Introduc)on to Signal Processing

Why Digi/ze?
Part 1: Introduc)on to Signal Processing
1. Classifica/on and Characteris/cs of Signals 1. To solve noise issues
2. Sampling Theorem and Aliasing
2. To reduce hardware costs
3. Difference Equa/ons for FIR and IIR Filters
4. Convolu/on and Correla/on 3. To store and reuse informa/on easily
Classifica)on and Characteris)cs of Signals 4. To achieve faster and more reliable data
transfer
Electronic Signal
• a /me varying voltage, current or EM wave that 5. To reduce system’s weight and size
poten/ally provides informa/on on the status of
a physical system, or conveys a message between
observers, among other possibili/es
Digi)za)on

• taking analog informa/on and encoding it into

1s and 0s so that computers can store, process,
and transmit such informa/on

✓ Sampling – process of measuring the

instantaneous values (samples) of con/nuous-
/me signal in a discrete form

✓ Quan/zing – process of discre/zing the sample

points of the sampled signal with a number of
quan/za/on levels

Analog signals – con/nuous-wave signals that

change with /me period and thus represented
with the sine waves

Digital signals – discrete in nature, represented

with square waves.
Common Basic Electronic Signals Spectrum
• used to classify something, or suggest that it
can be classified, in terms of its posi/on on a
scale between two extreme or opposite points
o Signal Spectrum – describes a signal's
magnitude and phase characteris/cs as a
func/on of frequency
o System Spectrum – describes how the
system changes signal magnitude and
phase as a func/on of frequency

The Audio Spectrum

Other Classifica/on of Signals

The Electromagne/c Spectrum

System
• may be referred to any set of
components, which func/on in
interrelated manner for a common cause
or objec/ve.
• has the following characteris/cs:
✓ Organiza/on
✓ Interac/on
✓ Interdependence
✓ Integra/on
✓ Central Objec/ve
Proper/es of a System Signal analysis is the study of collec/ng,
• Memory understanding, and deducing informa/on and
• Casualty intelligence from various signals
• Stability
• Inver/bility Signal processing covers the analysis,
• Time-Invariance modifica/on, and synthesis of signals.
• Linearity

Time Response
• occurs when the output of a system for an
input varies with respect to /me
• consists of two parts:
- Transient Response
- Stead-State Response

Basic Opera/ons of Signal Processing

Frequency Response
• quan/ta/ve measure of the output
spectrum of a system, that provides useful
insights into its stability and performance
characteris/cs
• system response to sinusoidal inputs at
varying frequencies
Digital Signal Processing (DSP) is the
mathema/cs, the algorithms, and the techniques
used to manipulate signals a`er they have been
converted into a digital form. This includes variety
of goals, such as:
1. Enhancement of Visual Images
2. Data Compression for Storage and
Transmission
3. Speech Recogni/on & Genera/on

Why DSP?
- highly linear (within limited dynamic
range), allowing noise to be control easily
a`er ini/al quan/za/on
- complex algorithms fit into a single chip,
while allowing flexibility for easy
adjustment of parameters via so`ware
- insensi/ve to component tolerances,
aging, environmental condi/ons,
electromagne/c interference
Sampling Theorem and Aliasing EXAMPLE:

Sampling Theorem A digital communica/on system is designed to

• specifies the minimum-sampling rate at ensure accurate representa/on of the audio
which a con/nuous-/me signal needs to signal to be transmiked with a maximum
be uniformly sampled so that the original frequency of 4 kHz. Calculate the minimum
signal can be completely recovered or sampling rate and the required bandwidth for the
reconstructed by these samples alone system to accurately capture and reproduce the
• states that a signal has to be sampled at audio signal without aliasing.
least with twice the frequency of the
original signal

𝑓S > 𝑓N = 2𝐵 or 𝑓N ≥ 2𝑓MAX and 𝑓S ≥ 2𝐵

• also known as the Shannon-Nyquist

Theorem (or Nyquist Theorem)
Informa=on Theory studies the quan/fica/on,
Sampling converts the independent variable storage, and communica/on of informa/on. It is
(/me) from con/nuous to discrete the mathema/cal treatment of the concepts,
parameters and rules governing the transmission
of messages through communica/on systems
How the Sample/Hold Process Works

