Signal Operations and Elementary Signals

Introduction

Signals are mathematical representations of physical phenomena, measurements, messages, or control variables. In electronics and communication engineering, a signal may represent voltage, current, sound pressure, temperature, position, or a transmitted information waveform. Signals and systems analysis becomes manageable when complicated waveforms are expressed in terms of elementary signals and manipulated using standard operations.

This chapter develops the principal operations performed on continuous-time and discrete-time signals: amplitude scaling, time shifting, time reversal, compression, expansion, differentiation, integration, and discrete differences. It also introduces the unit impulse, unit step, ramp, exponential, sinusoidal, and complex exponential signals. These concepts form the foundation for convolution, Fourier analysis, Laplace transforms, system modeling, and digital signal processing.

Learning Objectives

After studying this chapter, the reader should be able to:

• distinguish continuous-time and discrete-time signals;
• describe and sketch elementary signals;
• apply amplitude scaling, time shifting, reversal, compression, and expansion;
• determine the effect of differentiation and integration on common signals;
• use impulses and steps to represent switching and discontinuities;
• calculate first and second differences of discrete-time signals;
• explain the relationship between real sinusoids and complex exponentials;
• identify practical uses and limitations of signal operations.

1. Signal Descriptions and Notation

A continuous-time signal is defined for every real value of time and is written as x(t), where t is measured in seconds. A discrete-time signal is defined only at integer-indexed instants and is written as x[n], where n is an integer. A sampled signal may be written as x[n] = x(nT_s), where T_s is the sampling interval in seconds.

The independent variable is not necessarily time. For example, x(r) may represent a spatial waveform, and x[k] may represent a sequence indexed by sample number. In this chapter, t denotes continuous time and n denotes a discrete index unless stated otherwise.

A signal may be real-valued or complex-valued. Physical voltages and currents are usually real-valued, whereas complex signals are extremely useful mathematical representations in modulation, filtering, and frequency analysis.

2. Elementary Signals

2.1 Unit impulse

The continuous-time unit impulse, or Dirac delta, δ(t), is defined by its sampling property rather than as an ordinary finite-valued function:

Here x(t) is assumed to be continuous at t = t₀. The impulse is zero away from t = 0 and has unit area:

Its physical dimension is inverse time, so that δ(t)dt is dimensionless. A shifted and scaled impulse Aδ(t − t₀) has area A and is located at t = t₀.

The discrete-time unit impulse is denoted δ[n]:

Unlike the continuous-time impulse, the discrete impulse is an ordinary sequence. It has the selection property

Impulses are important because many signals and systems can be described in terms of their response to an impulse.

2.2 Unit step

The continuous-time unit step is

The value at t = 0 depends on convention. In many theoretical contexts u(0) = 1/2 is used, but the value at one isolated point does not affect ordinary integration. The shifted step u(t − t₀) changes from zero to one at t = t₀.

The discrete-time unit step is

A rectangular pulse can be formed from two steps:

The pulse has amplitude A and duration t₂ − t₁. In discrete time, the corresponding sequence is A[u[n − n₁] − u[n − n₂]].

2.3 Ramp

The unit ramp is defined by

It is zero for t < 0 and increases linearly with slope one for t > 0. A delayed ramp is (t − t₀)u(t − t₀). The ramp and step are related through differentiation and integration:

and, with zero initial condition,

The variable τ is a dummy integration variable. A ramp of slope A is A t u(t), where A has units appropriate to the signal divided by time.

2.4 Exponential signals

A real exponential is written

where A is the value at t = 0 and α has units s⁻¹. If α < 0, the signal decays; if α > 0, it grows. For a decaying exponential x(t) = A e^{−t/τ}, τ is the time constant in seconds. At t = τ, the signal has fallen to e⁻¹ ≈ 0.368 of its initial value.

The discrete-time exponential is

where a is dimensionless. It decays for |a| < 1, remains constant in magnitude for |a| = 1, and grows for |a| > 1 as n increases. If a is negative, the sign alternates.

2.5 Sinusoids

A continuous-time sinusoid may be expressed as

where A is amplitude, ω₀ is angular frequency in radians per second, and φ is phase in radians. The frequency in hertz is f₀ = ω₀/(2π), and the period is T₀ = 2π/ω₀, provided ω₀ ≠ 0.

For example, 5 cos(200πt + π/4) has amplitude 5, angular frequency 200π rad/s, frequency 100 Hz, period 0.01 s, and phase π/4.

A discrete-time sinusoid is

where Ω₀ is normalized angular frequency in radians per sample. Discrete-time frequency is periodic: sinusoids with frequencies Ω₀ and Ω₀ + 2πk, where k is any integer, produce the same sequence. A discrete sinusoid is periodic only when Ω₀/(2π) is rational.

