DNS-3

The DNS-3 dataset is used to evaluate speech enhancement methods and contains a variety of noisy speech recordings for benchmarking performance.

Papers using DNS-3 (3)

HALO: Half-Frame-Rate Adaptive Learnable Operator for Lightweight STFT-Based Speech Enhancement2026I-DCCRN-VAE: An Improved Deep Representation Learning Framework for Complex VAE-based Single-channel Speech Enhancement2025Investigation of Speech and Noise Latent Representations in Single-channel VAE-based Speech Enhancement2025