微型扬声器异音信号产生机理及分析方法研究

发布时间：2018-01-28 00:14

本文关键词： 微型扬声器短时傅里叶变换异音产生机理共振频率　出处：《天津科技大学》2015年硕士论文　论文类型：学位论文

【摘要】：纯音检测是微型扬声器产品质量检测的重要环节——检测微型扬声器在额定功率下,采用正弦扫频激励时,扬声器的响应信号是否存在异常。目前,在电声行业,用人耳检听方法检测扬声器异音故障仍然占主导地位。由于人工检听的方法具有一定的主观性,而且长期检听还会伤害听音员的听觉神经。为此,随着工业化进程的推进,微型扬声器异音故障的自动检测已成为电声行业亟待解决的问题。结合微型扬声器自身特点,本文对微型扬声器异音信号产生机理做出进一步分析和研究。在对扬声器响应信号做时频分析的基础上,对扬声器共振频率附近区域和全局高阶能量进行统计,并以此区分扬声器的好坏。为了实现扬声器异音故障检测,运用扫频仪发出的正弦扫频信号激励微型扬声器,经由传声器将声压信号转变为电信号,数据采集卡将模拟电信号转变为数字量并送入计算机处理。计算机对采集到的信号做短时傅里叶变换,获得信号的时间—频率联合分布,提取特征并进行数理统计,根据合格扬声器和故障扬声器的特征分布,确定阈值。本文的主要研究成果包括：1.给出微型扬声器异音故障的产生原因并对产生机理进行了研究；2.结合时频图中的能量分布状况,提出了基于共振频率附近区域高阶能量的异音检测方法；3.给出了基于短时傅里叶变换的扬声器共振频率的计算；4.完成了检测系统软硬件模块的集成。实验表明,基于共振频率附近区域高阶能量的异音检测方法能够较好的识别微型扬声器的好坏,在短时傅里叶变换的基础上进行共振频率求解的方法可行。
[Abstract]:Pure tone detection is an important link in the product quality detection of micro loudspeakers. At present, the response signal of the loudspeaker is abnormal when it is excited by sine sweep frequency under rated power. In the electro acoustic industry, the method of human ear test is still dominant in detecting the abnormal sound of loudspeakers. Because of the subjective nature of the method of manual audiometry, and the long-term detection of hearing will also harm the auditory nerve of the listener. For this reason. With the advancement of industrialization, the automatic detection of abnormal sound faults of micro loudspeakers has become an urgent problem to be solved in the electroacoustic industry, combining with the characteristics of micro loudspeakers. Based on the time-frequency analysis of the loudspeaker response signal, the region near the loudspeaker resonance frequency and the global high-order energy are analyzed. In order to detect the abnormal sound of loudspeaker, the sinusoidal sweep signal is used to excite the miniature loudspeaker, and the sound pressure signal is transformed into electric signal through microphone. The data acquisition card converts analog signals into digital signals and sends them to the computer for processing. The computer performs short time Fourier transform on the collected signals to obtain the time-frequency joint distribution of the signals. Feature extraction and mathematical statistics, according to qualified loudspeaker and fault loudspeaker feature distribution. The main research results in this paper include: 1. The causes of abnormal sound faults of micro loudspeakers are given and the mechanism of producing them is studied. 2. Considering the energy distribution in the time-frequency map, a new method based on the high-order energy near the resonance frequency is proposed. 3. The calculation of loudspeaker resonance frequency based on short time Fourier transform is given. 4. The software and hardware modules of the detection system are integrated. The experiment shows that the method based on the high order energy near the resonance frequency can better identify the micro loudspeaker. The method of solving resonance frequency based on short time Fourier transform is feasible.
【学位授予单位】：天津科技大学
【学位级别】：硕士
【学位授予年份】：2015
【分类号】：TN643

【相似文献】