the rank of Professor. His research interests are in the field of sig- nal processing, particularly in sparse representation, sparse recovery, sampling, multirate systems, filter banks, transforms, wavelets, and their applications in signal analysis, compression, processing, and communications. His pioneering research on integer- coefficient transforms and pre-/post-filtering operators has been adopted as critical components of Microsoft Windows Media Video 9 and JPEG XR—the latest international still-image compres- sion standard ISO/IEC 29199–2. He is currently a regular consultant for the US Army Research Laboratory, Adelphi, MD. He was the codirector (with Prof. J. L. Prince) of the 33rd Annual Conference on Information Sciences and Systems (CISS’99), Baltimore, in March 1999. In the summer of 2002, he was an ASEE/ONR Summer Faculty Research Fellow at the Naval Air Warfare Center Weapons Division (NAWCWD), China Lake, CA. He has served as Associate Editor of the IEEE TRANSACTIONS ON SIGNAL PROCESSING as well as the IEEE TRANSACTIONS ON IMAGE PROCESSING. He was a former member of the IEEE Technical Committee on Signal Processing Theory and Methods (SPTM TC) and is a current member of the IEEE Image Video and Multidimensional Signal Processing (IVMSP) Technical Committee. He received the NSF CAREER award in 2001, the William H. Huggins Excellence in Teaching Award from The Johns Hopkins University in 2007, and the Capers and Marion McDonald Award for Excellence in Mentoring and Advising in 2009. CHAPTER 9 Yufang Bao has been an Assistant Professor in the Department of Mathematics and Computer Science at UNC Fayetteville State University (UNCFSU) since 2007. She is also a scholar of the Center of Defense and Homeland Security (CDHS) at UNCFSU. She received her first Ph.D. degree in prob- ability/statistics from Beijing Normal University (BNU), Beijing, China, and her second Ph.D. degree in Electrical Engineering from North Carolina State University (NCSU), Raleigh, NC. Her research focus was in probability/ statistics. She subsequently directed her research into the area of applying mathematics in signal/image processing and analysis. Her contributions in signal/image processing included algorithm development that bridged stochastic diffusion and multi-scale wavelet theory with scale space analysis methods for image denoising and segmenta- tion. Between 2002 and 2007, she has worked at the VA Center, UCSF, CA and then at the University of Miami, School of Medicine, FL, both as a research scientist focusing on statistical image reconstruction in Frequency domain with MR spectroscopy imaging, and with parallel
- 20. xli Authors Biography MR image reconstruction using mathematical modeling. Currently, her research interests are in applying mathematics to statistical digital signal/image processing and analysis, mathematical modeling, and their applications. Hamid Krim (ahk@ncsu.edu) received his degrees in ECE from University of Washington and Northeastern University. He was a Member of Technical Staff at AT&T Bell Labs, where he has conducted research and development in the areas of telephony and digital communication systems/subsystems. Following an NSF postdoctoral fellowship at Foreign Centers of Excellence, LSS/ University of Orsay, Paris, France, he joined the Laboratory for Information and Decision Systems, Massachusetts Institute of Technology, Cambridge, MA as a Research Scientist and where he was performing and supervising research. He is presently Professor of Electrical Engineering in the ECE Department, North Carolina State University, Raleigh, leading the Vision, Information and Statistical Signal Theories and Applications group. His research interests are in statistical signal and image analysis and mathematical modeling with a keen emphasis on applied problems in classification and rec- ognition using geometric and topological tools. CHAPTER 10 Lisandro Lovisolo was born in Neuquen, Argentina, but considers himself brazilian. He received the Electronics Engineering degree from Universidade Federal do Rio de Janeiro, in 1999, the M.Sc. degree in Electrical Engineering in 2001, and the D.Sc. degree in Electrical Engineering both from Universidade Federal do Rio de Janeiro (COPPE/UFRJ). Since 2003 he has been with the Department of Electronics and Communications Engineering (the under- graduate department), UERJ. He has also been with the Postgraduate in Electronics Program, since 2008. His research interests lie in the fields of digital signal and image processing and communications. Eduardo A. B. da Silva was born in Rio de Janeiro, Brazil. He received the Electronics Engineering degree from Instituto Militar de Engenharia (IME), Brazil, in 1984, the M.Sc. degree in Electrical Engineering from Universidade Federal do Rio de Janeiro (COPPE/UFRJ) in 1990, and the Ph.D. degree in Electronics from the University of Essex, England, in 1995. In 1987 and 1988 he was with the Department of Electrical Engineering at Instituto Militar de Engenharia, Rio de Janeiro, Brazil. Since 1989 he has been with the Department of Electronics Engineering (the undergraduate department), UFRJ. He has also been with the Department of Electrical Engineering (the graduate studies department), COPPE/UFRJ, since 1996. His research interests lie in the fields of digital signal and image processing, especially signal compression, digital television, wavelet transforms, mathematical morphology, and applications to telecommunications.
- 21. xlii Authors Biography CHAPTER 11 Suleyman Serdar Kozat received the B.S. degree with full scholarship and high honors from Bilkent University, Turkey. He received the M.S. and Ph.D. degrees in Electrical and Computer Engineering from University of Illinois at Urbana Champaign, Urbana, IL, in 2001 and 2004, respectively. After graduation, he joined IBM Research, T.J. Watson Research Center, Yorktown, NY as a Research Staff Member in Pervasive Speech Technologies Group, where he focused on problems related to statistical signal process- ing and machine learning. While doing his Ph.D., he was also working as a Research Associate at Microsoft Research, Redmond, WA, in Cryptography and Anti-Piracy Group. He holds several patent applications for his works performed in IBM Research and Microsoft Research. Currently, he is an Assistant Professor at the electrical and electronics engineering department, Koc University, Turkey. He coauthored more than 50 papers in refereed high impact journals and conference proceedings and has several patent applications. Overall, his research interests include intelligent systems, adaptive filtering for smart data analyt- ics, online learning, and machine learning algorithms for signal processing. He has been serving as an Associate Editor for the IEEE Transactions on Signal Processing and he is a Senior Member of the IEEE. He has been awarded IBM Faculty Award by IBM Research in 2011, Outstanding Faculty Award by Koc University in 2011, Outstanding Young Researcher Award by the Turkish National Academy of Sciences in 2010, ODTU Prof. Dr. Mustafa N. Parlar Research Encouragement Award in 2011 and holds Career Award by the Scientific Research Council of Turkey, 2009. He has won several scholarships and medals in international and national science and math competitions. Andrew C. Singer received the S.B., S.M., and Ph.D. degrees, all in Electrical Engineering and Computer Science, from the Massachusetts Institute of Technology. Since 1998, he has been on the faculty of the Department of Electrical and Computer Engineering at the University of Illinois at Urbana- Champaign, where he is currently a Professor in the ECE department and the Coordinated Science Laboratory. During the academic year 1996, he was a Postdoctoral Research Affiliate in the Research Laboratory of Electronics at MIT. From 1996 to 1998, he was a Research Scientist at Sanders, A Lockheed Martin Company in Manchester, New Hampshire, where he designed algo- rithms, architectures, and systems for a variety of DOD applications. His research interests include signal processing and communication systems. He was a Hughes Aircraft Masters Fellow and was the recipient of the Harold L. Hazen Memorial Award for excellence in teaching in 1991. In 2000, he received the National Science Foundation CAREER Award, in 2001 he received the Xerox Faculty Research Award, and in 2002 he was named a Willett Faculty Scholar. He has served as an Associate Editor for the IEEE Transactions on Signal Processing and is a member of the MIT Educational Council, and of Eta Kappa Nu and Tau Beta Pi. He is a Fellow of the IEEE. In 2005, he was appointed as the Director of the Technology Entrepreneur Center (TEC) in the College of Engineering. He also co-founded Intersymbol Communications, Inc., a venture-funded
- 22. xliii Authors Biography fabless semiconductor IC company, based in Champaign Illinois. A developer of signal processing enhanced chips for ultra-high speed optical communications systems, Intersymbol was acquired by Finisar Corporation (NASD:FNSR) in 2007. He serves on the board of directors of a number of technology companies and as an expert witness for electronics, communications, and circuit- related technologies. CHAPTER 12 Vítor H. Nascimento was born in São Paulo, Brazil. He obtained the B.S. and M.S. degrees in Electrical Engineering from the University of São Paulo in 1989 and 1992, respectively, and the Ph.D. degree from the University of California, Los Angeles, in 1999. From 1990 to 1994 he was a Lecturer at the University of São Paulo, and in 1999 he joined the faculty at the same school, where he is now an Associate Professor. One of his papers received the 2002 IEEE SPS Best Paper Award. He served as an Associate Editor for the IEEE Signal Processing Letters from 2003 to 2005, for the IEEE Transactions on Signal Processing from 2005 to 2008, and for the EURASIP Journal on Advances in Signal Processing from 2006 to 2009, and was a member of the IEEE-SPS Signal Processing Theory and Methods Technical Committee from 2007 to 2012. Since 2010 he is the chair of the São Paulo SPS Chapter. His research interests include signal processing theory and applications, robust and nonlinear estimation, and applied linear algebra. Magno T. M. Silva was born in São Sebastião do Paraíso, Brazil. He received the B.S., M.S., and Ph.D. degrees, all in electrical engineering, from Escola Politécnica, University of São Paulo, São Paulo, Brazil, in 1999, 2001, and 2005, respectively. From February 2005 to July 2006, he was an Assistant Professor at Mackenzie Presbyterian University, São Paulo, Brazil. Since August 2006, he has been with the Department of Electronic Systems Engineering at Escola Politécnica, University of São Paulo, where he is currently an Assistant Professor. From January to July 2012, he worked as a Postdoctoral Researcher at Universidad Carlos III de Madrid, Leganés, Spain. His research interests include linear and non-linear adaptive filtering. CHAPTER 14 Ambuj Tewari is with the Department of Statistics, University of Michigan, Ann Arbor. He has served on senior program committees of the confer- ences on Algorithmic Learning Theory (ALT), Conference on Learning Theory (COLT), and Neural Information Processing Systems (NIPS). His work has received both the student paper award (2005) and the best paper award (2011) at COLT. He received his M.A. in Statistics (2005) and Ph.D. in Computer Science (2007) from the University of California at Berkeley where his advisor was Peter Bartlett. He was a research Assistant Professor in Toyota Technological Institute at Chicago (2008–2010), an Assistant Professor
- 23. xliv Authors Biography (part-time) in the Department of Computer Science, University of Chicago (2008–2010), and a postdoctoral fellow in the Institute for Computational Engineering and Sciences, University of Texas at Austin (2010–2012). He has also been a Visiting Researcher at Microsoft Research, Redmond. Peter L. Bartlett is with the Division of Computer Science and Department of Statistics, University of California at Berkeley, and with the School of Mathematical Sciences, Queensland University of Technology. He is the co- authorofthebook“LearninginNeuralNetworks:TheoreticalFoundations.”He has served as Associate Editor of the journals Machine Learning, Mathematics of Control Signals and Systems, the Journal of Machine Learning Research, the Journal of Artificial Intelligence Research, and the IEEE Transactions on Information Theory. He was awarded the Malcolm McIntosh Prize for Physical Scientist of the Year in Australia for his work in statistical learning theory. He was a Miller Institute Visiting Research Professor in Statistics and Computer Science at U.C. Berkeley in Fall 2001, and a Fellow, Senior Fellow, and Professor in the Research School of Information Sciences and Engineering at the Australian National University’s Institute for Advanced Studies (1993–2003). CHAPTER 15 Barbara Hammer received her Ph.D. in Computer Science in 1995 and her venia legendi in Computer Science in 2003, both from the University of Osnabrueck, Germany. From 2000 to 2004, she was a leader of the junior research group “Learning with Neural Methods on Structured Data” at University of Osnabrueck before accepting an offer as professor for Theoretical Computer Science at Clausthal University of Technology, Germany, in 2004. Since 2010, she is holding a professorship for Theoretical Computer Science for Cognitive Systems at the CITEC cluster of excellence at Bielefeld University, Germany. Several research stays have taken her to Italy, UK, India, France, the Netherlands, and the USA. Her areas of expertise include hybrid systems, self-organizing maps, clustering, and recurrent networks as well as applications in bioin- formatics, industrial process monitoring, or cognitive science. She is leading the task force “Data Visualization and Data Analysis” of the IEEE CIS Technical Committee on Data Mining, and the Fachgruppe Neural Networks of the GI. CHAPTER 16 John Shawe-Taylor is a Professor at the Department of Computer Science, University College London (UK). His main research area is Statistical Learning Theory, but his contributions range from Neural Networks, to Machine Learning, to Graph Theory. He has published
- 24. xlv Authors Biography over 150 research papers. He obtained a Ph.D. in Mathematics at Royal Holloway, University of London in 1986. He subsequently completed an M.Sc. in the Foundations of Advanced Information Technology at Imperial College. He was promoted to Professor of Computing Science in 1996. He moved to the University of Southampton in 2003 to lead the ISIS research group. He was appointed the Director of the Center for Computational Statistics and Machine Learning at University College, London in July 2006. He has coordinated a number of Europeanwide projects investigating the theory and practice of Machine Learning, including the NeuroCOLT projects. He is currently the scientific coordinator of a Framework VI Network of Excellence in Pattern Analysis, Statistical Modeling and Computational Learning (PASCAL) involving 57 partners. Shiliang Sun received the B.E. degree in automatic control from the Depart- ment of Automatic Control, Beijing University of Aeronautics and Astronautics in 2002, and the Ph.D. degree in pattern recognition and intel- ligent systems from the State Key Laboratory of Intelligent Technology and Systems, Department of Automation, Tsinghua University, Beijing, China, in 2007. In 2004, he was entitled Microsoft Fellow. Currently, he is a Professor at the Department of Computer Science and Technology and the Founding Director of the Pattern Recognition and Machine Learning Research Group, East China Normal University. From 2009 to 2010, he was a Visiting Researcher at the Department of Computer Science, University College London, working within the Center for Computational Statistics and Machine Learning. He is a member of the PASCAL (Pattern Analysis, Statistical Modelling and Computational Learning) network of excellence, and on the editorial boards of multiple international journals. His research interests include machine learning, pattern recognition, computer vision, natural language processing, and intelligent trans- portation systems. CHAPTER 17 Konstantinos Slavakis received the M.E. and Ph.D. degrees in electrical and electronic engineering from Tokyo Institute of Technology (TokyoTech), Tokyo, Japan, in 1999 and 2002, respectively. For the period from 2004 to 2006, he was with TokyoTech as a JSPS PostDoc, and from 2006 to 2007, he was a Postdoc in the Department of Informatics and Telecommunications, University of Athens, Greece. From 2007 to 2012, he served as an Assistant Professor at the Department of Informatics and Telecommunications, University of Peloponnese, Tripolis, Greece. Currently, he is a Research Associate at the University of Minnesota, Digital Technology Center.
