TELKOMNIKA ISSN: 1693-6930
A Methodology for Characterizing Real-Time Multimedia Quality of… Yoanes Bandung 1537
2.3.1. Simulating Multimedia Delivery
In the simulation, we were sending audio and video packet along with background traffic. Parameters configured for the delivery of audio were data rate and size of the transmitted
audio. In this research, we used data rate which representing Vorbis technology. Same as the delivery of audio, data rate for video delivery can use the size required for the delivery of
compressed video such as VP8 [14] . Based on this data rate, we determine the length of each packet and the interval between packets. The size of audio and video packets used in the
simulation depends on the data rate and duration of simulation and was calculated using the equation 3.
∗ ∗
3 For mechanism of audio and video delivery, we installed a VoIP application in the first
host as a sender and VoIP application in the second host as a receiver. Similarly, the video delivery used UDP applications installed together with VoIP applications which the first host as
server and the second host as a client. Audio and video packets were transmitted simultaneously from the first host to the second host. We included background traffic so that
approached real conditions in the field. Background traffic was set up using UDP traffic from source to destination. In addition, the delivery of the packets were limited to maximum delivery
time of 100 seconds.
Simulation of multimedia delivery was repeated using several different data rates. Multimedia packets which were consisted of audio and video were sent over the network with
the lowest bandwidth allocation up to the highest during the measurements. Parameters of end- to-end delay and packet loss were also configured according to network characteristics in the
field. 2.3.2. Characterizing Multimedia QoS
In characterizing multimedia QoS, we measured and analyzed the network parameters using objective testing. After measuring and analyzing the multimedia packet delivery, we
investigated the characteristics of multimedia QoS using PESQ [15] and PSNR. PESQ was used to estimate the audio quality and PSNR was used to estimate the video quality. Then we
applied users perception expressed in Mean Opinion Score MOS to characterize multimedia QoS. The correlation between PSNR and MOS can be seen in Table 1 while PESQ using MOS
scale. MOS value was calculated using the E-model formulation [16] . The first step in E-model formulation is calculating the rating factor .
. 4
. .
. .
5 ln
6 The equation for aims to describe disturbance factors on the network that affect
multimedia QoS. Rating factor is calculated using equation 4. The variable of is a factor of
quality degradation caused by delay. The variable of is a factor of quality degradation caused by compression technique and packet loss
. The constants in equation 4, 5 and 6 are the recommended numbers by [16] which are not changed because of the network conditions. To
calculate MOS ITU-T P.800 based on the estimated -value, there are provisions presented in equation 7. ITU-T recommendation G.107 [16] has established a classification of user
satisfaction based on rating factor of the E-model formula and the estimated MOS as shown in Table 2.
, R R R
R ,
7
ISSN: 1693-6930
TELKOMNIKA Vol. 14, No. 4, December 2016 : 1534 – 1544
1538 Table 1. Correlation between
PSNR and MOS Table 2. Relation between MOS and
User Satisfaction
PSNR MOS
upper limit R-Value
lower limit MOS
lower limit User satisfaction
37 4,5
90 4,34 Very
satisfied 32
− 37 3,5
80 4,03 Satisfied
26 – 31 2,5
70 3,60
Some users dissatisfied 20 – 25
1,5 60
3,10 Many users dissatisfied
20 0,5
50 2,58
Nearly all users dissatisfied
2.4. Validating Important Issues of the Methodology