一种语音信号重建算法

为了减轻因信包丢失而造成的语音失真,提出了一种基于双边线性预测的信包丢失隐藏算法。这种方法利用丢失信包的前一信包或邻接信包(在后一信包可获得的情况下)预测丢失信包,通过线性加权双边线性预测的样点获得最终的重建信号,使用重叠相加和幅度调整操作平滑重建信号和真实信号之间的边界。经过非正式试听和ITU-T P.862协议所推荐的PESQ算法测试,该算法的重建语音信号质量与其他四种流行重建算法相比,有了较为明显的改善。     

用时域及幅度域修正方法提升基于线性预测的信包丢失后重建语音信号质量

本文介绍对一种基于线性预测方法的信包丢失重建算法[1]的改进研究.通过对基于线形预测方法的原始重建信号在时域和幅度域上的进一步处理,重建信号在相位连续性和样点幅度变化上的表现更加合理.经过非正式试听和ITU-T P.862协议所推荐的PESQ[2]算法客观测量,大量改进后测试语音信号的质量比原来均有提高.     

用线性预测法实现气声语音的重建

本文介绍了一种用线性预测参数实现气声语音重建的方法。首先提取气声语音的线性预测参数，用这些参数建立信号产生模型，然后进行合成，合成时加入基音周期，使其恢复到正常语音。     

1.8kb/s的多带线性预测语音编码算法

设计了一种数码率为１．８ｋｂ／ｓ的多带线性预测（ＭＢＬＰ）语音压缩编码算法。该算法采用基于谐振结构的线性预测分析和对激励信号采用多带处理的方法。试验结果表明，本算法提供了相当于码率为２．４ｋｂ／ｓ美国联邦声码器标准ＭＥＬＰ的重建语音质量，具有较高的清晰度和自然度。     

一种高质量的4 Kb/s RCELP语音编码算法

给出一种高质量的4Kb／s更新式码激励线性预测(RCELP)语音编码算法。该算法的编码器帧长为20ms,主要特点是使用了从自适应激励信号中分析得到的码本作为固定码本,采用预测式两级分裂矢量量化器量化线谱对(LSP)参数。主观试听表明,该算法的MOS值为3．67,其语音质量与32Kb／s ADPCM基本相当。     

LPC与LSF转换算法的比较研究

线性预测法是语音信号处理中的核心技术。在语音信号的处理中,常常需要将线性预测的LPC系数与LSF参数相互转换。本文根据Chebyshev多项式求根法,研究了几种由LPC求解LSF的算法,分析了它们各自的特点及相互关系。分析并推导了由LSF求解LPC的算法。     

高质量4kb/s FS-ACELP语音编码算法及性能

4kb/s有限状态代数码激励线性预测语音编码算法FS-ACELP是一种具有延时较短,合成语音质量高、算法复杂度较低的语音编码算法.在线性预测(LP)参数量化上,利用了语音帧内和帧间的相关性,对线谱对(LSP)参数使用预测式分裂式矢量量化,获得很高的量化效率.在自适应码本搜索上,采用了有限状态控制分数延时搜索的算法.在保证合成语音质量的同时,有效地降低了运算量.对于随机码本,采用了具有多模结构的代数码本,提高语音合成质量.对于激励码序列的增益,采用了预测式矢量量化,有效地提高了量化精度.经非正式听音测试,4kb/s FS-ACELP的合成语音质量超过了北美8kb/s VSELP,接近G.729 8kb/s CS-ACELP,MOS分约为3.9.     

高质量4kb/s FS-ACELP语音编码算法及性能

4kb/s有限状态代数码激励线性预测语音编码算法FS-ACELP是一种具有延时较短、合成语音质量高、算法复杂度较低的语音编码算法.在线性预测(LP)参数量化上,利用了语音帧内和帧间的相关性,对线谱对(LSP)参数使用预测式分裂式矢量量化,获得很高的量化效率.在自适应码本搜索上,采用了有限状态控制分数延时搜索的算法,在保证合成语音质量的同时,有效地降低了运算量.对于随机码本,采用了具有多模结构的代数码本,提高语音合成质量.对于激励码序列的增益,采用了预测式矢量量化,有效地提高了量化精度.经非正式听音测试,4kb/s FS-ACELP的合成语音质量超过了北美8kb/s VSELP,接近G.729 8kb/s CS-ACELP,MOS分约为3.9.     