- 25. xlvi Authors Biography He serves as an Associate and Area Editor of the IEEE Transactions on Signal Processing. His research interests include applications of convex analysis and computational algebraic geometry to signal processing, machine learning, array, and multidimensional systems problems. Pantelis Bouboulis received the M.Sc. and Ph.D. degrees in informatics and telecommunications from the National and Kapodistrian University of Athens, Greece, in 2002 and 2006, respectively. From 2007 till 2008, he served as an Assistant Professor in the Department of Informatics and Telecommunications, University of Athens. Since 2008, he teaches mathe- matics in Greek High Schools. His current research interests lie in the areas of machine learning, fractals, wavelets, and image processing. CHAPTER 18 Franz Pernkopf received his M.Sc. (Dipl. Ing.) degree in Electrical Engineering at Graz University of Technology, Austria, in summer 1999. He earned a Ph.D. degree from the University of Leoben, Austria, in 2002. In 2002 he was awarded the Erwin Schrödinger Fellowship. He was a Research Associate in the Department of Electrical Engineering at the University of Washington, Seattle, from 2004 to 2006. Currently, he is Associate Professor at the Laboratory of Signal Processing and Speech Communication, Graz University of Technology, Austria. His research interests include machine learning, dis- criminative learning, graphical models, feature selection, finite mixture mod- els, and image- and speech processing applications. Robert Peharz received his M.Sc. degree in Telematics at Graz University of Technology (TUG) in 2010. He currently pursues his Ph.D. studies at the SPSC Lab, TUG. His research interests include probabilistic graphical mod- els, sparse coding, nonnegative matrix factorization, and machine learning in general, with applications to signal processing, audio engineering, and com- puter vision. Sebastian Tschiatschek received the B.Sc. degree and M.Sc. degree in Electrical Engineering at Graz University of Technology (TUG) in 2007 and 2010, respectively. He conducted his Master thesis during a one-year stay at ETH Zürich, Switzerland. Currently, he is with the Signal Processing and Speech Communication Laboratory at TUG where he is pursuing the Ph.D. degree. His research interests include Bayesian networks, informa- tion theory in conjunction with graphical models and statistical pattern recognition.
- 26. xlvii Authors Biography CHAPTER 19 A. Taylan Cemgil (M’04) received his Ph.D. (2004) from SNN, Radboud University Nijmegen, the Netherlands. Between 2004 and 2008 he worked as a postdoctoral researcher at Amsterdam University and the Signal Processing and Communications Laboratory, University of Cambridge, UK. He is cur- rently an Associate Professor of Computer Engineering at Bogazici University, Istanbul, Turkey. He is a member of the IEEE MLSP Technical Committee and an Associate Editor of IEEE Signal Processing Letters and Digital Signal Processing. His research interests are in Bayesian statistical methods and infer- ence, machine learning, and audio signal processing. CHAPTER 20 Dao Lam is a Ph.D. Candidate at Missouri University of Science and Technology, Rolla, MO. He received the B.S. degree from Post and Telecommunications Institute of Technology, Ho Chi Minh, Vietnam in 2003. He got his M.S. from Waseda University, Japan in 2008. His research interests are image pro- cessing, robotics, supervised and unsupervised learning, and computational intelligence. Donald Wunsch is the Mary K. Finley Missouri Distinguished Professor at Missouri University of Science & Technology (Missouri S&T). Earlier employers were: Texas Tech University, Boeing, Rockwell International, and International Laser Systems. His education includes: Executive MBA— Washington University in St. Louis, Ph.D., Electrical Engineering—University of Washington (Seattle), M.S., Applied Mathematics (same institution), B.S., Applied Mathematics—University of New Mexico, and Jesuit Core Honors Program, Seattle University. Key research contributions are: Clustering; Adaptive Resonance and Reinforcement Learning architectures, hardware and applications; Neurofuzzy regression; Traveling Salesman Problem heuristics; Robotic Swarms; and Bioinformatics. He is an IEEE Fellow and previous INNS President, INNS Fellow and Senior Fellow 07—present, and served as IJCNN General Chair, and on several Boards, including the St. Patrick’s School Board, IEEE Neural Networks Council, International Neural Networks Society, and the University of Missouri Bioinformatics Consortium. He has produced 16 Ph.D. recip- ients in Computer Engineering, Electrical Engineering, and Computer Science; has attracted over $8million in sponsored research; and has over 300 publications including nine books. His research has been cited over 6000 times.
- 27. xlviii Authors Biography CHAPTER 21 Andrzej Cichocki received the M.Sc. (with honors), Ph.D. and Dr.Sc. (Habilitation) degrees, all in electrical engineering, from Warsaw University of Technology in Poland. Since 1972, he has been with the Institute of Theory of Electrical Engineering, Measurement and Information Systems, Faculty of Electrical Engineering at the Warsaw University of Technology, where he obtained a title of a full Professor in 1995. He spent several years at University Erlangen-Nuerenberg in Germany, at the Chair of Applied and Theoretical Electrical Engineering directed by Professor Rolf Unbehauen, as an Alexander-von-Humboldt Research Fellow and Guest Professor. In 1995–1997 he was a team leader of the laboratory for Artificial Brain Systems, at Frontier Research Program RIKEN (Japan), in the Brain Information Processing Group. He is currently senior team leader and the head of the Cichocki Laboratory for Advanced Brain Signal Processing, at RIKEN Brain Science Institute in Japan. CHAPTER 22 Xueyuan Zhou received his Ph.D. degree in computer science from the University of Chicago in 2011. His research interests include statistical machine learning theory and application in non- linear high dimensional data. Mikhail Belkin is an Associate Professor in the Department of Computer Science and Engineering at the Ohio State University. He received his Ph.D. degree in mathematics from the University of Chicago in 2003. His research interests include a range of theoretical questions concerning the computational and statistical limits of learning and mathematical foundations of learning structure from data. Yannis Kopsinis received the B.Sc. degree from the Department of Informatics and Telecommunications, University of Athens, Greece, in 1998 and his Ph.D. degree in 2003 from the same department. From January 2004 to December 2005 he has been a research fellow with the Institute for Digital Communications, School of Engineering and Electronics, the University of Edinburgh, UK. From January 2006 to September 2009 he was a senior researcher in the same department. From January 2012 to April 2013 he was a Ramon Y Cajal Fellow in the School of Applied Physics, University of Granada, Spain. He currently holds a Marie Curie Intra-European fellowship
- 28. xlix Authors Biography on sparse online learning. His current research interests include adaptive signal processing, time- frequency analysis, and compressed sensing. Konstantinos Slavakis received the M.E. and Ph.D. degrees in elec- trical and electronic engineering from Tokyo Institute of Technology (TokyoTech), Tokyo, Japan, in 1999 and 2002, respectively. For the period of 2004–2006, he was with TokyoTech as a JSPS PostDoc, and from 2006 to 2007, he was a Postdoc in the Department of Informatics and Telecommunications, University of Athens, Greece. From 2007 to 2012, he served as an Assistant Professor at the Department of Informatics and Telecommunications, University of Peloponnese, Tripolis, Greece. Currently, he is a Research Associate at the University of Minnesota, Digital Technology Center. He serves as an Associate and Area Editor of the IEEE Transactions on Signal Processing. His research interests include applications of convex analysis and computational algebraic geometry to signal processing, machine learning, array, and multidimensional systems problems. CHAPTER 24 José C. Principe is currently a Distinguished Professor of Electrical and Biomedical Engineering at the University of Florida, Gainesville, USA. He is BellSouth Professor and Founder and Director of the University of Florida Computational Neuro-Engineering Laboratory (CNEL). He is an IEEE fellow and AIMBE fellow, and is the past Editor-in-Chief of the IEEE TRANSACTIONS ON BIOMEDICAL ENGINEERING, past President of the International Neural Network Society, former Secretary of the Technical Committee on Neural Networks of the IEEE Signal Processing Society, and a former member of the Scientific Board of the Food and Drug Administration. He is involved in biomedical signal processing, in particular, the electroencephalogram (EEG) and the modeling and applications of adaptive systems. Badong Chen received his Ph.D. degree in Computer Science and Technology from Tsinghua University, Beijing, China, in 2008. He was a Postdoctoral Associate at the University of Florida Computational NeuroEngineering Laboratory (CNEL) during the period October, 2010 to September, 2012. He is currently a Professor at the Institute of Artificial Intelligence and Robotics, Xi’an Jiaotong University, Xi’an, China. His research interests are in statistical signal processing, information theoretic learning, online kernel learning, and their applications in cognition and neuroscience.
- 29. l Authors Biography Luis G. Sanchez Giraldo was born in 1983 in Manizales, Colombia. He received the B.S. in electronics engineering and M.Eng. in industrial automation from Universidad Nacional de Colombia in 2005 and 2008, respectively, and his Ph.D. in electrical and computer engineering from University of Florida in 2012. Between 2004 and 2008, he was appointed as a research assistant at the Control and Digital Signal Processing Group (GCPDS) at Universidad Nacional de Colombia. During his Ph.D. studies he worked as a research assistant at the Computational Neuro-Engineering Laboratory (CNEL) at University of Florida. His main research interests are in machine learning and signal processing. CHAPTER 25 Enes Makalic was born in 1980. He received the Bachelor of Computer Science (Honors) degree in 2002 and the Ph.D. degree in 2007, both from Monash University, Australia. His research interests include information theoretic model selection using Minimum Message Length (MML) and Minimum Description Length (MDL) theories of statistical inference. He currently holds a Postdoctoral position with The University of Melbourne, Australia. Daniel F. Schmidt was born in 1980. He received the Bachelor of Digital Systems (Honors) degree in 2002 and the Ph.D. degree in 2008, both from Monash University, Australia. His research interests are primarily information theoretic approaches to statistical inference, model selection, and estimation. He currently holds a Postdoctoral position with The University of Melbourne, Australia. Abd-Krim Seghouane received his Ph.D. degree from University of Paris sud (Paris XI) in 2003. His research interests are within statistical signal and image processing. He is currently a Senior Research Fellow within the department of Electrical and Electronic Engineering, The University of Melbourne. CHAPTER 26 George Tzanetakis is an Associate Professor and Canada Research Chair in Computer Analysis of Audio and Music at the Department of Computer Science with cross-listed appointments in ECE and Music at the University of Victoria, Canada. In 2011 he was a visiting scientist at Google Research. He received his Ph.D. in Computer Science at Princeton University in 2002 and was a Post-Doctoral fellow at Carnegie Mellon University in 2002–2003. His research spans all stages of audio content analysis such as feature extrac- tion, segmentation, classification with specific emphasis on music information retrieval. He is also the primary designer and developer of Marsyas an open
- 30. li Authors Biography source framework for audio processing with specific emphasis on music information retrieval applications. His pioneering work on musical genre classification received a IEEE signal pro- cessing society young author award and is frequently cited. More recently he has been explor- ing new interfaces for musical expression, music robotics, computational ethnomusicology, and computer-assisted music instrument tutoring. These interdisciplinary activities combine ideas from signal processing, perception, machine learning, sensors, actuators, and human-computer interaction with the connecting theme of making computers better understand music to create more effective interactions with musicians and listeners.
