8th Speech in Noise Workshop, 7-8 January 2016, Groningen

Cross-talk cancellation and its potential benefits to speech understanding in background noise to patients with bilateral bone-anchored hearing aids

Robert W J Mcleod(a)
Cardiff University, GB

John Culling(b)
Cardiff University, GB

(a) Presenting
(b) Attending

Bilateral bone-anchored hearing aids (BAHAs) suffer from cross-talk within the skull, but two BAHAs could be utilised to create a cross-talk cancellation system. In order to achieve this, sound vibrations from each BAHA reaching the contralateral cochlea must be cancelled by an out-of-phase signal of the same level from the ipsilateral BAHA. We have developed a psychoacoustic method for calculating the cancellation level and phase needed for cross-talk cancellation, shown how these values can be employed to increase speech understand in background noise and demonstrated the potential benefits to patients. Participants with normal hearing wore two B71 bone transducers (BT) (one on each mastoid) and bilateral ER2 earphones. Both BTs were stimulated with the same pure tone and the level and phase adjusted in the right BT in order to cancel all perceived sound at the right ear. Participants could reliably pinpoint the level and phase necessary for cancellation between 1500 and 8000 Hz. Speech reception thresholds (SRTs) were obtained with and without cross-talk cancellation based on these measurements, which were used to modify the noise signal and add it to the speech signal presented to the right BT. The noise from the left BT was cancelled at the right cochlear by the modified noise signal on that side, resulting in a mean benefit of 15.6 dB in SRT. In order to get an indication of real-world benefit, head related impulse responses (HRIR) at 0° and 90° from a BAHA 4 were used to simulate a simple listening situation that was then presented either using earphones or using the BTs. The difference between the two gives an indication of the deleterious effect of cross-talk. SRTs were measured with speech at 0°and noise at 0° or 90°. Because the cross-talk cancellation only addresses the distortion of ILDs by cross-talk, the relative spatial unmasking benefit from ILDs and ITDs was measured; the level and phase components of the HRIR from the 90° and 0° were exchanged to produce ILD-only and ITD-only conditions. Spatial release from masking due to ILDs recorded from BAHA microphones was 4.4 dB better with the ER2s, than with the BTs, indicating the maximum potential benefit of cross-talk cancellation.

Last modified 2016-05-12 14:22:09