2.6 Complex exponentials

Euler’s identity connects exponentials and sinusoids:

where j = √−1. Consequently,

The complex exponential e^{(α + jω₀)t} has magnitude e^{αt} and phase ω₀t. It therefore represents a sinusoid whose amplitude changes exponentially when α ≠ 0. Complex notation simplifies differentiation, integration, and the analysis of linear time-invariant systems.

3. Basic Signal Operations

3.1 Amplitude scaling

Multiplication by a constant changes the amplitude but not the time locations of signal features:

If |A| > 1, magnitudes increase; if 0 < |A| < 1, magnitudes decrease; if A < 0, the signal also undergoes a polarity reversal. The same principle applies to y[n] = A x[n]. Units must be considered: if x is measured in volts and A is dimensionless, y is also measured in volts.

3.2 Time shifting

A delay of t₀ seconds is produced by

A positive t₀ shifts the waveform to the right. An advance is x(t + t₀). In discrete time, x[n − n₀] delays a sequence by n₀ samples when n₀ is a positive integer.

For a sinusoid,

Thus a time delay appears as a phase delay of ω₀t₀ radians. This relationship is valid for a sinusoid at a specified frequency and is central to phase analysis.

3.3 Time reversal

Time reversal, or folding, is obtained by replacing t with −t:

A feature originally at t = t₀ appears at t = −t₀. In discrete time, y[n] = x[−n]. Reversal is especially important in convolution, where one signal is reversed before being shifted.

For a sinusoid, cos(−ω₀t + φ) = cos(ω₀t − φ), so reversal changes the sign of phase but not the amplitude or frequency. For an even signal x(t) = x(−t), reversal has no effect; for an odd signal x(t) = −x(−t), reversal changes its sign.

3.4 Time compression and expansion

The operation

changes the time scale. If |a| > 1, the signal is compressed by a factor |a|; if 0 < |a| < 1, it is expanded by a factor 1/|a|. If a < 0, reversal occurs in addition to scaling.

For x(t) = cos(ω₀t), y(t) = cos(ω₀at), so the new angular frequency is |a|ω₀. Time compression increases frequency, whereas expansion decreases frequency. In discrete time, x[an] is directly defined without interpolation only when an is an integer for every relevant n. Therefore discrete-time expansion and compression require care and may involve insertion or deletion of samples.

3.5 Combined operations

Operations should be interpreted from the argument of the signal. For example,

means that the original signal is compressed by a factor of 2 and shifted to the right by 2 seconds. The expression x(2t) − 4 would mean something entirely different: it subtracts 4 from the amplitude rather than shifting time.

A reliable procedure is to factor the argument as a(t − t₀), identify the scaling a, and then identify the shift t₀. Amplitude factors outside the argument are handled separately.

4. Differentiation and Integration

For a differentiable signal, differentiation emphasizes rapid changes:

If x(t) is measured in volts, y(t) is measured in volts per second. Differentiation of common signals gives

and

The derivative of a sinusoid has the same frequency, amplitude multiplied by |ω₀|, and a phase shift of π/2, with sign determined by the chosen sine or cosine representation.

The derivative of a step includes an impulse:

Therefore a sudden change in a waveform produces an impulse in its derivative. More generally, if a signal jumps by magnitude C at t = t₀, its derivative contains Cδ(t − t₀), in addition to the derivative of its continuous portions.

Integration accumulates signal area:

Here y(t₀) is the initial value. The units of y are the units of x multiplied by seconds. Integration smooths abrupt variations and converts an impulse into a step. With zero initial condition,

Differentiation and integration are inverse operations only when initial conditions and discontinuities are treated correctly.

5. Discrete Differences and Accumulation

Differentiation is not directly defined for a discrete sequence, but finite differences provide an analogous operation. The first backward difference is

It measures the change between adjacent samples. A forward difference is x[n + 1] − x[n]. The backward difference is often convenient because it uses the current and previous samples.

If the sampling interval is T_s, an approximation to the continuous derivative is

This approximation becomes more accurate as T_s decreases, provided x(t) is sufficiently smooth and noise is not excessive.

The discrete-time accumulator is

or, with an initial index n₀,

A running sum is the discrete counterpart of integration. Unlike an ideal continuous integrator, a digital accumulator has finite word length and may overflow or accumulate quantization error.

6. Worked Examples

Example 1: Interpreting a transformed signal

Let x(t) be a triangular pulse and define y(t) = −2x(3t + 6).

Rewrite the argument:

Thus the waveform is compressed by a factor of 3, advanced by 2 seconds, and multiplied by −2. The negative amplitude factor reflects the waveform vertically and doubles its magnitude. The order of these interpretations is conceptually useful, although the amplitude scaling commutes with time operations.

