Large-scale training to increase speech intelligibility for hearing-impaired listeners in novel noises

• Published in: The Journal of the Acoustical Society of America Vol. 139; no. 5; pp. 2604 - 2612
• Main Authors: Chen, Jitong, Wang, Yuxuan, Yoho, Sarah E., Wang, DeLiang, Healy, Eric W.
• Format: Journal Article
• Language: English
• Published: United States Acoustical Society of America 01-05-2016
• Subjects: Acoustic Stimulation; Adult; Aged; Algorithms; Auditory Threshold; Case-Control Studies; Electric Stimulation; Female; Hearing Aids; Hearing Loss, Sensorineural - diagnosis; Hearing Loss, Sensorineural - physiopathology; Hearing Loss, Sensorineural - psychology; Hearing Loss, Sensorineural - rehabilitation; Humans; Male; Middle Aged; Neural Networks (Computer); Noise - adverse effects; Perceptual Masking; Persons With Hearing Impairments - psychology; Persons With Hearing Impairments - rehabilitation; Signal Processing, Computer-Assisted; Sound Spectrography; Speech Acoustics; Speech Communication; Speech Intelligibility; Speech Perception; Young Adult
• ISSN: 0001-4966; 1520-8524
• DOI: 10.1121/1.4948445

Supervised speech segregation has been recently shown to improve human speech intelligibility in noise, when trained and tested on similar noises. However, a major challenge involves the ability to generalize to entirely novel noises. Such generalization would enable hearing aid and cochlear implant users to improve speech intelligibility in unknown noisy environments. This challenge is addressed in the current study through large-scale training. Specifically, a deep neural network (DNN) was trained on 10 000 noises to estimate the ideal ratio mask, and then employed to separate sentences from completely new noises (cafeteria and babble) at several signal-to-noise ratios (SNRs). Although the DNN was trained at the fixed SNR of
−
2 dB, testing using hearing-impaired listeners demonstrated that speech intelligibility increased substantially following speech segregation using the novel noises and unmatched SNR conditions of 0 dB and 5 dB. Sentence intelligibility benefit was also observed for normal-hearing listeners in most noisy conditions. The results indicate that DNN-based supervised speech segregation with large-scale training is a very promising approach for generalization to new acoustic environments.