Frequency Response and Filters GATE ECE Notes + Formulas + PYQs

Introduction

A real signal is rarely made of one pure frequency. Speech contains many tones, ECG contains slow biological variations plus noise, and a communication signal occupies a band of frequencies. Frequency response helps us understand how a system affects each part of that mixture.

In Signals and Systems, this topic connects Fourier analysis with practical system design. Once you know the frequency response, you can predict whether a system amplifies, attenuates, delays, distorts, or cleans a signal.

Why It Matters

Engineering problem solved

Filters remove noise without destroying the useful signal band.
Receivers separate one communication channel from nearby channels.
Audio systems shape bass, midrange, and treble using frequency-selective behavior.
Sensors use filtering to suppress power-line hum, vibration noise, and high-frequency interference.

Exam and interview value

GATE ECE problems often ask filter type from magnitude response.
University exams test cutoff frequency, bandwidth, phase response, and ideal filters.
Interviews check whether you can explain why convolution in time becomes multiplication in frequency.

Prerequisites

LTI systems and impulse response.
Convolution of signals.
Fourier Transform and sinusoidal steady-state idea.
Magnitude and phase of complex numbers.
Basic low-pass, high-pass, band-pass, and band-stop filter vocabulary.

Basic Intuition

Imagine a security gate that allows some people to enter and blocks others. A filter is similar, but its decision is based on frequency. Low-frequency components may pass, high-frequency components may be reduced, or only a middle band may be selected.

Frequency response is the rulebook of that gate. It tells how much each frequency is scaled and how much its phase is shifted.

A filter does not understand the signal name. It only responds to frequency content.

Input Spectrum, Filter Response, and Output Spectrum Diagram Here

Animated Frequency Selection Visualization

Core Theory

Frequency response of an LTI system

For an LTI system with impulse response $$h(t)$$, the frequency response is the Fourier Transform of $$h(t)$$.

$$H(j\omega)=\int_{-\infty}^{\infty}h(t)e^{-j\omega t}dt$$

For discrete-time systems, the frequency response is evaluated as $$H(e^{j\omega})$$.

Input-output relation

$$Y(j\omega)=X(j\omega)H(j\omega)$$

This is the central reason frequency response is powerful. Time-domain convolution becomes frequency-domain multiplication, so system action becomes easier to visualize.

Magnitude and phase

$$H(j\omega)=|H(j\omega)|e^{j\angle H(j\omega)}$$

Magnitude response tells gain or attenuation. Phase response tells delay or phase shift. A system can preserve amplitude but still distort waveform shape if phase is not handled properly.

Working Principle

A filter works by giving different gains to different frequency components. The output spectrum is formed by multiplying the input spectrum with the filter response.

Step 1: Break into frequencies

Fourier analysis views the input as a combination of sinusoids.

Step 2: Apply system response

Each frequency is scaled by magnitude response and shifted by phase response.

Step 3: Recombine output

The modified frequency components combine to create the output signal.

Animated Filter Magnitude and Phase Response Visualization

Formula Explanation

Continuous-time frequency response

$$H(j\omega)=\mathcal{F}\{h(t)\}$$

The impulse response contains the full time-domain identity of an LTI system.

Discrete-time frequency response

$$H(e^{j\omega})=\sum_{n=-\infty}^{\infty}h[n]e^{-j\omega n}$$

This is the DTFT of the discrete-time impulse response.

Output spectrum

$$Y(\omega)=X(\omega)H(\omega)$$

The filter shapes the input spectrum by multiplication.

Bandwidth

$$BW=f_H-f_L$$

Bandwidth is the useful range of frequencies passed by a filter or channel.

Diagram Explanation Placeholder

The best visual for this topic shows three aligned graphs: input spectrum, filter magnitude response, and output spectrum. Frequencies inside the passband remain strong, while stopband frequencies shrink.

Magnitude Response and Passband-Stopband Diagram Here

Phase Response and Group Delay Diagram Here

Animated Input Spectrum Through Filter Visualization

Frequency Behavior

Filter names describe which part of the spectrum is allowed to pass. Ideal filters have sharp boundaries, but practical filters transition gradually and introduce phase effects.

Filter type	Passes	Rejects	Typical use
Low-pass	Low frequencies	High-frequency noise	Anti-aliasing, smoothing
High-pass	High frequencies	DC and slow drift	Audio coupling, edge emphasis
Band-pass	A selected band	Below and above the band	Communication channel selection
Band-stop	Outside a selected band	One unwanted band	50/60 Hz hum removal

Real Engineering Implementation

In hardware, filters are built using resistors, capacitors, inductors, op-amps, or switched-capacitor circuits. In DSP, filters are implemented as algorithms that calculate each output sample from input samples and stored past values.