- 31. 1 CHAPTER Introduction to Signal Processing Theory Isabela F. Apolinário and Paulo S.R. Diniz Program of Electrical Engineering and the Department of Electronics and Computer Engineering, COPPE/Poli, Universidade Federal do Rio de Janeiro, Brazil 1.01.1 Introduction Signal processing is a key area of knowledge that finds applications in virtually all aspects of modern life. Indeed the human beings are employing signal processing tools for centuries without realizing it [1]. In present days the younger generation might not be able to understand how one can live without carrying a mobile phone, traveling long distances without an almost self piloted airplane, exploring other planets without human presence, and utilizing a medical facility without a wide range of diagnostic and intervention equipments. Signal processing consists of mapping or transforming information bearing signals into another form of signals at the output, aiming at some application benefits. This mapping defines a continuous or analog system if it involves functions representing the input and output signals. On the other hand, the system is discrete or digital if its input and output signals are represented by sequences of numbers. Signal processing theory is very rich and as a reward it has been finding applications in uncountable areas, among which we can mention bioengineering, communications, control, surveillance, environ- ment monitoring, oceanography, and astronomy, just to mention a few. In particular, the enormous advance in digital integrated circuit technology, responsible for the computing and information revo- lutions that we are so used today, enabled the widespread use of Digital Signal Processing Systems. These systems allowed advances in fields such as: speech, audio, image, video, multirate processing, besides in several applications of wireless communications and digital control. This chapter is an overview of the classical signal processing tools whose details are encountered in numerous textbooks such as: [1–15]. The topics treated in the following chapters entail the description of many signal processing tools, usually not covered in the classical textbooks available in the market. As a result, we believe that the readers can benefit from reading many of these chapters, if not all, in order to deepen and widen their current knowledge as well as to start exploiting new ideas. 1.01.2 Continuous-time signals and systems The processing of signals starts by accessing the type of information we want to deal with. Many signals originating from nature are continuous in time and as such, the processing involved must include analog signal acquisition systems followed by the implementation of a continuous-time system if the processing Academic Press Library in Signal Processing. http://dx.doi.org/10.1016/B978-0-12-396502-8.00001-2 © 2014 Elsevier Ltd. All rights reserved. 3
- 32. 4 CHAPTER 1 Introduction to Signal Processing Theory x (t) y (t) t t H{·} FIGURE 1.1 Continuous-time system. remains analog all the way. Alternatively, assuming the original continuous-time signal contains limited spectral information, we can sample it in order to generate a sequence representing the continuous-time signal unambiguously1. In this latter form one can benefit from the advanced digital integrated circuit technology to process the signal. However, many real life applications still includes continuous-time signal processing at least at the acquisition and actuator phases of the processing. A continuous-time system maps an analog input signal represented by x(t) into an output signal represented by y(t), as depicted in Figure 1.1. A general representation of this mapping is y(t) = H{x(t)}, where H{·} denotes the operation performed by the continuous-time system. If the system is linear and time-invariant (LTI), as will be described in Chapter 2, there are several tools to describe the mapping features of the system. Typically, a general description of the systems consists of a differential equation which in turn can be solved and analyzed by employing the Laplace transform. A key feature of the LTI systems is their full representation through their impulse responses. The behavior of a nonlinear system is more complicated to analyze since it can not be fully characterized by its impulse response. Another important tools to deal with the continuous-time signals which are periodic and non- periodic are the Fourier series and Fourier transform, respectively. As will be discussed in Chapter 2, the Fourier and Laplace transforms are closely related, being both essential tools to describe the behavior of continuous-time signals when applied to LTI systems. The Laplace transform is suitable to represent a time-domain non-periodic function of a continuous and real (time) variable, resulting in a frequency-domain non-periodic function of a continuous and complex frequency variable. That is, X(s) = ∞ −∞ x(t)e−st dt ⇐⇒ x(t) = 1 2π eσt ∞ −∞ X(σ + jω)ejωt dω. The Fourier transform is employed to represent a time-domain non-periodic function of a continuous and real (time) variable with a frequency-domain non-periodic function of a continuous and imaginary 1As will be seen in Chapter 5, we can completely reconstruct a continuous-time band-limited signal x(t) from its sampled version x(n) if the sampling frequency is chosen correctly.
- 33. 1.01.3 Discrete-Time Signals and Systems 5 frequency variable. The Fourier transform is described by X() = ∞ −∞ x(t)e−jt dt ⇐⇒ x(t) = 1 2π ∞ −∞ X()ejt d. The representation of a time-domain periodic function of a continuous and real (time) variable is performed by the Fourier series. In the frequency-domain, the representation consists of a non-periodic function of a discrete and integer frequency variable, as following described: X(k) = 1 T T 0 x(t)e−j(2π/T )kt dt ⇐⇒ x(t) = ∞ k=−∞ X(k)ej(2π/T )kt . 1.01.3 Discrete-time signals and systems A discrete-time signal is represented by a sequence of numbers and can be denoted as x(n) with n ∈ Z, where Z is the set of integer numbers. The samples x(n) might represent numerically the amplitude of a continuous-time signal sample at every T seconds or can be originated from a discrete information. In many applications, the sampling period T represents the time interval between two acquired samples of the continuous-time signal. However, in other situations it might represent the distance between two sensorsoftwopixelsofanimage,ortheseparationbetweentwoantennas,justtomentionafewexamples. Discrete-time systems map input sequences into output sequences. A general representation of this mapping is y(n) = H{x(n)}, where H{·} denotes the operation performed by the discrete-time system. Figure 1.2 depicts the input- to-output mapping of sequences. According to the properties of H{·}, the discrete-time system might be LTI. This way, it benefits from a number of analysis and design tools such as frequency-domain representations. However, there are many applications employing nonlinear, time-varying, and even non-causal systems [2,5,7,8,11,12,16]. If the system is linear and time-invariant, as will be described in Chapter 2, there are several tools to describe the mapping features of the system. Typically, a general description of discrete-time sys- tems through a difference equation can be solved and analyzed by employing the z transform. Also a x(n) y(n) n n H{·} FIGURE 1.2 Discrete-time signal representation.
- 34. 6 CHAPTER 1 Introduction to Signal Processing Theory discrete-time LTI system is fully described by its impulse response, whereas a nonlinear system can not be fully characterized by its impulse response. The z transform is the key tool to represent a time-domain non-periodic function of a discrete and integer variable through a frequency-domain non-periodic function of a continuous and complex frequency variable. That is, X(z) = ∞ n=−∞ x(n)z−n ⇐⇒ x(n) = 1 2πj C X(z)zn−1 dz. The discrete-time Fourier transform (DTFT) represents a time-domain non-periodic function of a discrete and integer variable by a periodic function of a continuous frequency variable in the frequency domain as follows: X(ejω ) = ∞ n=−∞ x(n)e−jωn ⇐⇒ x(n) = 1 2π ∞ −∞ X(ejω )ejωn dω. Finally, the discrete Fourier transform (DFT) is the right tool to represent a time-domain periodic function of a discrete and integer variable through a periodic function of a discrete and integer frequency variable in the frequency domain. The DFT is defined as X(k) = N−1 n=0 x(n)e−j(2π/N)kn ⇐⇒ x(n) = 1 N N−1 k=0 X(k)ej(2π/N)kn . The DFT plays a key role in digital signal processing since it represents a finite-length sequence in the time domain through a finite-length sequence in the frequency domain. Since both domains utilize sequences, this feature makes the DFT a natural choice for time-frequency representation of information in a digital computer. A little mental exercise allows us to infer that the DFT is the perfect representation in the frequency domain of a finite-length sequence in the time domain, if we interpret the latter as a period of a periodic infinite-length sequence. In addition, by employing appropriate zero-padding in two finite-length sequences, the product of their DFTs represents the DFT of their linear convolution (see Chapter 3), turning the DFT a valuable tool for LTI system implementation. There is a plethora of applications for the DFT in signal processing and communications and some of them can be accessed in the references [2–13]. Often, whenever a DSP system is LTI, it can be described through a difference equation as follows: N i=0 ai y(n − i) + M l=0 bl x(n − l) = 0. The above description is suitable for implementation in digital computers. The difference equation represents a causal LTI system if its auxiliary conditions correspond to its initial conditions and those are zeros [2]. Assuming a0 = 1, the above equation can be rewritten as y(n) = − N i=1 ai y(n − i) + M l=0 bl x(n − l).
- 35. 1.01.4 Random Signals and Stochastic Processes 7 This class of system has infinite impulse response (i.e., y(n) = 0 when n → ∞) and, as such, it is called IIR system or filter. The nonrecursive system generates the output signal from past input samples, that is, y(n) = M l=0 bl x(n − l). In this case, the resulting system has finite impulse response and is known as FIR systems or filters. 1.01.4 Random signals and stochastic processes In most practical applications, we have to deal with signals that can not be described by a function of time, since their waveforms follow a random pattern [17,18]. Take, for example, the thermal noise inherent to any material. Despite the lack of exact knowledge of the signal values, it is possible to analyze and extract information the signals contain by employing the mathematical tools available to deal with random signals. Chapter 4 will present a compact and yet clarifying review of random signals and stochastic processes which are crucial to understand several concepts that will be discussed in the more advanced chapters. The theory starts by defining a random variable as a mapping of the result of a random process experiment onto the set of numbers. A random variable X is a function that assigns a number x to every outcome of a random process experiment. An example is given in Figure 1.3(a), in which each outcome of a throw of a dice, numbers 1–6, corresponds to a determined value, x1–x6, respectively. x1 X x 1 x3 x2 x4 t x 1 ( t ) x 2 ( t ) x 3 ( t ) x n ( t ) t t t {X } (b) (a) FIGURE 1.3 Examples of random variable and stochastic process. (a) Random variable. (b) Stochastic process.
- 36. 8 CHAPTER 1 Introduction to Signal Processing Theory x (t) y (t) t t H{·} FIGURE 1.4 Filtering of a random signal. The stochastic process is a rule to describe the time evolution of the random variable depending on the random process experiment, whereas the set of all experimental outcomes is its domain known as the ensemble. An example is given in Figure 1.3(b), in which, at any time instant t0, the set formed by the output samples of the stochastic process {X}, {x1(t0), x2(t0), x3(t0), . . . , xn(t0)}, represents a random variable. A single outcome xi (t) of {X} is a random signal. Since random signals do not have a precise description of their waveforms, we have to characterize them either via measured statistics or through a probabilistic model. In general, the first- and second-order statistics (mean and variance, respectively) are sufficient for characterization of the stochastic process, particularly due to the fact that these statistics are suitable for measurements. As an illustration, Figure 1.4 shows a random signal as an input of a highpass filter. We can observe that the output signal is still random, but clearly shows a faster changing behavior given that the low frequency contents of the input signal were attenuated by the highpass filter. 1.01.5 Sampling and quantization A digital signal processing system whose signals originated from continuous-time sources includes several building blocks, namely: an analog-to-digital (A/D) converter; a digital signal processing (DSP) system; a digital-to-analog (D/A) converter; and a lowpass filter. Figure 1.5 illustrates a typical digital signal processing setup where: • The A/D converter produces a set of samples in equally spaced time intervals, which may retain the information of the continuous-time signal in the case the latter is band limited. These samples are converted into a numeric representation, in order to be applied to the DSP system. • The DSP performs the desired mapping between the input and output sequences. • The D/A converter produces a set of equally spaced-in-time analog samples representing the DSP system output. • The lowpass filter interpolates the analog samples to produce a smooth continuous-time signal. Chapter 5 will discuss in detail the conditions which a continuous-time signal must satisfy so that its sampled version retains the information of the original signal, dictated by the sampling theorem. This theorem determines that a band limited continuous-time signal can be theoretically recovered from its sampled version by filtering the sequence with an analog filter with prescribed frequency response. It is also important to mention that, while processing the signals in the digital domain, these are subject to quantization errors, such as: roundoff errors originated from internal arithmetic operations
- 37. 1.01.5 Sampling and Quantization 9 0 T 2T t x(t) n n t t x(t) x(n) x(n) y(n) y(n) y (t) i y (t) i y(t) y(t) converter A/D converter D/A filter Lowpass system DSP 0 ... 12 ... 12T ... 12 1 2 0 1 2 FIGURE 1.5 DSP system. performed in the signals; deviations in the filter response due to finite wordlength representation of the multiplier coefficients inherent to the signal processing operation; and errors due to representation of the acquired continuous-time signals with a set of discrete levels. The actual implementation of a DSP system might rely on general purpose digital machines, where the user writes a computer software to implement the DPS tasks. This strategy allows fast prototyping and testing. Other mean is to use special-purpose commercially available CPUs, known as Digital Signal
- 38. 10 CHAPTER 1 Introduction to Signal Processing Theory x (t) t x (n) n n x Q (n) x Q (n) n t x(t) t x(t) (a) (b) (c) (f) (e) (d) ˆ ˆ FIGURE 1.6 A digital signal generation. (a) Original continuous-time signal. (b) A/D converter: sampling. (c) A/D converter: quantization. (d) Digital signal. (e) D/A converter.(f) Recovered continuous-time signal. Processors, which are capable of implementing sum of product operations in a very efficient manner. Yet another approach is to employ special purpose hardware tailored for the given application. Figure 1.6 depicts a continuous-time signal, its digitized version, and its recovered continuous-time representation. We can notice, from Figures 1.6(a) to 1.6(f), the effects of a low sampling frequency and a small number of quantization steps on the A/D and D/A processes. Figure 1.7 is a detailed example of the steps entailing an A/D and a D/A conversion, where in Figure 1.7(a) a continuous-time signal is depicted along with its equally spaced samples. These samples are then quantized as illustrates Figure 1.7(b). Assuming the quantized samples are converted into an analog signal through a zero-order hold, the output of the D/A converter becomes as illustrated in Figure 1.7(c). The sampled-and-held continuous-time signal is lowpass filtered in order to recover the original continuous-time signal. As can be observed in Figure 1.7(d), the recovered signal resembles the original where the difference originates from the quantization effects and the nonideal lowpass filter at the D/A converter output. 1.01.6 FIR and IIR filter design The general form of an FIR transfer function is given by H(z) = M l=0 bl z−l = H0z−M M l=0 (z − zl). Since all the poles of the FIR transfer function are placed at z = 0, it is always stable. As a result, the transfer function above will always represent a stable filter. The main feature of the FIR filters is
- 39. 1.01.6 FIR and IIR Filter Design 11 0 0.02 0.04 0.06 0.08 0.1 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 time(s) Amplitude x(t) x(n) (a) 0 0.02 0.04 0.06 0.08 0.1 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 time(s) Amplitude x(t) xQ (n) (b) 0 0.02 0.04 0.06 0.08 0.1 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 time(s) Amplitude x(t) x ZOH (t) 0 0.02 0.04 0.06 0.08 0.1 −1 −0.8 −0.6 −0.4 −0.2 0 0.2 0.4 0.6 0.8 1 time(s) Amplitude x(t) x rec (t) (d) (c) FIGURE 1.7 A real digital signal generation. (a) A/D converter: sampling. (b) A/D converter: quantization. (c) D/A converter: zero-order hold. (d) D/A converter: lowpass filter. the possibility of designing structurally induced linear-phase filters. On the other hand, an FIR filter requires a higher number of multiplications, additions and storage elements when compared to their IIR counterparts in order to satisfy a prescribed specification for the magnitude response. There are many methods to design an FIR filter [19], ranging from the simple ones, such as the window and frequency sampling methods, to more efficient and sophisticated, that rely on some kind of optimization method [20–22]. The simple methods lead to solutions which are far from opti- mum in most practical applications, in the sense that they either require higher order or cannot sat- isfy prescribed specifications. A popular optimization solution is the minimax method based on the Remez exchange algorithm, which leads to transfer functions with minimum order to satisfy prescribed specifications.
