許多的因素決定了VoIP的聲音品質,包括語音編解碼器(speech codec)的選擇、封裝(packetization)、封包遺失(packet loss)、延遲(delay)、延遲變異(jitter)以及網路服務模式。平均意見評分 (MOS) 是語音通話的品質分數,它是由許多的受話者在已經控制的環境條件下的主觀測試的結果。E-model是ITU-T制定的標準,它的傳輸評價因素 (transmission rating factor) R可以直接對應到估算的MOS來評估VoIP的通話品質。 在資源有限的接取網路中,使用適當的VoIP編解碼器和適當的網路服務模式可以讓VoIP訊務獲得較好的效能。我們藉由ns-2網路模擬器來模擬VoIP訊務,而且在我們的模擬劇本中考慮使用三種常用的語音編解碼器及三種網路服務模式。我們的實驗結果顯示差異化服務(DiffServ)網路確實能改善VoIP的服務品質(QoS),而VoIP編解碼器G.711在DiffServ網路中有最好的效能表現。在Best-effort網路中,VoIP編解碼器G.729有最好的效能表現。
The voice quality of Voice over Internet Protocol (VoIP) is determined by many factors, including the choice of speech codec, packetization, packet loss, delay, delay variation (jitter), and the network service model. Mean Opinion Score (MOS), the quality score of a voice call is determined by a large number of listeners and measured by subjective testing under controlled conditions. The transmission rating factor R of the E-model, standardized by ITU-T, can be directly mapped to an estimated MOS to evaluate the quality of a VoIP call. In a resource limited access network, using proper VoIP codec and proper network service model can achieve better performance of VoIP traffic. We simulated the VoIP traffic by Network Simulator version 2 (ns-2) and considered three widely used speech codecs and three network service models in our simulation scenarios. The results of our simulations show that the adoption of Differentiated Services (DiffServ) network can improve the Quality of Service (QoS) of VoIP, and the VoIP codec G.711 has the best performance in DiffServ network. In Best-effort network, the VoIP codec G.729 has the best performance.