Analog implementation

RC filters for simple low-pass and high-pass behavior.
RLC filters for resonance and band selection.
Active filters using op-amps for gain and sharper response.

Digital implementation

FIR filters for stable linear-phase designs.
IIR filters for efficient sharp responses.
Real-time DSP in audio, telecom, radar, and embedded systems.

Real-World Applications

Anti-aliasing filters before analog-to-digital converters.
Audio equalizers and noise reduction systems.
Channel selection in radio and communication receivers.
ECG and EEG signal conditioning in biomedical instruments.
Image processing filters for sharpening and smoothing.
Vibration analysis and fault detection in industrial machines.
Control-system noise rejection and sensor signal conditioning.

Solved Examples

Example 1: Output spectrum

A signal has frequency components at $$100\,Hz$$ and $$5\,kHz$$. It passes through an ideal low-pass filter with cutoff $$1\,kHz$$.

$$100\,Hz\ \text{passes},\qquad 5\,kHz\ \text{is rejected}$$

The output keeps the slow component and removes the high-frequency component.

Example 2: Bandwidth

A band-pass filter passes frequencies from $$2\,kHz$$ to $$8\,kHz$$.

$$BW=f_H-f_L=8\,kHz-2\,kHz=6\,kHz$$

The useful frequency range is $$6\,kHz$$ wide.

Example 3: Frequency response from impulse response

For $$h(t)=e^{-at}u(t)$$ where $$a>0$$:

$$H(j\omega)=\int_0^\infty e^{-at}e^{-j\omega t}dt=\frac{1}{a+j\omega}$$

Magnitude decreases as $$\omega$$ increases, so the system behaves like a low-pass response.

Common Mistakes

Looking only at magnitude response and ignoring phase distortion.
Confusing cutoff frequency with bandwidth.
Assuming ideal brick-wall filters exist exactly in practical circuits.
Calling every frequency-selective system a low-pass filter without checking the passband.
Forgetting that output spectrum is input spectrum multiplied by system response.
Mixing Hz and rad/s without the conversion $$\omega=2\pi f$$.

Interview Questions

What is frequency response in an LTI system?
Why is the impulse response enough to determine frequency response?
What is the difference between magnitude response and phase response?
How does a low-pass filter remove high-frequency noise?
What is bandwidth and why does it matter in communication systems?
Why are ideal filters impossible to realize exactly?
How are FIR and IIR filters different in practical DSP?

Exam Notes

For LTI systems, remember $$Y(\omega)=X(\omega)H(\omega)$$.
Magnitude response decides passband and stopband behavior.
Phase response decides delay and waveform distortion.
Low-pass filters are common before ADC sampling.
Bandwidth of a band-pass filter is $$f_H-f_L$$.
For distortionless transmission, magnitude should be constant and phase should be linear over the signal band.

Revision Summary

Frequency response is the Fourier Transform of impulse response.
Filters shape signals by passing some frequencies and attenuating others.
Magnitude response controls gain; phase response controls timing shift.
Convolution in time becomes multiplication in frequency.
Cutoff frequency separates passband and stopband behavior.
Practical filters have transition bands and non-ideal phase behavior.

Practice Questions

Conceptual

Explain frequency response using the idea of a frequency gate.
Why does phase response matter even if magnitude response looks correct?
Why is an anti-aliasing filter usually low-pass?
What makes a system distortionless over a signal band?

Numerical

A band-pass filter has $$f_L=300\,Hz$$ and $$f_H=3.4\,kHz$$. Find bandwidth.
For $$H(j\omega)=1/(2+j\omega)$$, describe whether the response is low-pass or high-pass.
An input has components at $$50\,Hz$$, $$1\,kHz$$, and $$20\,kHz$$. Which components pass through an ideal low-pass filter of cutoff $$2\,kHz$$?
Convert cutoff frequency $$f_c=1\,kHz$$ into angular frequency $$\omega_c$$.

MCQs

For an LTI system, frequency response is Fourier Transform of: impulse response / input only / output only / noise.
A band-stop filter rejects: one band / all low frequencies only / all high frequencies only / no frequency.
Distortionless transmission requires: constant magnitude and linear phase / zero magnitude / random phase / infinite bandwidth.