scte PRESENTS
The CHANGING FACE OF STREAMING Panel 2: Edge Compute & AI Edge computing and AI will shape the next phase of streaming Steve Miller-Jones, Product Strategy
Mike Gannon of Vecima focused on AI’s role in video optimisation, arguing that artificial intelligence
Director at Netskrt, explained how edge computing is already transforming content delivery. Unlike traditional
Edge computing and artificial intelligence are set to play a growing role in the future of media delivery, but speakers in the Compute & AI session agreed that the technology must ultimately serve one purpose: improving what viewers see on screen. Opening the session, independent consultant and SCTE executive member Ian Nock challenged delegates to define “the edge” – a term widely used across the industry but rarely agreed upon. Nock concluded there is no single definition. “Someone’s edge is somebody else’s centre, and someone’s centre is somebody else’s edge. It all depends on who you are and where you are.” Instead, Nock described the edge as “the logical location for your service delivery” that provides the quickest and most appropriate experience for the user.
is beginning to solve one of streaming’s oldest compromises: balancing quality against bandwidth. Rather than replacing encoders, AI can improve incoming video before it reaches them, removing noise, correcting colour and reconstructing damaged images. “The encoder can only work with what it’s given. AI rebuilds that source material into something much closer to mezzanine quality before encoding even begins. The result is significantly lower bitrates while maintaining – or even improving – picture quality.” He said early operator trials suggest streams can be reduced from around 6Mbps to 2.5Mbps while preserving visual quality, potentially reducing bandwidth, transport and storage costs simultaneously.
CDNs built around large data centres, Netskrt deploys software directly inside internet service provider networks, allowing content to be served much closer to viewers. He said live sport is changing the economics of streaming because huge audiences now watch the same event simultaneously, creating traffic peaks far greater than those generated by on- demand viewing. “The variation between the average traffic during the day and the peak event is much bigger than it used to be. The challenge isn’t that more people are watching video overall, it’s that they’re all watching one thing at exactly the same time.” To cope with this, the company’s software continually analyses network conditions and dynamically chooses the best delivery path for every user. “We use a lot of real- time data from our edge locations. We must understand whether that network is performing well for the users it serves and, if it isn’t, change our decisions in real time.”
He also cautioned against unrealistic expectations of artificial intelligence.
Hamid El Gharbaoui of test automation specialist Witbe argued that the industry has become too focused on
“AI is a wonderful thing, but it doesn’t think. It isn’t intelligent and it doesn’t create anything. It looks creative because we’re very good at pattern matching, but it still relies on information created by people. It’s a fantastic, elevated search engine, but knowing things is not what AI is good at. You still need experts.”
engineering metrics and not enough on the viewer’s actual experience.
16
SEPTEMBER 2026 Volume 48 No.3
Made with FlippingBook - Online magazine maker