Signal Classifications, Mathematical Operations & System Properties

1. 5 Key Ideas: Signal Architecture & System Axioms

Idea 1: Continuous vs Discrete Signals & The Elementary Singularity Toolkit

In electrical and computer engineering, a signal is a mathematical function of one or more independent variables that conveys information about the state or behavior of a physical system:

• Continuous-Time (CT) Signals $x(t)$: The independent variable $t$ is continuous over the real numbers ($t \in \mathbb{R}$). Examples: analog voltages, acoustic pressure, ambient temperature.
• Discrete-Time (DT) Signals $x[n]$: The independent variable $n$ is defined strictly at integer indices ($n \in \mathbb{Z}$). Represented as indexed sequences $x[n]$, typically obtained by sampling a continuous signal at uniform intervals $T_s$: $x[n] = x(n T_s)$.

Engineering signal analysis relies on a foundational set of elementary building blocks:

1. Unit Step Function (Heaviside Function): $u(t) = \begin{cases} 1, & t > 0 \\ 0, & t < 0 \end{cases} \quad \text{and} \quad u[n] = \begin{cases} 1, & n \ge 0 \\ 0, & n < 0 \end{cases}$ Models instantaneous switching on of a circuit or DC voltage excitation.

2. Dirac Delta Function (Unit Impulse): The continuous unit impulse $\delta(t)$ is a generalized function (distribution) defined by its integral properties rather than instantaneous point values: $\delta(t) = 0 \quad \text{for } t \ne 0, \quad \int_{-\infty}^{\infty} \delta(t)\,dt = 1$ It is the formal distributional derivative of the unit step function: $\delta(t) = \frac{du(t)}{dt} \iff u(t) = \int_{-\infty}^{t} \delta(\tau)\,d\tau$ In discrete time, the unit impulse $\delta[n]$ (Kronecker delta) is simply: $\delta[n] = \begin{cases} 1, & n = 0 \\ 0, & n \ne 0 \end{cases}$

3. Unit Ramp Function $r(t)$: $r(t) = \int_{-\infty}^{t} u(\tau)\,d\tau = t \cdot u(t) = \begin{cases} t, & t \ge 0 \\ 0, & t < 0 \end{cases}$

4. Standard Pulse Geometries:

   • Rectangular Pulse: $\text{rect}\left(\frac{t}{T}\right) = u\left(t + \frac{T}{2}\right) - u\left(t - \frac{T}{2}\right)$ (unit height, width $T$, centered at origin).
   • Triangular Pulse: $\Lambda\left(\frac{t}{T}\right) = \left(1 - \frac{|t|}{T}\right) \text{rect}\left(\frac{t}{2T}\right)$ (base $2T$, peak height 1 at $t=0$).
   • Normalized Sinc Pulse: $\text{sinc}(t) = \frac{\sin(\pi t)}{\pi t}$ (zero crossings at all non-zero integers $t = \pm 1, \pm 2, \dots$).

Idea 2: Sifting, Sampling & Distributional Calculus of the Dirac Delta

The Dirac delta impulse possesses powerful analytical properties essential for convolution, Fourier analysis, and sampling theory:

1. Sampling / Multiplication Property: Multiplying any continuous function $x(t)$ by an impulse concentrated at $t = t_0$ isolates the value $x(t_0)$: $x(t)\,\delta(t - t_0) = x(t_0)\,\delta(t - t_0)$

2. Sifting (Filtering) Property: Integrating the product of $x(t)$ and $\delta(t - t_0)$ extracts the scalar value of $x(t)$ at the impulse singularity: $\int_{-\infty}^{\infty} x(t)\,\delta(t - t_0)\,dt = x(t_0)$ Proof: $\int_{-\infty}^\infty x(t)\delta(t - t_0)dt = \int_{-\infty}^\infty x(t_0)\delta(t - t_0)dt = x(t_0) \int_{-\infty}^\infty \delta(t - t_0)dt = x(t_0) \cdot 1 = x(t_0)$.

3. Time-Scaling Property of Impulses: For any non-zero real constant $a \ne 0$: $\delta(at) = \frac{1}{|a|} \delta(t) \quad \text{and} \quad \delta(at - b) = \frac{1}{|a|} \delta\left(t - \frac{b}{a}\right)$ Notice the absolute value: compressing an impulse by factor $a$ magnifies its integral intensity by $1/|a|$, and reflection $\delta(-t) = \delta(t)$ confirms that the impulse is an even function.

