Exploration of a New Active Noise Cancellation Headset Based on Spectral Contrast Enhancement Technique

Abstract

The purpose of this project is to provide a new and more cost-effective Active Noise Cancellation headset for hearing impaired patients for improving speech recognition. This project mainly adopts the Active Noise Cancellation technique based on the LMS algorithm and the spectral contrast enhancement technique of sound. After comparing the commercially available Active Noise Cancellation techniques and spectral contrast enhancement techniques, the algorithms in this project were selected according to the needs and limitations of realistic signal processing components. Also, SSCE is used as the base algorithm of the spectral contrast technique, and FxLMS and NLMS are used as the base algorithm of the Active Noise Cancellation technique, respectively, and improvements and optimizations are made on this basis. Finally, this topic is based on the MATLAB SIMULINKF simulation platform to design computer simulation experiments, and the performance of the algorithm is measured empirically. The results show that Active Noise Cancellation can effectively reduce the random noise of non-human voices in the audio, enhance the signal-to-noise ratio of the original audio, and provide a basis for the signal enhancement of spectral contrast enhancement techniques in specific human voice bands when tested offline. However, the Active Noise Cancellation system or spectral enhancement techniques both will cause a certain distortion of the audio signal. The distortion of the signal can only be mitigated to a certain extent by the modeling of the propagation channel and can not be well improved.