It takes a lot more than a stable WiFi connection to deliver that video to your TV, smart phone or tablet at the quality that you’re looking for—let alone finding a way to watch without buffering!
Behind the scenes, the focus used to be on the quality-of-service provided to users. Technicians would base key metrics off of bandwidth, package drop rate and network delay. Today, the concerns have morphed, instead centering around how users will perceive the quality of the stream and if the creative intent of the producers themselves has been properly preserved during the distribution process.
But what if there was a framework that could be used by everyone from engineers to executives to determine exactly where things are falling apart in the distribution channel? This problem is what one man affectionately known as “The Professor” hopes to have solved.
Dr. Zhou Wang, SSIMWAVE Chief Science Officer and co-founder, is proposing a unified end-to-end framework that will determine Quality-of-Experience, a metric that measures “the overall acceptability of an application or service as perceived subjectively by the end-user.” This QoE is meant to take the place of the now outdated QoS metric. Underneath that framework is the Structural Similarity (the SSIM) algorithm that earned him an Engineering Emmy Award back in 2015. It’s now become the most-widely used QoE algorithm in the world, seeing more than 44K academic citations as of July 2018. He’d be quick to tell you that he didn’t do it alone, looking to his other co-founders and PhDs Abdul Rehman and Kai Zeng.
The group will be on-hand at next week’s SCTE-ISBE Cable-Tec Expo in Atlanta. Wang will be presenting his paper on the solution at Monday’s “How to Enhance QoE for Video and WiFi Services” panel at 2:15 pm. The group has exhibited in the past, granting them an opportunity to bring together the scientists behind the tech with the client companies.
“In this world, the Professor and Abdul have become mini-rock stars,” SSIMWAVE VP, Marketing, Saj Jamal said, describing how other tech professionals have literally asked for selfies with the duo.
Wang’s idea may seem fairly obvious. With the rapid growth of video distribution, why not create a more accurate way to determine the visual QoE? In reality, it’s taken 20 years to bring that idea to life. It was in a 1998 paper on video quality assessment where Wang said that a video’s quality should not be determined by measuring a signal. Instead, it should be measured by how a human being sees video. While his peers discouraged him from pursuing the idea, he kept pursuing it. While he would call it an “interesting problem,” Jamal still tells the Prof that “an interesting problem to you is ‘let’s find the most difficult problem on the planet and solve it.’”
The easiest and best solution to ensure smooth, faithful video delivery would be to place human inspectors at every stage, from encoding to routing to rendering and everything in between. Of course, that’s far from viable.
CEO Rehman and Zeng, who were Wang’s students, worked together on how to make that base idea work in the digital age while also future-proofing as much as possible. The end result was what they now call the SSIMPLUS Live Monitor.
The team created objective QoE monitoring probes that should “see” and “behave” like human inspectors, perceiving each and every pixel like humans. They then assign a specific score for the stream at that point in the process. If the score drops at any point in the process, engineers can quickly identify an issue within the stream and fix it, sometimes before a viewer would even notice.
At this point, it’s all about adoption. This form of video monitoring is attempting to shift how an entire industry has approached the issue for decades. SSIMWAVE told CFX that SSIMPLUS has already been deployed to millions of US subs, and if the Prof’s enthusiasm has anything to do with it, things will only grow from here.
The original article was posted on Cablefax.