SpecMNet: Spectrum mend network for monaural speech enhancement. (15th June 2022)

• Record Type: Journal Article
• Authors: Fan, Cunhang; Zhang, Hongmei; Yi, Jiangyan; Lv, Zhao; Tao, Jianhua; Li, Taihao; Pei, Guanxiong; Wu, Xiaopei; Li, Sheng
• Is Part Of: Applied acoustics. Volume 194(2022)
• Publication Date: 2022-06-15
• Subjects: Monaural speech enhancement -- Speech distortion -- Spectrum mend network -- SI-SNR -- BLSTM; Acoustical engineering -- Periodicals; Periodicals; 620.2
• DOI: 10.1016/j.apacoust.2022.108792
• Languages: English
• ISSNs: 0003-682X

Abstract

Highlights: To the best of our knowledge, this is the first work to apply the spectrum mend method to address the speech distortion problem. We propose the SpecMNet algorithm to mend the pre-enhanced spectrum by the weighted original spectrum so that it can retrieve the lost speech information. Experiments are conducted on TIMIT + (100 Nonspeech Sounds and NOISEX-92) datasets. Experimental results prove that our proposed SpecMNet method can significantly improve the performance of speech enhancement and it is effective to alleviate the speech distortion problem. Abstract: Speech enhancement methods usually suffer from speech distortion problem, which leads to the enhanced speech losing so much significant speech information. This damages the speech quality and intelligibility. In order to address this issue, we propose a spectrum mend network (SpecMNet) for monaural speech enhancement. The proposed SpecMNet aims to retrieve the lost information by mending the weighted enhanced spectrum with weighted original spectrum. More specifically, the proposed algorithm consists of pre-enhancement network and the mend network. The main task of pre-enhancement network is to acquire the pre-enhanced spectrum so that it can remove the most of the noise signals. Because of the speech distortion problem, it loses a great deal of speech components. While the original spectrum has no speech information lost. Therefore, we utilize the original spectrum to mend the pre-enhanced spectrum by adding these two weighted spectrums so that the lost speech information can be retrieved. Then the mend network is used to predict mend weights for these two spectrums. Finally, the mended spectrum is used as the enhanced output. Our experiments are conducted on the TIMIT + (100 Nonspeech Sounds and NOISEX-92) datasets. Experimental results demonstrate that our proposed SpecMNet approach is effective to alleviate the speech distortion problem.