TITRE Adaptive Packet Video Streaming Over IP Networks - LaBRI

4.3.2 Video Object-Based Rate Adaptation Algorithm
Let S be a set of MPEG-4 AVOs containing n AVOs O , j with j∈{1, 2…n}. Without loss of
generality, we assume that these objects are sorted in a decreasing order of priority score. Each
object O j may consist of m j layers (m j ≥1). Note that lower layers within an object have higher
priorities than higher layers.
Let P be the function that returns the relative priority score of a particular object or layer.
Without loss of generality, we assume that:
∀j,1
≤
∀j,1
≤
j < n : P(
O
j+
1
) ≤ P(
O
j < n,
∀l,1
≤ l < m
j
j
)
: P(
L
j,
l+
1
) < P(
L
j,
l
)
(Eq. 9)
L j,l is the Layer number l of the Object O j
By using Eq. 9 we can construct an Audio-Visual Entity set called E composed of all object
layers ordered by their priorities.
E= {L 1,1 , L 1,2 …L 1,m1, L 2,1 , L 2,2 …L 2,m2 , …, L n,1 , L n,2 …L n,mn }. We will note E as follows:
E= {e 1 , e 2 , ...,e W } with w= |E|=∑
n
m j
j=
1
Note that if two objects have the same priority score, then the associated layers of an object
have the same priority score as the object (in relation to other objects) with the lower layers having
higher priorities than higher layers.
At time i t , the function R i gives the instantaneous transmission rate of an audio-visual entity.
For example, the audio-visual entity e p has an instantaneous transmission rate equal to R i (e p ), and
the object O j has the instantaneous transmission rate equal to R i (O ). j
Our object-based quality adaptation mechanism operates as follows: The server evaluates the
network state from the information gathered (i.e. RTT and loss rate) at time t , i then computes the
allowed sending rate R TCP using Eq. 1. The server tries to send as much as possible of the audiovisual
entities without exceeding R TCP taking into consideration entities priorities. Details of the
adding and the dropping process will be presented in section 4.3.2.1 and 4.3.2.2 respectively.
Assume that we have an MPEG-4 scene composed of four audio-visual objects: O 1 , O 2 , O 3
and O 4 . Assume that O 1 is composed of a single layer, and that each of O 2 , O 3 and O 4 is composed
of three layers (one base layer and two enhancement layers). Also, as presented in Figure 4-24, we
assume that the associated priorities are as follows:
• O1 is the most important,
• O2 and O3 have the same priority score,
• O4 is the less important.
Then, E= { L 1,1 , L 2,1 , L 3,1 , L 2,2 , L 3,2 , L 2,3 , L 3,3 , L 4,1 , L 4,2 , L 4,3 } = {e 1 ,e 2 , …, e 10 }. Here w=10.
94