语音编码技术及其硬件实现

压缩编码技术是无线语音通信的关键技术之一。介绍了语音编码技术的基本概念及分类,并选用AMBE多带激励压缩编码算法,通过单片机控制专用语音压缩DSP芯片,提出了一种适合低速无线语音、数据实时通信的系统解决方案,完成了硬件、软件设计,实现了低速率下的语音、数据的同步大气传输。测试结果显示,在语音编码速率为2．4kbps以下时,仍然可以得到音质较好的语音输出。     

基于听觉感知的语音稀疏表示及压缩感知*

本文针对语音信号稀疏表示及压缩感知问题,将听觉感知引入稀疏系数筛选过程,用掩蔽阈值筛选重要系数,以得到更符合听觉感受的语音稀疏表示。通过对一帧浊音信号分别采用掩蔽阈值和能量阈值方法进行系数筛选对比实验,结果表明掩蔽阈值法具有更好的稀疏表示效果。为验证听觉感知对语音压缩感知性能的影响,与能量阈值法对照对测试语音进行压缩感知观测和重构,通过压缩比、信噪比、主观平均意见分等主客观指标评价其性能,结果表明,掩蔽阈值法可有效地提高压缩比且保证重构语音具有较高的主观听觉质量。     

ODEON仿真与实际测量语言传输指数实验对比

近年来音质仿真技术的快速发展为语言传输指数STI的预测提供了一个潜在的解决方案。但这种方法的有效性如何,则是在使用该技术之前应该考虑的问题。本文对3个房间内音质仿真与实际测量STI进行实验对比,研究表明:在仿真模型与实际空间的声学等效较为准确的情况下,使用音质仿真软件ODEON计算得到的STI误差较小;混响时间的变化在背景噪声较高时可能会对仿真STI的准确性带来显著影响,随着混响时间的增加,仿真与实际测量STI的差值可能变大;信噪比的变化并不会给仿真STI的准确性带来显著影响;仿真脉冲响应与实际测量脉冲响应的频谱有一定差别,时域上的反射声序列也不相同,但这些差别对仿真STI的影响并不大;仿真过程中比较容易产生的信噪比误差对仿真STI产生了较大影响。由于影响音质仿真结果的因素较多,仿真模型与实际空间的声学等效也比较复杂,尤其是对于没有实际参照校准的房间来说,想要获得准确的STI预测结果是比较困难的。     

美国物理化学家朗缪尔在中日2国讲学概况

欧文·朗缪尔是1932年诺贝尔化学奖得主。1934年,朗缪尔受邀前往日本和中国北平学术演讲,与我国物理学家、化学家进行座谈。他的精彩演讲给我国学者留下了深刻的印象。依据朗缪尔文集等文献和当时报刊的报道,力图重现朗缪尔中日讲学之盛况。     

Impact of Statistical Multiplexing on Voice Quality in Cellular Networks

This paper examines the quality of transmission of voice over cellular, packet-switched networks. The medium access mechanism in the uplink is simulated under various statistical multiplexing scenarios in order to assess the effect of front-end clipping on voice quality. Moreover, the simulation is implemented in a real-time demonstration platform utilized to acquire subjective indicators of voice quality by performing Mean Opinion Score (MOS) tests. Results from the MOS tests are reported, and an analysis of the obtained speech samples is presented. Finally, the results are summarized and potential further directions for the simulation tool and the speech models are discussed.     

