MOSA-Net Whisper features

This is a demo of MOSA-Net+, an improved version of MOSA- NET that predicts human-based speech quality and intelligibility. MOSA-Net+ uses Whisper to generate cross-domain features. The model employs a CNN- BLSTM architecture with an attention mechanism and is trained using a multi-task learning approach to predict subjective listening test scores.
MOSA-Net+ was tested in the noisy-and-enhanced track of the VoiceMOS Challenge 2023, where it obtained the top-ranked performance among nine systems full paper

If the model contributes to your research please cite the following work:

R. E. Zezario, S. -W. Fu, F. Chen, C. -S. Fuh, H. -M. Wang and Y. Tsao, "Deep Learning-Based Non-Intrusive Multi-Objective Speech Assessment Model With Cross-Domain Features," in IEEE/ACM Transactions on Audio, Speech, and Language Processing, vol. 31, pp. 54-70, 2023, doi: 10.1109/TASLP.2022.3205757.

R. E. Zezario, Y.-W. Chen, S.-W. Fu, Y. Tsao, H.-M. Wang, C.-S. Fuh, "A Study on Incorporating Whisper for Robust Speech Assessment," IEEE ICME 2024, July 2024, (Top Performance on the Track 3 - VoiceMOS Challenge 2023)"

demo contributed by @wetdog