Overview

This project focuses on enhancing audio similarity detection using machine learning and signal processing techniques. By leveraging Python libraries like Librosa and Resemblyzer, the system extracts voice embeddings and computes cosine similarity between audio files. The solution addresses the challenge of distinguishing between similar voices while accounting for noise variations. Advanced feature extraction techniques such as MFCC and chroma features are employed to improve accuracy. The model is fine-tuned to identify the same speaker's voice across noisy and clean audio samples.

Features

• Audio preprocessing and feature extraction using Librosa
• Voice embedding generation with Resemblyzer
• Cosine similarity calculation for audio comparison
• MFCC and chroma feature extraction to enhance comparison accuracy
• Robust to noisy datasets
• Visualization of waveforms, spectrograms, and similarity scores

Requirements

Python 3.7 or later is recommended.

Installation

Install the required libraries by running

  pip install numpy

  pip install librosa

  pip install seaborn

  pip install matplotlib

  pip install resemblyzer

Download or clone the repository from GitHub.

  git clone https://github.com/5W4PN1L/Audio-Matching-Tool.git

Usage

1. Save the audio files in .wav format.
2. Place your target audio file in the project directory and name it aud1.wav, or modify the script to use your own file name.
3. Place reference audio files (e.g., audio1.wav, audio4.wav) in the same directory, or modify the script to use your own file name.
4. Run the audiomatching.py script to compute the similarity between the target and reference audio files.

To execute the script:

  python audiomatching.py

Authors

• B.Swapnil