摘要: | 近年來半導體發展迅速,因此許多數位的家電產品誕生,而數位家庭的概念開始施行於日常生活中。舊有傳輸界面,如HDMI及DVI,採用有線方式傳輸。然而當一個家庭的數位產品越來越多(如setup-box、Blue-ray disc player、PS3)相互連接變成一個重要的議題。另外,影像品質的提昇讓高解析度畫面的影像開始成為新的主流。雖然無線網路速度不斷提昇,但是速度仍然不夠許多影像的同時傳輸。 因此,本子計畫基於這些現象,提出「適用於數位家庭之多訊源可延展影像壓縮系統設計」。此系統結合可延伸影像處理單元、影像預測機制、進階的嵌入式壓縮核心設計、以及Video MUX。除了支援多種大小的影像格式(譬如1080p-1920x1080、720p-720x480、DVD-720x480)的相互轉換,多來源視訊的編碼外,另外增強的預測機制能進一步提昇壓縮倍率,且能夠在硬體上平行處理的壓縮演算法以及輸出切換機制更能達成多影像來源的編碼。另外解碼端只需要一精簡的解碼架構,因此縮小了整體硬體成本。經由跟其他架構的整合以及連接,除了軟硬體部份溝通的建立,也可以建立一完整之無線高畫質多來源影像傳輸系統。藉由此計畫所開發之壓縮,將可以解決高畫質及多來源視訊在無線網路上傳輸的瓶頸,並實現數位家庭的概念。 The advances in VLSI technology brought up rapid increase of home media products, and a new concept, digit home, is developed and accepted in our lives. Traditional devices are connected via wires, such as HDVI, DVI, etc. However, wires are expansive; the transmission distance will further increase the cost of wires. With more and more digital products in a house, interconnection between devices becomes a nightmare. On the other hand, high resolution video increases the data bandwidth during transmission, and concurrent video transmission is still a heavy burden for wireless network. Therefore, in this project, the “multi source scalable video compression” scheme is introduced. The compression engine combines scalable video processing unit, video prediction scheme, advanced embedded compression engine, and a video MUX. The introduced compression engine support higher compression ratio, higher video quality in near lossless mode, and the ability of arbitrary video size conversion (such as 1080p, 720p, DVD). Additional prediction scheme increases the compression ratio. Multi source video encoding is supported by multiple compression engine and additional video MUX. In the wireless receiver end, a simple decoding algorithm is suffix to uncompress the multi source video bit-stream. The combination of the sub-project with other sub-projects leads to a complete high definition wireless video transmission system, and the proposed system can achieve high data-rate video transmission. 研究期間:9908 ~ 10007 |