Crafting Better Generative Music in Games with Deep Learning

Explore how deep learning can revolutionize procedural music generation in video games through this 46-minute conference talk from ADC 2024. Learn the fundamentals of machine learning before diving into peer-reviewed research on using deep learning to generate additional layers for pre-composed game music and create expressive MIDI performances during gameplay. Discover ethical and responsible approaches to AI-powered game audio tools that can empower composers and music designers rather than replace them, potentially reducing burnout while providing novel player experiences. Examine why procedural music generation has lagged behind other forms of procedural content creation in games, and understand how emerging AI technologies can address these challenges when implemented thoughtfully. Gain insights from cutting-edge PhD research that demonstrates practical applications of deep learning in game audio, including methods for enhancing existing musical compositions and creating dynamic, responsive soundtracks that adapt to player actions in real-time.