@misc{10481/80569,
year = {2021},
month = {3},
url = {https://hdl.handle.net/10481/80569},
abstract = {This paper presents a dual-channel speech enhance- ment framework that effectively integrates deep neural net- work (DNN) mask estimators. Our framework follows a beamforming-plus-postfiltering approach intended for noise reduction on dual-microphone smartphones. An extended Kalman filter is used for the estimation of the relative acous- tic channel between microphones, while the noise estimation is performed using a speech presence probability estimator. We propose the use of a DNN estimator to improve the prediction of the speech presence probabilities without making any assump- tion about the statistics of the signals. We evaluate and compare different dual-channel features to improve the accuracy of this estimator, including the power and phase difference between the speech signals at the two microphones. The proposed in- tegrated scheme is evaluated in different reverberant and noisy environments when the smartphone is used in both close- and far-talk positions. The experimental results show that our ap- proach achieves significant improvements in terms of speech quality, intelligibility, and distortion when compared to other approaches based only on statistical signal processing.},
organization = {Spanish Ministry of Science and Innovation Project No. PID2019-104206GB- I00/AEI/10.13039/501100011033},
organization = {Spanish Ministry of Uni- versities through the National Program FPU (grant reference FPU15/04161)},
publisher = {Proceedings of IBERSPEECH 2021},
keywords = {Dual-microphone smartphone},
keywords = {Beamforming},
keywords = {Extended Kalman filter},
keywords = {Speech presence probability},
keywords = {Deep Neural Network},
title = {Dual-channel eKF-RTF framework for speech enhancement with DNN-based speech presence estimation},
doi = {10.21437/IberSPEECH.2021-7},
author = {Martín Doñas, Juan M. and Peinado Herreros, Antonio Miguel and López Espejo, Iván and Gómez García, Ángel Manuel},
}