Overview of speech quality metrics in terms of automated evaluation of signal denoising in a presence of non-stationary noise

Duzinkiewicz K, Koszewski D, Pietrusinska K, Trella P. Overview of speech quality metrics in terms of automated evaluation of signal denoising in a presence of non-stationary noise. In: AES Convention 149. Audio Engineering Society; 2020. Paper 625. Available from: https://aes.org/publications/elibrary-page/?id=20911

@inproceedings{Duzinkiewicz2020_20911,
  author = {Duzinkiewicz, Karol and Koszewski, Damian and Pietrusinska, Kamila and Trella, Pawel},
  title = {{Overview of speech quality metrics in terms of automated evaluation of signal denoising in a presence of non-stationary noise}},
  booktitle = {AES Convention 149},
  note = {Paper 625},
  year = {2020},
  month = oct,
  publisher = {Audio Engineering Society},
  url = {https://aes.org/publications/elibrary-page/?id=20911}
}

TY  - CPAPER
TI  - Overview of speech quality metrics in terms of automated evaluation of signal denoising in a presence of non-stationary noise
AU  - Duzinkiewicz, Karol
AU  - Koszewski, Damian
AU  - Pietrusinska, Kamila
AU  - Trella, Pawel
T2  - AES Convention 149
M1  - Paper 625
PY  - 2020
DA  - 2020/10/06
UR  - https://aes.org/publications/elibrary-page/?id=20911
PB  - Audio Engineering Society
LA  - en
AB  - Recent developments in neural network-based speech enhancement ask for a robust subjective metric that can be used for comparing performance of a different noise suppression algorithms (for non-stationary noises), that would closely match costly and time-consuming subjective user tests. The article describes results of a comparison between subjective scores obtained using MUSHRA methodology vs. automated evaluation with objective metrics i.e. POLQA, 3QUEST, STOI & ESTOI, on a set of recordings processed by 2 different denoising algorithms for close & far speaker distance. Correlation coefficient is calculated between subjective scores and examined metrics. The results are based on recordings simulated using an in-house simulation toolchain, based on impulse responses from actual laptop device used in low reverb quiet room.
ER  -