高品质音频信号的变换域稀疏压缩与精准重构研究及其在移动终端中的应用

The high definition audio has high speed rate mass data in generally,and the real-time audio signals are always compressed by the high compression rate and lossy encoding methods, since the restriction of bandwidth and storage capacity in the mobile terminal. These encoding methods make the high definition audio signals unreachable by the user, let alone the feeling of immersive with these lossy audio signals. In our research, we want to study the basis functions which can transform the audio signals from the time domain to the others domain first. According to the study, we want to find out the condition that make the mass audio signals sparse in the spatial domain. After the sparse transform,we study the sparsity of the high definition audio signals in the spatial domain. And this kind of audio signals will be compressed then by the compressive method.Third, we do some research of precise reconstruction for the compressed signals in the receiver. In the meantime, the analysis of the audio signal quality is studied when it is passed through the wireless channel at the condition of cascade coding and diversity transmitting. There is no doubt that the compensation algorithms are also studied to eliminate the hazards brought about by the channel transmission. All in all, we do the research in the areas of audio transform algorithm in the spatial domain, precise reconstruction algorithm and joint source-channel coding for compensating the damaged audio signals. In our study, we want to find out the reason why high definition audio signals are sparse and can be compressed in the spatial domain, and why this type signals can be reconstructed lossless first. Second, we want to find the transformation condition so that the high definition audio can be transformed. Third, we want to realize the key algorithms that can compress and reconstruct in the lossless way. Furthermore, the compensation algorithms which will eliminate the hazards brought about by the channel transmission are also studied. At last, we hope the proposed technology can be applied for storing and transmitting the high definition audio signals in the mobile device to meet the users' strong desire for high definition audio signals.

高品质音频一般是高码率的海量数据, 由于移动终端受带宽和存储容量的限制，目前其上的实时音频播放信号都是采用高压缩率的有损方法，用户无法获得高品质的音频信号，离"身临其境"的体验有很大距离。本课题通过对音频信号的变换域基函数研究，找出能使音频信号稀疏转换的条件和空间域，通过变换域的稀疏转化，使高品质音频信号被置换成异域中的稀疏信号并被压缩再在接收端被精准重构。同时，分析级联编码和多样性传输信道对音频信号在整个移动信道传输中受损情况并建立相关补偿模型，通过空间域的变换和精准重构技术的研究，结合信道的联合编码，研究对受损音频的补偿方法。本项目拟以一种全新的观点揭示出音频信号通过变换域进行稀疏、压缩再被精准重构的机理、变换条件和核心算法，以及级联编码及多样性信道对音频的损伤及补偿方法，探索将研究成果应用于移动终端实现高品质音频信号再现的技术路线，为在移动终端上的应用提出创新性的解决方法。

报告围绕高品质音频信号如何在变换域实现稀疏变换压缩和精准重构的关键问题上所开展的研究工作, 通过对整个研究工作的回顾和总结，详细汇报了项目的执行情况，如何按照任务书的具体要求，逐项开展研究工作。报告详细阐述了研究工作的内容、进展和取得的研究成果，以及在人才培养、国际交流和合作、项目成果转化及应用等方面所做的工作。整个项目的研究工作采用了理论研究和实验验证相结合的方式，重点研究高品质音频在变换域压缩中的基本原理和创新方法，通过提出有效的实现算法和方法流程，努力揭示出音频信号的变换域压缩和精准重构的本质和实现机理，并结合国际上最新的研究理论和技术趋势，同时面对移动终端上的具体应用，努力探索更好更新的解决方法和实现手段，为研究成果的进一步实用化打下理论研究和技术实现的基础。. 报告详细总结本项目研究中所提出的各种方法和算法，部分研究成果已经在国际著名期刊或国际学术会议上发表。在本项目的研究过程中，课题组共发表了论文25篇，培养了博士和硕士研究生共13名，由本项目衍生出的研究成果所提出的发明专利申请1项，获科技奖2项，部分研究成果已经得到了技术转化并应用到了实际，取得了良好的社会和经济效益。. 本研究项目已经全部完成了任务书所提出的任务，报告同时也指出了尚存在的问题和不足，在后续的研究工作中将继续探索和改进，我们将继续努力并争取将本项目所取得的研究成果向实际应用转化，为中国乃至世界的信息技术的创新和发展继续我们的贡献和努力。