Ferkans — Interactive Telecom Tutor

Historical Note: Joseph Fourier and the Heat Equation

In 1807, Joseph Fourier shocked the Paris Academy by claiming that any periodic function could be decomposed into a sum of sines and cosines. Lagrange and Laplace were sceptical. It took over a century (Dirichlet, Riemann, Lebesgue, Carleson) to make the claim rigorous, but the engineering utility was immediate — and has only grown since.

Definition:
Fourier Series

A periodic signal $x(t)$ with fundamental period $T_0$ and fundamental frequency $f_0 = 1/T_0$ can be expressed as

$x(t) = \sum_{k=-\infty}^{\infty} c_k\, e^{j 2\pi k f_0 t}$

where the Fourier coefficients are

$c_k = \frac{1}{T_0} \int_{T_0} x(t)\, e^{-j 2\pi k f_0 t}\,dt.$

Properties:

$c_0$ is the DC (average) value of $x(t)$ .
For real $x(t)$ : $c_{-k} = c_k^*$ (conjugate symmetry).
Parseval's theorem for periodic signals: $P_x = \frac{1}{T_0}\int_{T_0}|x(t)|^2\,dt = \sum_{k=-\infty}^{\infty}|c_k|^2$ .

Definition:
Continuous-Time Fourier Transform (CTFT)

For an energy signal $x(t)$ , the Fourier transform is

$X(f) = \mathcal{F}\{x(t)\} = \int_{-\infty}^{\infty} x(t)\,e^{-j2\pi f t}\,dt$

and the inverse Fourier transform is

$x(t) = \mathcal{F}^{-1}\{X(f)\} = \int_{-\infty}^{\infty} X(f)\,e^{j2\pi f t}\,df.$

$X(f)$ is in general complex: $X(f) = |X(f)|\,e^{j\angle X(f)}$ where $|X(f)|$ is the magnitude spectrum and $\angle X(f)$ is the phase spectrum.

We use the ordinary frequency variable $f$ (Hz) rather than the angular frequency $\omega = 2\pi f$ (rad/s). This avoids factors of $2\pi$ in the transform pair and is standard in communications.

,

Existence Conditions

The Fourier transform exists (in the classical sense) if $x \in L^1(\mathbb{R})$ , i.e., $\int_{-\infty}^{\infty}|x(t)|\,dt < \infty$ . For power signals (e.g., sinusoids, $\delta(t)$ ) we extend the definition using distributions/generalised functions, giving pairs like $e^{j2\pi f_0 t} \leftrightarrow \delta(f - f_0)$ .

Example: Rectangular Pulse and Sinc Function

Find the Fourier transform of the rectangular pulse $x(t) = A\,\mathrm{rect}(t/T)$ , where $\mathrm{rect}(u) = 1$ for $|u| \le 1/2$ and $0$ otherwise.

Solution

Direct computation

$X(f) = \int_{-T/2}^{T/2} A\,e^{-j2\pi f t}\,dt = A \cdot \frac{e^{-j2\pi f \cdot (-T/2)} - e^{-j2\pi f \cdot T/2}}{j2\pi f} = AT\,\frac{\sin(\pi f T)}{\pi f T}.$ $

Sinc notation

Defining $\mathrm{sinc}(u) = \sin(\pi u)/(\pi u)$ , we have

$\mathrm{rect}(t/T) \leftrightarrow T\,\mathrm{sinc}(fT).$

The sinc function has zeros at $f = k/T$ , $k \ne 0$ . Wider pulses produce narrower spectra — an instance of the time–frequency uncertainty principle. $\blacksquare$

Theorem: Key Properties of the Fourier Transform

Let $x(t) \leftrightarrow X(f)$ and $y(t) \leftrightarrow Y(f)$ . Then:

Property	Time domain	Frequency domain
Linearity	$ax(t) + by(t)$	$aX(f) + bY(f)$
Time shift	$x(t - t_0)$	$e^{-j2\pi f t_0} X(f)$
Frequency shift (modulation)	$e^{j2\pi f_0 t} x(t)$	$X(f - f_0)$
Time scaling	$x(at)$	$\frac{1}{\|a\|} X(f/a)$
Convolution	$(x * y)(t)$	$X(f) Y(f)$
Multiplication	$x(t) y(t)$	$(X * Y)(f)$
Differentiation	$\frac{d}{dt} x(t)$	$j2\pi f\, X(f)$
Conjugation	$x^*(t)$	$X^*(-f)$
Duality	$X(t)$	$x(-f)$

The modulation and convolution properties are the most important for communications. Modulation shifts spectra — it's how we place baseband signals onto a carrier. Convolution becomes multiplication — it's how we analyse filters in the frequency domain.

