CS304-SpeechRecognition

The class is taken at DKU, taught by Prof. Ming Li

The group project by Jingheng Huan and Gezhi Wang

Description

This project is designed to capture, process, and visualize audio data, specifically focusing on the extraction and plotting of Mel-Frequency Cepstral Coefficients (MFCC). It includes a set of Python scripts that handle audio capture, MFCC computation, and visualization of audio signals and features.

Installation

Before running the project, ensure you have Python installed on your system. You will also need the following libraries:

• numpy
• scipy
• matplotlib
• wave

Install these using pip:

pip install numpy scipy matplotlib wave

Usage

To use this project, run the main.py script:

python main.py

This will start the audio capture process, compute the MFCC, and plot the waveform, spectrogram, and cepstrum, saving each plot as an image.

Files and Functions

• audio_capture.py: Handles the audio recording functionality.
• audio_utils.py: Contains utility functions for audio processing.
• config.py: Configuration settings for the project.
• main.py: The main script that orchestrates the capture and plotting process.
• plotting.py: Contains functions to plot the audio waveform, spectrogram, and cepstrum.

Contributing

If you'd like to contribute to this project, please fork the repository and create a pull request with your features or fixes.

License

This project is licensed under the MIT License - see the LICENSE.md file for details.

Acknowledgments