基于神经网络的线性预测语音编码算法

语音压缩是多媒体通信技术的重要环节,线性预测编码(LPC)技术是参数编码技术的重要内容,线性预测是语音信号处理中最有效的方法之一.文中从LPC原理入手,阐述了最佳LPC系数的计算,针对目前自相关法和协方差等存在着估计误差的特点,提出一种基于神经网络的线性预测算法,最后通过实验数据证明这种方法既提高了解的精度,又保证了系统的稳定性.     

与VoIP声码器结合的回声消除器

NLMS算法是回声消除器中最常用的算法之一,然而语音信号的强相关性使NLMS(归一化最小均方)算法的收敛速度变慢。鉴于VoIP(网络电话)常用的低速率声码器大多基于LP(线性预测)编码,给出了一种与声码器结合的回声消除器,通过利用从解码端获得的LP系数和激励信号,省去了对去相关滤波器系数的计算,并且不再需要专门的前置滤波器为远端信号去相关,同时又改善了回声消除器性能。仿真结果表现出更高的收敛速度和回声抵消量。     

基于瞬时相位差和深度学习的丢包隐藏方法

实时IP 语音通信在数据包会丢失的情况下,语音质量会受到严重影响。为了恢复传输过程中丢失的语音信息,本文提出了一种基于瞬时相位差(Instantaneous Phase Deviation, IPD)和深度神经网络(Deep Neural Network, DNN)的丢包隐藏 (Packet Loss Concealment, PLC)方法。在训练阶段,将语音的对数功率谱(Log Power Spectrum, LPS)和IPD作为训练DNN的输入特征,以学习从接收包到丢失包的映射关系;在重构阶段,将丢包前接收到的语音包送入训练好的DNN中,恢复出丢失包的语音。实验结果表明,在不同丢包率下,所提方法的性能优于传统的基于LPS和DNN的PLC方法。      

基于机会式网络编码的高效广播传输算法

为了提高无线网络广播传输的效率,针对单跳无线网络提出了采用编码方法的广播传输算法。在传统的无线广播传输模型的基础上,分别实现了基于机会式网络编码的单组合分组广播传输算法和多组合分组广播传输算法。它们采用不同的策略选择多个丢失分组编码组合成重传分组,并通过从编码组合数据分组中恢复丢失分组的方式来提高广播传输的吞吐量。仿真结果表明,新算法在不同无线信道传输模型下相比已有的算法有效地降低了广播传输所需的传输带宽。     

Single-ended parametric voicing-aware models for live assessment of packetized VoIP conversations

The perceptual quality of VoIP conversations depends tightly on the pattern of packet losses, i.e., the distribution and duration of packet loss runs. The wider (resp. smaller) the inter-loss gap (resp. loss gap) duration, the lower is the quality degradation. Moreover, a set of speech sequences impaired using an identical packet loss pattern results in a different degree of perceptual quality degradation because dropped voice packets have unequal impact on the perceived quality. Therefore, we consider the voicing feature of speech wave included in lost packets in addition to packet loss pattern to estimate speech quality scores. We distinguish between voiced, unvoiced, and silence packets. This enables to achieve better correlation and accuracy between human-based subjective and machine-calculated objective scores.     

