sfi

NDP

ENTERPRISE IRELAND

NCNRC

Communications Network Research Institute

VIDAS

Mark Davis Claus Jürgensen Stuart Wallace
Goal: To develop a prototype Statistical Video Quality Analysis tool that employs objective metrics.
Team: Mark Davis (PI), Nikki Cranley (LI), Claus Jürgensen, Stuart Wallace
Project duration: January 2007 - March 2008
Funding Agency: Enterprise Ireland (under Commercialization Fund Proof of Concept program)

Background:

Video streaming services are becoming an increasingly important revenue stream for triple-play and 3G network operators. In particular, IPTV and Video on Demand (VoD) are the driving forces in triple-play services. For such services to succeed it is necessary that the service can be delivered with acceptable end user Quality of Service. Quality of Service for multimedia streaming applications can be decomposed into two key areas: the Quality of Delivery (QoD) and the Quality of Experience (QoE). QoD relates to the end-to-end delivery of the multimedia stream over the network in terms of reliability, packet loss, throughput, and delay. The QoE can be measured in terms of the measured distortion of the decoded multimedia stream. The relationship between QoD and QoE is complex owing to a large number of independent factors including the encoding configuration, packetization schemes, error concealment and correction techniques, as well as the nature of the transmission impairments. Moreover, there is a significant degree of uncertainty regarding the impact of a transmission impairment event introduced as a consequence of the spatio-temporal nature of video streaming.

Objective:

The objective of this project is to develop a prototype Statistical Video Quality Analyzer tool for use by network operators and multimedia service providers to perform video quality analysis and performance evaluation of multimedia streaming services over a diverse range of network technologies. This tool will remove the need to perform lengthy subjective video quality assessment and network trials which are costly in terms of time and resources and which have a poor accuracy and repeatability.

Motivation:

The received video quality strongly depends on which packets in the video stream are lost in the network. These variables include whether the lost packets contain crucial motion vector information, whether the packet belongs to an I, P, or B frame, and whether the packet contains the data for a central part of the image or the edge of the image. In addition to this there are error propagation effects whereby distortions and errors can propagate through subsequent frames due to the inherent frame dependency.

2

We are developing a prototype system that will emulate the impairments caused by poor network conditions on the video stream and measuring the received video quality.

4

Key Benefits

Prototype Details/Requirements

Download:

Flash presentation outlining the functionality of the prototype (HTML)
Flash presentation outlining the functionality of the prototype (SWF)