1.06 Typical Signal Processing Operations

In engineering practice, raw physical signals are rarely usable in their original form. To extract information, transmit data over long distances, or suppress noise, we must perform Typical Signal Processing Operations {standard mathematical manipulations applied to waveforms within a communication system}.

These standard operations form the physical building blocks of transmitters, channels, and receivers.

---

1.7.1 Correlation

Correlation {a mathematical measure of similarity between two waveforms as a function of a relative time shift} is used to compare a reference signal with one or more target signals to determine how closely they match.

Classifications of Correlation

1. Cross-correlation ($R_{xy}$ or $r_{xy}$): Measures the similarity between two different signals, $x(t)$ and $y(t)$, as one is shifted in time relative to the other.
   • Continuous-Time Formula: $R_{xy}(\tau )={\int }_{-\infty }^{\infty }x(t)y(t+\tau )dt$
   • Discrete-Time Formula: $r_{xy}[l]={\sum }_{n=-\infty }^{\infty }x[n]y[n-l]$
2. Auto-correlation ($R_{xx}$ or $r_{xx}$): Measures the similarity of a signal with a time-shifted version of itself. It is highly effective at extracting periodic signals heavily corrupted by noise.
   • Continuous-Time Formula: $R_{xx}(\tau )={\int }_{-\infty }^{\infty }x(t)x(t+\tau )dt$
   • Discrete-Time Formula: $r_{xx}[l]={\sum }_{n=-\infty }^{\infty }x[n]x[n-l]$

High-Yield Applications of Correlation

• Radar & Sonar Echo Detection: Measuring the time-of-flight delay ($\tau$) of a reflected echo to calculate the exact distance to an object.
• Signal-in-Noise Extraction: Pulling extremely weak communication signals out of channel noise.
• Pattern Matching: Biometric fingerprint matching and voice/image recognition algorithms.
• Cross-spectral analysis: Mapping similarities across frequency bands.

---

1.7.2 Filtering

Filtering {a frequency-selective operation that passes desired spectral components of a signal while suppressing unwanted ones} is used to clean up signals and block out-of-band interference.

The 5 Core Filter Types:

1. Low-Pass Filter (LPF): Passes all frequencies below a specified cutoff frequency (${\Omega }_c$) and attenuates those above it. [GRAPH: Ideal LPF frequency response |H(j\Omega)| showing a flat passband of height 1 from -\Omega_c to +\Omega_c, and a stopband of 0 for |\Omega| > \Omega_c]
2. High-Pass Filter (HPF): Passes all frequencies above a specified cutoff frequency (${\Omega }_c$) and blocks those below it. [GRAPH: Ideal HPF frequency response |H(j\Omega)| showing a stopband of 0 for |\Omega| < \Omega_c, and a flat passband of height 1 for |\Omega| > \Omega_c]
3. Band-Pass Filter (BPF): Passes frequencies within a specific frequency band (${\Omega }_{c1}$ to ${\Omega }_{c2}$) and attenuates all frequencies outside this band. [GRAPH: Ideal BPF frequency response |H(j\Omega)| showing a flat passband of height 1 in the intervals [\Omega_{c1}, \Omega_{c2}] and [-\Omega_{c2}, -\Omega_{c1}], and 0 elsewhere]
4. Band-Stop Filter (BSF): Blocks frequencies within a specific frequency band (${\Omega }_{c1}$ to ${\Omega }_{c2}$) and passes all frequencies outside it. [GRAPH: Ideal BSF frequency response |H(j\Omega)| showing a stopband notch of 0 in the intervals [\Omega_{c1}, \Omega_{c2}] and [-\Omega_{c2}, -\Omega_{c1}], and a passband of height 1 elsewhere]
5. Notch Filter: A specialized, extremely narrow band-stop filter designed to reject a single, highly specific frequency component.
   • Example: Used in biomedical systems (ECG/EEG machines) to suppress the 50 Hz/60 Hz power-line hum without distorting sensitive physiological signals.

Filtering in LTI Systems

In the time domain, filtering is modeled by convolving the input signal $x(t)$ with the filter’s impulse response $h(t)$: $y(t)=x(t)\ast h(t)$ In the frequency domain, this complex convolution simplifies to simple algebraic multiplication with the transfer function $H(j\omega )$: $Y(j\omega )=X(j\omega )H(j\omega )$

---

1.7.3 Modulation and Demodulation

• Modulation {the process of varying one or more parameters of a high-frequency carrier wave in proportion to a low-frequency message signal} shifts baseband frequencies to higher bands [1.7.3].
• Demodulation {the reverse process of extracting the original baseband message signal from the received modulated wave} restores the signal at the receiver.

The Physical Importance of Modulation

Baseband signals (like human speech, 20 Hz to 20 kHz) have extremely long wavelengths ($\lambda =c/f$). Because an antenna’s physical size must be proportional to the wavelength ($\approx \lambda /4$), transmitting raw baseband signals would require antennas kilometers tall!

Modulation translates the message to a high-frequency carrier (e.g., 100 MHz), reducing the wavelength to meters and allowing compact, practical antennas. It also enables Frequency-Division Multiplexing to prevent multiple transmitters from interfering.

Mathematical Model of Amplitude Modulation (AM):

$s(t)=m(t)\cos ({\omega }_{c}t)$ where $m(t)$ is the low-frequency modulating message signal, $\cos ({\omega }_{c}t)$ is the high-frequency carrier wave, and $s(t)$ is the resulting modulated signal.

