The field of music and audio technologies is moving towards more intuitive and interactive systems, enabling non-experts to engage with music creation and audio processing. Researchers are exploring new interfaces and models that allow for embodied and explainable interactions, making it easier for users to generate and manipulate music and audio. Noteworthy papers in this area include: SonicMaster, which introduces a unified generative model for music restoration and mastering with text-based control. Live Music Models, which presents a new class of generative models for music that produce a continuous stream of music in real-time with synchronized user control. TofuML, which employs a physical and spatial interface to make machine learning concepts more accessible and engaging for non-expert users. These innovative approaches are advancing the field and opening up new possibilities for music and audio creation, processing, and interaction.
Advancements in Interactive Music and Audio Technologies
Sources
DeformTune: A Deformable XAI Music Prototype for Non-Musicians
TofuML: A Spatio-Physical Interactive Machine Learning Device for Interactive Exploration of Machine Learning for Novices
How Would It Sound? Material-Controlled Multimodal Acoustic Profile Generation for Indoor Scenes
Fine-Tuning Text-to-Speech Diffusion Models Using Reinforcement Learning with Human Feedback
SonicMaster: Towards Controllable All-in-One Music Restoration and Mastering
Live Music Models
Towards Hallucination-Free Music: A Reinforcement Learning Preference Optimization Framework for Reliable Song Generation
Estimating Musical Surprisal from Audio in Autoregressive Diffusion Model Noise Spaces