谐波复合音畸变产物产生模型-中国医学物理学杂志.pdf

doi: 10.3969/j.issn. 1005-202X.2015.05.006 －639 － 谐波复合音畸变产物产生模型 1 2 2 1 王 健 ，关 添 ，吴默村 ，杨文霞 1. 深圳信息职业技术学院电子与通信学院，广东 深圳 518172 ；2. 清华大学深圳研究生院生物医学工程研究中心，广东 深圳 518055 【摘 要】由于人耳具有压缩非线性，声音进入人耳后会产生畸变产物。为了深入地研究畸变产物对听觉音高感知的 影响，本文建立了一个畸变产物产生模型，该模型产生的畸变产物频率低于谐波复合音主成分的频率。模型的实现过程 包括：计算谐波复合音的不同次幂、低通滤波、使用模拟退火遗传算法搜索最佳权重、加权求和。使用兴奋模式验证模型 的正确性，观察兴奋模式上的突出峰，计算峰谷差，匹配对应的畸变产物频率。结果发现畸变产物产生模型能够产生正 确的畸变产物，畸变产物能够增强谐波复合音的可分离性，干扰不可分离谐波复合音的音高感知。本文提出在听觉音高 感知实验中，为了减小畸变产物对听觉音高感知的影响，必须使用背景噪声掩蔽畸变产物。 【关键词】谐波复合音；音高感知；畸变产物；模拟退火遗传算法 【中图分类号】R318.18 【文献标识码】A 【文章编号】1005-202X （2015 ）05-0639-04 Model of generating distortion products of harmonic complex tone 1 2 2 1 WANG Jian , GUAN Tian , WU Mo-cun , YANG Wen-xia 1. School of Electronics and Communication, Shenzhen Institute of Information Technology, Shenzhen 518172, China; 2. Research Center of Biomedical Engineering, Graduate School at Shenzhen, Tsinghua University, Shenzhen 518055, China Abstract: Because of the compression nonlinearity of human ears, distortion products (DPs) are generated after sounds stimulate the cochlear. A model of generating DPs was established in this paper to further study on the impacts of DPs on auditory pitch perception. The frequencies of DPs were lower than those of primary components of the harmonic complex tone. The implementations of the model included calculating different powers of the harmonic complex tone, low- pass filtering, searching the optimal weight by the simulated annealing genetic algorithm, weighting and summing the input signal. The excitation patterns were applied to verify the accuracy of this model. Salient peaks in excitation patterns were observed, and peak-to-