SCALABLE VIDEO TRANSPORT OVER IP NETWORKS

[attachment=17871]

Introduction
1.1 Background and Related Works
With the advances in video compression and networking techniques, the last ten
years have witnessed an explosive growth of video applications over the Internet. Es-
pecially video streaming applications, such as YouTube, are used every day by people
around the world. However, the service model of the current best-effort Internet was
never engineered to handle video traffic and, as a result, video applications still suffer
from varying and unpredictable network conditions, in terms of bandwidth, packet
loss and delay.


2.2 Preliminaries
2.2.1 MC-FGS Video Coding
The video coder employed in this work is based on the MPEG-4 fine granular-
ity scalable (FGS) video coder. The FGS video coding scheme was developed to
cope with dynamic channel conditions and increasingly heterogeneous network en-
vironments [50]. It partitions the compressed bitstream into a base layer (BL) and
a progressively (embedded) encoded FGS enhancement layer (EL) such that the EL
can be truncated at any arbitrary location. The BL carries the most important infor-
14
mation, such as coding mode, low-frequency coefficients and motion vectors (MVs),
while subsequent ELs contain information that can improve the final video quality if
reliably received.


2.2.2 The Unequal Loss Protection (ULP) Scheme
As discussed above, scalable video bitstreams, such as MC-FGS, are highly struc-
tured and prioritized. However, the current Internet provides only a simple first-in-
first-out (FIFO) queuing/scheduling policy regardless of the priority and content of
packets. So it is important to convert a scalable, prioritized bitstream into an ar-
bitrary number of equally important packets that can be decoded independently so
that it could better ”match” the Internet.