Skip to content

Glossary

Terms used in the book, with the chapter where each is developed

A

Aliasing
The appearance of a frequency above half the sampling rate as a lower frequency after sampling; it cannot be undone after sampling (Chapter 9).
Amplitude modulation (AM)
Multiplying a message by a sinusoidal carrier, which moves the message spectrum to the carrier frequency; conventional AM also transmits the carrier (Chapters 6, 14).
Analog signal
A signal whose amplitude can take any value in a continuous range (Chapter 1).
Anti-aliasing filter
An analog low-pass filter placed before a sampler to remove frequencies above half the sampling rate (Chapter 9).

B

Bandwidth
The width of the band of frequencies that holds most of a signal's energy, or that a system passes; several definitions exist, such as the 3 dB bandwidth (Chapters 6, 12).
BIBO stability
Bounded-input, bounded-output stability: every bounded input gives a bounded output. For an LTI system it holds if and only if the impulse response is absolutely summable or integrable (Chapters 3, 4).
Bilinear transform
The substitution \(s = \frac{2}{T}\frac{1 - z^{-1}}{1 + z^{-1}}\), which turns a stable analog filter into a stable digital filter with warped frequency axis (Chapter 12).
Bode plot
Magnitude in decibels and phase in degrees plotted against frequency on a logarithmic axis (Chapter 12).
Butterfly
The basic FFT operation \(E[k] \pm W_N^kO[k]\), which forms two outputs from two inputs with one multiplication (Chapter 8).
Butterworth filter
A filter with maximally flat passband, \(|H|^2 = 1/(1 + (\omega/\omega_c)^{2N})\) (Chapter 12).

C

Causal system
A system whose output never depends on future inputs; for an LTI system, \(h = 0\) for negative time (Chapters 3, 4).
Circular convolution
Convolution of two length-\(N\) sequences in which shifts wrap around modulo \(N\); it corresponds to multiplying DFTs (Chapter 8).
Comb filter
A filter whose magnitude response has regularly spaced peaks and dips, produced by adding a delayed copy of a signal (Chapter 14).
Complex exponential
The signal \(e^{st}\) or \(z^n\); an eigenfunction of every LTI system (Chapters 2, 4).
Conjugate symmetry
The property \(X(-j\omega) = X^*(j\omega)\) (or \(a_{-k} = a_k^*\)) of the spectrum of every real signal (Chapters 5 to 8).
Convolution
The operation \(y = x * h\), \(y[n] = \sum_k x[k]h[n-k]\) or \(y(t) = \int x(\tau)h(t - \tau)d\tau\), that gives the output of an LTI system (Chapter 4).
Corner frequency
The frequency at which the straight-line asymptotes of a first-order Bode plot meet, 3 dB from the true curve (Chapter 12).

D

Decibel (dB)
A logarithmic measure of gain, \(20\log_{10}|H|\) (Chapter 12).
Decimation
Keeping every \(M\)th sample of a sequence, \(x[Mn]\); it must be preceded by low-pass filtering to avoid aliasing (Chapters 2, 9).
Difference equation
An equation relating the present output of a DT system to past outputs and present and past inputs (Chapters 3, 11).
Dirichlet conditions
Conditions on a periodic signal under which its Fourier series converges to it at points of continuity and to the midpoint at jumps (Chapter 5).
Discrete Fourier transform (DFT)
The transform \(X[k] = \sum_{n=0}^{N-1}x[n]e^{-j2\pi kn/N}\) of a finite sequence; it samples the DTFT at \(N\) frequencies (Chapter 8).
Discrete-time Fourier transform (DTFT)
The transform \(X(e^{j\Omega}) = \sum_n x[n]e^{-j\Omega n}\); always periodic with period \(2\pi\) (Chapter 7).
Discrete-time signal
A signal defined only at integer values of its index, \(x[n]\) (Chapter 1).

E

Eigenfunction
A signal that passes through a system unchanged in shape, only scaled; complex exponentials are eigenfunctions of LTI systems (Chapter 4).
Energy signal
A signal with finite, non-zero energy \(\int|x|^2dt\); its average power is zero (Chapter 1).
Envelope detector
A simple AM receiver that follows the peaks of the modulated carrier (Chapter 14).
Even signal
A signal with \(x(-t) = x(t)\) (Chapter 1).

F

Fast Fourier transform (FFT)
An algorithm that computes the DFT in about \(\frac{N}{2}\log_2N\) multiplications instead of \(N^2\) (Chapter 8).
FIR filter
A filter with a finite impulse response; non-recursive, always stable, and able to have exactly linear phase (Chapters 11, 12).
Fourier series
The representation of a periodic signal as a sum of harmonically related complex exponentials (Chapter 5).
Fourier transform
The representation of an aperiodic signal as a continuous superposition of complex exponentials, \(X(j\omega) = \int x(t)e^{-j\omega t}dt\) (Chapter 6).
Frequency response
The factor \(H(j\omega)\) or \(H(e^{j\Omega})\) by which an LTI system multiplies a complex exponential; the Fourier transform of the impulse response (Chapters 5 to 7).

G

Gibbs phenomenon
The overshoot of about 9% of the jump that persists near a discontinuity in partial Fourier sums (Chapter 5).
Group delay
The negative derivative of phase with respect to frequency; the delay of a narrow band of frequencies (Chapter 12).

