MELP声码器的算法研究及实现

为了满足数字通信及其它商业应用的需求,语音压缩编码技术得到了迅速发展。特别是低码率语音编码的研究具有十分重要的现实意义。在现有的语音编码研究中,混合激励线性预测编码(MELP)是一种比较好的方法。对MELP编解码算法的原理进行简要分析,讨论如何在MATLAB上实现该算法,并研究其关键技术,最后对测试结果进行分析和比较。     

超低速率MELP语音编码算法研究

在语音编码算法中,混和激励线性预测(MELP)算法因为能更好的模拟自然语言特征,在低速率上能合成较高质量的语音,而成为现代低速率语音编码中最有潜力的算法之一。但在无线通信、卫星通信以及军用和保密通信中,信道带宽成为一个突出的问题,因此对更低速率语音压缩编码技术乃至超低速率的语音压缩编码技术的研究是非常有必要的。针对语音通信中关于极低速率的要求,深入分析了现今的几种基于MELP的低速率语音编码算法,对其原理以及关键技术进行了归纳总结,并对语音质量进行了比较。     

MELP／CELP混合语音编码方法研究

介绍了混合激励线性预测（MELP）的算法特点和采用的几种关键技术。在MELP基础上,针对MELP对周期性不强的语音帧（过渡帧、清音帧等）不能准确描述其激励形状,提出了采用MELP／CELP互补的混合编码方案,并在MELP中采用了相位对齐的方法。MATLAB仿真发现8kb／s的这种混合编码器可以获得与32kb／s编码速率的ADPCM相近的语音质量。     

语音信号的线性预测分析原理及算法分析

通过对LPC(线性预测编码)的研究,介绍语音信号的线性预测分析原理,详细分析用来求解线性预测方程的自相关法、协方差法的原理和计算方法,对算法进行比较,并用Matlab对实际语音信号进行线性预测编码实验。     

变阶数全极点声道模型及其在MPLPC中的应用仿真

介绍了声道的级联声管全极点数学模型，针对声道复杂度的非平稳性，提出了改进的变阶全极点模型和这种模型在多脉冲线性预测语音编码（MPLPC）中的应用。用Matlab对该方法进行了仿真，得到了良好的合成语音质量。     

2．4kbps混合激励线性预测编码方法研究

分析LPC编码方法缺点的基础上，介绍了混合激励线性预测（MELP）编码方法，着重分析了它的一些新特性。进而，设计了2．4kbps的MELP编码器，并用Matlab进行了仿真。仿真结果证明该编码器可以获得在主观听音方面接近于4．8kbps的CELP的合成语音。     

基于Matlab的LFM脉冲压缩仿真

基于Matlab平台以线性调频信号为例通过仿真研究了雷达信号处理中的脉冲压缩技术。在对线性调频信号时域波形进行仿真的基础上介绍了数字正交相干检波技术。最后基于匹配滤波算法对雷达回波信号进行了脉冲压缩仿真,仿真结果表明采用线性调频信号可以有效地实现雷达回波信号脉冲压缩,提高雷达的距离分辨力。     

ACELP算法在PC上实时播放的实现

对TETRA系统中所采用的ACELP（Algebraic CELP）语音压缩算法进行了分析,该算法是基于CELP的改进算法,在语音编码中具有较强的代表性。在Windows XP环境下,利用VS2005实现将声卡采集到的数据进行压缩和解压缩,延时10s后通过声卡播放出来。实现过程用到许多Windows核心程序开发知识,并且给出了关键部分的源代码。对语音编码研究人员和程序员都有一定的借鉴意义。     

水下双工语音通信系统的设计与实现

蛙人语音通信是水声领域研究的热点问题。利用多帧联合处理以及矢量量化技术实现了1.2kbps 速率的混合激励线性预测语音压缩编码,为适应水声信道的特点,采用正交频分复用技术对信源进行调制,并加入了同步、信道估计、RS纠错码等相关技术,以TMS320DM642和TLV320AIC23B等芯片为核心搭建硬件平台,设计并实现了一套水下实时双工语音通信系统。水池试验表明,在20 m距离上,合成语音清晰连贯并可分辨出说话人为谁,系统性能稳定。     

