A simple python script for transcribing and translating videos locally, combining Faster-Whisper transcription capability with Argos Translate/Opus-MT translation engine.

No internet required* - everything is processed locally!
*after the dependencies and models are downloaded

• Python 3.10.18 (specifically this version due to dependency constraints)
• FFmpeg: For audio extraction from videos
• NVIDIA GPU: For CUDA acceleration (optional, CPU is supported but slower)

# 1. Navigate to project
cd goober

# 2. Initiate and activate virtual environment
uv venv --python 3.10.18
./venv/Scripts/activate

# 3. Sync dependencies
uv sync

# 4. Run interactive mode
uv run python main.py

This project uses uv for fast Python package management:

# Install uv if you don't have it
curl -LsSf https://astral.sh/uv/install.sh | sh  # bash
# or
powershell -ExecutionPolicy Bypass -c "irm https://astral.sh/uv/install.ps1 | iex"  # PowerShell

# Clone this repository
git clone https://github.com/narendnp/goober

# Navigate to project directory
cd goober

# Create a virtual environment
uv venv --python 3.10.18

# Activate the virtual environment
./venv/Scripts/activate  # bash
# or
./venv/Scripts/activate.ps1  # PowerShell

# Install dependencies
uv sync

If you plan to use Opus-MT translation engine, you need to install NLTK data:

# Enter Python environment
uv run python

# In Python interpreter:
>>> import nltk
>>> nltk.download('punkt_tab')
>>> exit()

For more information, go here.

For GPU acceleration, ensure you have:

• NVIDIA GPU with CUDA support
• CUDA toolkit installed
• PyTorch with CUDA support (already included in dependencies)

Inside the virtual environment, run the main script:

uv run python main.py

The script will prompt you for:

• Video path: Path to your input video file
• Source language: Language code (e.g., en, fr, ja) or auto for detection
• Target language: Desired translation language (e.g., id, en, es)
• Silence duration: Minimum silence duration for VAD (default: 500ms)
• Threshold: VAD sensitivity (0.1-1.0, default: 0.5)
• Translation engine: argos (faster) or opus (more accurate)

You can also run the script directly from the command line:

uv run src/tl_argos.py "path/to/video.mp4" \
  --language auto \
  --to en \
  --vad-ms 500 \
  --vad-threshold 0.5

uv run src/tl_opus.py "path/to/video.mp4" \
  --language auto \
  --to en \
  --vad-ms 500 \
  --vad-threshold 0.5

• --model: Whisper model size (large-v3, distil-large-v3, medium, small)
  • Larger models are more accurate but slower
  • Default: large-v3
• --device: Processing device (cuda for GPU, cpu for CPU)
• --compute-type: Precision level (float16, int8_float16)
• --language: Source language code or auto for detection
• --to: Target language code (required)
• --beam-size: Beam search size for transcription (default: 5)

• --vad-ms: Minimum silence duration in milliseconds (default: 500)
• --vad-threshold: Speech detection sensitivity (0.1-1.0, default: 0.5)
• --no-vad: Disable VAD filtering

• --batch-size: Translation batch size (default: 32)

Faster-Whisper, Argos Translate, and Opus-MT generally support a wide array of languages.

These are some of the popular language codes:

• English: en
• Indonesian: id
• Spanish: es
• French: fr
• German: de
• Japanese: ja
• Chinese: zh
• Korean: ko
• Arabic: ar

Please refer to the respective library's documentation for more details.

The tool generates two subtitle files on the directory of the video file:

1. Original transcription: {video_name}.orig.srt
2. Translated subtitles: {video_name}.{target_lang}.srt

1. Failed to build/install fasttext libary (Windows)

• Try installing it using the pre-built wheel (credit to FKz11)

  Inside this repo's directory, run:

  uv pip install https://github.com/FKz11/fasttext-0.9.3-windows-wheels/releases/download/0.9.3/fasttext-0.9.3-cp310-cp310-win_amd64.whl

• Re-run uv sync

Q: Why is it named goober?
A: Because I can see how gooners would use this to generate subtitles to watch JAV. (I know this doesn't explain it but I just think it's funny).

Q:

A: No.

MIT License. See LICENSE file for details.

• Faster-Whisper: Speech-to-text transcription
• Argos Translate: Fast translation library
• Opus-MT: High-quality translation models
• FFmpeg: Audio/video processing
• UV: Ultra-fast Python dependency management