IPTV QOE: Picture Imperfect

With network operators pouring princely sums into IPTV, the one thing they cannot afford to hear is that the picture quality isn't as good as what someone might get from a cable or satellite TV provider. In this, which goes by the term quality of experience (QOE), the telcos and others that have gotten into IPTV are both fortunate and unfortunate. They're in luck because there are currently more than a dozen vendors offering gear that purportedly can compute the subscriber QOE. But this is also what makes things tricky for the telcos.

Because IPTV is relatively new, these QOE measurement vendors are hoping to ride the wave of IPTV's predicted rapid growth. More than a few of the vendors have for some time been providing tools for determining the quality of data and voice services. But this does not mean they also have the best solution for measuring video quality. Some specialize in analyzers that are useful for running tests in the lab; others have probes for in-service monitoring; and some others have both analyzers and probes. Some have software that can determine the quality of the video signal received at the set-top box; others claim to have tools that ascertain the quality of the video sent from the set-top box to the subscriber's TV. And some vendors even claim that their video quality measurement techniques take into consideration the human visual system, what the human eye can and cannot perceive – something that's probably not fully understood yet by science.

To make matters more challenging for network operators, there is no international standard for what's called a no-reference measurement, and there's not going to be one for some time. This is the measurement method best suited to measure the quality of an in-service video stream – where the measurement solution computes the quality of a video signal, at some point in the network or even past the set-top box, without having any knowledge of the quality of the original signal received upstream at the headend.

It's okay and even necessary for these QOE measurement vendors to put substantial effort into marketing their solutions. But what's often missing from the marketing effort is a clear picture of how a particular vendor's solution can help the network operator understand if there is an issue with the quality of the video.

If I were a network operator, here are the kinds of things I'd want to know: If I rigorously test in the lab, will this reduce the need to deploy probes in the network? Is there really an advantage for me if one solution decodes at the header level versus one that decodes pixels? What can I know about the video quality if the video stream is encrypted? Since there is a correlation between QOE and network quality of service (QOS), does this reduce the need to measure QOE if I have the right QOS?

Perhaps the answers are there. It's just that they are buried under all the tech-speak and data sheets.

— Sam Masud, Senior Analyst, Heavy Reading

Be the first to post a comment regarding this story.
Sign In