NOISE REDUCTION IN MODERN HEARING AIDS – SHORT-TERM MEASUREMENTS USING SPEECH

Smeds K, Bergman N, Nyman T.

ORCA Europe, Stockholm, Sweden.

Modern hearing aids normally incorporate noise reduction algorithms. The input to the hearing aid is analyzed, and the gain is adjusted based on estimates of the signal-to-noise ratio in the various compression channels. There are no standard measurements available that can describe how the noise reduction works, and, as a clinician, it is very difficult to know if noise reduction algorithms from various manufacturers operate in the same way.

Measurements of long-term average gain reductions have shown that modern hearing aids use noise reduction algorithms that vary substantially. Differences are seen in the frequency dependence of the gain reduction, in how the reduction varies with signal-to-noise ratio and presentation level, and in if the gain reduction is dependent on the programmed hearing loss or not. But, do these long-term average measurements give a fair description of what the hearing aid is doing for a speech signal in the short-term perspective?

The current study aims at showing how the noise reduction algorithms of contemporary hearing aids function for real speech in noise.

Coupler gain measurements in an acoustic test chamber were performed using ten modern hearing aids and a real speech signal in stationary noise. Recordings of the input to and output from the hearing aids, with the noise reduction switched on and off, were used to calculate short-term average gain reductions. The measurements show how the gain reductions depend on signal-to-noise ratio and presentation level. The results will be presented both in moving pictures and in static graphs.

The results show that the hearing aid manufacturers have chosen to build their noise reduction algorithms based on completely different principles. Hearing aids that seemed to have similar properties when long-term averages were studied proved to have different effects in the short-time perspective.

The clinical significance of the results will be discussed.

Email: karolina.smeds@orca-eu.info