Sampling Period
• the fixed gap between the samples of the
discre/zed signal
• reciprocal of the Nyquist Rate (𝑓N) or the
sampling rate at which the data in the
message signal should neither be lost nor
it should get over-lapped as described by
the Sampling Theorem
EXAMPLE: A wireless communica/on system is
designed to transmit digital data over a channel
Sampling Rate and Bandwidth in Wireless
with a signal-to-noise ra/o (SNR) of 20 dB. The
Communica/ons
system uses a sampling rate of 10,000 samples
per second. Calculate the maximum bandwidth of
the channel according to the Shannon-Hartley
theorem.
Aliasing An/-Aliasing Filter Design : Filter Parameters

• phenomenon of a 1. Passband
high-frequency 2. Stopband
component in the 3. Akenua/on
spectrum of a 4. Inser/on loss
signal, taking on 5. Impedance
the iden/ty of a 6. Ripple
low-frequency 7. Shape Factor (or Bandwidth Ra/o)
component in the 8. Quality Factor (Q)
spectrum of its 9. Poles and Zeros
sampled version 10. Envelope/Time Delay
• correc/ve measures to reduce its effect 11. Roll-off Rate (or Akenua/on Rate)
include: 12. Cutoff Frequency (-3dB Frequency)
o sampling the filtered signal at a rate 13. Center Frequency (f0)
slightly higher than the Nyquist rate 14. Bandwidth (BW)
o use of an an=-aliasing filter at the
transmiker side to eliminate
unwanted high frequency components

An/-Aliasing Filter Design: Types of Filter

Based on Frequency Response

EXAMPLE:
A digital data acquisi/on system is designed to
Based on Order sample an analog signal at a rate of 10,000
samples per second. The analog signal contains
frequency components up to 4 kHz. To avoid
aliasing during the sampling process, design an
an/-aliasing filter for the sampling process.

Based on Topology
Pulse Code Modula=on (PCM) is a standardized Quan)za)on Parameters
method used in telephone network to change an
analog signal to a digital one. Bit Depth
à The analog signal is first sampled at • no. of bits of informa/on per sample
sample rate of 8 kHz, then each sample is • directly corresponds to the resolu/on of
quan/zed into 1 of 256 levels and then each sample
encoded into digital 8-bit words.
Representa)on Levels
• discrete amplitudes of quan/zed output
• also known as reconstruc6on levels

Resolu)on
• number of representa6on levels over the
range of analog values

Quantum
• spacing between two adjacent
representa/on levels
• also known as step-size

Quan)za)on Error
• difference between an input value and its
quan/zed value

Dynamic Range
• ra/o of the highest amplitude to the
lowest amplitude of the quan/zed signal

Quan)za)on and Encoding

EXAMPLE: Standard Audio Digi)za)on Parameters

Calculate the number of quan/za/on levels

required for PCM encoding to ensure an adequate
fidelity during signal reconstruc/on of an analog
signal that has a bandwidth of 4 kHz at a rate of
20,000 samples per second. Assuming the signal
voltage range to be -5V to +5V, solve also for the
resolu/on, quan/za/on error, and dynamic range
of the quan/za/on process.

EXAMPLE: Digital Image Parameters

• Resolu=on – number of pixels in an image
A PCM system is used to digi/ze an analog audio or the measure of width and height of the
signal having a voltage range of -2V to +2V. image in terms of pixels
Determine the system's resolu/on at 8-bit • Pixels (Picture Elements) – likle dots that
quan/za/on for PCM encoding and write the make up the images on video frames
coded value of the given por/on of the signal • Color Depth (Bit Depth) – number of bits
used to indicate the color of a single pixel
• Aspect Ra=o – ra/o of image width to its
height, denoted by a colon in between
• Frame Rate (frames per second or FPS) is
the rate at which consecu/ve images
(frames) appear on a display
Resolu/on and Pixels

Screen Size

Aspect Ra/o

Frame Rate
• number of frames per second (FPS)
• frequency (rate) at which consecu/ve
images (frames) are captured or displayed.
Standard Video Digi/za/on Parameters