古琴的声功率级测试研究

古琴是中华文化中地位崇高的乐器,位列中国传统文化四艺"琴棋书画"之首,但目前古琴声功率级还未有过科学的量化测定。本文将古琴声音解构成散音、泛音、按音三种音色,从其声学结构中的激励系统的角度出发,研究并测试古琴在室内音乐厅内用不同力度的激励方式演奏单音、音阶和经典乐曲时的声功率级、动态范围其及1/3倍频带声功率级谱。测试结果通过数字化图示研究分析,对完善我国民族乐器的基础声学数据库有积极作用,并为进一步研究民族乐器发音特性、民族乐队编制以及民族音乐厅堂音质设计提供重要参考。

Analysis and optimization of power spectrum on EBPSK modulation in throughput-efficient wireless system

In order to satisfy increasingly greater demand for the performance of communication systems, a throughput efficient wireless system based on the extended binary phase shift keying （EBPSK） modulation is presented. Simultaneously, corresponding analysis of power spectra is also given with a brief process. The optimal waveform is proposed without useful information loss, by removing linear spectra presenting periodic components. On this basis, the reasonable definition of bandwidth is discussed, which indicates that the EBPSK belongs to the category of the ultra narrow band （UNB） throughput-efficient communication. Meanwhile, the modulation parameters＇ effects on bandwidth, transmission rate and transmission performance are analyzed. Results illustrate the validity of theoretical analysis and spectrum optimization. Results also prove that this UNB system can obtain good bit error rate （BER） performance with high spectra efficiency.

Feature Extraction of Acoustic/Seismic Signal of Target

A microphone and a seismic sensor always become a basic unit of UGS(unattended ground sensors) system. The mechanism of acoustic and seismic property of target and its propagation are described. The acoustic and seismic signals of targets are analyzed with time frequency distribution according to its non stationary property. Narrow band energy function (NEF) and local power spectral density (LPSD) are proposed to extract features for target identification. Experiment results show that local power spectral density indicates corresponding target clearly.

基于改进EWT的超高层建筑模态参数识别

文章基于改进经验小波变换(empirical wavelet transform,EWT)、自然激励技术(natural excitation technique,NExT)及归一化的希尔伯特变换(normalized Hilbert transform,NHT),发展一种超高层建筑的模态参数识别方法。该方法首先使用基于Burg算法的自回归功率谱替代傅里叶频谱划分频谱区间,改进傅里叶频谱分割过程,以此构造经验小波滤波器组,将结构响应信号自适应地分解为一系列固有模态函数(intrinsic mode function,IMF),再利用NExT得到单分量信号的自由衰减响应,最后通过NHT和曲线拟合识别结构的阻尼比和自振频率。利用该方法对数值模型和台风“海棠”影响下台北101大楼的结构响应进行了模态参数识别分析。研究结果表明,文中提出的模态参数识别方法用于超高层建筑模态参数识别具有有效性、准确性和适用性,该方法可以利用非线性、非平稳、动态响应小且噪声水平高的结构响应信号,准确地估计超高层建筑的固有频率和阻尼比。

基于带通滤波片的激光二极管自混频纳米颗粒干涉测量技术

激光二极管自混频技术可用于纳米颗粒的测量,采用多通道带通滤波技术分别测量了180,100,60nm三种纳米颗粒,获得自混频信号的频带功率谱(PS-BPF)。作为比较,对实时提取的自混频信号进行傅里叶变换得到功率谱密度(PSD)函数。分别对PSD函数和PS-BPF反演计算得到颗粒粒径分布。结果显示:两种方法得到的颗粒粒径分布基本一致,并与商用仪器测试结果吻合。对两种处理方式的模式矩阵分析表明,多通道带通滤波技术明显优于PSD函数处理方式。