改进的混合激励线性预测编码算法

针对混合激励线性预测编码中子带声音强度的硬判决导致激励源欠精细问题,将子带声音强度视为5维的模糊特征矢量,用改进的LBG算法设计码本并用5bit对其作矢量量化;以精细量化的子带声音强度调制带通滤波器,以此获取精细的混合激励信号,最终达到改善合成语音质量的目标。仿真实验表明:改进算法能有效地改善合成语音的自然度。     

强噪声环境下语音增强算法的比较研究

介绍语音增强的原理和从强噪声背景中提取语音信号的方法,并对基于减谱法的增强算法、基于自适应滤波法的增强算法和基于小波变换的增强算法进行对比研究。鉴于语音增强算法的两个目标即增强语音的清晰度与理解度并不是相关联的,有时甚至相互矛盾,因此任何一个语音增强算法都是根据不同的应用做适当的选择和折衷。基于此,对三种算法进行仿真实验的比较研究,实验结果证明应用小波变换的语音增强算法比其它的方法更有效。     

Applying and Comparison of Chaotic-Based Permutation Algorithms for Audio Encryption

This research presents, and clarifies the application of two permutation algorithms, based on chaotic map systems, and applied to a file of speech signals. They are the Arnold cat map-based permutation algorithm, and the Baker’s chaotic map-based permutation algorithm. Both algorithms are implemented on the same speech signal sample. Then, both the premier and the encrypted file histograms are documented and plotted. The speech signal amplitude values with time signals of the original file are recorded and plotted against the encrypted and decrypted files. Furthermore, the original file is plotted against the encrypted file, using the spectrogram frequencies of speech signals with the signal duration. These permutation algorithms are used to shuffle the positions of the speech files signals’ values without any changes, to produce an encrypted speech file. A comparative analysis is introduced by using some of sundry statistical and experimental analyses for the procedures of encryption and decryption, e.g., the time of both procedures, the encrypted audio signals histogram, the correlation coefficient between specimens in the premier and encrypted signals, a test of the Spectral Distortion (SD), and the Log-Likelihood Ratio (LLR) measures. The outcomes of the different experimental and comparative studies demonstrate that the two permutation algorithms (Baker and Arnold) are sufficient for providing an efficient and reliable voice signal encryption solution. However, the Arnold’s algorithm gives better results in most cases as compared to the results of Baker’s algorithm.     

噪声谱估计算法对语音可懂度的影响

噪声谱估计是单通道语音增强算法的关键步骤,当前大部分语音增强算法旨在提高语音质量,提高语音可懂度的算法却很少。在传统的单通道语音增强算法中,语音质量的提高往往是以牺牲语音的可懂度为代价的。对目前主流的几种噪声谱估计算法对语音可懂度影响进行分析。在不同噪声背景、不同信噪比情况下进行噪声谱估计,并采用谱减法对含噪语音信号作去噪处理,对比分析不同噪声、不同信噪比下增强前后语音的短时客观可懂度(Short-Time Objective Intelligibility,STOI)值,最后根据信噪比,对比分析了不同噪声环境下,语音增强前后语音能量高于噪声能量的时频块所占比例。实验表明,相比其他噪声估计算法,最小统计(Minima Statistics,MS)算法由于保留了更多的以语音能量为主的时频块,使得去噪后的语音有较高的可懂度。     

基于改进的Kalman滤波的语音增强算法

传统的kalman滤波方法在推导过程中假定观测噪声为白噪声。通常对于有色噪声需要用白噪声激励的方法予以模拟,并且需要以牺牲运算量作为代价。本文提出了一种改进的基于kalman滤波的语音增强算法,可以处理白噪声和有色噪声情况,不需要增加计算量,仿真结果表明了该算法对有色噪声的语音增强性能要优于基于传统kalman滤波方法。     

MoTransFrame: Model Transfer Framework for CNNs on Low-Resource Edge Computing Node

Deep learning technology has been widely used in computer vision, speech recognition, natural language processing, and other related fields. The deep learning algorithm has high precision and high reliability. However, the lack of resources in the edge terminal equipment makes it difficult to run deep learning algorithms that require more memory and computing power. In this paper, we propose MoTransFrame, a general model processing framework for deep learning models. Instead of designing a model compression algorithm with a high compression ratio, MoTransFrame can transplant popular convolutional neural networks models to resources-starved edge devices promptly and accurately. By the integration method, Deep learning models can be converted into portable projects for Arduino, a typical edge device with limited resources. Our experiments show that MoTransFrame has good adaptability in edge devices with limited memories. It is more flexible than other model transplantation methods. It can keep a small loss of model accuracy when the number of parameters is compressed by tens of times. At the same time, the computational resources needed in the reasoning process are less than what the edge node could handle.     

