應用於數位典藏長久保存之可重組式視訊轉碼技術

Reconfigurable Video Transcoding for Long-term Preservation in Digital Archives

近年來，由於數位典藏技術與應用的發展，各公、私立典藏單位已經收集大量的文字、影像、視訊的典藏資料，以利促進人文與社會、產業與經濟的發展。藉由結合數位典藏資料庫以及網際網路服務，對於整體資訊社會的影響能夠擴大，並落實數位典藏資訊的知識化與社會化，建構及發展知識社會。在數位典藏資料中視訊的資料量極大，為有效壓縮資料及方便管理，目前視訊資料大多以MPEG編碼標準或其他商業公司自有的編碼格式（例如，Microsoft WMV-9或RealNetwork等）壓縮並儲存。然而，我們可以預期，隨者科技之進步，未來仍然不斷有新的視訊壓縮技術及標準會被提出。在此趨勢之下，目前的典藏在未來數十年後是否仍能被有效的擷取是一個必須探討的問題。因此我們需要發展一個具有高使用彈性之泛用型視訊轉碼器，可以進行不同視訊編碼標準、解析度及位元率之轉換。此外此視訊轉碼器也要具備可擴充性，可以很容易地加入未來所發展之新的視訊編碼標準。本文敘述我們所發展的可重組之視訊轉碼器。此轉碼器採用MPEG所發展的Reconfigurable Video Coding（RVC）技術，可以動態地及彈性地建構編、解碼器，以便於進行不同視訊編碼標準間之轉換及未來新的視訊標準之擴充，達到長久保存及應用數位視訊檔案的目的。

With the advance and fast deployment of multimedia compression technology, many public and private sectors have been archiving a large number of digital media, including texts, images, and videos. In digital archives, digital video data consumes the majority of storage space and processing expenses. In order to reduce the storage capacity of video data, most video programs usually compress video data with MPEG CODECS or another proprietary CODECS. (e.g., Microsoft WMV-9 or RealNetwork). As a result, digital archives always contain video data compressed by various coding standards. At present, an existing problem is that current digital archives might not be extracted effectively in 10 years later. We can also expect that more and more new video coding standards will be developed to achieve higher compression efficiency and more functionality in the future. Therefore, the variety and heterogeneity of digital video contents in digital archives should increase greatly. For this reason, a general-purposed reconfigurable video transcoder is proposed for various coding standards, resolution and bit rate. In addition, the transcorder should be functional extendable and adaptable to new video coding standards. This article describes the proposed reconfigurable video transcoder that adopts the new MPEG Reconfigurable Video Coding (RVC) technology upon which a decoder can be built dynamically and flexibly. We will analyze various video coding standards, and then implement the related functional units. Based on the repository of functional units, our video transcoder can accommodate various input formats and output formats by dynamically assembling and reconfiguring the functional units.