多人線上遊戲(Massive Multi-player Online Games, MMOGs)已經成為網際網路中最熱門的應用之一。本實驗室先前的研究中，提出一種基於P2P環境的多人線上遊戲的分散式浸入式語音交談(Distributed Immersive Voice Communication, DIVC)技術。而DIVC研究觀察出人類在交談的模式底下，一個人只會專注傾聽某一比較在意的聲音，而會忽略掉其他比較不在意的聲音。基於此觀察，DIVC被設計成為一個考慮頻寬限制和傳輸延遲的分散式語音傳輸系統，具有很高的實用價值。 本論文將DIVC的技術構想實作在本實驗室另一研究成果VAST之上。 VAST為P2P網路架構的資料傳送技術，我們運用VAST架構建立語音傳輸功能並且使用適應式k-ary樹演算法依據各使用者的頻寬不同來轉傳語音資料。聲音部份採用DirectX中的Direct Sound套件來實現。最後我們實現一個多人線上遊戲的雛型 ? 虛擬導覽系統，讓使用者間可進行語音通訊，我們也加入浸入式的設計，讓聲音大小與左右聲道音量隨聲音來源的位置改變。此外，在語音通話時系統能判斷聲音音量大小，在原始聲音音量超過某臨界值時才會送出聲音資料以降低頻寬的使用。 我們使用科技接受模型(Technology Acceptance Model, TAM)來評估我們的系統。TAM目的為提出一般化理論，以解釋並預測科技使用之影響。經過實際使用及調查訪談結果分析，使用者認為系統中加入浸入式語音可快速完成導覽及方便溝通，有82%的使用者對於本系統相當滿意。The massively multi-player online game (MMOG) has been one of the popular Internet applications. One previous study proposed MMOG distributed immersive voice communication (DIVC) system in the P2P environment. The study observed that a human in conversation only focuses on a particular sound and ignores the other sounds. Based on the observation, the study considers a distributed voice transmission technology with bandwidth limitation and propagation delay constraint to design DIVC to have good performances, as shown in experimental results. In this thesis, we implement DIVC based on VAST, which is a data transmission technology for the P2P network architecture. We use VAST to establish a voice transmission system, and use the adaptive k-ary tree algorithm based on the bandwidth of each user to forward voice data. We use Direct Sound package to implement the prototype of DIVC as a virtual tour system. We also improve the immersion aspect of DIVC by adjusting voice volume according to the distance and the left-right relationship of the speaker and the listener. Furthermore, we also add a mechanism to send voice data packet only when the volume of a source voice is larger than a threshold value to reduce bandwidth consumption. We use the technology acceptance model (TAM) to evaluate our implementation. TAM is a general model to explain and predict the impact of the use of a technology. The TAM analysis results show that users think DIVC is helpful and that 82% of users think our system is very easy to use.