- 40. 12 CHAPTER 1 Introduction to Signal Processing Theory 0 500 1000 1500 2000 −120 −100 −80 −60 −40 −20 0 Frequency (Hz) Magnitude response (dB) Minimax method WLS method 0 200 400 600 800 1200 −40 −30 −25 −20 −15 −10 −5 5 10 Frequency (Hz) Magnitude response (dB) Minimax method WLS method 1 2 Αr + δp − δp f f (b) (a) FIGURE 1.8 Typical magnitude responses of FIR filters: Minimax and WLS. (a) Magnitude response. (b) Specifications for a lowpass filter using minimax design. Nowadays, with the growing computational power, it is possible to design quite flexible FIR filters by utilizing weighted least squares (WLS) solutions. These filters are particularly suitable for multirate systems, since the resulting transfer functions can have decreasing energy in their stopband [2]. The WLS method enables the design of filters satisfying prescribed specifications, such as the maximum deviation in the passband and in part of the stopband, while minimizing the energy in a range of frequencies in the stopband. As can be observed in Figure 1.8, for a given filter order, the minimax design leads to lower maximum stopband attenuation, whereas the WLS solution leads to lower stopband energy. Note there are solutions in between where some maximum values of stopband ripples can be minimized while the energy of the remaining ones is also minimized. It is possible to observe that the WLS solution does not satisfy the prescribed specifications given that it has the same filter order as the minimax solution. The typical transfer function of an IIR filter is given by H(z) = Y(z) X(z) = M l=0 bl z−l 1 + N i=1 ai z−i , which can be rewritten in a form explicitly showing the poles positions as H(z) = H0 M l=0 (1 − z−1zl) N i=0 (1 − z−1 pi ) = H0zN−M M l=0 (z − zl) N i=0 (z − pi ) . IIR filters are usually designed by using well-established analog filter approximations, such as But- terworth, Chebyshev, and elliptic methods (Figure 1.9) The prototype analog transfer function is then
- 41. 1.01.7 Digital Filter Structures and Implementations 13 0 500 1000 1500 2000 −100 −90 −80 −70 −60 −50 −40 −30 −20 −10 0 Frequency (Hz) Magnitude response (dB) Elliptic approximation Butherworth approx. Chebyshev Type I approx. FIGURE 1.9 Typical magnitude response of an IIR filter. transformed into a digital transfer function by using an adequate transformation method. The most widely used transformation methods are bilinear and the impulse invariance methods [2]. 1.01.7 Digital filter structures and implementations From a closer examination of the FIR and IIR transfer functions, one can infer that they can be realized with three basic operators: adder, multiplier and delay (represented by z−1). For example, Figure 1.10 shows a linear-phase FIR filter realization of odd order. Linear-phase filters have symmetric or anti-symmetric impulse response and the symmetry should be exploited in order to minimize the number of multipliers in the filter implementation, as illustrates Figure 1.10. Figure 1.11 depicts a possible direct-form realization of an IIR transfer function. This realization requires the minimum number of multiplications, adders and delays to implement the desired IIR transfer function. There are many alternative structures to implement the FIR and IIR filters. For example, IIR transfer functions can be implemented as a product or as a summation of lower order IIR structures by starting with their description as: H(z) = m k=1 γ0k + γ1kz−1 + γ2kz−2 1 + m1kz−1 + m2kz−2 = h p 0 + m k=1 γ p 1k z + γ p 2k z2 + m1kz + m2k , respectively. The right choice for a filter realization must take into consideration some issues, such as: modularity, quantization effects, design simplicity, power consumption, etc. [2]. Chapter 6 will address
- 42. 14 CHAPTER 1 Introduction to Signal Processing Theory z−1 z−1 z−1 z−1 z−1 × × × + + + + x(n) y(n) b0 b1 b2 FIGURE 1.10 Odd-order linear-phase FIR filter structure with M = 5. advanced methods for FIR and IIR filter realization and implementation, where elegant and creative solutions are introduced in a clear manner. This chapter will also discuss strategies to implement digital filters efficiently. 1.01.8 Multirate signal processing In digital signal processing it is easy to change the sampling rate of the underlying sequences by a ratio of integer values. This feature is highly desirable whenever we want to merge information of systems employing distinct sampling rates. In these cases, those rates should be made compatible [6,9,10]. Systems that internally employ multiple sampling rates are collectively referred to as multirate systems. In most cases, these systems are time-varying or, at most, periodically time-invariant. The basic operations of the multirate systems are the decimation and interpolation. The right compo- sition of these operations allows arbitrary rational sampling rate changes. If this sampling rate change is arbitrary and non-rational, the only solution is to recover the bandlimited continuous-time signal x(t) from its samples x(m), and then re-sample it with a different sampling rate, thus generating a distinct discrete-time signal x(n). If we assume that a discrete-time signal signal x(m) was generated from an continuous-time signal y(t) with sampling period T1, so that x(m) = y(mT1), with m = . . . , 0, 1, 2 . . ., the sampling theorem requires that y(t) should be bandlimited to the range [− π T1 , π T1 ]. The sampled continuous-time signal is given by: y (t) = ∞ m=−∞ x(m)δ(t − mT1).
- 43. 1.01.8 Multirate Signal Processing 15 z−1 × + × + z−1 × + z−1 × + × × × x(n) b0 b1 b2 − a1 − a2 − aN bM y(n) FIGURE 1.11 Direct-Form IIR filter structure, for M = N. The spectrum of the sampled signal is periodic with period 2π T1 . The original continuous-time signal y(t) can be recovered from y(t) through an ideal lowpass filter. This interpolation filter, whose impulse response is denoted as h(t), has ideal frequency response H(ejω) as follows: H(ejω ) = 1, ω ∈ [− π T1 , π T1 ] 0, otherwise, so that y(t) = y (t) ∗ h(t) = 1 T1 ∞ m=−∞ x(m)sinc π T1 (t − mT1). By re-sampling y(t) with period T2, we obtain the discrete-time signal in the new desired sampling rate as follows: x(n) = 1 T1 ∞ m=−∞ x(m)sinc π T1 (nT2 − mT1), where x(n) = y(nT2), with n = . . . , 0, 1, 2 . . .. In this general equation governing sampling rate changes, there is no restriction on the values of T1 and T2. However, if T2 T1 and aliasing is to be avoided, the interpolation filter must have a zero gain for ω / ∈ [− π T2 , π T2 ].
- 44. 16 CHAPTER 1 Introduction to Signal Processing Theory In the case the sampling rate change corresponds to a ratio of integer numbers, all we need is simple decimation and interpolation operations. The decimation operation, or down-sampling, consists of retaining every Mth sample of a given sequence x(m). Since we are disposing samples from the original sequence, either the signal is sufficiently limited in band or the signal has to be filtered by a lowpass filter so that the decimated signal retains the useful information of the original signal. Decimation is a time-varying operation, since, if x(m) is shifted by m0, the output signal will, in general, differ from the unshifted output shifted by m0. Indeed, decimation is a periodically time-invariant operation, which consist of an extra degree of freedom with respect to the traditional time-invariant systems. The interpolation, or up-sampling, of a discrete-time signal x(m) by a factor of L consists of inserting L − 1 zeros between each of its samples. In the frequency domain, the spectrum of the up-sampled signal is periodically repeated. Given that the spectrum of the original discrete-time signal is periodic with period 2π, the interpolated signal will have period 2π L . Therefore, in order to obtain a smooth interpolated version of x(m), the spectrum of the interpolated signal must have the same shape of the spectrum of x(m). This can be obtained by filtering out the repetitions of the spectra beyond [−π L , π L ]. Thus, the up-sampling operation is generally followed by a lowpass filter. The interpolation is only periodically time-invariant and does not entail any loss of information. Figure 1.12 illustrates discrete-time signal decimation and interpolation operations. As can be observed in Figure 1.12(a), the decimation factor of two widens the spectrum of the sampled sig- nal in comparison to the new sampling rate. Figure 1.12(b) depicts the effect of increasing the sampling rate of a given sequence by two, where it can be seen that the frequency contents of the original signal repeats twice as as often and appears narrower in the new sampling rate. Chapter 7 will further discuss the theory and practice of multirate signal processing. This chapter will show how sophisticated real-life signal processing problems can be addressed starting from the basic theory. In particular, the authors address the design of flexible communications systems incorporating several advanced signal processing tools [4,23–25]. 1.01.9 Filter banks and wavelets In many applications it is desirable to decompose a wideband discrete-time signal into several non- overlapping narrowband subsignals in order to be transmitted, quantized, or stored. In this case, each narrow-band channel can have its sampling rate reduced, since the subband signals have lower bandwidth than their originator signal. The signal processing tool that performs these tasks is the analysis filter bank. The analysis filters, represented by the transfer functions Hi (z), for i = 0, 1, . . . , M − 1, consist of a lowpass filter H0(z), bandpass filters Hi (z), for i = 1, 2, . . . , M −2, and a highpass filter HM−1(z). Ideally, these filters have non-overlapping passbands, whereas their spectrum combination should cover the overall spectrum of the input signal. The outputs of the analysis filters, denoted as xi (n), for i = 0, 1, . . . , M−1, have together M times the number of samples of the original signal x(n). However, since each subband signal occupies a spectrum band M times narrower than the input signal, they can be deci- mated so the number of samples in the subbands does not increase. Indeed, if the input signal is uniformly split into subbands, we can decimate each xi (n) by a factor of L smaller or equal to M without generating unrecoverable aliasing effects. Figure 1.13 shows the general configuration of a maximally (or critically) decimated analysis filter, where the decimation factor is equal to the number of subbands, i.e., L = M.
