Inspiration for our work sometimes comes from everyday conversation. I can remember a key discussion with someone unconnected to hearing aids or audiology. I explained why and how we use speech intelligibility tests to evaluate our hearing aids. Their spontaneous and not so naïve reaction was that speech intelligibility is difficult to measure because it depends on who is talking.
I could associate this comment to feedback from hearing aid users about their experiences when watching television. For the news broadcast, as an example, some hearing aid users say that they understand the main presenter quite well but they have difficulties when different people are randomly interviewed. This experience combines two effects: the disadvantage of unforeseen multi-talkers and the advantage of knowing the talker.
Effects on speech intelligibility
Research has studied these effects and shows that speech intelligibility is reduced when different talkers are used (Kirk, 1997) or when the talker is not known (Souza, 2013). But how do these effects impact the measurement and the evaluation of the benefit of amplification with hearing aids? Is the use of a speech test with a single talker adequate to cover a broad range of listening situations?
Speech intelligibility tests
We can use speech intelligibility tests to compare the performance of different hearing aids, algorithms, or settings. However, when the test conditions are not challenging enough, then we will have a ceiling effect. That means we might not have a chance to find differences that might exist, because all the test results are too good on average. In this case, we have no difference in the test results but we cannot conclude that the tested conditions are equivalent. This situation is often encountered in clinical practice, where hearing aid users report subjective differences between hearing aids but the differences cannot be measured with a standard speech intelligibility test.
Multi-talker speech intelligibility tests
Using multi-talker speech material for the evaluation of different solutions might create more challenging test conditions, especially for hearing-impaired listeners. It will also better reflect daily life listening conditions. So, we developed multi-talker test material to recreate more challenging and realistic test conditions for the evaluation of our products.
Evaluation results
We evaluated the multi-talker effect on the benefit of amplification and found interesting results: the measured benefit of amplification depends on the test material. The benefit of amplification is reduced when switching the speech material from a single talker to a multi-talker. This change should increase the chance of finding useful results when comparing different test conditions.
Interested to learn more? Our research was just published in the latest edition of the journal Speech Communication. You can find our article here.
The selection of a speech test should ensure that it has intrinsic qualities such as objectivity, reliability, and precision, but it should also reflect different aspects of daily life. It also raises the question of generalizing the results or how the results will be translated into daily life of the hearing aid user. These questions might not have a simple answer, but we hope that our article will generate some thoughts and questions for our daily practice.
References
Kirk, K. I., Pisoni, D. B., & Miyamoto, R. C. (1997). Effects of Stimulus Variability on Speech Perception in Listeners with Hearing Impairment. Journal of Speech Language and Hearing Research, 40(6), 1395.
Souza, P., Gehani, N., Wright, R., & McCloy, D. (2013). The Advantage of Knowing the Talker. Journal of the American Academy of Audiology, 24(8), 689–700.