论文部分内容阅读
通过对声音的主观评价与客观分析而建立的主观感受数学模型,在许多领域都有重要的应用.本文采用多元线性回归分析手段对水下噪声音色属性建立回归模型,提取音色特征并改善水下目标的识别效果.首先,在前期水下噪声音色属性主观评价实验的基础上,将构成音色属性空间的5个成分的评价分值作为回归分析中的因变量,提取大量与听觉感知相关的听觉特征作为自变量;然后,通过相关分析和改进的逐步筛选法,挑选出反映音色属性的“最优”自变量子集;最后,利用向后剔除回归分析和水下目标识别实验,确定适当的音色模型,并通过假设检验证明该线性模型不仅正确有效,而且能改善水下目标识别效果.
Subjective perception mathematical model established through subjective evaluation and objective analysis of sound has important applications in many fields.In this paper, a regression model of underwater noise sound attributes is established by using multivariate linear regression analysis to extract the sound color features and improve the underwater Firstly, based on the subjective evaluation experiments of underwater acoustic noise attributes in the previous period, the evaluation scores of the five components that constitute the tonal attributes space are taken as the dependent variables in the regression analysis to extract a large number of auditory-related auditory Feature as an independent variable; then, through the correlation analysis and the improved step-by-step screening method, the subset of “optimal” independent variables that reflect the tonal properties are selected; finally, the regression analysis and underwater target recognition experiments are used to determine Appropriate tone color model, and through hypothesis testing proves that the linear model is not only correct and effective, but also can improve underwater target recognition effect.