T-61.3010 Digital Signal Processing 22.1.

2007
and Filtering

Introduction
• Examples of signals: Speech, music,
picture, and video signals

1
Signals and • A signal is a function of independent
Signal Processing
variable, e.g., time, distance, position,
temperature, pressure etc.
• Most signals are generated by natural
means
• Signals may be generated synthetically
or by computer simulation
© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 2

Signal Processing Characterization of Signals

• One-dimensional (1-D) signal:
• A signal carries information and the
– Function of a single independent variable,
objective of signal processing is to e.g., speech signal, s(t)
extract the information carried by the • Two-dimensional (2-D) signal:
signal, i.e., – Two independent variables, e.g., image s(x,y)
Signal processing is concerned with • Multidimensional signal:
the mathematical representation of – Black and white video signal is a 3-D signal, two
the signal and the algorithmic spatial variables and time, i.e., v(x,y,t)
operation carried out to extract the – Color video signal has three channels of 3-D
signals (RGB), i.e., u(x,y,t)=[r(x,y,t) g(x,y,t) b(x,y,t)]T
information present
© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 3 © 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 4

Characterization of Signals Elementary Signal Processing Operations

(a) Analog signal:
Continuous in both
time and amplitude
• Scaling: y[n] = ax[n]
y[n] = x[n − n0 ]
(b) Digital signal:
Discrete in both time • Delay:
and amplitude
(c) Sampled-data • Addition: y[n] = ax1[n] + bx2 [n] − cx3 [n]
signal:
Discrete-time and
continuous- • Multiplication: y[n] = x1[n]x2 [n]
amplitude signal
(d) Quantized boxcar
signal:
Complex signal processing operations are
Continuous-time implemented by combining two or more
and discrete-
amplitude signal
elementary operations
© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 5 © 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 6

Mitra: Chapter 1 / Copyright © 2007

Olli Simula 1
T-61.3010 Digital Signal Processing 22.1.2007
and Filtering

Filtering Convolution
• Filtering is used to pass certain frequency ∞
components in a signal through the system y[n] = ∑ x[k ]h[n − k ] = x[n] * h[n]
k = −∞
without any distortion and to block other
frequency components
• Linear filtering is described in time-domain x[n] h[n] y[n]
by the convolution operation
∞ +∞
y[n] = ∑h[n − k]x[k]
k =−∞
y (t ) = ∫ x(τ )h(t − τ )dτ = x(t ) * h(t )
−∞

© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 7 © 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 8

Example: Filtering of a Signal Output of Different Filters

• Consider an input signal consisting of three
sinusoidal components of frequencies 50
Hz, 110 Hz, and 210 Hz
(c) Cutoff frequency at 150 Hz
(b) Cutoff frequency at 80 Hz

(d) Cutoff frequencies at 80 Hz and 150 Hz (d) Cutoff frequencies at 80 Hz and 150 Hz

© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 9 © 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 10

Modulation and Demodulation Modulation and Demodulation

• In amplitude modulation, the amplitude of a • The modulated signal y(t), is composed of two
high-frequency sinusoidal signal Acos(Ω0t) ,called sinusoidal signals at frequencies Ω0+Ω1 and Ω0-Ω1
the carrier signal, is varied by the low-frequency A A
Y ( jΩ) = X ( j(Ω + Ω0 )) + X ( j(Ω − Ω0 ))
bandlimited signal x(t)= cos(Ω1t) (with Ω1<<Ω0), 2 2
called the modulating signal, generating a high-
frequency signal y(t), called the modulated • Spectrum of the
modulating signal x(t)
signal, according to
y(t ) = Ax(t ) cos(Ω0t) = Acos(Ω1t ) cos(Ω0t)
• Spectrum of the
A A modulated signal y(t)
= cos((Ω0 + Ω1 )t ) + cos((Ω0 − Ω1 )t )
2 2
© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 11 © 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 12

Mitra: Chapter 1 / Copyright © 2007

Olli Simula 2
T-61.3010 Digital Signal Processing
and Filtering
Modulation and Demodulation
x(t) x y(t)
Acos(Ω0 t)
Modulator
• The envelope of the modulated carrier is the waveform
of the modulating signal
© 2007 Olli Simula
Example: Speech Waveform
© 2007 Olli Simula
Telephone and Dialing Systems
Pressing of each button generates a
unique two-tone signal, e.g.,
pressing ”4” corresponds to
frequencies 770 Hz and 1209 Hz
© 2007 Olli Simula
Mitra: Chapter 1 / Copyright © 2007
Olli Simula
r(t)
Lowpass
y(t) x (A/2)x(t)
filter
cos(Ω0 t)
Demodulator
T-61.3010 / Mitra: Chapter 1 13
(a) Sentence-length segment
(b) Magnified version of the
voiced segment (letter A)
(c) Magnified version of the
voiced segment (letter S)
T-61.3010 / Mitra: Chapter 1 15
(1) The two tones are separated using
lowpass and highpass filters
(2) Then their frequencies are detected
using bandpass filters
T-61.3010 / Mitra: Chapter 1 17
22.1.2007
Examples of Typical Signals
• Medical signals: ECG (electrocardiography) and
EEG (Electroencephalogram) signals
(a) A typical ECG trace and (b) one cycle of a ECG waveform
© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 14
Signal Processing Applications
• Sound recording applications
– Compressors and limiters
– Expanders and noise gates
– Equalizers and filters
– Noise reduction systems
– Delay and reverberation systems
• Telephone and dialing applications
• Electronic music synthesis
• Echocancellation in telephone systems
© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 16
Echo Cancellation
Signal paths in a telephone network: Echo supression detects the direction of
(a) Transmission path, (b) echo path for the conversation and blocks the opposite path;
talker, and (c) echo path for the listener Echo canceler is an adaptive filter
© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 18
3
T-61.3010 Digital Signal Processing 22.1.2007
and Filtering

System Block Diagram

Why Digital Signal Processing ?
• 17th century: Finite difference methods,
numerical integration methods, numerical
interpolation
• 1950s: Simulation of analog signal
processing methods
• 1960s: DSP started to be a field by itself due
to an increase of computational power and
applications, e.g., space research
• 1965: Invention of the FFT algorithm !

© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 19 © 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 20

Advantages of DSP
• Accuracy and stability
• Flexibility due to (re)programmability
• Implementation on VLSI and ASIC
• Functions that are not possible in
analog signal processing
• Digitalization …

© 2007 Olli Simula T-61.3010 / Mitra: Chapter 1 21

Mitra: Chapter 1 / Copyright © 2007

Olli Simula 4