/voice-gender-with-extractor

This repository is dedicated to training a machine learning model to classify voices as male or female based on their acoustic properties.

Primary LanguageJupyter NotebookMIT LicenseMIT

Voice Gender Recognition

Overview

This project focuses on training a machine learning model to classify voices as male or female based on their acoustic properties. The dataset used for training consists of 36,168 voice samples collected from a Yandex contest. The best-performing model achieves an accuracy of 98% during cross-validation.

Dataset Information

You can download the raw dataset, which includes .wav files and a corresponding .csv file with labels, from the following link: Raw Dataset.

Alternatively, a pre-processed version of the dataset, containing 3.5k rows (due to hardware limitations), is available here: Pre-processed Dataset.

Acoustic Properties

The following acoustic properties are measured for each voice sample:

  • meanfreq: Mean frequency (in kHz)
  • sd: Standard deviation of frequency
  • median: Median frequency (in kHz)
  • Q25: First quantile (in kHz)
  • Q75: Third quantile (in kHz)
  • IQR: Interquantile range (in kHz)
  • skew: Skewness
  • kurt: Kurtosis
  • sp.ent: Spectral entropy
  • sfm: Spectral flatness
  • mode: Mode frequency
  • centroid: Frequency centroid
  • peakf: Peak frequency (frequency with highest energy)
  • meanfun: Average fundamental frequency measured across the acoustic signal
  • minfun: Minimum fundamental frequency measured across the acoustic signal
  • maxfun: Maximum fundamental frequency measured across the acoustic signal
  • meandom: Average dominant frequency measured across the acoustic signal
  • mindom: Minimum dominant frequency measured across the acoustic signal
  • maxdom: Maximum dominant frequency measured across the acoustic signal
  • dfrange: Range of dominant frequency measured across the acoustic signal
  • modindx: Modulation index, calculated as the accumulated absolute difference between adjacent measurements of fundamental frequencies divided by the frequency range

Technology Stack

The project utilizes the following libraries and tools:

  • Pandas: For data manipulation and analysis
  • NumPy: For numerical operations
  • Seaborn and Matplotlib: For data visualization
  • Scikit-learn: For machine learning algorithms and model evaluation
  • XGBoost: For gradient boosting algorithms
  • Librosa: For audio processing and feature extraction
  • Concurrent: For parallel processing
  • Logging: For logging errors and warnings

Environment Setup

  1. Create a Virtual Environment: 
    python -m venv env
  2.  source env/bin/activate
  3.  pip install -r requirements.txt