Proof

Time-shift property

$\mathcal{F}\{x(t - t_0)\} = \int x(t - t_0) e^{-j2\pi ft}\,dt$ . Substituting $u = t - t_0$ : $= \int x(u) e^{-j2\pi f(u + t_0)}\,du = e^{-j2\pi f t_0} X(f)$ . $\blacksquare$

Modulation property

$\mathcal{F}\{e^{j2\pi f_0 t} x(t)\} = \int x(t) e^{-j2\pi(f - f_0)t}\,dt = X(f - f_0)$ . $\blacksquare$

Theorem: Parseval's Theorem (Rayleigh's Energy Theorem)

For an energy signal $x(t) \leftrightarrow X(f)$ ,

$E_x = \int_{-\infty}^{\infty} |x(t)|^2\,dt = \int_{-\infty}^{\infty} |X(f)|^2\,df.$

The quantity $|X(f)|^2$ is the energy spectral density (ESD): $\Phi_x(f) = |X(f)|^2$ , so that $E_x = \int \Phi_x(f)\,df$ .

Energy is conserved between time and frequency representations. This is the Fourier analogue of the Pythagorean theorem: the "length" (energy) of a signal does not change when we express it in a different orthonormal basis.

Proof

$\int |x(t)|^2\,dt = \int x(t) x^*(t)\,dt = \int x(t) \left[\int X^*(f) e^{-j2\pi ft}\,df\right] dt = \int X^*(f) \left[\int x(t) e^{-j2\pi ft}\,dt\right] df = \int X^*(f) X(f)\,df = \int |X(f)|^2\,df$ . $\blacksquare$

,

Essential Fourier Transform Pairs

$x(t)$	$X(f)$	Notes
$\delta(t)$	$1$	Flat spectrum
$1$	$\delta(f)$	Duality
$e^{j2\pi f_0 t}$	$\delta(f - f_0)$	Single frequency
$\cos(2\pi f_0 t)$	$\frac{1}{2}[\delta(f-f_0)+\delta(f+f_0)]$	Two spectral lines
$\mathrm{rect}(t/T)$	$T\,\mathrm{sinc}(fT)$	Pulse shaping
$\mathrm{sinc}(Wt)$	$\frac{1}{W}\mathrm{rect}(f/W)$	Ideal LPF (duality)
$e^{-at}u(t),\;a>0$	$\frac{1}{a + j2\pi f}$	Causal exponential
$e^{-a\|t\|},\;a>0$	$\frac{2a}{a^2 + (2\pi f)^2}$	Two-sided exponential
$e^{-\pi t^2}$	$e^{-\pi f^2}$	Gaussian (self-dual)

Why This Matters: Modulation Is Frequency Shifting

The modulation property $e^{j2\pi f_c t} x(t) \leftrightarrow X(f - f_c)$ is the mathematical core of every wireless system. A baseband signal $x(t)$ with bandwidth $W$ is multiplied by a carrier $e^{j2\pi f_c t}$ to shift its spectrum to a band centred at $f_c$ . At the receiver, multiplication by $e^{-j2\pi f_c t}$ shifts it back to baseband — this is down-conversion. Without the modulation property, frequency-division multiplexing, OFDM, and virtually all modern wireless protocols would not exist.

Definition:
Bandwidth

The bandwidth $W$ of a signal or system quantifies the range of frequencies it occupies or passes. Several definitions are used:

Absolute bandwidth: the support of $X(f)$ (zero outside $[-W, W]$ ). Finite only for band-limited signals.
3 dB bandwidth: the range of $f$ where $|X(f)|^2 \geq \frac{1}{2} |X(f)|^2_{\max}$ .
Null-to-null bandwidth: distance between the first zeros of $|X(f)|$ around the main lobe. For $\mathrm{sinc}(fT)$ , this equals $2/T$ .
RMS (Gabor) bandwidth: $W_{\text{rms}}^2 = \frac{\int f^2 |X(f)|^2\,df}{\int |X(f)|^2\,df}$ .

There is no single "correct" definition of bandwidth. In wireless standards, the 3 dB bandwidth or the occupied bandwidth (containing 99% of power) is most common.

Theorem: Time–Frequency Uncertainty Principle

For any signal $x(t) \leftrightarrow X(f)$ with finite energy,

$\Delta t \cdot \Delta f \geq \frac{1}{4\pi}$

where $\Delta t$ and $\Delta f$ are the RMS durations in time and frequency:

$\Delta t^2 = \frac{\int t^2 |x(t)|^2\,dt}{\int |x(t)|^2\,dt}, \qquad \Delta f^2 = \frac{\int f^2 |X(f)|^2\,df}{\int |X(f)|^2\,df}.$

Equality holds if and only if $x(t)$ is a Gaussian pulse $x(t) = C e^{-\alpha t^2}$ .

You cannot make a signal simultaneously short in time and narrow in frequency. This is the communications analogue of the Heisenberg uncertainty principle. It places a fundamental limit on time–frequency resolution and explains why pulse shaping always involves a time–bandwidth trade-off.