M-Mode Color Doppler Ultrasonic Imaging of Vertical Tongue Movement During Articulatory Movement

To observe and estimate the movement of the tongue, ultrasonic investigation is the most harmless real-time monitoring procedure for analyzing articulatory movements. Color Doppler ultrasonic imaging is special in that it can only sample a moving target, and it can indicate the velocity and direction of the target by color and brightness in real time. This study assessed and demonstrated the validity of M-mode color Doppler ultrasonic imaging to observe the movements of the tongue during syllable repetition tasks performed by normal subjects and dysarthric patients, those affected by amyotrophic lateral sclerosis, cerebellar ataxia, Parkinsonism, and polymyopathy. When the transducer was set below the jaw, upward movement was indicated by a blue signal and downward movement was indicated by a red one on the screen of the ultrasound machine. We also measured the velocity of the tongue by contrast scale classified by 15 degrees. Thus, we could observe vertical tongue movement by a color-coded pattern after quantitative analysis. The Doppler signal patterns of normal subjects were verified by simultaneous video x-ray fluorography recordings. The findings for dysarthric patients corresponded well with previously reported features analyzed by other methods. Therefore, color Doppler ultrasonic imaging of the tongue is a useful procedure to researchers for clinical speech and voice studies.     

汉语耳语音孤立字识别研究

耳语音识别有着广泛的应用前景,是一个全新的课题.但是由于耳语音本身的特点,如声级低、没有基频等,给耳语音识别研究带来了困难.本文根据耳语音信号发音模型,结合耳语音的声学特性,建立了一个汉语耳语音孤立字识别系统.由于耳语音信噪比低,必须对其进行语音增强处理,同时在识别系统中应用声调信息提高了识别性能.实验结果说明了MFCC结合幅值包络可作为汉语耳语音自动识别的特征参数,在小字库内用HMM模型识别得出的识别率为90.4%.     

Acoustic Signal Typing for Evaluation of Voice Quality in Tracheoesophageal Speech

SUMMARY: Because of the aperiodicity of many tracheoesophageal voices, acoustic analysis of the tracheoesophageal voice is less straightforward than that of the normal voice. This study presents the development and testing of an acoustic signal typing system based on visual inspection of a narrow-band spectrogram that can be used by researchers for classification of voice quality in tracheoesophageal speech. In addition to this classification system, a selection of acoustic measures [median fundamental frequency, standard deviation of fundamental frequency, jitter, percentage of voiced (%Voiced), harmonics-to-noise ratio (HNR), glottal-to-noise excitation (GNE) ratio, and band energy difference (BED)] was computed to provide more insight into the acoustic components of tracheoesophageal voice quality. For clinical relevance, relationships between the acoustic signal types and an overall judgment of the voice were investigated as well. Results showed that the four acoustic signal types form a good basis for performing more acoustic analyses and give a good impression of the overall quality of the voice.     

Post Recognition Speech Localization

How much does knowledge regarding a certain spoken word or phrase help with its localization? This is a very fundamental question for speech processing, and will be partially addressed in this paper. In particular, this work will utilize prior information regarding the contents of a speech signal in order to improve the artificial localization of it using Time delay of arrival (TDOA) between two microphones. The prior information, which is used to develop a very simple frequency-selective phase transform (FPT), increases the effective SNR by only using a subset of the highest SNR frequencies in the Phase Transform. Simulations in a reverberant environment show that the proposed approach can more robustly and accurately localize speech sources. For 20 ms signal segments, it is shown that using a subset of 45 percent of available speech frequency bins is superior to using 30, 60, or 100, where using 100 corresponds to the standard Phase Transform.     

Elastically coupled molecular motors

We study the influence of filament elasticity on the motion of collective molecular motors. It is found that for a backbone flexibility exceeding a characteristic value (motor stiffness divided through the mean displacement between attached motors), the ability of motors to produce force reduces as compared to rigidly coupled motors, while the maximum velocity remains unchanged. The force-velocity-relation in two different analytic approximations is calculated and compared with Monte-Carlo simulations. Finally, we extend our model by introducing motors with a strain-dependent detachment rate. A remarkable crossover from the nearly hyperbolic shape of the Hill curve for stiff backbones to a linear force-velocity relation for very elastic backbones is found. With realistic model parameters we show that the backbone flexibility plays no role under physiological conditions in muscles, but it should be observable in certain in vitro assays. Received: 27 November 1997 / Revised: 20 February 1998 / Accepted: 27 March 1998