TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI
You also want an ePaper? Increase the reach of your titles
YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.
14<br />
12<br />
our proposal<br />
classical approach<br />
The AU’s losses depenency<br />
Succesive AUs lost<br />
10<br />
8<br />
6<br />
4<br />
2<br />
0<br />
0 5 10 15 20<br />
Time(s)<br />
Figure 4-20: Correlated AU’s losses.<br />
4.2.3.2.2 Scenarios 2<br />
In this scenario, we have compared between two configurations. The first is RTP payload<br />
format with UEP scheme (Configuration A), and the second is without UEP scheme<br />
(Configuration B). For each configuration, we vary gradually the network load to get more<br />
information on the behavior of the different mechanisms.<br />
In order to highlight the efficiency of the UEP scheme, we compute the number of AU that<br />
can be decoded at the Client. Figure 4-21 shows this result of comparison between the decoded<br />
object ratios. The X-axis represents the throughput of the background traffic. As expected, the<br />
quantity of the AVOs decoded at the receiver side decreases when the network load increases<br />
because it entails more packet losses.<br />
1<br />
0.9<br />
No Error Protection<br />
FEC-based UEP<br />
Decoded Object Ratio<br />
0.8<br />
0.7<br />
0.6<br />
0.5<br />
0.4<br />
1.5 2 2.5 3 3.5 4 4.5<br />
Background Traffic (Mbit/s)<br />
Figure 4-21: Decoded object ratio vs. background traffic throughput<br />
<strong>Packet</strong> loss is accented by using our FEC-based UEP because UEP increases the MPEG-4<br />
packet-stream throughput by 7 %. For this reason, there is more packet losses with UEP<br />
configuration, for a given network load. However, the redundant UEP information better recovers<br />
90