RE-USE: A universal speech enhancement model for diverse degradations, sampling rates, and languages.

Upload or record a noisy clip, then click Enhance to listen to the result and view its spectrogram. (ref: https://huggingface.co/spaces/rc19477/Speech_Enhancement_Mamba)

Note 1: For bandwidth extension, the performance may be affected by the characteristics of the input data, particularly the cutoff pattern. A simple solution is to apply low-pass filtering beforehand.

Note 2: When processing long input audio, out-of-memory (OOM) errors may occur. To address this, use the chunk-wise inference implementation provided on the Hugging Face.