- 45. 1.01.9 Filter Banks and Wavelets 17 x̄(n) m x(m ) x̄(n) M 0 1 2 12 X (ejω) 2π −2π p − p X (ejω) 2π −2π 2ωp −2ωp n 0 1 2 6 3 4 5 x(m ) (a) Signal decimation with M = 2. 2π 2 − 4π 2 3 x(m ) 5 4 n x(m ) x̄(n) L 0 1 2 12 X (ejω) 2π −2π ωp −ωp X (ej ) − 2π 2 ωp 2 − ωp 2 x̄(n) m 0 1 2 6 4π 2 (b) Signal interpolation with L = 2. FIGURE 1.12 Signal decimation and interpolation. (a) Signal decimation with M = 2. (b) Signal interpolation with L = 2. Whenever the input signal is split in subbands and decimated, if L ≤ M, it is always possible to recover the input signal by properly designing the analysis filters in conjunction with the synthesis filters Gi (z), for i = 0, 1, . . . , M −1. The synthesis filters are placed after interpolators and they have the task of smoothing the up-sampled signals. Figure 1.14 illustrates the general configuration of the synthesis filter bank. A key feature of the synthesis filter bank is to cancel out or reduce the aliasing effects. If the signals in the subbands are not modified, the filter bank output y(n) can be a delayed copy signal of x(n). The cascade connection of the analysis and synthesis filter banks satisfying this condition is called a perfect reconstruction filter bank. The cascade of an M-channel synthesis filter bank with an M-channel analysis filter bank gives rise to a transmultiplex system. The transmultiplex model is widely used in communications to model multiuser, multicarrier, and multiple access systems [4,23,24]. Chapter 7 will utilize multirate signal processing and transmultiplexers to discuss a design procedure to be applied in cognitive radio. Chapter 8 will present several methods for designing the analysis filters Hi (z) and the synthesis filters Gi (z), such that perfect reconstruction, as well as other features, are met. This chapter will also discuss
- 46. 18 CHAPTER 1 Introduction to Signal Processing Theory M M M M H0(z) H1(z) H2(z) HM − 1(z) |H 0 (e jω )| |H 1 (e jω )| |H 2 (e jω )| |H M − 1 (e jω )| x(n) x0(m) x1(m) x2(m) xM − 1(m) FIGURE 1.13 Analysis Filter Bank. the design of filter banks with the subbands divided in octave bands, giving rise to the discrete-time wavelet series. Figure 1.15 shows a four-band filter bank design and its effect in a composed signal. Figure 1.16 illustrates how a four-band synthesis filter bank performs the recomposition of the input signal starting from the decimated subband signals. 1.01.10 Discrete multiscale and transforms Discrete-time signals and systems can be characterized in the frequency domain by their Fourier trans- forms. One of the main advantages of discrete-time signals is that they can be processed and represented
- 47. 1.01.10 Discrete Multiscale and Transforms 19 π π π π + G0(z) G1(z) G2(z) GM − 1(z) ω G 0 (e jω ) G 1 (e jω ) G 2 (e jω ) ω ω G M − 1 (e jω ) ω x(n) x0(m) x1(m) x2(m) xM − 1(m) M M M M FIGURE 1.14 Synthesis Filter Bank. in digital computers. However, when we examine the definition of the discrete-time Fourier transform, X(ejω ) = ∞ n=−∞ x(n)e−jωn , we notice that such a characterization in the frequency domain depends on the continuous variable ω. This implies that the Fourier transform, as it is, is not suitable for the processing of discrete-time signals in digital computers. We need a transform depending on a discrete-frequency variable that, if possible, preserves the handy interpretations of the Fourier-based tools, which retains important information. This can be obtained from the Fourier transform itself in a very simple way, by sampling uniformly the continuous-frequency variable ω as long as the input sequence has finite length, otherwise the time-domain signal will be affected by aliasing. Using this strategy it is possible to obtain a mapping of a signal depending on a discrete-time variable n to a transform depending on a discrete-frequency
- 48. 20 CHAPTER 1 Introduction to Signal Processing Theory m x 0 (m ) H i (e jω ) π ω H0 H1 H2 H3 π 4 π 2 3π 4 m x 3 (m ) m x 2 (m ) m x 1 (m ) 4 4 4 4 n y (n) Analysis Filte r Bank FIGURE 1.15 Analysis filter bank design. variable k, leading to another interpretation for the DFT. Unfortunately, the DFT has its limitations in representing more general class of signals as will be following discussed. The wavelet transform of a function belonging to L2{R}, the space of the square integrable functions, is the function decomposition in a basis composed by expansions, compressions, and translations of a single mother function ψ(t), called wavelet. The wavelet transform can be defined as x(t) = ∞ m=−∞ ∞ n=−∞ cm,nψm,n(t) cm,n = ∞ −∞ x(t)ψ∗ m,n(t) dt, where ψm,n(t) = 2−m/2 ψ(2−m t − n) ψm,n(t) = 2−m/2 ψ(2−m t − n). This pair of equations defines a biorthogonal wavelet transform which are characterized by two wavelets: the analysis wavelet, ψ(t), and the synthesis wavelet, ψ(t). Any function x(t) ∈ L2{R} can be
- 49. 1.01.10 Discrete Multiscale and Transforms 21 x 2 (m ) m m x 1 (m ) m x 0 (m ) x 3 (m ) m n x(n − N0) 4 4 4 4 |G i (e jω )| G0 G1 G2 G3 4 2 3 4 Synthesis Filter Bank + x (n − N 0 ) FIGURE 1.16 Synthesis filter bank design. decomposed as a linear combination of contractions, expansions, and translations of the synthesis wavelet ψ(t). The weights of the expansion can be computed via the inner product of x(t) with expan- sions, contractions, and translations of the analysis wavelet ψ(t). Functions ψm,n(t) do not comprise an orthogonal set, so neither do the functions ψm,n(t). However, functions ψm,n(t) are orthogonal to ψm,n(t) so that ψm,n(t), ψk,l(t) = δ(m − k)δ(n − l). A byproduct of the development of the wavelet transforms is the so called wavelet series, allowing the representation of signals in the time and frequency domains through sequences. The cascade of two- band filter banks can produce many alternative maximally decimated decompositions. Of the particular interest is the hierarchical decomposition achieving an octave-band decomposition, in which only the lowpass band is further decomposed. This configuration gives rise to the widely used wavelets series. As will be discussed in Chapter 10, there are many transform-based tools in signal processing theory available to meet the requirement of different applications. The classical Fourier based transforms have been used for centuries, but their efficiency in dealing with nonstationary signals is limited. With wavelet series, for example, it is possible to obtain good resolution in time and frequency domains, unlike the Fourier-based analysis. Chapter 9 will include a myriad of transform solutions tailored for distinct practical situations. Figure 1.17 illustrates a wavelet series representation of a composed signal, where the input signal is decomposed by an octave-band filter bank representing a wavelet series.
- 50. Exploring the Variety of Random Documents with Different Content
- 54. The Project Gutenberg eBook of Misinforming a Nation
- 55. This ebook is for the use of anyone anywhere in the United States and most other parts of the world at no cost and with almost no restrictions whatsoever. You may copy it, give it away or re-use it under the terms of the Project Gutenberg License included with this ebook or online at www.gutenberg.org. If you are not located in the United States, you will have to check the laws of the country where you are located before using this eBook. Title: Misinforming a Nation Author: Willard Huntington Wright Release date: December 20, 2019 [eBook #60985] Most recently updated: October 17, 2024 Language: English Credits: Produced by WebRover, MWS and the Online Distributed Proofreading Team at http://www.pgdp.net (This file was produced from images generously made available by The Internet Archive) *** START OF THE PROJECT GUTENBERG EBOOK MISINFORMING A NATION ***
- 56. MISINFORMING A NATION BOOKS BY MR. WRIGHT MISINFORMING A NATION MODERN PAINTING: Its Tendency and Meaning WHAT NIETZSCHE TAUGHT THE MAN OF PROMISE THE CREATIVE WILL IN PREPARATION MODERN LITERATURE PRINCIPLES OF ÆSTHETIC FORM AND ORGANIZATION Misinforming a Nation
- 57. by Willard Huntington Wright New York B. W. Huebsch MCMXVII COPYRIGHT, 1917, BY B. W. HUEBSCH PRINTED IN THE UNITED STATES OF AMERICA
- 58. CONTENTS CHAPTER PAGE I Colonizing America 1 II The Novel 24 III The Drama 52 IV Poetry 68 V British Painting 85 VI Non-British Painting 102 VII Music 122 VIII Science 148 IX Inventions, Photography, Æsthetics 160 X Philosophy 174 XI Religion 195 XII Two Hundred Omissions 218 MISINFORMING A NATION
- 59. I COLONIZING AMERICA The intellectual colonization of America by England has been going on for generations. Taking advantage of her position of authority—a position built on centuries of æsthetic tradition—England has let pass few opportunities to ridicule and disparage our activities in all lines of creative effort, and to impress upon us her own assumed cultural superiority. Americans, lacking that sense of security which long-established institutions would give them, have been influenced by the insular judgments of England, and, in an effort to pose as au courant of the achievements of the older world, have adopted in large degree the viewpoint of Great Britain. The result has been that for decades the superstition of England’s pre-eminence in the world of art and letters has spread and gained power in this country. Our native snobbery, both social and intellectual, has kept the fires of this superstition well supplied with fuel; and in our slavish imitation of England—the only country in Europe of which we have any intimate knowledge—we have de-Americanized ourselves to such an extent that there has grown up in us a typical British contempt for our own native achievements. One of the cardinal factors in this Briticization of our intellectual outlook is the common language of England and America. Of all the civilized nations of the world, we are most deficient as linguists. Because of our inability to speak fluently any language save our own, a great barrier exists between us and the Continental countries. But no such barrier exists between America and England; and consequently there is a constant exchange of ideas, beliefs, and opinions. English literature is at our command; English criticism is familiar to us; and English standards are disseminated among us without the impediment of translation. Add to this lingual
- 60. rapprochement the traditional authority of Great Britain, together with the social aspirations of moneyed Americans, and you will have both the material and the psychological foundation on which the great edifice of English culture has been reared in this country. The English themselves have made constant and liberal use of these conditions. An old and disquieting jealousy, which is tinctured not a little by resentment, has resulted in an open contempt for all things American. And it is not unnatural that this attitude should manifest itself in a condescending patronage which is far from being good-natured. Our literature is derided; our artists are ridiculed; and in nearly every field of our intellectual endeavor England has found grounds for disparagement. It is necessary only to look through British newspapers and critical journals to discover the contemptuous and not infrequently venomous tone which characterizes the discussion of American culture. At the same time, England grasps every opportunity for foisting her own artists and artisans on this country. She it is who sets the standard which at once demolishes our individual expression and glorifies the efforts of Englishmen. Our publishers, falling in line with this campaign, import all manner of English authors, eulogize them with the aid of biased English critics, and neglect better writers of America simply because they have displeased those gentlemen in London who sit in judgment upon our creative accomplishments. Our magazines, edited for the most part by timid nobodies whose one claim to intellectual distinction is that they assiduously play the parrot to British opinion, fill their publications with the work of English mediocrities and ignore the more deserving contributions of their fellow-countrymen. Even our educational institutions disseminate the English superstition and neglect the great men of America; for nowhere in the United States will you find the spirit of narrow snobbery so highly developed as in our colleges and universities. Recently an inferior British poet came here, and, for no other reason apparently save that he was English, he was made a professor in one of our
- 61. large universities! Certainly his talents did not warrant this appointment, for there are at least a score of American poets who are undeniably superior to this young Englishman. Nor has he shown any evidences of scholarship which would justify the honor paid him. But an Englishman, if he seek favors, needs little more than proof of his nationality, whereas an American must give evidence of his worth. England has shown the same ruthlessness and unscrupulousness in her intellectual colonization of America as in her territorial colonizations; and she has also exhibited the same persistent shrewdness. What is more, this cultural extension policy has paid her lavishly. English authors, to take but one example, regard the United States as their chief source of income. If it were the highest English culture—that is, the genuinely significant scholarship of the few great modern British creators—which was forced upon America, there would be no cause for complaint. But the governing influences in English criticism are aggressively middle-class and chauvinistic, with the result that it is the British bourgeois who has stifled our individual expression, and misinformed us on the subject of European culture. No better instance of this fact can be pointed to than the utterly false impression which America has of French attainments. French genius has always been depreciated and traduced by the British; and no more subtle and disgraceful campaign of derogation has been launched in modern times than the consistent method pursued by the English in misinterpreting French ideals and accomplishments to Americans. To England is due largely, if not entirely, the uncomplimentary opinion that Americans have of France—an opinion at once distorted and indecent. To the average American a French novel is regarded merely as a salacious record of adulteries. French periodicals are looked upon as collections of prurient anecdotes and licentious pictures. And the average French painting is conceived as a realistic presentation of feminine nakedness. So deeply rooted are these conceptions that the very word “French” has become, in the American’s vocabulary, an adjective signifying all manner of sexual
- 62. abnormalities, and when applied to a play, a story, or an illustration, it is synonymous with “dirty” and “immoral.” This country has yet to understand the true fineness of French life and character, or to appreciate the glories of French art and literature; and the reason for our distorted ideas is that French culture, in coming to America, has been filtered through the nasty minds of middle-class English critics. But it is not our biased judgment of the Continental nations that is the most serious result of English misrepresentation; in time we will come to realize how deceived we were in accepting England’s insinuations that France is indecent, Germany stupid, Italy decadent, and Russia barbarous. The great harm done by England’s contemptuous critics is in belittling American achievement. Too long has bourgeois British culture been forced upon the United States; and we have been too gullible in our acceptance of it without question. English critics and English periodicals have consistently attempted to discourage the growth of any national individualism in America, by ridiculing or ignoring our best æsthetic efforts and by imposing upon us their own insular criteria. To such an extent have they succeeded that an American author often must go to England before he will be accepted by his own countrymen. Thus purified by contact with English culture, he finds a way into our appreciation. But on the other hand, almost any English author—even one that England herself has little use for—can acquire fame by visiting this country. Upon his arrival he is interviewed by the newspapers; his picture appears in the “supplements”; his opinions emblazon the headlines and are discussed in editorials; and our publishers scramble for the distinction of bringing out his wares. In this the publishers, primarily commercial, reveal their business acumen, for they are not unaware of the fact that the “literary” sections of our newspapers are devoted largely to British authors and British letters. So firmly has the English superstition taken hold of our publishers that many of them print their books with English spelling. The reason for this un-American practice, so they explain, is that the books may be ready for an English edition without resetting. The
- 63. English, however, do not use American spelling at all, though, as a rule, the American editions of English books are much larger than the English edition of American books. But the English do not like our spelling; therefore we gladly arrange matters to their complete satisfaction. The evidences of the American’s enforced belief in English superiority are almost numberless. Apartment houses and suburban sub-divisions are named after English hotels and localities. The belief extends even to the manufacturers of certain brands of cigarettes which, for sale purposes, are advertised as English, although it would be difficult to find a box of them abroad. The American actor, in order to gain distinction, apes the dress, customs, intonation and accent of Englishmen. His great ambition is to be mistaken for a Londoner. This pose, however, is not all snobbery: it is the outcome of an earnest desire to appear superior; and so long has England insisted upon her superiority that many Americans have come to adopt it as a cultural fetish. Hitherto this exalted intellectual guidance has been charitably given us: never before, as now, has a large fortune been spent to make America pay handsomely for the adoption of England’s provincialism. I refer to the Encyclopædia Britannica which, by a colossal campaign of flamboyant advertising, has been scattered broadcast over every state in the union. No more vicious and dangerous educational influence on America can readily be conceived than the articles in this encyclopædia. They distort the truth and disseminate false standards. America is now far enough behind the rest of the civilized world in its knowledge of art, without having added to that ignorance the erroneous impressions created by this partial and disproportioned English work; for, in its treatment of the world’s progress, it possesses neither universality of outlook nor freedom from prejudice in its judgments—the two primary requisites for any work which lays claim to educational merit. Taken as a whole, the Britannica’s divisions on culture are little more than a brief for British art and science—a brief fraught
- 64. with the rankest injustice toward the achievements of other nations, and especially toward those of America. The distinguishing feature of the Encyclopædia Britannica is its petty national prejudice. This prejudice appears constantly and in many disguises through the Encyclopædia’s pages. It manifests itself in the most wanton carelessness in dealing with historical facts; in glaring inadequacies when discussing the accomplishments of nations other than England; in a host of inexcusable omissions of great men who do not happen to be blessed with English nationality; in venom and denunciation of viewpoints which do not happen to coincide with “English ways of thinking”; and especially in neglect of American endeavor. Furthermore, the Britannica shows unmistakable signs of haste or carelessness in preparation. Information is not always brought up to date. Common proper names are inexcusably misspelled. Old errors remain uncorrected. Inaccuracies abound. Important subjects are ignored. And only in the field of English activity does there seem to be even an attempt at completeness. The Encyclopædia Britannica, if accepted unquestioningly throughout this country as an authoritative source of knowledge, would retard our intellectual development fully twenty years; for so one-sided is its information, so distorted are its opinions, so far removed is it from being an international and impartial reference work, that not only does it give inadequate advice on vital topics, but it positively creates false impressions. Second- and third-rate Englishmen are given space and praise much greater than that accorded truly great men of other nations; and the eulogistic attention paid English endeavor in general is out of all proportion to its deserts. In the following chapters I shall show specifically how British culture is glorified and exaggerated, and with what injustice the culture of other countries is treated. And I shall also show the utter failure of this Encyclopædia to fulfill its claim of being a “universal” and “objective” reference library. To the contrary, it will be seen that the Britannica is a narrow, parochial, opinionated work of dubious scholarship and striking unreliability.
