Counteracting Acoustic Disturbances in Human Speech Communication

Document type: Dissertations
Full text:
Author(s): Nils Westerlund
Title: Counteracting Acoustic Disturbances in Human Speech Communication
Series: Blekinge Institute of Technology Doctoral Dissertation Series
Year: 2006
Issue: 2
ISBN: 91-7295-085-4
ISSN: 1653-2090
Publisher: Blekinge Institute of Technology
City: Karlskrona
Organization: Blekinge Institute of Technology
Department: School of Engineering - Dept. of Signal Processing (Sektionen för teknik – avd. för signalbehandling)
School of Engineering S- 372 25 Ronneby
+46 455 38 50 00
http://www.tek.bth.se/
Authors e-mail: nwe@bth.se
Language: English
Abstract: A signal can be said to be any information bearing unit or action carrying a message from a sender to a receiver. This definition covers a vast number of human and non-human actions, ranging from flirtation to satellite communication. This thesis deals with increasing the quality of one of the most ubiquitous human-to-human signals: Speech. Surrounding noise is a severe obstacle to relaxed speech communication. Cars, industry and many everyday machines emit high noise levels that render personal communication difficult, degrade our mental and physical ability and may cause nausea, vertigo, fatigue and temporary or permanent hearing loss. Indeed, exposure to noise of a high sound pressure level is a major contributor to often irreversible sensorineural hearing impairment, i.e. Noise Induced Hearing Loss (NIHL).

This doctoral thesis deals with three different approaches for facilitating human speech communication. First, methods for adaptively controlling acoustic feedback — commonly denoted "howling" — in hearing aids are developed and subsequently evaluated. Howling is a very common problem in hearing aids and a stable and robust feedback eliminator would serve many hearing aid users. The proposed method detects tonal components, i.e. howling, in the signal path of the hearing aid. If such a component is detected, the feedback control system invokes different countermeasures to adaptively cancel the howling. Second, a method for speech quality and intelligibility enhancement is described. Instead of focusing on noise suppression, the method acts as a speech booster: Frequency bands containing a usable amount of speech energy are boosted. All other frequency bands remain unchanged. This results in an increased Signal-to-Noise Ratio (SNR) and the elimination of background artefacts which may be present in some other noise reduction algorithms. Furthermore, speech distortion is kept to a negligible level. Finally, a compact receiver unit, based on bone conduction, is designed and evaluated. The unit is placed inside the external auditory canal of a user and picks up bone conducted speech from the user's speech organ. This solution, in combination with a pair of active hearing protectors, yields several advantages and allows the user to reliably communicate in environments where extremely high sound pressure levels are present.
Subject: Signal Processing\Speech Enhancement
Keywords: Speech Enhancement, Noise Reduction, Hearing Aids, Adaptive Feedback Control, Ear-Mic, Active Noise Control
Note: Also published at: http://www.its.bth.se/staff/nwe/thesis/thesis.pdf
URN: urn:nbn:se:bth-00311
Edit