Two-Stage Method of Speech Denoising by Long Short-Term Memory Neural Network

Nasretdinov, Rauf; Ilyashenko, I. D.; Lependin, A. A.

doi:10.1007/978-3-030-94141-3_8

Cited by 3 publications

(1 citation statement)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Recently, a large number of neural network methods for improving speech quality have been proposed, working both in real time [13,16] and offline [14]. The first type mainly includes models based on recurrent or convolutional neural networks [17], the second type includes models using self-attention layers [18].…”

Section: Model For Speech Enhancementmentioning

confidence: 99%

Speech enhancement augmentation for robust speech recognition in noisy environments

Nasretdinov,

Lependin,

Ilyashenko

2024

ITM Web Conf.

Self Cite

View full text Add to dashboard Cite

Abstract. The use of augmentations as a data enrichment method has become an important element in improving the performance of speech recognition systems. To work effectively in noisy conditions, augmentation is usually used to simulate the presence of background noise. However, the quality of speech recognition on samples pre-processed by noise reduction models does not increase. This paper proposes a new approach to speech data augmentation when training ASR systems, intended for their joint use with models for speech enhancement. It was based on the creation of several additional data samples containing speech samples processed by the enhancement model. The proposed approach was tested on the E-Branchformer neural network model using data from the Librispeech set. The quality of speech samples was assessed using the DNSMOS metric. By means of a 100-hour sample of clean speech samples it was shown that the proposed augmentation allows for an improvement in the WER metric of more than 4% in absolute value compared to the generally accepted approach based on adding noisy speech samples. Experiments on 960-hour data demonstrated the robustness of this approach as the training set size increased.

show abstract