声学的

Based on＂The Data Bank System of the Standard Phonetic Features of Different Level in PSC＂, the thesis analyzes the variants of affricative versus fricative in syllables: z, c ,s - zh, ch, sh - j, q, x which are obscure in physiology; analyzes timbre phonetic variants aroused from the different match of seven main simple vowels and initials as well as the tonal patterns with the help of drawing the synthetic spectrograph and pitch spectrograph. So that the data can be offered to describe phonemes in 396 syllables in common use according to the acoustic analysis.

本文在"普通话水平测试各等级标准语言特征数据库系统"的支持下，通过制作396个常用音节语音样本的音高图和综合语图，考察在普通话教学和测试中从生理角度难以准确把握的一组仅有细微差异的塞擦音和擦音声母z、c、s-zh、ch、sh-j、q、x在音节拼合中的变体；考察7个主要单元音在与不同声母拼合构成的音节中音质上的变化模式；考察四个声调的变化形式，力图从语音声学物理特征角度的分析中为普通话常用音节音位语音规范度的描写提供客观的数据说明。

Then the boundary element integral equation of interior and exterior form is deduced in detail, also the form with corner coefficient. The significance for numerical calculation and principle of the singular integral is analyzed, and a non-isoparametric transformation method is presented to calculate weak singular integral and Cauchy integral, the method presented provides us a very simple way to computer the two kinds of singular integral of Helmholtz boundary integral equation, and it is easy to program in computer. After the difficulty of the calculation for multi-frequency of Helmholtz boundary element is explained, a method named SECHIEF (Series Expansion Combined Helmholtz Integral Equation Formulation), which is focused on the computational efficiency, is presented.

对结构声辐射的边界积分方程的内部形式与外部形式进行了详细的推导，给出了角点系数的计算方法与边界积分方程的形式，在此基础上，分析了奇异积分产生的原理及其对数值计算的重要性，提出了一种计算奇异积分的非等参(来源：Ae8a8BC论文网www.abclunwen.com)单元的变换方法，该方法给Helmholtz 声学边界积分方程中的弱奇异积分与Cauchy 奇异积分的计算以及编程提供了极大便利。

Based on this, a new WENO difference scheme which based on Dispersion-Relation-Preserving relation is developed, and representative test cases with this scheme for computational aeroacoustics problems has been implemented and compared in order to test capability of wave capturing; In addition, WENO schemes generally do not converge at high order in the presence of contact discontinuity of Euler equations, so a conservative front tracking technique coupling WENO schemes and Level Set method to simulate the translating density profile is presented here, and numerical simulation with this technique for representative test case has been implemented and results show the desired accuracy.

本文研究了高阶精度加权基本无振荡格式及其在双曲守恒律方程中的应用，在此基础上作了两个方面的工作：一是针对高频声波问题构造出一种基于保色散关系的WENO有限差分格式，并对计算气动声学问题的代表性算例进行了大量数值实验，比较了该格式捕捉波数的能力；另外，针对高阶WENO格式在处理Euler方程的接触间断时精度有所降低的问题，研究了利用界面追踪技术Level Set方法和高阶激波捕捉WENO格式相结合的一种守恒追踪方法，并且给出有代表性的密度滑移面问题的算例，得到一致高阶精度的数值模拟结果。

The word-to-sound rule and the structure of a speech database are modified during the implementation of TTS system, and then a filter algorithm based on part-of-speech is proposed to solve the polyphone problem. Prosodic research on Chinese characters plays a great role in speech recognition and speech synthesis system. In order to improve the naturalness of a speech synthesis system, the relationship among the sentence structure, the length of a sentence and the breath boundaries is analyzed and a breath boundary assigning algorithm based on such relationship is put forward. In addition, the author analyzes the process of a speech synthesis system and builds a prosodic markup language suitable for Chinese speech synthesis.

呼吸群边界划分作为汉语韵律的一种表现方式，它在语音感知和文语转换系统的性能方面起着十分重要的作用，作者在对汉语句法结构、句子长度和和呼吸群划分之间关系的研究基础上，设计并实现了基于句子长度和句法结构的呼吸群边界划分算法；为更好地描述文语转换过程，作者提出了适合汉语语音合成的SSML韵律标注语言，该标注语言不仅从声学层上对语音的韵律进行描述，而且从语言学层上标注句法结构信息，最后本文对基于语言理解的韵律规则用SSML进行了描述并用实验验证了语言学层面的标注对系统自然度的影响。

In order to decrease the error rate of speech recognition, speaker adaptation techniques are often used to adjust speaker-dependent acoustic models. MLLR (Maximum Likelihood Linear Regression) and MAP are two of the most popular techniques in recent years. MLLR uses the technique of regression trees. It calculates the transform matrix for each leaf node of the tree. This makes it possible to use fewer sentences to decrease the error rate of HMM-based speech recognition. However, while we examined the recognition result, we found that although the overall error rate decreased, but the error rate of certain confusable phones was higher. In order to solve this problem, we propose the use MPE (Minimum Phone Error Discriminative Training) to solve this problem.

