Skip to content

Latest commit

 

History

History
123 lines (95 loc) · 3.05 KB

README.md

File metadata and controls

123 lines (95 loc) · 3.05 KB

AudioViz

AudioViz is a powerful Python-based tool that transforms audio content into a sequence of AI-generated images, creating visual narratives from sound. Using state-of-the-art AI models from OpenAI and Anthropic, it analyzes audio content and generates corresponding visual representations.

Features

  • Smart Audio Processing

    • Automatic audio compression and segmentation
    • Support for various audio formats
    • Intelligent handling of large audio files

  • Advanced Content Analysis

    • Speech-to-text transcription using OpenAI's Whisper
    • Context-aware text analysis
    • Automated visual style guide generation

  • AI Image Generation

    • High-quality image generation using DALL-E 3
    • Consistent visual styling across sequences
    • Customizable image parameters (size, quality, style)

Use Cases

  • YouTube Content Creation
  • Podcast Visualization
  • Educational Content
  • Story Visualization
  • Music Video Creation
  • Speech Visualization

Requirements

Essential Dependencies

  • Python 3.11+
  • FFmpeg (Required for audio processing)
  • OpenAI API key
  • Anthropic API key (optional, for Claude-based text analysis)

FFmpeg Installation

  • macOS (using Homebrew): 
    brew install ffmpeg
  • Ubuntu/Debian: 
    sudo apt-get update
sudo apt-get install ffmpeg
  • Windows: Download from FFmpeg official website

Setup

  1. Clone the repository:

    git clone https://github.com/aLVINlEE9/AudioViz.git
cd AudioViz

  2. Install dependencies using Poetry:

    poetry install

  3. Create .env file:

    OPENAI_API_KEY=your_openai_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key  # Optional

Usage

Basic usage example:

from audio_viz import AudioVisualizer

config = {
    "openai_api_key": "your_openai_api_key",
    "text_analysis": {
        "provider": "openai",  # or "anthropic"
        "max_tokens": 1000,
        "max_segments": 5
    },
    "image_generation": {
        "size": "1024x1024",
        "quality": "standard",
        "style": "vivid"
    }
}

visualizer = AudioVisualizer(config, verbose=True)
visualizer.process_audio("input.wav", "output_directory")

Configuration Options

Text Analysis

  • provider: Choose between "openai" or "anthropic"
  • max_tokens: Maximum tokens for LLM response
  • max_segments: Target number of visual segments

Image Generation

  • size: "1024x1024", "1024x1792", or "1792x1024"
  • quality: "standard" or "hd"
  • style: "vivid" or "natural"

Roadmap

  • Enhanced prompting system for more detailed and accurate scene descriptions
  • Video output support (combining audio with generated images)
  • Additional image generation models support (Midjourney, Stable Diffusion)
  • Custom style presets and templates
  • Batch processing support
  • Multiple language support

License

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

Support

For support, please open an issue in the repository or contact the maintainers.