Fourier Integrals and Fourier Transforms
Exercise 1 : Fourier Series
For the following periodic function:
$$ f(t) = \begin{cases} -1 & \text{for } -\frac{T}{2} \leq t < 0 \\ 1 & \text{for } 0 < t \leq \frac{T}{2} \end{cases} $$Calculate the Fourier series for
$$ f(t) = \begin{cases} 0 & \text{for } -\frac{T}{2} \leq t < -\frac{t_0}{2} \\ -1 & \text{for } -\frac{t_0}{2} \leq t < 0 \\ 1 & \text{for } 0 < t \leq \frac{t_0}{2} \\ 0 & \text{for } \frac{t_0}{2} < t \leq \frac{T}{2} \end{cases} $$Exercise 2 : Limiting Spectrum (T \to \infty)
Now perform the limiting process for $T \to \infty$ and obtain the amplitude spectrum.
Exercise 3 : Spectral Sketches for Varying Periods
Sketch the frequency spectrum of the Fourier series for $T = t_0$ and $T = 2t_0$, $T = 4t_0$ and $T = 8t_0$ and then the amplitude spectrum of the Fourier integral.
Exercise 4 : Complex Fourier Transform of the Pulse
Perform the Fourier transform in the complex representation for the function given in exercise 1 and obtain the amplitude function and the amplitude spectrum:
$$ f(t) = \begin{cases} -1 & \text{for } -\frac{t_0}{2} \leq t < 0 \\ 1 & \text{for } 0 < t \leq \frac{t_0}{2} \end{cases} $$Exercise 5 : Amplitude Spectra for Different $t_0$
Sketch the function and the amplitude spectrum for the preceding exercise for $t_0 = 1$, $t_0 = 2$ and $t_0 = 4$.
Exercise 6 : Shifted Pulse — Arbitrary Position
Determine amplitude function and amplitude spectrum for the function of the preceding exercise in an arbitrary position:
$$ f(t) = \begin{cases} 1 & \text{for } t_1 < t \leq t_1 + \frac{t_0}{2} \end{cases} $$Exercise 7 : Convolution and Triangular Pulse
Let $g(t)$ be the rectangular pulse of width $t_0$ and unit amplitude centered at the origin (the pulse from Exercise 4 with $t_1=0$).
- Compute the convolution $h(t)=g(t)*g(t)$ and show that $h(t)$ is a triangular pulse. Give an explicit expression for $h(t)$.
- Using the Fourier transform, derive the amplitude function $H(\omega)$ and show how it relates to $G(\omega)$ (the transform of $g$).
- Sketch the amplitude spectrum of $h(t)$ and compare it to the squared magnitude $|G(\omega)|^2$.
Exercise 8 : Parseval’s Theorem — Energy Equality
For the rectangular pulse $g(t)$ of width $t_0$ and unit amplitude:
- Compute the signal energy $E=\int_{-\infty}^{\infty} |g(t)|^2,dt$ directly.
- Compute the energy using the Fourier transform via Parseval’s theorem and verify both results agree.
Exercise 9 : Modulation and Frequency Shifting
Let $g(t)$ be a finite-duration signal with Fourier transform $G(\omega)$.
- Show that the modulated signal $g_m(t)=g(t)\cos(\omega_0 t)$ has a Fourier transform consisting of two shifted copies of $G(\omega)$ and write the expression for $G_m(\omega)$.
- For the rectangular pulse from Exercise 4, compute and sketch the amplitude spectrum of $g_m(t)$ for $\omega_0 = 2\pi,,4\pi$.
Exercise 10 : Time Scaling and Bandwidth
Consider the time-scaled signal $g_a(t)=g(at)$ for real nonzero $a$.
- Derive the relation between $G_a(\omega)$ and $G(\omega)$, and explain how time compression/expansion affects the amplitude spectrum and effective bandwidth.
- For the rectangular pulse with $t_0=1$, sketch amplitude spectra for $a=0.5,;1,;2$.
Exercise 11 : Sampling, Aliasing and the Poisson Summation Formula
Let $g(t)$ be bandlimited with transform $G(\omega)$ supported in $|\omega|\leq \Omega_m$.
- Describe ideal sampling at period $T_s$ (multiplying by a comb) and derive the sampled-spectrum using the Poisson summation formula.
- Determine the sampling condition to avoid aliasing and illustrate with sketches when $T_s$ is too large and when $T_s$ satisfies the Nyquist requirement.
Exercise 12 : Gaussian Pulse — Transform and Properties
Let $g(t)=e^{-\alpha t^2}$ with $\alpha>0$.
- Compute the Fourier transform $G(\omega)$ in closed form (you may use standard integrals) and show that it is a Gaussian in $\omega$.
- Discuss the time–bandwidth product for this pulse and show that scaling $\alpha$ trades time-spread for frequency-spread.
Exercise 13 : Hilbert Transform and Analytic Signal
Define the Hilbert transform $\mathcal{H}{g}(t)$ and the analytic signal $g_a(t)=g(t)+i,\mathcal{H}{g}(t)$.
- For a simple cosine $g(t)=\cos(\omega_0 t)$, compute the Hilbert transform and the analytic signal explicitly.
- For a one-sided rectangular pulse (nonzero only for $t>0$), discuss qualitatively how the Hilbert transform affects the phase of the Fourier spectrum.
Exercise 14 : Windowing, Spectral Leakage and Resolution
Take the infinite-duration sinusoid $s(t)=\cos(\omega_0 t)$ and multiply it by a finite window $w(t)$ (rectangular, Hamming, and Hann windows of duration $T_w$).
- For each window, compute or sketch the resulting amplitude spectrum and point out main-lobe width and side-lobe behavior.
- Discuss how window choice and window length $T_w$ affect frequency resolution and leakage. Provide sketches for $T_w=1,;2,;4$ and $\omega_0=5$.