Audio Processing Repository

This repository contains a collection of scripts, utilities, and a Jupyter Notebook for exploring and processing audio signals. It is designed for educational purposes and provides a hands-on approach to understanding audio signal processing concepts such as sine waves, Fourier Transform, Short-Time Fourier Transform (STFT), and Mel Frequency Cepstral Coefficients (MFCC).

---

Repository Structure

Files and Directories

• audioprocessing.ipynb: A comprehensive Jupyter Notebook that walks through various audio processing techniques, including:
  • Generating sine waves.
  • Visualizing waveforms.
  • Applying Fourier Transform and STFT.
  • Extracting MFCCs.
• audiofiles/: Contains sample audio files for analysis and experimentation.
  • notes/: Includes .wav files such as C2_C4.wav and G6_A6.wav for pitch and frequency analysis.
• images/: Contains visual aids and diagrams used in the notebook, such as:
  • Fourier Transform illustrations.
  • Mel Scale and MFCC process diagrams.
• utils/: A collection of Python utility scripts for audio processing:
  • plot_waveform.py: Functions to plot waveforms.
  • read_audio_and_plot.py: Reads audio files and plots their waveforms.
  • record_audio_and_plot.py: Records audio and visualizes it.
  • save_and_play_signal.py: Saves and plays generated audio signals.
• pyproject.toml: Defines the project dependencies and metadata.
• README.md: This file, providing an overview of the repository.
• uv.lock: Dependency lock file.

---

Features

1. Audio Signal Generation:

   • Generate sine and cosine waves with customizable frequencies, amplitudes, and phases.
   • Combine multiple signals to create complex waveforms.

2. Visualization:

   • Plot time-domain signals.
   • Visualize frequency-domain representations using FFT and STFT.

3. Audio Analysis:

   • Extract and analyze Mel Frequency Cepstral Coefficients (MFCC).
   • Understand the sensitivity of human hearing to different frequencies.

4. Interactive Learning:

   • Step-by-step explanations and visualizations in the Jupyter Notebook.
   • Links to external resources for deeper understanding.

---

Installation

Prerequisites

• Python 3.13 or higher.
• Recommended: A virtual environment to manage dependencies.

Steps

1. Install libraries using uv

   uv sync

---

Usage

Running the Jupyter Notebook

1. Open audioprocessing.ipynb and follow the cells step-by-step.

---

Concepts Covered

1. Sine Waves:

   • Mathematical representation:

   

2. Fourier Transform:

   • Convert time-domain signals to frequency-domain.

3. Short-Time Fourier Transform (STFT):

   • Analyze how frequencies change over time.

4. Mel Frequency Cepstral Coefficients (MFCC):

   • Extract features for audio classification and speech recognition.

---

Resources

Libraries

• Librosa: Audio and music signal analysis.
• SoundFile: Reading and writing sound files.

Tutorials

• Mel-Spectrogram and MFCCs | Lecture 72 (Part 1) | Applied Deep Learning
• Mel Frequency Cepstral Coefficients (MFCC) Explained

Research Papers

• McFee, Brian, et al. “librosa: Audio and music signal analysis in python.”

---