Stable Audio Open

The 'Stable Audio Open' dataset/benchmark contains audio samples used to evaluate the effectiveness of generative audio models in achieving controllable outputs, such as intensity, pitch, and beats, while maintaining generation quality.

Papers using Stable Audio Open (3)

Low-Resource Guidance for Controllable Latent Audio Diffusion2026Music Boomerang: Reusing Diffusion Models for Data Augmentation and Audio Manipulation2025Fast Text-to-Audio Generation with Adversarial Post-Training2025