由於無線網路技術的蓬勃發展，加上多媒體應用快速普及與資料編解碼技術成熟，使得無線網路結合影音多媒體以提供各式多樣化服務已是通訊技術中一項重要的必然趨勢。但由於無線通道的傳送品質具有明顯的不可靠性，使得影音多媒體服務在無線網路上的傳送充滿挑戰性。無線網路中通常用前向錯誤更正（Forward Error Correction，FEC）降低無線通道中的高位元錯誤率（Bit Error Rate，BER）。FEC通過發送多餘資訊進行錯誤恢復，糾正越高的BER需要越多的多餘資訊。對於無線網路中變動較大的BER來說，採用更正錯誤能力低的編碼在遇到高BER時沒有顯著改善，採用更正錯誤能力高的編碼在低BER情況下又浪費頻寬，合理的做法是使用動態的FEC來適應BER的變化。為了提高前向錯誤更正（FEC）編碼在影像傳輸上的效能，我們提出了一種改進的適應性前向錯誤更正傳輸機制（AFEC）。新機制中，發送端收到回饋資訊後，再決定是否使用錯誤更正碼（FEC）傳輸，同時會根據網路的環境狀況，來調整資料的附加長度（Redundancy），且不會為了等待回饋資訊而停止資料傳送。實驗上的比較結果，我們比較了不同類型的實驗影片，AFEC的影像明顯比傳統不加FEC方法的影像品質好。在PSNR的計算方面，在PER為0.2時，多出10 db以上。而與固定FEC的比較，顯示在接近相同影像品質的表現，AFEC所花費的封包冗餘率明顯比較少，以在PER為0.005為例，AFEC只增加不到6%的冗餘封包，固定FEC則仍然需要增加100%的冗餘封包。同時本文提出平均畫面損壞率的概念，並且估計錯誤率。使用者可以根據所需要的影像品質條件，來設定AFEC的最佳的臨界值。 With the advance of wireless network technologies, wireless multimedia applications can be widely used in the near future. The video transmission over wireless networks is an important and inevitable trend by providing handheld devices with the telecommunication technique. However, the wireless networks are not trustworthy and reliable. The video multi-media over the wireless network service will be filled with challenges. The forward error correction (FEC) usually reduces the errors of the wireless networks with high bit error rate (BER). The FEC sends out redundant data packets to recover the data errors. If we want to correct data during the wireless networks with higher BER, we need to send more redundant information. If the BER of wireless networks is high, the low-strength error correction coding does not offer notable improvement. On the other hand, the high-strength of error correction coding in low BER would waste network bandwidth. It is reasonable that using dynamic FEC adapts the variety of BER. For raising the power of forward error correction for video delivery over wireless networks, we present an adaptive forward error correction (AFEC) mechanism. In our AFEC mechanism, if the sender receives feedback information, it will decide whether the forward error correction transmission should be used or not. The AFEC would adjust the additional length (redundancy) of data according to the BER of wireless networks. The sender would not stop transferring data for waiting the feedback information. The result of the experiments reveals that the AFEC mechanism can improve the quality of video delivery over bad situations. The experiment comparison result shows that the image quality of AFEC obviously better than the image quality without FEC method. When PER is 0.2, The PSNR of AFEC mechanism is higher without FEC mechanism more than 10 db. The result shows that the image quality of AFEC mechanism almost the same with fixed FEC mechanism. However, AFEC mechanism cost less redundant packets obviously. When PER is 0.005, the AFEC mechanism only increase 6% redundancies. In the same situation, Fixed FEC mechanism still increase 100% redundancies. The AFEC provides a new method for transferring online video. We also suggest a mean frame corruption rate which estimates a video error rate. The user can set the optimal threshold value for AFEC according to different video quality level.
