自适应反馈取消的R3920的节奏半导体

　　本资料说明了反馈取消在最新数字助听器中提供的功能对节奏的®R3920安森美半导体。反馈取消不同于其他反馈减少策略，因为它不减少前进正常工作条件下助听器的增益。相反，该算法减少反馈通过制定

　　真助听器反馈信号的内部估计然后减去这个估计从麦克风信号。如果内部估计是准确的，那么实际反馈被取消，助听器不会尖叫。正确估计真正的助听器反馈信号是必不可少的正确操作算法。由于反馈信号随时间而变化算法必须对内部进行连续的调整模型，以确保外部反馈准确匹配。在半导体上的反馈消除器中，这些调整自动发生，不需要探针信号或其他用户干预可能干扰助听器的操作。最新的反馈消除算法安森美半导体新机制增加保持音频质量的适应速度。这机制自动决定是否当前反馈条件需要大或小的调整内反馈−路径模型。大调整用于快速消除声反馈。当大不需要调整，选择较小的调整最大化音频质量。下面的部分描述反馈消除操作更详细地强调影响反馈的因素检测器的性能。文章的结论与可用于优化反馈的方法的讨论检测器的性能。这些方法允许用户关于半导体数字放大器的增加产品性能。

　　PRINCIPLE OF OPERATION Adaptive Feedback Cancellation （AFC）， from ON Semiconductor， is a true feedback cancellation （also known as phase cancellation） algorithm. The algorithm operates by subtracting an internal estimate of the hearing-aid feedback signal from the microphone signal. The internal feedback signal is obtained by passing the hearing-aid output signal through an internal， feedback-path model. The feedback-path model in AFC is intended to match the impulse response， or time response， of the external feedback path. The feedback-path impulse response is the time signal that would be observed at the microphone output if an impulsive signal were produced by the hearing aid. As an example， the measured impulse response of a typical BTE hearing aid is shown in Figure 1−a. This plot shows an initial period of approximately 0.5 − 0.75 ms with little or no energy. This time window corresponds to the analog-to-digital and digital-to-analog converter delays as well as any acoustic delays due to transducers， tubing， sound propagation through the air， etc. Following the initial delay is the main part of the feedback-path response which， in this example， persists for more than 2 ms. The net effect of the feedback canceller algorithm is to remove a certain portion of the feedback path impulse response. Since computational resources inside a hearing-aid signal processor are limited， only a portion of the total feedback path response can be cancelled. To achieve maximum effectiveness of the canceller， the cancelled part of the response should be that portion of the feedback path with the largest energy. This ensures that the net response has the lowest energy for the given length of canceller. For example， if the cancellation window is restricted to 1ms， then for the feedback path of Figure 1−a， the internal feedback model should look like the plot of Figure 1−b. When this is subtracted from the feedback−path response， the net feedback path is as shown in Figure 1−c. The corresponding frequency response is shown in Figure 2. This figure illustrates that the net effect of the feedback canceller in this example is to reduce the peak response of the feedback path by approximately 12 dB. Consequently， the feedback canceller in this example should provide approximately 12 dB of added stable gain （ASG） that is achieved without any impact on the forward gain of the hearing aid.