Proof

Sketch of proof

Apply the Cauchy–Schwarz inequality $|\langle f, g \rangle|^2 \le \|f\|^2 \|g\|^2$ to $f(t) = t\,x(t)$ and $g(t) = x'(t)$ . Using integration by parts and the differentiation property $\mathcal{F}\{x'(t)\} = j2\pi f X(f)$ , one obtains the stated bound. Equality in Cauchy–Schwarz requires $g = \lambda f$ , i.e., $x' = \lambda t x$ , whose solution is a Gaussian. $\blacksquare$

,

Time–Frequency Duality: $\mathrm{rect} \leftrightarrow \mathrm{sinc}$

Watch the fundamental duality between time and frequency. As the rectangular pulse widens (

T

increases), its sinc spectrum narrows — and vice versa. This is the uncertainty principle in action.

The rect–sinc transform pair animated as the pulse width

T

sweeps from 0.3 to 3.0. Wider pulse

\to

narrower spectrum.

Fourier Transform Pairs Explorer

Select a signal to see its time-domain waveform $x(t)$ and frequency-domain spectrum $X(f)$ side by side. Adjust parameters to observe how time-domain changes affect the spectrum (and vice versa).

Parameters

Signal type

Width / frequency parameter1

Common Mistake: Angular Frequency vs. Ordinary Frequency

Mistake:

Mixing up $\omega = 2\pi f$ and $f$ conventions, leading to missing or extra factors of $2\pi$ in transform pairs and filter designs.

Correction:

Pick one convention and stick with it. This book uses $f$ (Hz) throughout, so the Fourier kernel is $e^{-j2\pi ft}$ and the inverse kernel is $e^{j2\pi ft}$ — both without any $1/(2\pi)$ pre-factor. If you see $\omega$ in a reference, substitute $\omega = 2\pi f$ and simplify before combining with our formulae.

Quick Check

If $x(t) \leftrightarrow X(f)$ and $y(t) = x(t) e^{j2\pi f_0 t}$ , what is $Y(f)$ ?

$X(f) e^{j2\pi f_0 f}$

$X(f - f_0)$

$X(f + f_0)$

$X(f) + \delta(f - f_0)$

Correction:

X(f - f_0)

Correct. The modulation property states that multiplication by a complex exponential in time shifts the spectrum: $e^{j2\pi f_0 t} x(t) \leftrightarrow X(f - f_0)$ .

Quick Check

A signal $x(t)$ has energy $E_x = 5$ J. If we define $y(t) = 2x(t - 3)$ , what is $E_y$ ?

$5$ J

$10$ J

$20$ J

$40$ J

Correction:

20

J

Correct. Time-shifting does not change energy. Amplitude scaling by $a$ scales energy by $|a|^2$ . So $E_y = |2|^2 E_x = 4 \cdot 5 = 20$ J.

Time–Frequency Duality

The duality property states that if $x(t) \leftrightarrow X(f)$ , then $X(t) \leftrightarrow x(-f)$ . This means every time-domain statement has a frequency-domain mirror image:

$\mathrm{rect}(t/T) \leftrightarrow T\,\mathrm{sinc}(fT)$ implies $\mathrm{sinc}(Wt) \leftrightarrow \frac{1}{W}\mathrm{rect}(f/W)$
Convolution in time $\leftrightarrow$ multiplication in frequency, and vice versa
Short in time $\leftrightarrow$ wide in frequency (uncertainty principle)

Duality halves the number of transform pairs you need to memorise.

Fourier Transform

$X(f) = \int_{-\infty}^{\infty} x(t) e^{-j2\pi ft}\,dt$ . Decomposes a signal into its frequency components.

Bandwidth

The range of frequencies occupied by a signal or passed by a system. Common measures: absolute, 3 dB, null-to-null, RMS.

Sinc Function

$\mathrm{sinc}(u) = \sin(\pi u)/(\pi u)$ , with $\mathrm{sinc}(0) = 1$ . The Fourier transform of $\mathrm{rect}(t)$ .

Energy Spectral Density

$\Phi_x(f) = |X(f)|^2$ . Describes how the energy of a signal is distributed across frequency. Integrates to total energy $E_x$ .

Fourier Transform and Frequency-Domain Analysis

Historical Note: Joseph Fourier and the Heat Equation

Definition: Fourier Series

Definition: Continuous-Time Fourier Transform (CTFT)

Existence Conditions

Example: Rectangular Pulse and Sinc Function

Direct computation

Sinc notation

Theorem: Key Properties of the Fourier Transform

Time-shift property

Modulation property

Theorem: Parseval's Theorem (Rayleigh's Energy Theorem)

Proof

Essential Fourier Transform Pairs

Why This Matters: Modulation Is Frequency Shifting

Definition: Bandwidth

Theorem: Time–Frequency Uncertainty Principle

Sketch of proof

Time–Frequency Duality: rect↔sinc\mathrm{rect} \leftrightarrow \mathrm{sinc}rect↔sinc

Fourier Transform Pairs Explorer

Parameters

Common Mistake: Angular Frequency vs. Ordinary Frequency

Quick Check

Quick Check

Time–Frequency Duality

Fourier Transform

Bandwidth

Sinc Function

Energy Spectral Density

Definition:
Fourier Series

Definition:
Continuous-Time Fourier Transform (CTFT)

Definition:
Bandwidth

Time–Frequency Duality: $\mathrm{rect} \leftrightarrow \mathrm{sinc}$