在语音辨识技术的实际应用上，语者调适技术常被用做调整语者无关之声学模型，使其对於特定语者之辨识率有所提升，常见的语者调适技术有最大化相似度线性回归（Maximum Likelihood Linear Regression），其精神在於透过对语音辨识模型相近的mixture做群聚，再对各个群聚做调整，以期能达到利用少量语料，提升辨识率的目的，但其缺点为，在发音上相近的模型，其mixture原本就极为近似，若再分类为同一群聚做调整，容易因为使用者的发音习惯，而使模型偏向ㄓ或ㄗ，因此虽然整体的辨识率获得提升，但却造成混淆音的错误率上升。

Because of the existence of natural detecting blind area o f general radar on detecting flying object flying in veryˉ low altitude, it is so essential to put forward a suspended bomb t heory which would make use of sound wave's characˉ teristic parameters as its wavelength and spectrum to get the obje ct's direction and position, and to discuss its orientaˉ tion method and embattle model.

由于现有的常规雷达探测超低空飞行的来袭兵器有其固有的盲区，作为雷达的必要补充，利用声音的大小、频谱等特征能够反映声源的方位等特征参数的特性，提出了一种被动声学探测悬浮弹理论，并对其定位算法和布阵模式进行探讨，目的是防御拦截低空超声速巡航导弹。

In the Sixth Chapter, the Wavelet analysis method is introduced in discussing some acoustic imaging problems related to the physical background of the ultrasonic detection. In the light of the solution of some theoretical problem in the 2D Wavelet transmission of two components form, the enhancement and identification of the spotty and fibriform faults of the acoutic image, along with two acoustic image deblurring methods based respectively on the acoustic physical feature analysis and the acoustic image's spatial feature analysis of the object, are realized. Therefore the image degeneration problem resulting from the acoustic beam diffusion or the acoustic focus diffusion in the ultrasonic detection are effectively improved. The software for this purpose has also been developed.

第六章首次应用小波分析的方法，研究了与超声检测的物理背景有关的一些声象处理问题，在解决了二分量形式的二维小波变换中的一些理论问题的基础上，建立起了对声象中点状和线状缺陷目标的增强与分离方法，以及分别基于目标的声学物理特征分析和基于目标声象空间特征分析的二种声象弥散消除方法，实验分析结果表明，这种方法较好地解决了超声检测中由于声束截面或声焦点扩散所造成的成象退化问题。

The tidal river discharge observing system of this thesis gives a new method and experience of the hydrology observing to the river reach and cross section with complicated landform and flow, or busy shipping. This system has achieved 3 innovations: instruments for measurements which use the unmechanized rotor and continue measuring ADCP as the velocity mainframe; methods of velocity measurement which change the vertical velocity measurement to horizontal velocity measurement; the tidal discharge measurement which solve the problem that current method of discharge measurement can not achieve the accuracy of the criterion according to the continuous real time observing over the velocity of flow.

本论文研究的感潮河流流量监测系统，具有测验设备的创新，用可连续测验的非机械转子式的声学多普勒流速剖面仪为测速主机；流速测验方式的创新，改变以往测取垂向流速为水平流速的方式；潮流量测验方式的创新，对流速进行连续实时的监测，解决了用现行测流方法达不到规范要求精度的问题；为地形、水流复杂或是航运频繁的测验河段和断面，提供水文监测的新方法、方式和经验。

Golikov as the general and single-valued criterion for evaluating the acoustics of rooms and auditori- ums have been critically examined,with detailed discussionsconcerning:(1) the empirical foundation of the articulation-Q"functional relationship A= f,(2)the derivation of the"particular criterion"or the optimum re- verberation condition form Q",(3)the meaning of a time integration of the "noise-interference factoras viewed from the theorey of auditory perce- ption...

主要内容有：(1)以在不同的条件(如语言、测试人员、宣读方式以及声学场所等)之下的清晰度测验结果来验证 A=f函数的适用性，(2)判断了从Q&来推求以体积为参数的最佳混响时间的可能性，(3)，从听觉理论来讨论&混响干扰因子&对时间积分的意义，(4)讨论了原作者对&噪声因子&的看法，指出了&噪声因子&定义中的一些值得存疑之处。

The prototype system considers the geometry of the building, the position of the source and receiver, and the acoustic properties of the sound transmission (the sound absorption coefficient of the walls and the temperature of the environment). The system then builds the beam tree by beam tracing, and finds all possible paths of the transmission and reflection. All sound paths are then traced to compute the sound obtained by the receiver.

本系统利用建筑物的几何空间、音源与接收者的资讯，考虑建筑材料的吸音系数、环境的温度等声学相关属性，使用光束追踪法建立光束树，寻找各种可能的穿透与反射路径，建立声音的路径模型，并且计算声音的衰减，处理声音的音响输出，来模拟接收者所听到的声音与回响。

The split between the two groups can hardly be papered over.

这两个团体间的分歧难以掩饰。

This approach not only encourages a greater number of responses, but minimizes the likelihood of stale groupthink.

这种做法不仅鼓励了更多的反应，而且减少跟风的可能性。