Digital Signal Processing: Core • Why care about breaking down

Concepts and Transforms waves? In music, it helps analyze

notes; in engineering, it aids in
This study material covers fundamental
understanding electrical signals; in
topics in Digital Signal Processing, including
image processing, it's used for
Fourier Series, Fourier Transform, Z-
image compression.
Transform, Discrete-Time Fourier Transform
(DTFT), Discrete Fourier Transform (DFT), • History: Introduced by Jean-
and Frequency Response. Baptiste Joseph Fourier, a French
mathematician and physicist, in his
1. Fourier Series and Fourier Transform 1822 work, "Théorie analytique de la
Instructor: CK Caoile Course: Digital chaleur" (The Analytical Theory of
Signal Processing: Fourier Series and Heat).
Transform • Initial Application: Used to
Intended Learning Outcomes: At the end describe the flow of heat in objects.
of the lesson, you should be able to:
• Impact: Has a significant impact on
• Provide spectral representations for modern signal processing, electrical
periodic continuous-time signals. engineering, and various fields in
physics.
• Apply a limiting process to the
Fourier series. 1.2. Fourier Transform

• Evaluate the application of the • Definition: A mathematical tool

Fourier series and Fourier transform used for frequency analysis of
to a variety of standard continuous- signals.
time signals.
• Distinction from Laplace
1.1. Fourier Series Transform: While Fourier Transform
is for frequency analysis, Laplace
• Definition: A Fourier series is a Transform is a mathematical tool for
mathematical method for the analysis of circuits and systems.
representing a periodic function
as a sum of sine and cosine • Existence: The Fourier Transform
functions. can exist for Energy Signals, Power
Signals, and Impulse Related
• Purpose: It simplifies complex Signals.
periodic signals by breaking them
down into basic waveforms, making • Formulas:
them easier to analyze and o Forward Fourier
manipulate. Transform: $X(j\omega) =
• Analogy: Imagine recreating a song \int_{-\infty}^{\infty} x(t)e^{-
with a few musical notes; by picking j\omega t} dt$
the right notes and playing them o Inverse Fourier Transform:
correctly, you can approximate the
$x(t) = \frac{1}{2\pi} \int_{-
original melody, similar to how
\infty}^{\infty}
Fourier series breaks down complex
X(j\omega)e^{j\omega t}
waves.
d\omega$
 • Conditions for Existence • Uses:
(Dirichlet Conditions):
o Simplifies analysis of discrete
1. A signal should have a finite number Linear Time-Invariant (LTI)
of maxima and minima over any finite systems.
interval.
o Helps in stability analysis
2. A signal should have a finite number (poles inside the unit circle
of discontinuities over any finite interval. indicate a stable system).
3. A signal should be absolutely o Used in difference
integrable, i.e., $\int_{-\infty}^{\infty} |x(t)| dt equations (digital filter
< \infty$. design).
2. Z-Transform and Related Transforms o Provides a link between the
time domain and the
Instructor: CK Caoile Course: Digital
frequency domain.
Signal Processing: Z-Transform and
Related Transforms 2.2. Discrete-Time Fourier Transform
(DTFT)
Intended Learning Outcomes: At the end
of the lesson, you should be able to: • Formula: $X(e^{j\omega}) =
\sum_{n=-\infty}^{\infty} x[n]e^{-
• Apply Z-transform in analyzing
j\omega n}$. (Note: Source uses
discrete signals.
$X(z)$ in the formula's notation, but
• Differentiate between DTFT, DFT, the context clearly indicates DTFT
and ZT. and its continuous frequency
domain, making the $z$ an error in
• Use MATLAB/Python to compute notation and should be
and visualize transforms. $e^{j\omega}$).
2.1. Z-Transform (ZT) • Domain: Continuous frequency ($-
• Definition: The Z-Transform \pi \leq \omega \leq \pi$).
converts a discrete-time signal • Output: Periodic spectrum.
into a complex frequency domain
representation. It is a general tool • Use: Theoretical frequency analysis
for analyzing discrete signals in the of discrete-time signals.
frequency domain.
2.3. Discrete Fourier Transform (DFT)
• Formula: $X(z) = \sum_{n=-
• Formula: $X[k] = \sum_{n=0}^{N-1}
\infty}^{\infty} x[n]z^{-n}$.
x[n]e^{-j \frac{2\pi}{N} kn}$.
• Inverse Z-Transform: Retrieves the
• Domain: Discrete frequency
time-domain sequence from $X(z)$.
samples ($k = 0, \ldots, N-1$).
• Region of Convergence (ROC):
• Implementation: Implemented
Represents the values of $z$ where
efficiently using FFT (Fast Fourier
the transform converges. For a
Transform).
stable casual system, the ROC
would be, for example, $|z| > 0.5$ if • Use: Practical computation of
$x[n] = (0.5)^n u[n]$. spectra.
2.4. Comparison Table: DTFT vs DFT vs • It consists of two components:
ZT Magnitude response and Phase
response.
Z-
Feature DTFT DFT Transfor 3.2. Magnitude Response
m
• Notation: $|H(e^{j\omega})|$.
Input Finite or • Purpose: Shows how much each
Infinite Finite
length infinite frequency is amplified or
attenuated.
Frequenc Continuo Complex
Discrete
y axis us plane • Application: Used to identify filter
behavior (e.g., in a low-pass filter,
Expressio low frequencies pass with high
Output Infinite n with magnitude, while high frequencies
N samples
length samples poles/zer are attenuated).
os
3.3. Phase Response
System
Theoretic Practical • Notation: $\angle H(e^{j\omega})$.
analysis
al frequency
Main use (stability, • Purpose: Shows how each
spectral computati
difference frequency is shifted in time
analysis on
eqns) (delay/advance).