Lossy compression of TPC data and trajectory tracking efficiency for the ALICE experiment

In this paper a quasi-lossless algorithm for the on-line compression of the data generated by the Time Projection Chamber (TPC) detector of the ALICE experiment at CERN is described. The algorithm is based on a lossy source code modeling technique, i.e. it is based on a source model which is lossy if samples of the TPC signal are considered one by one; conversely, the source model is lossless or quasi-lossless if some physical quantities that are of main interest for the experiment are considered. These quantities are the area and the location of the center of mass of each TPC signal pulse, representing the pulse charge and the time localization of the pulse.

So as to evaluate the consequences of the error introduced by the lossy compression process, the results of the trajectory tracking algorithms that process data off-line after the experiment are analyzed, in particular, versus their sensibility to the noise introduced by the compression. Two different versions of these off-line algorithms are described, performing cluster finding and particle tracking. The results on how these algorithms are affected by the lossy compression are reported.

Entropy coding can be applied to the set of events defined by the source model to reduce the bit rate to the corresponding source entropy. Using TPC simulated data according to the expected ALICE TPC performance, the compression algorithm achieves a data reduction in the range of 34.2% down to 23.7% of the original data rate depending on the desired precision on the pulse center of mass.

The number of operations per input symbol required to implement the algorithm is relatively low, so that a real-time implementation of the compression process embedded in the TPC data acquisition chain using low-cost integrated electronics is a realistic option to effectively reduce the data storing cost of ALICE experiment.     

一种自适应噪声抵消系统的仿真与设计

自适应噪声抵消系统是基于自适应滤波原理的一种扩展,它能从被噪声干扰的环境中检测和提取有用信号,抑制或衰减噪声干扰,提高信号传递和接收的信噪比质量。研究常用噪声消除算法,并仿真验证系统设计和所选自适应算法的可行性,最后应用DSP技术设计一个自适应噪声抵消系统,使其能消除含噪语音信号中的背景噪声,达到提高语音信号质量的目的。     

A Distributed Covert Channel of the Packet Ordering Enhancement Model Based on Data Compression

Covert channel of the packet ordering is a hot research topic. Encryption technology is not enough to protect the security of both sides of communication. Covert channel needs to hide the transmission data and protect content of communication. The traditional methods are usually to use proxy technology such as tor anonymous tracking technology to achieve hiding from the communicator. However, because the establishment of proxy communication needs to consume traffic, the communication capacity will be reduced, and in recent years, the tor technology often has vulnerabilities that led to the leakage of secret information. In this paper, the covert channel model of the packet ordering is applied into the distributed system, and a distributed covert channel of the packet ordering enhancement model based on data compression (DCCPOEDC) is proposed. The data compression algorithms are used to reduce the amount of data and transmission time. The distributed system and data compression algorithms can weaken the hidden statistical probability of information. Furthermore, they can enhance the unknowability of the data and weaken the time distribution characteristics of the data packets. This paper selected a compression algorithm suitable for DCCPOEDC and analyzed DCCPOEDC from anonymity, transmission efficiency, and transmission performance. According to the analysis results, it can be seen that DCCPOEDC optimizes the covert channel of the packet ordering, which saves the transmission time and improves the concealment compared with the original covert channel.     

自复叠制冷技术发展现状

自复叠制冷因制冷温区较宽,在普冷、深冷领域具有十分广阔的应用前景。在介绍自复叠制冷技术的应用和原理的基础上,与单级压缩、两级压缩及复叠式制冷技术进行比较,指出了自复叠制冷技术的优势和特点。主要从自复叠制冷技术流程及混合制冷剂选择和配比两方面,对当前自复叠制冷技术的研究概况进行分析,并指出该技术在系统流程设计和混合工质的选择和配比方面未来的发展动向,为该技术的进一步研发和推广应用提供参考。