遍地即時多媒體系統與技術---子計畫五：多媒體傳輸安全機制與可調音訊編碼

Abstract

隨著網路的蓬勃發展，多媒體資訊傳播和搜尋變的非常的容易。相關的媒體傳送行為衍 生出許多重大的問題。諸如保護資料傳輸時的安全以免被竊取、限制資料不被任意散 播、或是透過可信賴的付費的機制來保障著作人以及付費使用者的權益等等，都使得數 位智慧財權管理(Digital Rights Management, DRM)，變成了一個重要的議題。此外，近 年來網路技術與設備趨於多樣化，使得透過網路存取資訊更加便利，但此種異質網路與 設備的結合也產生新的應用模式，因此多媒體編碼法也開始朝向多層次可調性編碼發 展。近年來網路音樂與語音應用快速成長，當面對異質網路時，可以發現亦有類似的應 用問題。但與可調視訊編碼不同，可調音訊與語音壓縮較少人研究，而目前也尚未有良 好效率的細微可調音訊編碼方法。 本計畫針對上述問題，分為三個主題進行探討。(1) IPMP 之研究與實作:我們從標準文獻 出發，配合主要的媒體保護概念及機制，深入探討在現有工具下的多媒體保護策略。(2) 可調性多媒體保護之研究: 此部份以調性壓縮特性為出發點，結合多媒體保護概念，提 出可行的架構。(3) 可調性音訊與語音編碼，此部份針對聲音訊號進行可調性編碼研究。 擬結合語音預估編碼法與人類聽覺特性，設計細緻可調的音訊編碼法。本子計畫為「遍 布即時多媒體系統與技術」的一部分，預定計畫全程為三年。

With the fast growing of the Internet, multimedia information distribution and retrieval becomes more and more popular. There are several important issues raised from the exchange of multimedia contents. For example, how to protect the content from unauthorized accesses, how to protect the content from illegal distribution, how to give the payment through a trusted mechanism. They are the hot topics in the area of digital rights management (DRM). Another issue is the heterogeneous inter-networking and device-connections due to inhomogeneous evolution of network and user terminals. In particular, the heterogeneous environment affects the delivery of audio and speech data. Although a few audio scalable coding schemes have been proposed to solve this problem, unlike video scalable coding, the scalable audio coding scheme is less studied and efficient algorithm is not mature yet. To investigate the above issues, this project includes three topics. (1) IPMP (Intellectual Property Protection and Management) system design and implementation: We will start from the MPEG IPMP standard, study the state-of-the-art DRM schemes, and propose a combination of the tools to protect contents. (2) Scalable multimedia protection: We make use of the scalable coding properties, and propose a structure that protects scalable-coded contents. (3) Scalable audio and speech coding: We will study the existing algorithms, and propose a scalable audio coding scheme that can efficiently encodes both audio and speech with fine granularity scalability. Essentially we like to incorporate the predictive techniques used in speech coding together with the audio psychoacoustic model. This sub-project is a part of the integrated project: Pervasive Multimedia System and Technology. Its entire duration is 3 years.