[DIAGRAM: Modulation block diagram showing input message m(t) and carrier source \cos(\omega_c t) multiplying in a mixer block to produce s(t) = m(t)\cos(\omega_c t)]

---

1.7.4 Transformation

Transformation {a mathematical mapping that shifts a signal from its original time-domain representation to an alternative frequency-domain coordinate system} simplifies system analysis by turning complex differential or difference equations into easy-to-solve algebraic equations.

graph LR
    Time[Time Domain] -- Fourier Transform --> Freq[Frequency Domain]
    Time -- Laplace Transform --> sDomain[s-Domain s = σ + jω]
    Time -- Z-Transform --> zDomain[z-Plane]

• Fourier Transform: Analyzes the steady-state frequency spectra of continuous-time signals. $F(j\omega )={\int }_{-\infty }^{\infty }f(t)e^{-j\omega t}dt$
• Laplace Transform: Generalizes the Fourier transform by introducing a damping envelope $e^{-\sigma t}$. Essential for analyzing transient responses in analog circuits and unstable LTI systems. $F(s)={\int }_{0^{-}}^{\infty }f(t)e^{-st}dt$
• Z-Transform: The discrete-time equivalent of the Laplace transform [10.01]. Maps discrete sequences to the complex z-plane. $X(z)={\sum }_{n=-\infty }^{\infty }x[n]z^{-n}$

---

1.7.5 Multiplexing and Demultiplexing

• Multiplexing {the simultaneous transmission of multiple independent message signals over a single shared physical channel} maximizes medium efficiency [1.7.5].
• Demultiplexing {separating the combined multiplexed signals back into their independent original forms at the receiver} routes signals to their proper destinations.

The Three Major Multiplexing Paradigms

1. Time-Division Multiplexing (TDM): Allocates different, non-overlapping time slots to each user over the entire channel bandwidth. [DIAGRAM: TDM frame structure showing sequential time slots t_1, t_2, t_3 occupied by User 1, User 2, User 3 repeating in time]
2. Frequency-Division Multiplexing (FDM): Allocates different, non-overlapping frequency bands (separated by guard bands to prevent adjacent-channel interference) to each user simultaneously. [DIAGRAM: FDM spectrum showing Channel 1, Channel 2, Channel 3 separated by unshaded guard bands in the frequency domain]
3. Code-Division Multiplexing (CDM): Assigns a unique mathematical code to each user, allowing multiple users to transmit simultaneously over the entire common frequency band without interference.

---

1. Common Mistakes That Cost Marks

• Convolving vs. Correlating: Students often confuse these two operations. Convolution requires time-reversing (folding) the sliding signal, whereas correlation does not fold the signal. $x(t)\ast h(t)={\int }_{-\infty }^{\infty }x(\tau )h(t-\tau )d\tau \text{vs.}{R}_{xy}(\tau )={\int }_{-\infty }^{\infty }x(t)y(t+\tau )dt$
• Notch vs. Band-Stop Filter: A Notch filter is not a standard band-stop filter. A Notch filter has an extremely narrow, sharp rejection band designed to suppress one specific frequency (such as power line noise) while leaving the rest of the spectrum untouched.
• Multiplexing vs. Modulation: Modulation is frequency translation for a single message, whereas Multiplexing is combining multiple distinct user messages to share a single channel.

---

2. Verbatim PYQ Bank

Potential Exam Questions (Syllabus/Lecture Alignment)

Q1. Define: (i) Correlation, (ii) Filtering, (iii) Modulation. State the physical importance of modulation.

• Grounded Answer Guidelines:
  • Define Correlation as a mathematical tool to compare signals and determine similarity.
  • Define Filtering as a frequency-selective operation designed to suppress unwanted spectral components.
  • Define Modulation as translating baseband message frequencies using a high-frequency carrier wave [1.7.3].
  • State that modulation is physically crucial because baseband signal wavelengths are too long, requiring impractically large antennas. Modulation reduces the wavelength, allowing compact antennas.

Q2. Distinguish between Time-Division Multiplexing (TDM) and Frequency-Division Multiplexing (FDM).

• Grounded Answer Guidelines:
  • TDM splits the shared medium by allocating non-overlapping time slots to each user over the entire channel bandwidth [1.7.5].
  • FDM splits the shared medium by allocating non-overlapping frequency bands separated by guard bands, allowing users to transmit simultaneously [1.7.5].

---

3. Self-Check Before Moving On

• Can you write down the exact continuous-time integrals for both cross-correlation and convolution, highlighting their algebraic differences?
• Do you know why a Notch filter is used instead of a standard Low-Pass Filter to remove power hum from an ECG signal? (To prevent removing high-frequency bio-potential data!)
• Can you explain the relationship between wavelength and antenna size, proving why modulation is necessary for wireless communication?

---

4. Source Citations

• Textbook: K. Deergha Rao, Signals and Systems, Section 1.7 (Typical Signal Processing Operations), pp. 30-32.
• Class Handnotes: Rabiul Sir Class Lectures, Week 3 (Lec 3, pg. 30) [Lec 3].
• Laboratory Guide: ECE 2108 Signals & Systems Lab Experiment 3 (Correlation & Convolution), pg. 30.