- 65. With the somewhat obscure history of the birth of the Eleventh Edition of the Encyclopædia Britannica, or with the part played in that history by Cambridge University and the London Times, I am not concerned. Nor shall I review the unethical record of the two issues of the Encyclopædia. To those interested in this side of the question I suggest that they read the following contributions in Reedy’s Mirror: The Same Old Slippery Trick (March 24, 1916). The Encyclopædia Britannica Swindle (April 7, 1916). The Encyclopædia Britannica Fake (April 14, 1916); and also the article in the March 18 (1916) Bellman, Once More the Same Old Game. Such matters might be within the range of forgiveness if the contents of the Britannica were what were claimed for them. But that which does concern me is the palpable discrepancies between the statements contained in the advertising, and the truth as revealed by a perusal of the articles and biographies contained in the work itself. The statements insisted that the Britannica was a supreme, unbiased, and international reference library—an impartial and objective review of the world; and it was on these statements, repeated constantly, that Americans bought the work. The truth is that the Encyclopædia Britannica, in its main departments of culture, is characterized by misstatements, inexcusable omissions, rabid and patriotic prejudices, personal animosities, blatant errors of fact, scholastic ignorance, gross neglect of non-British culture, an astounding egotism, and an undisguised contempt for American progress. Rarely has this country witnessed such indefensible methods in advertising as those adopted by the Britannica’s exploiters. The “copy” has fairly screamed with extravagant and fabulous exaggerations. The vocabulary of hyperbole has been practically exhausted in setting forth the dubious merits of this reference work. The ethics and decencies of ordinary honest commerce have been thrown to the wind. The statements made day after day were apparently concocted irrespective of any consideration save that of making a sale; for there is an abundance of evidence to show that the Encyclopædia was not what was claimed for it.
- 66. With the true facts regarding this encyclopædia it is difficult to reconcile the encomiums of many eminent Americans who, by writing eulogistic letters to the Britannica’s editor concerning the exalted merits of his enterprise, revealed either their unfamiliarity with the books in question or their ignorance of what constituted an educational reference work. These letters were duly photographed and reproduced in the advertisements, and they now make interesting, if disconcerting, reading for the non-British student who put his faith in them and bought the Britannica. There is no need here to quote from these letters; for a subsequent inspection of the work thus recommended must have sufficiently mortified those of the enthusiastic correspondents who were educated and had consciences; and the others would be unmoved by any revelations of mine. Mention, however, should be made of the remarks of the American Ambassador to Great Britain at the banquet given in London to celebrate the Encyclopædia’s birth. This gentleman, in an amazing burst of unrestrained laudation, said he believed that “it is the general judgment of the scholars and the investigators of the world that the one book to which they can go for the most complete, comprehensive, thorough, and absolutely precise statements of fact upon every subject of human interest is the Encyclopædia Britannica.” This is certainly an astonishing bit of eulogy. Its dogmatic positiveness and its assumption of infallibility caused one critic (who is also a great scholar) to write: “With all due respect for our illustrious fellow-countryman, the utterance is a most superlative absurdity, unless it was intended to be an exercise of that playful and elusive American humor which the apperceptions of our English cousins so often fail to seize, much less appreciate.” But there were other remarks of similar looseness at the banquet, and the dinner evidently was a greater success than the books under discussion. Even the English critics themselves could not accept the Britannica as a source for “the most comprehensive, thorough and absolutely precise statements on every subject of human interest.” Many legitimate objections began appearing. There is space here to quote
- 67. only a few. The London Nation complains that “the particularly interesting history of the French Socialist movement is hardly even sketched.” And again it says: “The naval question is handled on the basis of the assumption which prevailed during our recent scare; the challenge of our Dreadnought building is hardly mentioned; the menace of M. Delcassé’s policy of encirclement is ignored, and both in the article on Germany and in the articles on Europe, Mr. McKenna’s panic figures and charges of accelerated building are treated as the last word of historical fact.” The same publication, criticising the article on Europe, says: “There is nothing but a dry and summarized general history, ending with a paragraph or two on the Anglo-German struggle with the moral that ‘Might is Right.’ It is history of Europe which denies the idea of Europe.” Again, we find evidence of a more direct character, which competently refutes the amazing announcement of our voluble Ambassador to Great Britain. In a letter to the London Times, an indignant representative of Thomas Carlyle’s family objects to the inaccurate and biased manner in which Carlyle is treated in the Encyclopædia. “The article,” he says, “was evidently written many years ago, before the comparatively recent publication of new and authentic material, and nothing has been done to bring it up to date.... As far as I know, none of the original errors have been corrected, and many others of a worse nature have been added. The list of authorities on Carlyle’s life affords evidence of ignorance or partisanship.” “Evidently,” comments a shrewd critic who is not impressed either by the Ambassador’s panegyric or the photographed letters, “the great man’s family, and the public in general, have a reasonable cause of offense, and they may also conclude that if the Encyclopædia Britannica can blunder when handling such an approachable and easy British subject as Carlyle, it can be reasonably expected to do worse on other matters which are not only absolutely foreign, but intensely distasteful to the uninformed and prejudiced scribes to whom they seem to be so frequently, if not systematically, assigned.”
- 68. The expectation embodied in the above comment is more fully realized perhaps than the writer of those words imagined; and the purpose of this book is to reveal the blundering and misleading information which would appear to be the distinguishing quality of the Britannica’s articles on culture. Moreover, as I have said, and as I shall show later, few subjects are as “intensely distasteful” to the “uninformed and prejudiced” British critics as is American achievement. One finds it difficult to understand how any body of foreigners would dare offer America the brazen insult which is implied in the prodigal distribution of these books throughout the country; for in their unconquerable arrogance, their unveiled contempt for this nation—the outgrowth of generations of assumed superiority—they surpass even the London critical articles dealing with our contemporary literary efforts. Several of our more courageous and pro-American scholars have called attention to the inadequacies and insularities in the Britannica, but their voices have not been sufficiently far-reaching to counteract either the mass or the unsavory character of the advertising by which this unworthy and anti-American encyclopædia was foisted upon the United States. Conspicuous among those publications which protested was the Twentieth Century Magazine. That periodical, to refer to but one of its several criticisms, pointed out that the article on Democracy is “confined to the alleged democracies of Greece and their distinguished, if some time dead, advocates. Walt Whitman, Mazzini, Abraham Lincoln, Edward Carpenter, Lyof Tolstoi, Switzerland, New Zealand, Australia, Finland, Iceland, Oregon are unknown quantities to this anonymous classicist.” It is also noted that the author of the articles on Sociology “is not very familiar with the American sociologists, still less with the German, and not at all with the French.” The article is “a curious evidence of editorial insulation,” and the one on Economics “betrays freshened British capitalistic insularity.” In this latter article, which was substituted for Professor Ingram’s masterly and superb history of political economy in the Britannica’s Ninth Edition, “instead of a
- 69. catholic, scientific survey of economic thought, we have a ‘fair trade’ pamphlet, which actually includes reference to Mr. Chamberlain,” although the names of Henry George, Karl Marx, Friedrich Engels, John A. Hobson, and William Smart are omitted. The Eleventh Edition, concludes the Twentieth Century, after recording many other specimens of ignorance and inefficiency, “is not only insular; it betrays its class-conscious limitation in being woefully defective in that prophetic instinct which guided Robertson Smith in his choice of contributors to the Ninth Edition, and the contributors themselves in their treatment of rapidly changing subjects.” Robertson Smith, let it be noted, stood for fairness, progressiveness, and modernity; whereas the Britannica’s present editor is inflexibly reactionary, provincial, and unjust to an almost incredible degree. The foregoing quotations are not isolated objections: there were others of similar nature. And these few specimens are put down here merely to show that there appeared sufficient evidence, both in England and America, to establish the purely imaginary nature of the Britannica’s claims of completeness and inerrancy, and to reveal the absurdity of the American Ambassador’s amazing pronouncement. Had the sale of the Encyclopædia Britannica been confined to that nation whose culture it so persistently and dogmatically glorifies at the expense of the culture of other nations, its parochial egotism would not be America’s concern. But since this reference work has become an American institution and has forced its provincial mediocrity into over 100,000 American homes, schools and offices, the astonishing truth concerning its insulting ineptitude has become of vital importance to this country. Its menace to American educational progress can no longer be ignored. England’s cultural campaign in the United States during past decades has been sufficiently insidious and pernicious to work havoc with our creative effort, and to retard us in the growth of that self- confidence and self-appreciation which alone make the highest achievement possible. But never before has there been so
- 70. concentrated and virulently inimical a medium for British influence as the present edition of the Encyclopædia Britannica. These books, taken in conjunction with the methods by which they have been foisted upon us, constitute one of the most subtle and malign dangers to our national enlightenment and development which it has yet been our misfortune to possess; for they bid fair to remain, in large measure, the source of America’s information for many years to come. The regrettable part of England’s intellectual intrigues in the United States is the subservient and docile acquiescence of Americans themselves. Either they are impervious to England’s sneers and deaf to her insults, or else their snobbery is stronger than their self-respect. I have learned from Britishers themselves, during an extended residence in London, that not a little of their contempt for Americans is due to our inordinate capacity for taking insults. Year after year English animus grows; and to-day it is the uncommon thing to find an English publication which, in discussing the United States and its culture, does not contain some affront to our intelligence. It is quite true, as the English insist, that we are painfully ignorant of Europe; but it must not be forgotten that the chief source of that ignorance is England herself. And the Encyclopædia Britannica, if accepted as authoritative, will go far toward emphasizing and extending that ignorance. Furthermore, it will lessen even the meagre esteem in which we now hold our own accomplishments and potentialities; for, as the following pages will show, the Britannica has persistently discriminated against all American endeavor, not only in the brevity of the articles and biographies relating to this country and in the omissions of many of our leading artists and scientists, but in the bibliographies as well. And it must be remembered that broad and unprejudiced bibliographies are essential to any worthy encyclopædia: they are the key to the entire tone of the work. The conspicuous absence of many high American authorities, and the inclusion of numerous reactionary and often dubious English authorities, sum up the Britannica’s attitude.
