Description

Flow Visualizer is designed to help music producers create engaging visual content for their electronic music. The project leverages AI to analyze the emotional content of a song and automatically generate visuals that change in sync with the music. The goal is to provide a solution that enhances the emotional experience of the listener through dynamic and visually appealing animations. The user experience is intended to be immersive, transforming the way music is enjoyed both visually and emotionally.

Challenges, accomplishment and lessons learned

Challenges:
Learning TouchDesigner within a short timeframe during a hackathon to present the project prototype.
Translating emotions into colors, which we ultimately entrusted to generative AI due to the subjectivity of philosophical treatises on color and emotion.
Attempting to train a deep learning AI for emotion mapping based on incomplete online documentation.
Accomplishments:
Successfully overcoming the initial learning curve of TouchDesigner through collaboration and inspiration from online projects.
Establishing a partnership with Cyanite, a significant achievement that enhanced the project’s credibility and capabilities.
Lessons Learned:
The importance of teamwork and leveraging external resources for inspiration and guidance.
The value of adapting and finding innovative solutions when faced with subjective or incomplete information.

Technology

• Cyanite AI for high-level feature extraction
• Bing AI for image generation
• Python with Pylette for color palette extraction
• TouchDesigner as the graphic engine

Students

Guglielmo Fratticcioli: Focused on TouchDesigner and early AI development.

Guido Elli: Specialized in TouchDesigner development.

Elia Pirrello: Led the team and contributed to early TouchDesigner development.

Media