Physics-Informed Generative Audio for Interactive Game Environments

I aim to explore how AI-generated sound can respond to physical interactions and player context in games. The goal is to move towards adaptive audio that reflects what is happening in the game world, from an object's material and impact speed to the listener's position.

• Physics-informed sound generation

  Explore how material, collision velocity, contact type, and spatial position can guide generative audio models.

• Real-time game-engine integration

  Investigate low-latency integration in Unity, combining generative models, procedural synthesis, caching, and sample-based fallback.

• Player-adaptive audio

  Explore Audio Level of Detail to adapt sound generation to player attention, distance, and the importance of game events.

• Audio quality and player experience

  Plan to combine audio and system measurements with user studies of realism, immersion, responsiveness, and physical consistency.

This overview reflects my initial PhD research proposal. As I am at the beginning of my doctoral studies, the scope, methods, and research questions may evolve.