LTE系统上行链路调度算法研究

长期演进(Long Term Evolution,LTE)已经成为4G无线技术标准。目前,LTE分组调度的下行链路调度被大多数研究者研究,上行链路的研究相对较少。针对上行链路调度无法保证实时业务分组在延迟期限内传输,存在公平性较差、分组丢弃多的问题。因此,提出了一种新的上行链路调度算法。该算法根据实时业务的延迟约束条件建立目标整数线性规划模型,再根据目标整数线性规划模型进行调度。实验结果表明,该算法能保证实时业务分组在延迟期限内传输,适用于实时业务,能确保公平性,最小化分组丢弃,具有较好的适用性。     

Effects of Packet Losses in Waveform Coded Speech and Improvements Due to an Odd-Even Sample-Interpolation Procedure

We have studied the effects of random packet losses in digital speech systems based on 12-bit PCM and 4-bit adaptive DPCM coding. The effects are a function of packet lengthBand probability of packet loss PL. We have also studied tbe benefits of an odd-even sample-interpolation procedure that mitigates these effects (at the cost of increased decoding delay). The procedure is based on arranging a2B-block of codewords into twoB-sample packets, an odd-sample packet and an even-sample packet. If one of these packets is lost, the odd (or even) samples of the2B-block are estimated from the even (or odd) samples by means of adaptive interpolation. Perceptual considerations indicate that packet lengths most robust to losses are in the range 16-32 ms, irrespective of whether interpolation is used or not. With these packet lengths, tolerable PLvalues, which are strictly input-speech-dependent, can be as high as 2 to 5 percent without interpolation and 5 to 10 percent with interpolation. These observations are based on a computer simulation with three sentence-length speech inputs, and on informal listening tests.     

Media Synchronization and QoS Packet Scheduling Algorithms for Wireless Systems

Wireless multimedia synchronization is concerned with distributed multimedia packets such as video, audio, text and graphics being played-out onto the mobile clients via a base station (BS) that services the mobile client with the multimedia packets. Our focus is on improving the Quality of Service (QoS) of the mobile client's on-time-arrival of distributed multimedia packets through network multimedia synchronization. We describe a media synchronization scheme for wireless networks, and we investigate the multimedia packet scheduling algorithms at the base station to accomplish our goal. In this paper, we extend the media synchronization algorithm by investigating four packet scheduling algorithms: First-In-First-Out (FIFO), Highest-Priority-First (PQ), Weighted Fair-Queuing (WFQ) and Round-Robin (RR). We analyze the effect of the four packet scheduling algorithms in terms of multimedia packet delivery time and the delay between concurrent multimedia data streams. We show that the play-out of multimedia units on the mobile clients by the base station plays an important role in enhancing the mobile client's quality of service in terms of intra-stream synchronization and inter-stream synchronization. Our results show that the Round-Robin (RR) packet scheduling algorithm is, by far, the best of the four packet scheduling algorithms in terms of mobile client buffer usage. We analyze the four packet scheduling algorithms and make a correlation between play-out of multimedia packets, by the base station, onto the mobile clients and wireless network multimedia synchronization. We clarify the meaning of buffer usage, buffer overflow, buffer underflow, message complexity and multimedia packet delay in terms of synchronization between distributed multimedia servers, base stations and mobile clients.     

Accurate method for analysis of a packet speech multiplexer

A packet speech multiplexer is analysed using a fluid approximation for the flow of packets in and out of the multiplexer. Delay distributions and fractional packet loss are determined with a small amount of computation. Comparisons with a simulation using real speech show the analysis to be accurate.     

基于机会式网络编码的低时延广播传输算法

为了提高无线网络中数据包广播传输的效率,本文提出了一种基于机会式网络编码的广播传输算法.该算法在发送端按一定顺序选择不同终端的丢包,并采用异或运算编码重传包,在终端采用从重传包中解码数据包的方法恢复丢包.该算法优先恢复时间重要性较高的丢包,并使多个终端同时从单个重传包恢复其丢包,因此有效地提高了广播传输效率并降低了传输...     

Selective discarding procedure for improved tolerance to missing voice packets

A selective packet discarding procedure for voice communication is presented in which packets are labelled at the transmitter and, if necessary, discarded by network nodes according to their significance to reconstructed speech quality. Simulation and subjective tests show improved tolerance to missing packets in comparison with unselective methods where packets are discarded at random.<>