3. Frequency Response • Key Concepts:

Instructor: CK Caoile Course: Digital o Linear phase: Implies a
Signal Processing: Frequency Response constant delay for all
frequency components,
Intended Learning Outcomes: At the end avoiding distortion.
of the lesson, you should be able to:
o Nonlinear phase: Can
• Plot the frequency response of a cause signal distortion,
digital system. particularly in audio or
• Analyze system behavior based on images.
frequency plots. 3.4. Frequency Selectivity
3.1. What is Frequency Response? • Definition: Systems use frequency
• Definition: The frequency response selectivity to allow some
describes how a system modifies frequencies while blocking
the amplitude and phase of input others.
sinusoids at different frequencies. • Types of Filters:
• Formula: $H(e^{j\omega}) = o Low-pass (LPF): Allows
\sum_{n=-\infty}^{\infty} h[n]e^{- bass tones.
j\omega n}$. This is the DTFT of the
impulse response. o High-pass (HPF): Allows
treble.
 o Band-pass (BPF): Allows a • Significance: Shows how a system
specific band of frequencies. reacts to a sudden change,
revealing aspects like system
o Band-stop (Notch): Can
accumulation over time, settling
eliminate noise at one
time, overshoot, and stability.
specific frequency.
3.8. Stability Check
3.5. Bode Plots
• BIBO Stability (Bounded-Input
• Definition: Logarithmic plots that
Bounded-Output): A system is
show magnitude in decibels (dB)
stable if a bounded input always
and phase in degrees versus
produces a bounded output.
frequency.
Mathematically, the impulse
• Magnitude in dB Formula: $20 response must be absolutely
\log_{10} |H(e^{j\omega})|$. summable ($\sum
|\mathrm{h}[\mathrm{n}]| <
• Purpose: Useful for analyzing \infty$).
system behavior across wide
frequency ranges and observing • Z-Domain Criterion: A system is
how sharply filters "roll off". stable if all its poles are located
inside the unit circle in the Z-plane.
• Filter Characteristics:
• Types of Filters and Stability:
o LPF: Shows a flat passband
and a sharp drop at the cutoff o FIR (Finite Impulse
frequency. Response) filters: Always
stable because they have a
o HPF: Shows a rise after the finite response.
cutoff frequency.
o IIR (Infinite Impulse
o Clearly illustrates cutoff Response) filters: Must be
frequency and slope. carefully checked for stability
3.6. Impulse Response by analyzing the location of
their poles. For example, a
• Definition: The system's output pole at 0.95 is stable, while a
when the input is a unit impulse, pole at 1.05 is unstable.
$\delta[n]$. Pole-zero diagrams clearly
visualize this.
• Significance: It is the "system's
fingerprint" and completely 3.9. Practical Application with Python
characterizes an LTI system.
• Python libraries like sympy, numpy,
• Relation to Frequency Response: [Link], and [Link]
The frequency response is the DTFT can be used to compute and
of the impulse response. visualize transforms and frequency
responses.
3.7. Step Response
• Examples include symbolic Z-
• Definition: The system's output
Transform, numerical DTFT,
when the input is a unit step,
practical DFT, and plotting
$u[n]$.
magnitude/phase responses,
impulse responses, and step
responses.