- 71. However, as I have said, America, if the principal, is not the only country discriminated against. France has fallen a victim to the Encyclopædia’s suburban patriotism, and scant justice is done her true greatness. Russia, perhaps even more than France, is culturally neglected; and modern Italy’s æsthetic achievements are given slight consideration. Germany’s science and her older culture fare much better at the hands of the Britannica’s editors than do the efforts of several other nations; but Germany, too, suffers from neglect in the field of modern endeavor. Even Ireland does not escape English prejudice. In fact, it can be only on grounds of national, political, and personal animosity that one can account for the grossly biased manner in which Ireland, her history and her culture, is dealt with. To take but one example, regard the Britannica’s treatment of what has come to be known as the Irish Literary Revival. Among those conspicuous, and in one or two instances world-renowned, figures who do not receive biographies are J. M. Synge, Lady Gregory, Lionel Johnson, Douglas Hyde, and William Larminie. (Although Lionel Johnson’s name appears in the article on English literature, it does not appear in the Index—a careless omission which, in victimizing an Irishman and not an Englishman, is perfectly in keeping with the deliberate omissions of the Britannica.) Furthermore, there are many famous Irish writers whose names are not so much as mentioned in the entire Encyclopædia—for instance, Standish O’Grady, James H. Cousins, John Todhunter, Katherine Tynan, T. W. Rolleston, Nora Hopper, Jane Barlow, Emily Lawless, “A. E.” (George W. Russell), John Eglinton, Charles Kickam, Dora Sigerson Shorter, Shan Bullock, and Seumas MacManus. Modern Irish literature is treated with a brevity and an injustice which are nothing short of contemptible; and what little there is concerning the new Irish renaissance is scattered here and there in the articles on English literature! Elsewhere I have indicated other signs of petty anti-Irish bias, especially in the niggardly and stupid treatment accorded George Moore.
- 72. Although such flagrant inadequacies in the case of European art would form a sufficient basis for protest, the really serious grounds for our indignation are those which have to do with the Britannica’s neglect of America. That is why I have laid such emphasis on this phase of the Encyclopædia. It is absolutely necessary that this country throw off the yoke of England’s intellectual despotism before it can have a free field for an individual and national cultural evolution. America has already accomplished much. She has contributed many great figures to the world’s progress. And she is teeming with tremendous and splendid possibilities. To-day she stands in need of no other nation’s paternal guidance. In view of her great powers, of her fine intellectual strength, of her wide imagination, of her already brilliant past, and of her boundless and exalted future, such a work as the Encyclopædia Britannica should be resented by every American to whom the welfare of his country is of foremost concern, and in whom there exists one atom of national pride.
- 73. II THE NOVEL Let us inspect first the manner in which the world’s great modern novelists and story-tellers are treated in the Encyclopædia Britannica. No better department could be selected for the purpose; for literature is the most universal and popular art. The world’s great figures in fiction are far more widely known than those in painting or music; and since it is largely through literature that a nation absorbs its cultural ideas, especial interest attaches to the way that writers are interpreted and criticised in an encyclopædia. It is disappointing, therefore, to discover the distorted and unjust viewpoint of the Britannica. An aggressive insular spirit is shown in both the general literary articles and in the biographies. The importance of English writers is constantly exaggerated at the expense of foreign authors. The number of biographies of British writers included in the Encyclopædia far overweighs the biographical material accorded the writers of other nations. And superlatives of the most sweeping kind are commonly used in describing the genius of these British authors, whereas in the majority of cases outside of England, criticism, when offered at all, is cool and circumscribed and not seldom adverse. There are few British writers of any note whatever who are not taken into account; but many authors of very considerable importance belonging to France, Germany, Italy, Russia, and the United States are omitted entirely. In the Encyclopædia’s department of literature, as in other departments of the arts, the pious middle-class culture of England is carefully and consistently forced to the front. English provincialism and patriotism not only dominate the criticism of this department, but dictate the amount of space which is allotted the different nations. The result is that one seeking in this encyclopædia
- 74. adequate and unprejudiced information concerning literature will fail completely in his quest. No mention whatever is made of many of the world’s great novelists (provided, of course, they do not happen to be British); and the information given concerning the foreign authors who are included is, on the whole, meagre and biased. If, as is natural, one should judge the relative importance of the world’s novelists by the space devoted to them, one could not escape the impression that the literary genius of the world resides almost exclusively in British writers. This prejudiced and disproportionate treatment of literature would not be so regrettable if the Britannica’s criticisms were cosmopolitan in character, or if its standard of judgment was a purely literary one. But the criteria of the Encyclopædia’s editors are, in the main, moral and puritanical. Authors are judged not so much by their literary and artistic merits as by their bourgeois virtue, their respectability and inoffensiveness. Consequently it is not even the truly great writers of Great Britain who are recommended the most highly, but those middle-class literary idols who teach moral lessons and whose purpose it is to uplift mankind. The Presbyterian complex, so evident throughout the Encyclopædia’s critiques, finds in literature a fertile field for operation. Because of the limitations of space, I shall confine myself in this chapter to modern literature. I have, however, inspected the manner in which the older literature is set forth in the Encyclopædia Britannica; and there, as elsewhere, is discernible the same provincialism, the same theological point of view, the same flamboyant exaggeration of English writers, the same neglect of foreign genius. As a reference book the Britannica is chauvinistic, distorted, inadequate, disproportioned, and woefully behind the times. Despite the fact that the Eleventh Edition is supposed to have been brought up to date, few recent writers are included, and those few are largely second-rate writers of Great Britain. Let us first regard the gross discrepancies in space between the biographies of English authors and those of the authors of other
- 75. nations. To begin with, the number of biographies of English writers is nearly as many as is given all the writers of France and Germany combined. Sir Walter Scott is given no less than thirteen columns, whereas Balzac has only seven columns, Victor Hugo only a little over four columns, and Turgueniev only a little over one column. Samuel Richardson is given nearly four columns, whereas Flaubert has only two columns, Dostoievsky less than two columns, and Daudet only a column and a third! Mrs. Oliphant is given over a column, more space than is allotted to Anatole France, Coppée, or the Goncourts. George Meredith is given six columns, more space than is accorded Flaubert, de Maupassant and Zola put together! Bulwer-Lytton has two columns, more space than is given Dostoievsky. Dickens is given two and a half times as much space as Victor Hugo; and George Eliot, Trollope, and Stevenson each has considerably more space than de Maupassant, and nearly twice as much space as Flaubert. Anthony Hope has almost an equal amount of space with Turgueniev, nearly twice as much as Gorky, and more than William Dean Howells. Kipling, Barrie, Mrs. Gaskell, Mrs. Humphry Ward, and Felicia Hemans are each accorded more space than either Zola or Mark Twain.... Many more similar examples of injustice could be given, but enough have been set down to indicate the manner in which British authors are accorded an importance far beyond their deserts. Of Jane Austen, to whom is given more space than to either Daudet or Turgueniev, we read that “it is generally agreed by the best critics that Miss Austen has never been approached in her own domain.” What, one wonders, of Balzac’s stories of provincial life? Did he, after all, not even approach Miss Austen? Mrs. Gaskell’s Cranford “is unanimously accepted as a classic”; and she is given an equal amount of space with Dostoievsky and Flaubert! George Eliot’s biography draws three and a half columns, twice as much space as Stendhal’s, and half again as much as de Maupassant’s. In it we encounter the following astonishing specimen of criticism: No right estimate of her as an artist or a philosopher “can be formed without a steady recollection of her infinite capacity
- 76. for mental suffering, and her need of human support.” Just what these conditions have to do with an æsthetic or philosophic judgment of her is not made clear; but the critic finally brings himself to add that “one has only to compare Romola or Daniel Deronda with the compositions of any author except herself to realize the greatness of her designs and the astonishing gifts brought to their final accomplishment.” The evangelical motif enters more strongly in the biography of George Macdonald, who draws about equal space with Gorky, Huysmans, and Barrès. Here we learn that Macdonald’s “moral enthusiasm exercised great influence upon thoughtful minds.” Ainsworth, the author of those shoddy historical melodramas, Jack Sheppard and Guy Fawkes, is also given a biography equal in length to that of Gorky, Huysmans, and Barrès; and we are told that he wrote tales which, despite all their shortcomings, were “invariably instructive, clean and manly.” Mrs. Ewing, too, profited by her pious proclivities, for her biography takes up almost as much space as that of the “moral” Macdonald and the “manly” Ainsworth. Her stories are “sound and wholesome in matter,” and besides, her best tales “have never been surpassed in the style of literature to which they belong.” Respectability and moral refinement were qualities also possessed by G. P. R. James, whose biography is equal in length to that of William Dean Howells. In it there is quite a long comparison of James with Dumas, though it is frankly admitted that as an artist James was inferior. His plots were poor, his descriptions were weak, and his dialogue was bad. Therefore “his very best books fall far below Les Trois Mousquetaires.” But, it is added, “James never resorted to illegitimate methods to attract readers, and deserves such credit as may be due to a purveyor of amusement who never caters to the less creditable tastes of his guests.” In other words, say what you will about James’s technique, he was, at any rate, an upright and impeccable gentleman! Even Mrs. Sarah Norton’s lofty moral nature is rewarded with biographical space greater than that of Huysmans or Gorky. Mrs.