Example 2: Difference of a sequence

Suppose x[0] = 2, x[1] = 5, x[2] = 9, and x[3] = 8. Taking x[−1] = 0, the backward differences are

The negative final difference indicates a decrease from sample 2 to sample 3. If T_s = 0.01 s, the approximate derivative at n = 3 is −1/0.01 = −100 units per second.

Example 3: Differentiating a switched signal

Let x(t) = 4u(t) − 4u(t − 2). This is a rectangular pulse of amplitude 4 from t = 0 to t = 2, under the usual endpoint convention. Its derivative is

The positive impulse marks the rising edge and the negative impulse marks the falling edge. The total signed area of the derivative is zero, consistent with the signal returning to its initial value.

7. Applied Case: Edge Detection in a Digital Sensor

Background

Consider an illustrative industrial monitoring system that samples a conveyor-belt proximity sensor every 1 ms. The sensor output is approximately 0 V when no object is present and 5 V when an object is detected. Noise and switching imperfections produce small fluctuations.

Problem

The controller must identify the instant at which an object enters the sensing region, while avoiding false detections caused by small noise variations.

Analysis

A raw threshold detector declares an object present when x[n] exceeds 2.5 V. However, noisy samples may cross the threshold repeatedly. A first difference,

emphasizes the rapid transition. An additional threshold can require Δx[n] to exceed, for example, 1 V per sample. This combines level information with edge information.

Approach and solution

The controller first applies a short moving average to reduce high-frequency noise. It then computes the backward difference and identifies a rising edge when both conditions hold:

and

A falling-edge condition uses Δx[n] < −1 V. A lockout interval or state machine prevents multiple detections during one physical transition.

Outcome and learning points

The system uses amplitude scaling and thresholding to interpret sensor levels, delay and sampling concepts to locate events, and discrete differences to detect transitions. The example also illustrates a limitation: differentiation or differencing amplifies high-frequency noise. Practical designs therefore combine differences with filtering, hysteresis, adequate sampling, and carefully selected thresholds.

8. Limitations, Implementation, and Best Practices

Ideal impulses, discontinuities, and infinite bandwidth are mathematical abstractions. Real circuits have finite rise times, limited bandwidth, noise, and saturation. A measured impulse is usually a short pulse whose area approximates the ideal impulse area.

Time scaling in analog systems changes frequency and may require a system capable of operating over the new bandwidth. In digital systems, resampling requires interpolation for expansion and anti-alias filtering before decimation. Sampling frequency should be selected in relation to the highest significant signal frequency; otherwise aliasing can make time and frequency interpretations unreliable.

Differentiators are sensitive to noise, while integrators are sensitive to offsets and low-frequency errors. Hardware integrators may saturate, and digital accumulators may overflow. Engineers should specify initial conditions, units, sampling interval, numerical precision, and endpoint conventions.

A clear signal-processing workflow is:

1. state whether the signal is continuous or discrete;
2. identify units and independent-variable conventions;
3. separate amplitude operations from time-argument operations;
4. factor transformed arguments such as a(t − t₀);
5. account for discontinuities using impulses where appropriate;
6. verify limiting behavior, dimensions, and sample indexing;
7. test the result with a sketch, simulation, or measured data.

Artificial intelligence does not replace these fundamentals. In modern sensing systems, machine-learning models may classify waveforms, but preprocessing still commonly uses scaling, filtering, differences, steps, and spectral representations. Incorrect sampling or indexing can produce systematic errors that no classifier can reliably correct.

Chapter Summary

Signals may be continuous-time or discrete-time, real or complex, and may represent physical measurements or abstract mathematical quantities. The impulse selects signal values and represents idealized sudden events. The step models switching, the ramp models accumulated change, exponentials describe growth and decay, and sinusoids describe periodic behavior. Complex exponentials provide a compact representation of sinusoidal amplitude and phase.

Amplitude scaling changes magnitude and possibly polarity. Time shifting delays or advances a waveform, reversal folds it about the origin, and time scaling compresses or expands its temporal features. Differentiation emphasizes rapid changes and produces impulses at jumps; integration accumulates area and requires an initial condition. Discrete differences approximate differentiation, while accumulators approximate integration.

Correct interpretation depends on notation, units, sampling assumptions, endpoint conventions, and the order of operations. These elementary tools are directly applied in communication, control, instrumentation, sensor systems, and digital signal processing.

Key Terms

Signal: A mathematical representation of a varying physical or informational quantity.

Continuous-time signal: A signal defined for every real value of its independent variable.

Discrete-time signal: A sequence defined at discrete integer indices.

Unit impulse: An ideal signal with unit area and the sampling property.

Unit step: A signal that changes from zero to one at a specified origin.

Ramp: A signal that increases linearly after its origin.