I

IIR filter
A filter with an infinite impulse response, implemented recursively, with poles away from the origin (Chapters 11, 12).
Impulse response
The output of a system when the input is a unit impulse; it characterises an LTI system completely (Chapter 4).
Initial rest
The condition that a system's output is zero until its input becomes non-zero (Chapter 3).
Intersymbol interference (ISI)
Overlap of successive symbols in a digital link caused by the spreading effect of the channel (Chapter 14).
Invertible system
A system whose input can be recovered from its output (Chapters 3, 4).

L

Laplace transform
The transform \(X(s) = \int x(t)e^{-st}dt\), which generalises the Fourier transform to growing signals and unstable systems (Chapter 10).
Leakage
The spreading of a sinusoid's energy across many DFT bins when the record does not hold a whole number of its cycles (Chapter 8).
Linear phase
A phase response that is a straight line through the origin, corresponding to a pure delay (Chapters 6, 12).
Linear system
A system that obeys superposition: a weighted sum of inputs gives the same weighted sum of outputs (Chapter 3).
LTI system
A system that is both linear and time-invariant (Chapters 3, 4).

M

Matrix exponential
\(e^{At} = I + At + (At)^2/2! + \cdots\), which gives the free response of a state-space model (Chapter 13).
Memoryless system
A system whose output at each time depends only on the input at that time (Chapter 3).
Modulation
Multiplying a signal by a carrier to move its spectrum (Chapters 6, 14).

N

Notch filter
A filter that removes a narrow band around one frequency, made with zeros on the unit circle and poles just inside (Chapter 12).
Nyquist frequency
Half the sampling rate; the highest frequency a sampled signal can represent (Chapter 9).
Nyquist rate
Twice the highest frequency in a band-limited signal; sampling must be faster than this (Chapter 9).

O

Odd signal
A signal with \(x(-t) = -x(t)\) (Chapter 1).
OFDM
Orthogonal frequency-division multiplexing, which carries data on many subcarriers generated and separated by the FFT (Chapter 14).
Orthogonality
The property that harmonically related exponentials integrate (or sum) to zero against each other over a period; it gives the Fourier analysis formulas (Chapter 5).

P

Parseval's relation
The equality of energy (or power) computed in the time domain and in the frequency domain (Chapters 5 to 8).
Partial fractions
The expansion of a rational transform into simple terms that can be inverted from a table (Chapters 10, 11).
Periodic signal
A signal that repeats: \(x(t + T) = x(t)\) for all \(t\); the smallest such \(T > 0\) is the fundamental period (Chapter 1).
Phase margin
The extra phase lag a feedback loop can tolerate, at the frequency where the loop gain is 1, before becoming unstable (Chapter 14).
PID controller
A controller whose output is a weighted sum of the error, its integral and its derivative (Chapter 14).
Pole
A value of \(s\) or \(z\) at which a transfer function becomes infinite; poles govern stability and the natural modes of a system (Chapters 10, 11).
Power signal
A signal with finite, non-zero average power; its energy is infinite (Chapter 1).

Q

Quantization
Rounding signal values to a finite set of levels; with \(B\) bits the SNR for a full-scale sinusoid is about \(6.02B + 1.76\) dB (Chapter 9).

R

Region of convergence (ROC)
The set of \(s\) or \(z\) for which a Laplace or z-transform converges; it is part of the transform and distinguishes signals with the same formula (Chapters 10, 11).
Residue
The coefficient of a simple pole term in a partial-fraction expansion (Chapters 10, 11).

S

Sampling theorem
A signal band-limited to \(\omega_M\) is determined by samples taken at a rate \(\omega_s > 2\omega_M\) (Chapter 9).
Sifting property
\(\int x(t)\delta(t - t_0)dt = x(t_0)\) (Chapter 2).
Sinc function
\(\mathrm{sinc}(x) = \sin(\pi x)/(\pi x)\); the Fourier transform of a rectangular pulse and the interpolating function of ideal reconstruction (Chapters 2, 6, 9).
Spectrogram
A picture of the magnitude of the short-time Fourier transform against time and frequency (Chapter 14).
Spectrum
The description of a signal by the amplitudes and phases of its frequency components (Chapters 5, 6).
State
The smallest set of variables that, with future inputs, determines a system's future behaviour (Chapter 13).
Step response
The output of a system when the input is a unit step; for an LTI system, the running integral of the impulse response (Chapter 4).
System function
The transfer function of a DT system, \(H(z)\) (Chapter 11).

T

Time-invariant system
A system for which delaying the input delays the output by the same amount and changes nothing else (Chapter 3).
Transfer function
The Laplace (or z-) transform of the impulse response; the ratio of output to input transforms (Chapters 10, 11).

U

Unit impulse
\(\delta[n]\), equal to 1 at \(n = 0\) and 0 elsewhere; or \(\delta(t)\), the unit-area limit of ever-narrower pulses (Chapter 2).
Unit step
\(u(t)\) or \(u[n]\): zero before the origin and one after it (Chapter 2).

W

Window
A tapering sequence applied before a DFT to reduce leakage, or to an ideal impulse response in FIR design (Chapters 8, 12).

Z

Zero
A value of \(s\) or \(z\) at which a transfer function is zero (Chapters 10, 11).
z-transform
The transform \(X(z) = \sum x[n]z^{-n}\); on the unit circle it equals the DTFT (Chapter 11).
Zero-order hold
A reconstruction method that holds each sample value until the next one (Chapter 9).