---

Testing a video telephony system has a lot of aspects. One of the most important ones is the resulting video quality of the system. Problem is, there’s no easy way of doing that.

Two approaches exist today for video quality testing:

1. Testing against a known video reference
2. Testing the video without any reference

1. Testing against a known video reference

If we receive a video stream that is known in advance, we can check it against the known stream for discrepancies. This can be done, for example, with IPTV scenarios, where the IPTV servers distribute a movie and the testing clients check the quality of what is received against the known movie.

The problem is, in most other scenarios – this is simply irrelevant – think about real-time communications.

2. Testing the video without any reference

When sending video on today’s IP networks there are bound to be complications. These networks are unreliable and prone to errors, delays and packet re-ordering. By checking the characteristics of the network itself, we can learn things about the quality of the media being sent. For example, too many packet losses = low video quality.

How network conditions affect video quality

How do you score video quality?

Video quality is an essential part of a buzzword called Quality of Experience – in it, you measure the actual quality the user experiences. This means that any scoring system used on video streams needs to emulate the human vision – it should be subjective in a way.

This is exactly what we have achieved with our ProLab^TM Video Quality Testing Solution. This new solution was developed in our Human Vision research initiative and it is based on a set of algorithms that analyze the video stream and correlate it to human perception. I was actually one of the test subjects used for tweaking this system through a large scale survey.

Help us improve the technique further

We are now opening up the video quality survey internationally for people all over the world – we want to fine tune our technique based on a larger number of test subjects.

There are two separate surveys taking place, and you can join any one of them. Each takes about 15 minutes to complete. You should note that you’ll need to download about a gigabyte worth of video files, but you will be able to delete them the moment you complete the survey.

10 of the participants of the survey will receive a $100 Amazon Gift Card each.

To participate click on one of the links below:

• Survey number 1; FTP: ftp://rvsnvq:HumanVision@rvftp.radvision.com/VQS1
• Survey number 2; FTP: ftp://rvsnvq:HumanVision@rvftp.radvision.com/VQS2

The gift cards will be given by April 30, so hurry up!

If you need more information about this new product, please contact Elie Cohen: elic@radvision.com.