Browse Prior Art Database

Automatic Testing Platform for Hearing Aids Using Speaker Recognition and Speech Recognition

IP.com Disclosure Number: IPCOM000198713D
Original Publication Date: 2010-Aug-23
Included in the Prior Art Database: 2010-Aug-23
Document File: 3 page(s) / 473K

Publishing Venue

Siemens

Related People

Juergen Carstens: CONTACT

Abstract

Hearing aids have to be tuned and customized to the specific requirements of each individual. This has to be done after the analysis of the specific audiogram for that particular user. The testing of the hearing aids concerning the individual requirements after the fabrication poses several problems. The testing is mostly manual, cumbersome and often trained professionals are required. Thereby, an obvious disadvantage of the testing process is the subjectivity of the personnel. Currently there are several methods known for testing hearing aids, for example using voice detection.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 52% of the total text.

Page 1 of 3

Automatic Testing Platform for Hearing Aids Using Speaker Recognition and Speech Recognition

Idea: Antony Louis Piriyakumar Douglas, Ph.D., IN-Bangalore; Vishnu Swaminathan, Ph.D., IN- Bangalore; Ramasubramanian Viswanathan, Ph.D., IN-Bangalore

Hearing aids have to be tuned and customized to the specific requirements of each individual. This has to be done after the analysis of the specific audiogram for that particular user. The testing of the hearing aids concerning the individual requirements after the fabrication poses several problems. The testing is mostly manual, cumbersome and often trained professionals are required. Thereby, an obvious disadvantage of the testing process is the subjectivity of the personnel. Currently there are several methods known for testing hearing aids, for example using voice detection.

In the following an automated testing platform for hearing aids is proposed, using speaker recognition and speech recognition for evaluating the performance of hearing aids. For this purpose, the proposed testing platform compares automatically five signal chains among each other. A scheme of the proposed platform is shown in Figure 1. The five signal chains, defined as A, B, C, D and E (Figure 2), all contain the same acoustic input signal, in form of different speaker voices for the speaker recognition, different texts of same speakers for the speech recognition or music for the tone recognition. The signal chain A represents the original input signal, taken from an acoustic database, which feeds directly to the recognition system and is used as a reference value for the comparison. The signals B and C represent the input signal, as it is perceived by a person with a specific degree of hearing loss without (signal B) and with (signal C) the use of a hearing aid. The degree of hearing loss is determined by so called audiograms, taken from an acoustic database (Figure 3). The signal chain D corresponds to the signal perce...