Based on the digital technology, numerous algorhythm has been made recently, in the intention of reducing feedback in hearing aids. These algorhythms can be qualified by the amount of added stable gain. Objective and subjective measurements are both exist, that are used for determine the maximal stable gain, though, thorough knowledge of feedback phenomena is necessary for achieve any effectiveness in these measurements. In my thesis, a list of factors involved in feedback is presented, so as feedback detecting methods found in the literature are.
In a practical approach, I studied the feedback-canceller of ON Semiconductor’s processors in detail, whose operation is based on the modelling of the hearing aid’s feedback path and the reduction of the feedback path signal. The tuning of the model is implemented by a parameter called FBC acoudelay, which affects the efficiency of the feedback-canceller significantly. Therefore, the knowledge of the optimal value for each hearing aid type is strongly desired. To determine this optimal value, I studied the effect of the parameter on the hearing aids set to unstable and marginally stable state. According to my observations, the time-decay of the response signals provided relevant information about the effectiveness of the feedback-canceller. For that reason, I was concerned with investigating these features.
As a result of my work, a method for optimizing the feedback-canceller – even with the simplest tools – had been successfully created.