 |
DSP Methods for LD-DPS system
|
|
DSP Methods for LD-DPS system
|
The Kalman Filter based real-time method can measure contra-lateral DPOAE suppression dynamics directly in time domain[1]. Its effectiveness was validated on animal models and human subjects[2,3]. However, the speed when testing unquiet babies and the detectability when testing adults or seniors remain challenging[4], because of the degraded Signal-to-Noise Ratio (SNR). The purpose of this research is to improve the tests on these subjects. A new signal processing algorithms based on Adaptive ARMA and time-variant state equation modeling have been developed, and specifically designed probes as well as an amplifier system optimized for the algorithms are developed. Experiments on simulated and real recordings showed greatly improved testing speed and accuracy at low SNR. For a normal hearing adult, a clear DPOAE suppression dynamic curve can be observed in less than 10 seconds, making it a potential tool for assessing hearing system integrity, both sensorial and neural , through a quick acoustic-only test.
Contra-lateral DPOAE Suppression refers to the DPOAE level reduction when a contra-lateral stimuli (suppressor) is presented during DPOAE test. It is mediated by the medial olivocochlear system (MOCS), and can be used to assess the integrity of the cochlea as well as the neural pathway (MOSC).
In principle the DPOAE Suppression can be measured in real-time with narrowband filters. The issues are the conflict requirements for choosing the filters - DPOAE Suppression is a very weak signal buried in noise , to improve the SNR , the bandwidth should be chosen as narrow as possible; but to quickly and accurately track DPOAE Suppression changes , the filter must have a wide enough bandwidth.
The original Kalman filter based method 1 addresses the issues by adaptively adjust the filter bandwidth-when the algorithm detects a sudden change that does not fit the signal model, it temporally widens the bandwidth to quickly track the change, as soon as it catches the change it narrows the bandwidth to measure the steady DPOAE level more accurately. The method works well with high/moderate SNR, but performance degrades significantly in low SNR conditions .
The new method addresses the issues in a different approach - instead of solely relying on narrowing filter's bandwidth to reduce the DPOAE recording noise, it uses the auxiliary sensor channels to estimate the noise components in the main DPOAE recording channel and subtract it from the main channel. The channel data such processed have a higher SNR, and allow wider bandwidth filter to be used and thus better tracking performance.
The new method has significantly improved the signal processing performance at low SNR, and provides a robust , fast and direct means to view and measure DPOAE suppression dynamics.
1.9.6