Project tgt_20:

"On Training Targets for Deep Learning Approaches to Clean Speech Magnitude Spectrum Estimation"

Project status

A manuscript for this paper is currently under review for The Journal of the Acoustical Society of America.

Summary

• This study compares training targets for deep learning approaches to clean speech magnitude spectrum.
• The tgt_20 repository contains the code used to conduct the experiments for the project.
• The code utilises the Deep Xi repository (https://github.com/anicolson/DeepXi).

Citation

[1] A. Nicolson, K. K. Paliwal, On Training Targets for Deep Learning Approaches to Clean Speech Magnitude Spectrum Estimation, TechRxiv (2020).