- 77. Norton, we learn, “was not a mere writer of elegant trifles, but was one of the priestesses of the ‘reforming’ spirit.” One of her books was “a most eloquent and rousing condemnation of child labor”; and her poems were “written with charming tenderness and grace.” Great, indeed, are the rewards of virtue, if not in life, at least in the Encyclopædia Britannica. On the other hand, several English authors are condemned for their lack of nicety and respectability. Trollope, for instance, lacked that elegance and delicacy of sentiment so dear to the Encyclopædia editor’s heart. “He is,” we read, “sometimes absolutely vulgar—that is to say, he does not deal with low life, but shows, though always robust and pure in morality, a certain coarseness of taste.” Turning from the vulgar but pure Trollope to Charles Reade, we find more of this same kind of criticism: “His view of human life, especially of the life of women, is almost brutal ... and he cannot, with all his skill as a story-teller, be numbered among the great artists who warm the heart and help to improve the conduct.” (Here we have the Britannica’s true attitude toward literature. That art, in order to be great, must warm the heart, improve the conduct, and show one the way to righteousness.) Nor is Ouida to be numbered among the great uplifters. In her derogatory half-column biography we are informed that “on grounds of morality of taste Ouida’s novels may be condemned” as they are “frequently unwholesome.” Two typical examples of the manner in which truly great English writers, representative of the best English culture, are neglected in favor of those writers who epitomize England’s provincial piety, are to be found in the biographies of George Moore and Joseph Conrad, neither of whom is concerned with improving the readers’ conduct or even with warming their hearts. These two novelists, the greatest modern authors which England has produced, are dismissed peremptorily. Conrad’s biography draws but eighteen lines, about one-third of the space given to Marie Corelli; and the only praise accorded him is for his vigorous style and brilliant descriptions. In this superficial criticism we have an example of ineptitude, if not of
- 78. downright stupidity, rarely equaled even by newspaper reviewers. Not half of Conrad’s books are mentioned, the last one to be recorded being dated 1906, nearly eleven years ago! Yet this is the Encyclopædia which is supposed to have been brought up to date and to be adequate for purposes of reference! In the case of George Moore there is less excuse for such gross injustice (save that he is Irish), for Moore has long been recognized as one of the great moderns. Yet his biography draws less space than that of Jane Porter, Gilbert Parker, Maurice Hewlett, Rider Haggard, or H. G. Wells; half of the space given to Anthony Hope; and only a fourth of the space given to Mrs. Gaskell and to Mrs. Humphry Ward! A Mummer’s Wife, we learn, has “decidedly repulsive elements”; and the entire criticism of Esther Waters, admittedly one of the greatest of modern English novels, is that it is “a strong story with an anti-gambling motive.” It would seem almost incredible that even the tin-pot evangelism of the Encyclopædia Britannica would be stretched to such a length,—but there you have the criticism of Esther Waters set down word for word. The impelling art of this novel means nothing to the Encyclopedia’s critic; he cannot see the book’s significance; nor does he recognize its admitted importance to modern literature. To him it is an anti- gambling tract! And because, perhaps, he can find no uplift theme in A Mummer’s Wife, that book is repulsive to him. Such is the culture America is being fed on—at a price. Thomas Hardy, another one of England’s important moderns, is condemned for his attitude toward women: his is a “man’s point of view” and “more French than English.” (We wonder if this accounts for the fact that the sentimental James M. Barrie is accorded more space and greater praise.) Samuel Butler is another intellectual English writer who has apparently been sacrificed on the altar of Presbyterian respectability. He is given less than a column, a little more than half the space given the patriotic, tub-thumping Kipling, and less than half the space given Felicia Hemans. Nor is there any criticism of his work. The Way of all Flesh is merely mentioned in the list of his books. Gissing, another highly enlightened English writer, is
- 79. accorded less space than Jane Porter, only about half the space given Anthony Hope, and less space than is drawn by Marie Corelli! There is almost no criticism of his work—a mere record of facts. Mrs. M. E. Braddon, however, author of The Trail of the Serpent and Lady Audley’s Secret, is criticised in flattering terms. The biography speaks of her “large and appreciative public,” and apology is made for her by the statement that her works give “the great body of readers of fiction exactly what they require.” But why an apology is necessary one is unable to say since Aurora Floyd is “a novel with a strong affinity to Madame Bovary.” Mrs. Braddon and Flaubert! Truly a staggering alliance! Mrs. Henry Wood, the author of East Lynne, is given more space than Conrad; and her Johnny Ludlow tales are “the most artistic” of her works. But the “artistic” Mrs. Wood has no preference over Julia Kavanagh. This latter lady, we discover, draws equal space with Marcel Prévost; and she “handles her French themes with fidelity and skill.” Judging from this praise and the fact that Prévost gets no praise but is accused of having written an “exaggerated” and “revolting” book, we can only conclude that the English authoress handles her French themes better than does Prévost. George Meredith is accorded almost as much biographical space as Balzac; and in the article there appears such qualifying words as “seer,” “greatness,” and “master.” The impression given is that he was greater than Balzac. In Jane Porter’s biography, which is longer than that of Huysmans, we read of her “picturesque power of narration.” Even of Samuel Warren, to whom three-fourths of a column is allotted (more space than is given to Bret Harte, Lafcadio Hearn, or Gorky), it is said that the interest in Ten Thousand a Year “is made to run with a powerful current.” Power also is discovered in the works of Lucas Malet. The Wages of Sin was “a powerful story” which “attracted great attention”; and her next book “had an even greater success.” Joseph Henry Shorthouse, who is given more space than Frank Norris and Stephen Crane combined, possessed “high earnestness of purpose, a
- 80. luxuriant style and a genuinely spiritual quality.” Though lacking dramatic facility and a workmanlike conduct of narrative, “he had almost every other quality of the born novelist.” After this remark it is obviously necessary to revise our æsthetic judgment in regard to the religious author of John Inglesant. Grant Allen, alas! lacked the benevolent qualities of the “spiritual” Mr. Shorthouse, and—as a result, no doubt—he is given less space, and his work and vogue are spoken of disparagingly. One of his books was a succès de scandale “on account of its treatment of the sexual problem.” Mr. Allen apparently neither “warmed the heart” nor “improved the conduct” of his audience. On the other hand, Mrs. Oliphant, in a long biography, is praised for her “sympathetic touch”; and we learn furthermore that she was long and “honorably” connected with the firm of Blackwood. Maurice Hewlett has nearly a half-column biography full of praise. Conan Doyle, also, is spoken of highly. Kipling’s biography, longer than Mark Twain’s, Bourget’s, Daudet’s, or Gogol’s, also contains praise. In H. G. Wells’s biography, which is longer than that of George Moore, “his very high place” as a novelist is spoken of; and Anthony Hope draws abundant praise in a biography almost as long as that of Turgueniev! In the treatment of Mrs. Humphry Ward, however, we have the key to the literary attitude of the Encyclopædia. Here is an author who epitomizes that middle-class respectability which forms the Britannica’s editors’ standard of artistic judgment, and who represents that virtuous suburban culture which colors the Encyclopædia’s art departments. It is not surprising therefore that, of all recent novelists, she should be given the place of honor. Her biography extends to a column and two-thirds, much longer than the biography of Turgueniev, Zola, Daudet, Mark Twain, or Henry James; and over twice the length of William Dean Howells’s biography. Even more space is devoted to her than is given to the biography of Poe! Nor in this disproportionate amount of space alone is Mrs. Ward’s superiority indicated. The article contains the most fulsome praise, and we are told that her “eminence among latter-day women
- 81. novelists arises from her high conception of the art of fiction and her strong grasp on intellectual and social problems, her descriptive power ... and her command of a broad and vigorous prose style.” (The same enthusiastic gentleman who wrote Mrs. Ward’s biography also wrote the biography of Oscar Wilde. The latter is given much less space, and the article on him is a petty, contemptible attack written from the standpoint of a self-conscious puritan.) Thackeray is given equal space with Balzac, and in the course of his biography it is said that some have wanted to compare him with Dickens but that such a comparison would be unprofitable. “It is better to recognize simply that the two novelists stood, each in his own way, distinctly above even their most distinguished contemporaries.” (Both Balzac and Victor Hugo were their contemporaries, and to say that Thackeray stood “distinctly above” them is to butcher French genius to make an English holiday.) In Dickens’s biography, which is nearly half again as long as that of Balzac and nearly two and a half times as long as that of Hugo, we encounter such words and phrases as “masterpieces” and “wonderful books.” No books of his surpassed the early chapters of Great Expectations in “perfection of technique or in the mastery of all the resources of the novelist’s art.” Here, as in many other places, patriotic license has obviously been permitted to run wild. Where, outside of provincial England, will you find another critic, no matter how appreciative of Dickens’s talent, who will agree that he possessed “perfection of technique” and a “mastery of all the resources of the novelist’s art”? But, as if this perfervid rhetoric were not sufficiently extreme, Swinburne is quoted as saying that to have created Abel Magwitch alone is to be a god indeed among the creators of deathless men. (This means that Dickens was a god beside the mere mundane creator of Lucien de Rubempré, Goriot, and Eugénie Grandet.) And, again, on top of this unreasoned enthusiasm, it is added that in “intensity and range of creative genius he can hardly be said to have any modern rival.”
- 82. Let us turn to Balzac who was not, according to this encyclopædia, even Dickens’s rival in intensity and range of creative genius. Here we find derogatory criticism which indeed bears out the contention of Dickens’s biographer that the author of David Copperfield was superior to the author of Lost Illusions. Balzac, we read, “is never quite real.” His style “lacks force and adequacy to his own purpose.” And then we are given this final bit of insular criticism: “It is idle to claim for Balzac an absolute supremacy in the novel, while it may be questioned whether any single book of his, or any scene of a book, or even any single character or situation, is among the very greatest books, scenes, characters, situations in literature.” Alas, poor Balzac!—the inferior of both Dickens and Thackeray—the writer who, if the judgment of the Encyclopædia Britannica is to be accepted, created no book, scene, character or situation which is among the greatest! Thus are the world’s true geniuses disparaged for the benefit of moral English culture. De Vigny receives adverse criticism. He is compared unfavorably to Sir Walter Scott, and is attacked for his “pessimistic” philosophy. De Musset “had genius, though not genius of that strongest kind which its possessor can always keep in check”—after the elegant and repressed manner of English writers, no doubt. De Musset’s own character worked “against his success as a writer,” and his break with George Sand “brought out the weakest side of his moral character.” (Again the church-bell motif.) Gautier, that sensuous and un-English Frenchman, wrote a book called Mademoiselle de Maupin which was “unfitted by its subject, and in parts by its treatment, for general perusal.” Dumas père is praised, largely we infer, because his work was sanctioned by Englishmen: “The three musketeers are as famous in England as in France. Thackeray could read about Athos from sunrise to sunset with the utmost contentment of mind, and Robert Louis Stevenson and Andrew Lang have paid tribute to the band.” Pierre Loti, however, in a short biography, hardly meets with British approval. “Many of his best books are long sobs of remorseful memory, so personal, so intimate, that an English reader is amazed
- 83. to find such depth of feeling compatible with the power of minutely and publicly recording what is felt.” Loti, like de Musset, lacked that prudish restraint which is so admirable a virtue in English writers. Daudet, in a short and very inadequate biography, is written down as an imitator of Dickens; and in Anatole France’s biography, which is shorter than Marryat’s or Mrs. Oliphant’s, no adequate indication of his genius is given. Zola is treated with greater unfairness than perhaps any other French author. Zola has always been disliked in England, and his English publisher was jailed by the guardians of British morals. But it is somewhat astonishing to find to what lengths this insular prejudice has gone in the Encyclopædia Britannica. Zola’s biography, which is shorter than Mrs. Humphry Ward’s, is written by a former Accountant General of the English army, and contains adverse comment because he did not idealize “the nobler elements in human nature,” although, it is said, “his later books show improvement.” Such scant treatment of Zola reveals the unfairness of extreme prejudice, for no matter what the nationality, religion, or taste of the critic, he must, in all fairness, admit that Zola is a more important and influential figure in modern letters than Mrs. Humphry Ward. In the biography of George Sand we learn that “as a thinker, George Eliot is vastly [sic] superior; her knowledge is more profound, and her psychological analysis subtler and more scientific.” Almost nothing is said of Constant’s writings; and in the mere half- column sketch of Huysmans there are only a few biographical facts with a list of his books. Of Stendhal there is practically no criticism; and Coppée “exhibits all the defects of his qualities.” René Bazin draws only seventeen lines—a bare record of facts; and Édouard Rod is given a third of a column with no criticism. Despite the praise given Victor Hugo, his biography, from a critical standpoint, is practically worthless. In it there is no sense of critical proportion: it is a mere panegyric which definitely states that Hugo was greater than Balzac. This astonishing and incompetent praise is accounted for when we discover that it was written by Swinburne
- 84. who, as is generally admitted, was a better poet than critic. In fact, turning to Swinburne’s biography, we find the following valuation of Swinburne as critic: “The very qualities which gave his poetry its unique charm and character were antipathetic to his success as a critic. He had very little capacity for cool and reasoned judgment, and his criticism is often a tangled thicket of prejudices and predilections.... Not one of his studies is satisfactory as a whole; the faculty for the sustained exercise of the judgment was denied him, and even his best appreciations are disfigured by error in taste and proportion.” Here we have the Encyclopædia’s own condemnation of some of its material—a personal and frank confession of its own gross inadequacy and bias! And Swinburne, let it be noted, contributes no less than ten articles on some of the most important literary men in history! If the Encyclopædia Britannica was as naïf and honest about revealing the incapacity of all of its critics as it is in the case of Swinburne, there would be no need for me to call attention to those other tangled thickets of prejudices and predilections which have enmeshed so many of the gentlemen who write for it. But the inadequacy of the Britannica as a reference book on modern French letters can best be judged by the fact that there appears no biographical mention whatever of Romain Rolland, Pierre de Coulevain, Tinayre, René Boylesve, Jean and Jérôme Tharaud, Henry Bordeaux, or Pierre Mille. Rolland is the most gifted and conspicuous figure of the new school of writers in France to-day, and the chief representative of a new phase of French literature. Pierre de Coulevain stands at the head of the women novelists in modern France; and her books are widely known in both England and America. Madame Tinayre’s art, to quote an eminent English critic, “reflects the dawn of the new French spirit.” Boylesve stands for the classic revival in French letters, and ranks in the forefront of contemporary European writers. The Tharauds became famous as novelists as far back as 1902, and hold a high place among the writers of Young France. Bordeaux’s novels have long been familiar in translation even to American readers; and Pierre Mille holds very
- 85. much the same place in France that Kipling does in England. Yet not only does not one of these noteworthy authors have a biography, but their names do not appear throughout the entire Encyclopædia! In the article on French Literature the literary renaissance of Young France is not mentioned. There apparently has been no effort at making the account modern or up-to-date in either its critical or historical side; and if you desire information on the recent activities in French letters—activities of vital importance and including several of the greatest names in contemporary literature—you need not seek it in the Britannica, that “supreme” book of knowledge; for apparently only modern English achievement is judged worthy of consideration. Modern Russian literature suffers even more from neglect. Dostoievsky has less than two columns, less space than Charles Reade, George Borrow, Mrs. Gaskell, or Charles Kingsley. Gogol has a column and a quarter, far less space than that given Felicia Hemans, James M. Barrie, of Mrs. Humphry Ward. Gorky is allotted little over half a column, one-third of the space given Kipling, and equal space with Ouida and Gilbert Parker. Tolstoi, however, seems to have inflamed the British imagination. His sentimental philosophy, his socialistic godliness, his capacity to “warm the heart” and “improve the conduct” has resulted in a biography which runs to nearly sixteen columns! The most inept and inadequate biography in the whole Russian literature department, however, is that of Turgueniev. Turgueniev, almost universally conceded to be the greatest, and certainly the most artistic, of the Russian writers, is accorded little over a column, less space than is devoted to the biography of Thomas Love Peacock, Kipling, or Thomas Hardy; and only a half or a third of the space given to a dozen other inferior English writers. And in this brief biography we encounter the following valuation: “Undoubtedly Turgueniev may be considered one of the great novelists, worthy to be ranked with Thackeray, Dickens and George Eliot; with the genius of the last of these he has many affinities.” It will amuse, rather than
- 86. Welcome to Our Bookstore - The Ultimate Destination for Book Lovers Are you passionate about books and eager to explore new worlds of knowledge? At our website, we offer a vast collection of books that cater to every interest and age group. From classic literature to specialized publications, self-help books, and children’s stories, we have it all! Each book is a gateway to new adventures, helping you expand your knowledge and nourish your soul Experience Convenient and Enjoyable Book Shopping Our website is more than just an online bookstore—it’s a bridge connecting readers to the timeless values of culture and wisdom. With a sleek and user-friendly interface and a smart search system, you can find your favorite books quickly and easily. Enjoy special promotions, fast home delivery, and a seamless shopping experience that saves you time and enhances your love for reading. Let us accompany you on the journey of exploring knowledge and personal growth! ebookgate.com