4. Derivative of the Impulse (Doublet): The unit doublet $\delta'(t) = \frac{d\delta(t)}{dt}$ satisfies the derivative sifting property: $\int_{-\infty}^{\infty} x(t)\,\delta'(t - t_0)\,dt = -x'(t_0)$

Idea 3: Mathematical Operations on Independent Variables & Order Precedence

Modifying a signal's independent time axis is fundamental to radar ranging, audio pitch shifting, and communication modulation. Three primary elementary operations exist:

1. Time Shifting: $y(t) = x(t - t_0)$
   • If $t_0 > 0$: Time Delay — the waveform shifts to the RIGHT by $t_0$ seconds.
   • If $t_0 < 0$: Time Advance — the waveform shifts to the LEFT by $|t_0|$ seconds.
2. Time Scaling: $y(t) = x(at)$ ($a > 0$)
   • If $a > 1$: Time Compression — the waveform speeds up and shrinks horizontally by factor $a$.
   • If $0 < a < 1$: Time Expansion — the waveform slows down and stretches horizontally by factor $1/a$.
3. Time Reversal (Reflection): $y(t) = x(-t)$
   • Flips the waveform horizontally around the vertical axis $t = 0$.

The Combined Transformation Trap: $x(at - b)$

A common source of errors in exams is plotting a combined transformation $y(t) = x(at - b)$.

Method 1 (Factored Form — Shift First, Then Scale): Factor out the scaling coefficient $a$: $y(t) = x\left(a\left(t - \frac{b}{a}\right)\right)$

• Step 1: Shift $x(t)$ by $\frac{b}{a}$ (to the right if $b/a > 0$).
• Step 2: Scale by $a$ (and reflect if $a < 0$).

Method 2 (Scale First, Then Shift): Write directly as $x(at - b)$:

• Step 1: Scale $x(t)$ to obtain $x(at)$.
• Step 2: Replace $t$ with $t - \frac{b}{a}$! Notice: Shifting $x(at)$ by $b$ yields $x(a(t - b)) = x(at - ab)$, NOT $x(at - b)$! To obtain $x(at - b)$ after scaling, you must shift by $\frac{b}{a}$, not $b$.

Idea 4: Orthogonal Even & Odd Decomposition and Energy Conservation

Every real continuous or discrete signal $x(t)$ can be uniquely decomposed into the sum of an even component $x_e(t)$ and an odd component $x_o(t)$: $x(t) = x_e(t) + x_o(t)$

Where:

• Even Component $x_e(t)$: Symmetric about the vertical axis ($x_e(-t) = x_e(t)$): $x_e(t) = \frac{x(t) + x(-t)}{2}$
• Odd Component $x_o(t)$: Antisymmetric through the origin ($x_o(-t) = -x_o(t)$): $x_o(t) = \frac{x(t) - x(-t)}{2}$

Fundamental Properties of Even & Odd Signals

1. $x_o(0) = 0$ for any continuous odd signal (since $x_o(0) = -x_o(0) \implies 2x_o(0) = 0$).
2. $x_e(0) = x(0)$.
3. The integral of any odd function over symmetric limits is identically zero: $\int_{-A}^{A} x_o(t)\,dt = 0$
4. Product Rules:
   • $\text{Even} \times \text{Even} = \text{Even}$
   • $\text{Odd} \times \text{Odd} = \text{Even}$
   • $\text{Even} \times \text{Odd} = \text{Odd}$
5. Orthogonality & Energy Conservation: Even and odd functions are mutually orthogonal: $\int_{-\infty}^{\infty} x_e(t) x_o(t)\,dt = 0$ Consequently, the total energy of $x(t)$ is the direct sum of the energies of its even and odd parts: $E_x = E_{x_e} + E_{x_o}$

Idea 5: Energy vs Power Taxonomy & The 6 Foundational System Axioms

Energy vs Power Signal Classification

The instantaneous power dissipated across a normalized $1\text{-}\Omega$ resistor by voltage or current $x(t)$ is $p(t) = |x(t)|^2$.

• Total Signal Energy ($E$): $E = \int_{-\infty}^{\infty} |x(t)|^2\,dt \quad \text{(CT)} \quad \text{and} \quad E = \sum_{n=-\infty}^{\infty} |x[n]|^2 \quad \text{(DT)}$
• Average Signal Power ($P$): $P = \lim_{T \to \infty} \frac{1}{2T} \int_{-T}^{T} |x(t)|^2\,dt \quad \text{(CT)} \quad \text{and} \quad P = \lim_{N \to \infty} \frac{1}{2N + 1} \sum_{n=-N}^{N} |x[n]|^2 \quad \text{(DT)}$

The Three Mutually Exclusive Classes:

1. Energy Signal: $0 < E < \infty \implies P = 0$. (Finite-duration pulses, decaying exponentials $e^{-at}u(t)$).
2. Power Signal: $0 < P < \infty \implies E = \infty$. (Periodic waveforms, sine waves, square waves, unit step $u(t)$).
3. Neither Energy Nor Power Signal: $E = \infty$ and $P = \infty$. (Ramp functions $r(t) = t$, growing exponentials $e^{at}u(t)$).

