Whisper-Subtitle-Generator

The Whisper Subtitle Generator leverages OpenAI's Whisper model to generate subtitles from audio and video files. This Python-based tool supports multiple languages and employs advanced audio processing techniques to ensure high accuracy in transcription.

Features

• Support for multiple video and audio formats.
• Multilingual support with language auto-detection.
• Noise reduction and audio normalization.
• Utilizes GPU for accelerated processing (if available).
• Outputs in SRT and VTT subtitle formats.

Quick Start

git clone https://github.com/Xza85hrf/Whisper-Subtitle-Generator.git
cd Whisper-Subtitle-Generator
python -m venv .venv
# Windows:
.venv\Scripts\activate
# macOS/Linux:
source .venv/bin/activate
pip install -r requirements.txt

# GUI
python GUI.py

# CLI
python Generate_TL_Sub_Frm_Video.py --input_file "path/to/video.mp4" --output "path/to/output.srt"

Prerequisites

• Python 3.10 or newer.
• ffmpeg on PATH for video decoding. Install via gyan.dev builds on Windows, brew install ffmpeg on macOS, or sudo apt install ffmpeg on Debian-family Linux.
• GPU (optional): A CUDA-capable NVIDIA GPU accelerates Whisper substantially. The project falls back to CPU automatically if CUDA is not available — expect inference to be 5-10× slower without a GPU. See the PyTorch install guide to replace the default CPU wheel with a CUDA wheel matching your driver.

Usage

• GUI: python GUI.py — opens a Tkinter window for picking the input media and output directory.
• CLI:

  python Generate_TL_Sub_Frm_Video.py --input_file "path/to/video.mp4" \
      --output "path/to/output.srt"

  Run python Generate_TL_Sub_Frm_Video.py --help for all options.

Contributing

Contributions are welcome — please fork the repository and submit pull requests with your improvements.

License

Distributed under the MIT License. See LICENSE for more information.