Abstract
The development of the digits-in-noise (DIN) hearing test has traditionally been a resource-intensive and complex process. This thesis has successfully addressed this challenge by introducing an automated procedure using artificial intelligence (AI) techniques, named Aladdin (Automatic LAnguage-independent Development of the Digits-In-Noise test). The procedure significantly simplifies the creation of valid DIN tests across different languages. It leverages a high-quality text-to-speech (TTS) system and automatic speech recognition system (ASR) to generate and homogenize synthetic speech material. This approach has demonstrated promising results, with synthetic speech achieving comparable intelligibility to natural speech. The use of ASR for level corrections also showed strong correlations with human-derived corrections, indicating its potential for broader application in speech test development.
The research presented in this thesis also highlighted the relevance of extended high frequency (EHF) speech information in speech intelligibility in noise. We discussed the relevance for the creation of speech material for hearing tests, as well as the impact of using different headphones when conducting the tests. We also found that diotic and antiphasic DIN SRTs are independent of presentation level above a minimum level of 60 dB SPL in normal-hearing listeners, which is crucial for practical uses.
The Aladdin project represents a significant advancement in creating DIN tests in a much simpler and standardized way. Future research should focus on expanding the Aladdin tests to more languages and exploring its application to other types of speech tests. Additionally, integrating EHF measurements into clinical practice for specific populations and standardizing the use of synthetic speech in speech tests are important considerations for future studies.
Overall, the Aladdin procedure offers an efficient solution for developing DIN tests, potentially transforming hearing test development practices and improving accessibility to hearing screening globally.
Original language | English |
---|---|
Qualification | PhD |
Awarding Institution |
|
Supervisors/Advisors |
|
Award date | 13 Dec 2024 |
DOIs | |
Publication status | Published - 13 Dec 2024 |
Keywords
- digits-in-noise test
- antiphasic DIN test
- CVC word lists
- extended high frequencies
- hearing loss
- hearing screening
- hearing test