Generating-mel-spectrogram-from-the-videos: Convert Video Audio to Mel-Spectrogram Images

This repository features a Python script that extracts audio from video files and converts it into high-quality Mel-spectrogram images. Using Librosa for audio processing and Matplotlib for visualization, this tool is ideal for creating datasets or visualizing audio features for video-based projects.

Features

1. Video Processing: Extracts audio from video files in .mp4 format.

2. Mel-Spectrogram Generation:

Converts extracted audio signals to Mel-spectrograms.

Scales power values to decibels for better visualization.

3. Image Output: Saves the spectrograms as transparent .jpg images, ready for integration into datasets or analysis workflows.

4. Batch Processing: Automatically processes multiple video files in a folder.

Workflow

1. Input Videos: Provide a folder containing .mp4 video files.

2. Audio Extraction and Processing:

Extracts audio from each video.

Generates Mel-spectrograms from the audio signal.

3. Image Output: Saves each Mel-spectrogram as a .jpg image in the specified output folder.

Applications

Preparing labeled datasets for audio or video machine learning models.

Analyzing audio features extracted from videos.

Generating visual representations for audio debugging or presentations.

Requirements

Python

Librosa

Matplotlib

NumPy