Spectrogram visualization in Signal Processing - Mini Project: Build & Apply

Spectrogram Visualization

📖 Scenario: You have recorded a short audio signal and want to see how its frequency content changes over time. A spectrogram is a great way to visualize this. It shows frequencies on one axis, time on another, and the intensity of frequencies as colors.

🎯 Goal: Build a simple program that creates a spectrogram from a given audio signal and displays it as a color image.

📋 What You'll Learn

Create a sample audio signal as a numpy array

Set up parameters for the spectrogram calculation

Calculate the spectrogram using a signal processing function

Display the spectrogram using a plotting library

💡 Why This Matters

🌍 Real World

Spectrograms are used in audio analysis, speech recognition, and music visualization to understand how sound frequencies change over time.

💼 Career

Knowing how to create and interpret spectrograms is useful for roles in audio engineering, data science, and signal processing.

1

Create a sample audio signal

Create a numpy array called signal that contains a 1-second sine wave at 5 Hz sampled at 100 Hz. Use numpy's linspace to create time points and sin to generate the wave.

Signal Processing

# Your code here

Hint

Use np.linspace(0, 1, fs, endpoint=False) to create time points and np.sin(2 * np.pi * 5 * t) for the sine wave.

2

Set spectrogram parameters

Create a variable called nperseg and set it to 20. This will be the length of each segment for the spectrogram calculation.

Signal Processing

import numpy as np

fs = 100  # Sampling frequency in Hz

# Your code here


t = np.linspace(0, 1, fs, endpoint=False)  # Time vector for 1 second
signal = np.sin(2 * np.pi * 5 * t)  # 5 Hz sine wave

Hint

Set nperseg = 20 to define the segment length.

3

Calculate the spectrogram

Use scipy.signal.spectrogram to calculate the spectrogram of signal. Store the frequency array in frequencies, the time array in times, and the spectrogram matrix in Sxx. Use fs as the sampling frequency and nperseg as the segment length.

Signal Processing

import numpy as np
from scipy.signal import spectrogram

fs = 100  # Sampling frequency in Hz
nperseg = 20

# Your code here

t = np.linspace(0, 1, fs, endpoint=False)  # Time vector for 1 second
signal = np.sin(2 * np.pi * 5 * t)  # 5 Hz sine wave

Hint

Call spectrogram(signal, fs=fs, nperseg=nperseg) and unpack the result into frequencies, times, and Sxx.

4

Display the spectrogram

Use matplotlib.pyplot to display the spectrogram matrix Sxx as an image. Use plt.pcolormesh(times, frequencies, Sxx) to plot. Add labels for the x-axis as Time [sec] and y-axis as Frequency [Hz]. Finally, call plt.show() to display the plot.

Signal Processing

import numpy as np
from scipy.signal import spectrogram
import matplotlib.pyplot as plt

fs = 100  # Sampling frequency in Hz
nperseg = 20

# Your code here

t = np.linspace(0, 1, fs, endpoint=False)  # Time vector for 1 second
signal = np.sin(2 * np.pi * 5 * t)  # 5 Hz sine wave

frequencies, times, Sxx = spectrogram(signal, fs=fs, nperseg=nperseg)

Hint

Use plt.pcolormesh(times, frequencies, Sxx) to plot, then label axes with plt.xlabel and plt.ylabel. Finally, call plt.show().