Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement

Abstract: 

Deep-learning based speech enhancement systems have offered tremendous gains, where the best performing approaches use long short-term memory (LSTM) recurrent neural networks (RNNs) to model temporal speech correlations. These models, however, do not consider the frequency-level correlations within a single time frame, as spectral dependencies along the frequency axis are often ignored. This results in inaccurate frequency responses that negatively affect perceptual quality and intelligibility. We propose a deep-learning approach that considers temporal and frequency-level dependencies. More specifically, we enforce spectral-level dependencies within each spectral time frame through the introduction of a recurrent output layer that models the Markovian assumption along the frequency axis. We evaluate our approach in a variety of speech and noise environments, and objectively show that this recurrent spectral layer offers performance gains over traditional approaches. We also show that our approach outperforms recent approaches that consider frequency-level dependencies.

up
0 users have voted:

Paper Details

Authors:
Khandokar Md. Nayem, Donald S. Williamson
Submitted On:
13 October 2019 - 1:29pm
Short Link:
Type:
Poster
Event:
Presenter's Name:
Khandokar Md. Nayem
Paper Code:
143
Document Year:
2019
Cite

Document Files

Intra-Spectra Recurrent Output Layer

(24)

Subscribe

[1] Khandokar Md. Nayem, Donald S. Williamson, "Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4864. Accessed: Nov. 18, 2019.
@article{4864-19,
url = {http://sigport.org/4864},
author = {Khandokar Md. Nayem; Donald S. Williamson },
publisher = {IEEE SigPort},
title = {Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement},
year = {2019} }
TY - EJOUR
T1 - Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement
AU - Khandokar Md. Nayem; Donald S. Williamson
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4864
ER -
Khandokar Md. Nayem, Donald S. Williamson. (2019). Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement. IEEE SigPort. http://sigport.org/4864
Khandokar Md. Nayem, Donald S. Williamson, 2019. Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement. Available at: http://sigport.org/4864.
Khandokar Md. Nayem, Donald S. Williamson. (2019). "Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement." Web.
1. Khandokar Md. Nayem, Donald S. Williamson. Incorporating Intra-Spectral Dependencies With A Recurrent Output Layer For Improved Speech Enhancement [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4864