Blind Speech Separation and Dereverberation using Neural Beamforming

Pfeifenberger, Lukas; Pernkopf, Franz

doi:10.48550/arxiv.2103.13443

Cited by 1 publication

(1 citation statement)

References 47 publications

(93 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Besides speech enhancement, dereverberation is also a challenging task since it is hard to pinpoint the direct path signal and differentiate it from its copies, especially when reverberation is strong and non-stationary noise is also present. Some previous works have already been done on simultaneous multi-channel enhancement and dereverberation [10,11,12]. However, it is still a relatively hard task for the single-channel scenario due to the lack of spatial information.…”

Section: Introductionmentioning

confidence: 99%

Uformer: A Unet based dilated complex & real dual-path conformer network for simultaneous speech enhancement and dereverberation

Fu¹,

Liu²,

Li³

et al. 2021

Preprint

View full text Add to dashboard Cite

Complex spectrum and magnitude are considered as two major features of speech enhancement and dereverberation. Traditional approaches always treat these two features separately, ignoring their underlying relationship. In this paper, we propose Uformer, a Unet based dilated complex & real dual-path conformer network in both complex and magnitude domain for simultaneous speech enhancement and dereverberation. We exploit time attention (TA) and dilated convolution (DC) to leverage local and global contextual information and frequency attention (FA) to model dimensional information. These three sub-modules contained in the proposed dilated complex & real dual-path conformer module effectively improve the speech enhancement and dereverberation performance. Furthermore, hybrid encoder and decoder are adopted to simultaneously model the complex spectrum and magnitude and promote the information interaction between two domains. Encoder decoder attention is also applied to enhance the interaction between encoder and decoder. Our experimental results outperform all SOTA time and complex domain models objectively and subjectively. Specifically, Uformer reaches 3.6032 DNSMOS on the blind test set of Interspeech 2021 DNS Challenge, which outperforms all top-performed models. We also carry out ablation experiments to tease apart all proposed submodules that are most important.

show abstract

Section: Introductionmentioning

confidence: 99%