Amplitude scaling: Multiplication of a signal by a constant.

Time shift: Delay or advance of a waveform along its independent-variable axis.

Time reversal: Replacement of t by −t or n by −n.

Time scaling: Compression or expansion of a signal in time.

Exponential signal: A signal whose magnitude varies exponentially with time or index.

Sinusoid: A periodic sine or cosine waveform.

Complex exponential: An exponential with a complex exponent, representing magnitude and phase behavior.

Angular frequency: Frequency expressed in radians per second or radians per sample.

Phase: The angular displacement of a sinusoid relative to a reference.

Differentiation: An operation measuring instantaneous rate of change.

Integration: An operation accumulating signal area.

Finite difference: A discrete approximation to differentiation.

Accumulator: A discrete-time operation that forms a running sum.

Aliasing: Misinterpretation caused by insufficient sampling of a signal.

Review Questions

Short Answer Questions

1. Distinguish between a continuous-time signal and a discrete-time signal.
2. State the sampling property of the continuous-time unit impulse.
3. What is the effect of replacing x(t) by x(t − t₀) when t₀ > 0?
4. Define the first backward difference of a sequence.
5. Why does differentiation of a step produce an impulse?

Long Answer Questions

1. Explain amplitude scaling, time shifting, reversal, and time scaling, including the effect of each operation on a sinusoid.
2. Describe the unit impulse, unit step, and ramp and explain their derivative and integral relationships.
3. Derive the relationship between a continuous-time sinusoid and its complex-exponential representation.
4. Compare continuous-time integration with discrete-time accumulation, including initial conditions and practical limitations.
5. Explain why differentiation and finite differences can be useful for edge detection but problematic in noisy systems.

Analytical / Application Questions

1. For y(t) = 3x(2t − 8), identify the amplitude scaling, time scaling, and time shift relative to x(t).
2. A sampled sensor produces x[0] = 1, x[1] = 1.2, x[2] = 3.8, and x[3] = 4.0 at T_s = 2 ms. Calculate the backward differences and estimate the derivative at n = 2.
3. A communication waveform is delayed by 4 μs and has frequency 250 kHz. Determine the phase delay in radians for a sinusoidal component and explain why the result depends on frequency.

Multiple Choice Questions with Answers

Question 1

Which operation produces a delay of t₀ seconds for a continuous-time signal when t₀ > 0?

A. x(t + t₀)

B. x(t − t₀)

C. x(−t − t₀)

D. t₀x(t)

Correct Answer: B. x(t − t₀)

Question 2

The continuous-time unit impulse satisfies which property, assuming x(t) is continuous at t₀?

A. ∫x(t)δ(t−t₀)dt = 0

B. ∫x(t)δ(t−t₀)dt = x(t₀)

C. ∫x(t)δ(t−t₀)dt = t₀

D. ∫x(t)δ(t−t₀)dt = 1/x(t₀)

Correct Answer: B. ∫x(t)δ(t−t₀)dt = x(t₀)

Question 3

For y(t) = x(3t), the signal is

A. expanded by a factor of 3

B. delayed by 3 seconds

C. compressed by a factor of 3

D. amplitude-scaled by 3

Correct Answer: C. compressed by a factor of 3

Question 4

If x(t) = 2cos(100πt), its frequency is

A. 50 Hz

B. 100 Hz

C. 200π Hz

D. 100π Hz

Correct Answer: A. 50 Hz

Question 5

The first backward difference of x[n] is

A. x[n] + x[n−1]

B. x[n] − x[n−1]

C. x[n+1] − x[n]

D. x[n]x[n−1]

Correct Answer: B. x[n] − x[n−1]

Question 6

The derivative of 4u(t) − 4u(t−2) is

A. 4δ(t) + 4δ(t−2)

B. 4u(t) − 4u(t−2)

C. 4δ(t) − 4δ(t−2)

D. 4tδ(t−2)

Correct Answer: C. 4δ(t) − 4δ(t−2)

Question 7

If x(t) is measured in volts, the units of dx(t)/dt are

A. volts

B. seconds per volt

C. volts per second

D. dimensionless

Correct Answer: C. volts per second

Question 8

For x(t) = Ae^{−t/τ}, the value at t = τ is

A. A

B. 0

C. eA

D. A/e

Correct Answer: D. A/e

Question 9

Using Euler’s identity, e^{jθ} equals

A. sinθ − jcosθ

B. cosθ + jsinθ

C. cosθ + jθ

D. e^θ(cosθ + sinθ)

Correct Answer: B. cosθ + jsinθ

Question 10

If x[0] = 2, x[1] = 5, and x[2] = 9, with x[−1] = 0, what is Δx[2]?

A. 2

B. 3

C. 4

D. 9

Correct Answer: C. 4