An AI-Powered Speech Processing Toolkit and Open Source SOTA Pretrained Models, Supporting Speech Enhancement, Separation, and Target Speaker Extraction, etc.

A toolkit to calculate speech audio quality. Not affiliated with the original authors

Train no-reference speech quality estimators with multiple datasets via learned, per-dataset alignments.

Official baseline for ICASSP 2026 URGENT Challenge Track 2 (Speech Quality Assessment)

Streamlit Mechanical Turk (SMT) for the subjective evaluation of audio or speech generation models.

Automatic Speech Quality Assessment Pipeline

Performance of Objective Speech Quality Metrics on Languages Beyond Validation Data