基于听觉感知特性的双耳音频处理技术*

自20世纪30年代引入立体声以来,人类对逼真的听觉体验一直进行着孜孜不倦的追求。双耳音频处理技术基于人耳听觉感知特性,利用计算机和数字信号处理等技术在听者双耳鼓膜处模拟出与真实场景相同的声压,以期给人以“身临其境”的体验,一直是音频信号处理领域的重要研究内容,特别是近两年随着虚拟现实等应用的蓬勃发展,得到更多关注。该文主要围绕双耳音频处理技术中所涉及的关键环节:双耳录音、双耳合成、耳机重放、扬声器重放、头跟踪等领域,以及相关典型应用场景进行较为系统的介绍,最后给出总结与展望。     

Efficient equalization of multi-exciter distributed mode loudspeakers

An efficient digital equalization method is applied successfully to the problem of spectral equalization of multi-exciter distributed mode loudspeakers (DML). It is based on a chain of second-order sections of infinite impulse response parametric filters with very low computational cost. The method compensates for the measured multi-exciter DML response in order to achieve a desired frequency response. The sound radiation of these flat loudspeakers is a complex superposition of excited modes that vary strongly with frequency. Therefore, the characteristic multi-exciter DML spectrum is very irregular and is equalized with the method presented here for a natural, uncolored response. In multichannel systems, such as wave field synthesis (WFS), the use of efficient filters to equalize a large amount of drivers is an advantageous approach. The equalization process has been applied to two multi-exciter DML prototypes, comprising three and five exciters per panel. Both panel and exciter equalization have been addressed, which consequences on the filtered responses are discussed. Finally, some subjective assessments are carried out to optimize the order of the filter while maintaining the perceived quality of the equalization.     

直接辐射式扬声器系统的低频效率计算公式*

效率是扬声器系统的重要性能指标,然而常用的标称效率在低频范围并不适用。本文研究直接辐射式扬声器系统的低频效率,推导了基于TS参数的低频效率计算公式。针对封闭式扬声器系统和倒相式扬声器系统,开展实验验证,结果表明本文提出的计算公式得到的低频效率与实验测量得到的低频效率较为相符。进一步对比分析低频效率和标称效率,结果表明低频效率能够适用于低频,能准确地展示扬声器系统参数对效率的影响。利用本文提出的计算公式,有助于进一步优化提升扬声器系统的低频效率。     

Transient modal radiation of axisymmetric sources: Application to loudspeakers

The present study proposes a transient model of the acoustic radiation of axisymmetric structures. The pressure field is approximated by the Rayleigh integral corresponding to a monopole source distribution over the non-planar vibrating surface. The displacement field is expanded on the linear modes of the structure and a change of variables in the Rayleigh integral is then proposed in the case of a monotonic profile function to compute the Spatial Impulse Response associated to each mode of vibration efficiently. The results are compared to the formulation obtained in the case of planar and spherical sources. The method of calculation is then derived in the case of a typical loudspeaker profile (association of a truncated cone with a spherical cap). Finally, the present approach is used to estimate the nonlinear radiation pattern of a prototype loudspeaker and predictions are compared to measurements in anechoic room.     

Strategies for bass enhancement in Multiactuator Panels for Wave Field Synthesis

In this paper, an extension of the useful bandwidth in the low frequency band is applied to compensate the inherent poor bass response of Multiactuator Panels. These are a special type of flat panel loudspeakers that are commonly used to reproduce spatial audio under the Wave Field Synthesis system. The proposed algorithm combines two strategies: first, a dynamic electrical equalization, applied to additional exciters, which are carefully positioned to excite the lower frequency modes of the panel; second, a psycho-acoustical approach, taking profit of the behavior of human hearing based on the missing fundamental principle. For comparison purposes, the shelving and peak equalizations are also applied to the MAP prototypes developed in this paper. Objective and subjective results show that the combined approach results in an effective extension in the low frequency end of MAP with very low levels of distortion, outperforming conventional equalization methods, with better low frequency behavior, and hence better audio quality perception.     

基于最小二乘法的均匀扩声技术研究

针对目前室外及超大型室内空间的扩声应用中声场分布不均匀的问题,本文提出了一种基于最小二乘法声场重建方法的扩声技术。该技术通过对目标声场的逼近来计算线性扬声器阵列各通道的输入参数,实现扩声区域内声压级的均匀分布,同时约束非扩声区域的声能量以获得较好指向性。本文通过仿真研究算法各参数的改变对控制结果的影响,探讨不同扩声区域和目标所对应参数的选取方法。仿真和实验比较了该算法以及未经控制的声场、相移法波束控制的效果,证明该方法可以获得更好的声场均匀度。     

优化扬声器阵列的指向性与声功率级

提出了同时优化扬声器阵列的指向性和声功率级的方法。运用该方法可以优化扬声器阵列的输入电压分布,获得指向性和声功率级二者兼顾的扬声器阵列。本方法不必使用复杂的与频率相关的信号处理装置。模拟结果和实验测量结果表明该方法是可行的和有效的。     

使用噪声整形的多比特数字扬声器系统

本文讨论了ＰＣＭ数字式换能器的特点，采用了过采样和噪声整形技术对比特数字扬声器系统进行改进，给出了计算机模拟结果。     

扬声器中涡流的热效应

通过分析和比较两种扬声器涡流电学模型和两种扬声器热模型，提出一种分频段的非线性热模型类比线路，给出了对应的音圈稳态温度公式，并在高频段运用逐点测温的方法研究，论证了扬声器涡流随频率变化的规律及其对扬声器热效应的影响。