Solutions Address ABR Performance Challenges End to End
By Fred Dawson
November 11, 2016 – New approaches to generating quality assurance metrics essential to making video a real business in the OTT space cap a year of major developments that promise to shore up the monetization opportunities everyone is looking for.
From the outset of 2016, the combination of massive data flows from every corner of the Internet down to every last user and the big data analytics made possible by ever-more-powerful microprocessors has fueled a steady stream of QA solutions suited to operating in the adaptive bitrate (ABR) environment that dominates IP video consumption. Such capabilities provide content producers and distributors of every description means to ensure the caliber of performance they’re looking for is adhered to in the production, post-production and distribution processes as well as in the processes that enable dynamic placement of video ads in the content streams.
The breadth of these new efforts is illustrated by new advances from Tektronix and by a partnership in solutions integration between IneoQuest, a major player in video QoS (quality-of-service) with a traditional base on the NSP side, and Conviva, equally strong in QoE (quality of experience) solutions for content owners. These initiatives recognize that integration between QoS – enabling fast, often proactive measures against network-based threats to expected service performance – and QoE – tracking and analyzing data from every viewing session that’s pertinent to various business models – is essential to seizing opportunities that come with offering broadcast quality video in an IP-enabled business domain.
As described by Kurt Michel, senior marketing director at IneoQuest, this is what the combination of his firm’s and Conviva’s solutions is meant to accomplish. “We’re binging operational and content worlds together with an intelligent system that serves both sides,” Michel says. “Everyone needs accountability, and accountability needs metrics.”
Publishers and network operators alike have the end-to-end visibility that allows them to quickly address any streaming issues that negatively impact customer experience or advertising performance, he adds. When an error is detected upstream by IneoQuest’s FoQus platform, Conviva’s complementary Intelligent Control Platform immediately identifies the audience impact, he explains For example, when Conviva’s platform detects poor viewer experience, IneoQuest identifies if a quality issue is the cause and accurately points to the problem cause and location.
As previously reported, the recently introduced FoQus platform analyzes the availability of each asset, its quality in the pre- and post-encoder phases and the quality coming out of the CDN, matching that information with how many people are viewing the asset to assess the impact of any malfunction. Going beyond presentation of raw quality metrics based on PSNR (Peak Signal-to-Noise Ratio) or MSE (Mean Squared Error) to get at what the real viewing experience is, IneoQuest employs an on-the-fly execution of the Mean Opinion Score method of grading video, which relies on algorithms that reflect the actual responses of the human visual system.
Convivia, which works with top-tier OTT video broadcasters and operators like HBO, ESPN, Sky and Sony, leverages a real-time map of the Internet video delivery ecosystem to provide 3600 visibility across all users. The Intelligent Control Platform uses this information to maximize picture fidelity and eliminate playback delays and interruptions.
“Their analytics focus on what people are watching, how long the sessions last and what the ad experience is, including the volume of impressions and information about playback performance, across all end points,” Michel says. “We identify network issues. So it made sense to meld their analytics with what we do.”
The companies are developing APIs that can call out information from either platform to provide answers customers couldn’t get relying on one or the other alone. For example, Michel says, “If our system tells a customer there’s a problem with their encoders, they can get data from the Convivia system that lets them know how many people are affected, which helps determine whether the customer needs to add back-up encoding support.” This new level of video intelligence can be used to more accurately assess whether proactive action against an emerging problem is merited, he adds.
At Tektronix, the latest advances build on the incremental steps the company has been taking over the past three years toward creating an integrated quality assurance environment that brings the traditional deep content inspection capabilities of its Sentry platform into the ABR domain.
Notably, the recent enhancements to Tektronix’s Sentry ABR platform deliver end-to-end QoE monitoring for multiscreen content, says Charlie Dunn, general manager for the video product line at Tektronix.
“We integrate with DRMs and decrypt the content to perform analysis,” Dunn says. “We hit the origin servers and go through every manifest.” In the process, along with validating the manifests, the platform goes deep into the content to evaluate the actual viewing experience, he explains.
Along with QoE analysis that identifies frozen video, tiling, macroblocking, audio and other issues on each stream at each bitrate in real time, the platform performs PVQ (Perceptual Video Quality) analysis to generate Mean Opinion Scores. Tektronix has also enhanced its file-based Aurora quality-control (QC) platform, traditionally used to validate distribution readiness of VOD content, to catch the most common causes of ABR streaming problems and to perform PVQ analysis for reviewing stored content at the pre-distribution phase.
In addition, Dunn notes, Tektronix offers passive monitoring with Sentry Passive ABR probes that can be deployed anywhere in the network to enable real-time collection and reporting of key performance indicators at a given location, such as frequency of profile transitions, aggregate user session trends, IP sources and destinations requested. Combined, the processes performed by Aurora and Sentry ABR provide distributors the ability to address QoE issues wherever and whenever they occur with any given streamed video whether in live or on-demand scenarios.
The broad appeal of such solutions is reflected in the expanding customer base for the Sentry product line. “We now have over 20 customers for Sentry ABR worldwide, half of which are not network service providers,” Dunn says.
Farther contributing to a more integrated approach to quality control on the content producer side, Tektronix is offering its Prism hybrid SDI/IP media analysis platform, which diagnoses and correlates SDI and IP signal types in order to identify anomalies at both the content and IP layers in the production process for all formats, including different combinations of UHD with and without HDR (high dynamic range) and WCG (wide color gamut) enhancements. Notably, Prism along with other elements in the Tektronix portfolio was used by NBC Universal to enable control across production, post production, transmission and distribution workflows for its Summer Olympics coverage.
“We see these two worlds coming together as everything moves to IP,” Dunn says, speaking of the needs of traditional TV broadcasters and pay TV operators for end-to-end quality assurance in the OTT domain. “The goal is to get consistent QoE measurement across all points from production through distribution.”
While Sentry ABR users have not applied the platform to QA in the advertising space, the capabilities are there to use when needed, he says. As currently constituted, the platform combines real-time monitoring and alerting with historical auditing, allowing distributors to ensure proper function of insertion technology by identifying and correcting system errors when they occur.
One aspect of advertising in the ABR domain that remains to be addressed pertains to monitoring whether ABR “chunks” devoted to advertising content are in tune with the bitrates assigned to the primary content. “We don’t do that yet,” Dunn says.