FSD-50K

FSD-50K is a large-scale audio dataset that contains a variety of sound events, primarily used to evaluate the performance of neural networks in identifying and classifying single-source audio samples.

Papers using FSD-50K (3)

From Physics to Representation: Audio Learning with Synthetic Pre-training via Procedural Generation2026FSD50K-Solo: Automated Curation of Single-Source Sound Events2026DiT-Flow: Speech Enhancement Robust to Multiple Distortions based on Flow Matching in Latent Space and Diffusion Transformers2026