The 6 Core System Axioms

A system is a mathematical operator $\mathcal{T}$ that transforms an input signal $x(t)$ into an output signal $y(t) = \mathcal{T}\{x(t)\}$:

1. Linearity: The system satisfies the Principle of Superposition (both Additivity and Homogeneity): $\mathcal{T}\{a x_1(t) + b x_2(t)\} = a \mathcal{T}\{x_1(t)\} + b \mathcal{T}\{x_2(t)\} = a y_1(t) + b y_2(t)$ Necessary check: Zero input must produce zero output ($x=0 \implies y=0$). Any non-zero DC bias (e.g., $y = 2x + 3$) makes a system non-linear!
2. Time-Invariance: A time delay in the input produces an identical time delay in the output: $\text{If } \mathcal{T}\{x(t)\} = y(t), \quad \text{then } \mathcal{T}\{x(t - t_0)\} = y(t - t_0)$ A system is time-variant if time $t$ appears explicitly outside the input signal argument (e.g., $y(t) = t \cdot x(t)$ or $y(t) = x(2t)$).
3. Causality: The output $y(t)$ at any instant $t = t_0$ depends only on current and past input values ($x(t)$ for $t \le t_0$), never on future inputs ($t > t_0$). (All real-time physical systems are causal).
4. BIBO Stability: Every Bounded Input produces a Bounded Output: $|x(t)| \le M_x < \infty \implies |y(t)| \le M_y < \infty \quad \text{for all } t$
5. Memory (Static vs Dynamic):
   • Memoryless (Static): Output $y(t)$ at any instant $t$ depends strictly and exclusively on the input $x(t)$ at that exact same instant $t$ (e.g., $y(t) = R \cdot i(t)$).
   • Dynamic (With Memory): Output depends on past or future inputs (e.g., capacitors $v(t) = \frac{1}{C}\int_{-\infty}^t i(\tau)d\tau$, delays $y(t) = x(t - 1)$).
6. Invertibility: Distinct inputs produce distinct outputs. An inverse system $\mathcal{T}^{-1}$ exists such that $\mathcal{T}^{-1}\{\mathcal{T}\{x(t)\}\} = x(t)$. (Systems like $y(t) = x^2(t)$ or $y(t) = 0$ are non-invertible).

2. Key Formula Box

3. In-Depth Mathematical Derivations

Derivation 1: Rigorous Proof of the Sifting Property via Limiting Gaussian Distribution

Why does the Dirac delta extract $x(t_0)$ under integration?

Derivation 2: Orthogonal Energy Decomposition $E_x = E_{x_e} + E_{x_o}$

Why is the cross product of even and odd functions identically zero?

Derivation 3: Average Power of a Sinusoid $x(t) = A\cos(\omega_0 t + \theta)$

Why is the average power of a sinusoid equal to $A^2/2$, independent of frequency and phase?

4. Typical Laboratory Hardware Values & Engineering Benchmarks

5. Worked Examples: Step-by-Step Analytical Solutions

6. Common Exam Traps & Where Marks Are Lost

7. Interactive Circuit Verification & Laboratory Workbench

Experiment with signal transformations (delay, scale, reversal), observe orthogonal even and odd waveform slices, classify energy vs power signals, and evaluate the 6 system axioms in the live simulator below:

8. Parametric Practice Problem

9. Unit Mastery Quiz

10. References & Further Reading

1. Oppenheim, Alan V., Alan S. Willsky, and S. Hamid Nawab: Signals and Systems, 2nd Edition, Prentice Hall (Chapter 1: Signals and Systems).
2. Proakis, John G., and Dimitris G. Manolakis: Digital Signal Processing: Principles, Algorithms, and Applications, 4th Edition, Pearson.
3. Haykin, Simon, and Barry Van Veen: Signals and Systems, 2nd Edition, John Wiley & Sons.
4. Lathi, Bhagwandas P.: Linear Systems and Signals, 2nd Edition, Oxford University Press.
5. Presidency University Courseware: EEE2006 - Signals and Systems, Department of Electrical & Electronics Engineering, Bengaluru.