Quality monitoring, as most of the industry still practices it, is broken. Not broken in the sense that it fails to catch problems, but broken in the sense that it was built for a hardware era that no longer exists, and it burdens the humans running it with more noise than signal. That was the premise behind a recent webinar bringing together Adrian Roe, CEO and co-founder of Norsk, and Brenton Ough, CEO and co-founder of Touchstream, for a conversation about what monitoring looks like when it’s designed around a simple idea: measure once, read everywhere.
The discussion moved from the encoder, where quality data originates, through to the operations center and the audience-facing dashboards where that data actually gets used. Rather than walking through the session point by point, here are some of the key takeaways:
- Why “monitoring” has been solving the wrong problem. The instinct is to treat monitoring as a forensic tool, something that tells you what already went wrong. The discussion challenges that framing and asks what changes when the goal is instead making sure viewers never notice a problem in the first place.
- The physics of signal versus noise in operations centers. Every additional data point an operator has to watch is a data point that competes for their attention during the moment that actually matters. The session digs into where that trade-off breaks down, and why “more visibility” isn’t automatically the right goal.
- What happens when quality data travels with the stream instead of being recalculated at every stage. A recurring thread is the idea that quality metrics — VMAF scores, conformance levels, loudness compliance — don’t need to be rediscovered by every vendor at every stage of a delivery pipeline. The implications of that shift touch DRM, multi-vendor interoperability, and how monitoring tools are built and deployed in the first place.
- The gap between what’s technically possible and what organizations are culturally ready for. Both speakers have spent decades watching broadcasters and telcos adopt (or resist) new approaches to delivery infrastructure. The conversation gets into why some organizations move fast on this and others stay stuck, and it’s not always for the reasons you’d expect.
- Where AI actually earns its place in the monitoring stack (and where it doesn’t). Rather than treating AI as a blanket upgrade, the discussion is specific about what kind of problem it solves: turning a flood of technical events into something an operator can actually act on, without adding another layer of noise.
Along the way, the conversation surfaces a few real-world stories, including a 3 a.m. equipment failure with no good options and an operator problem that scaled from monitoring two streams to twenty, that illustrate why the stakes here aren’t abstract. They’re about stress on the people running these systems, and about what happens to a business when a channel goes dark and nobody can say why fast enough.
What ties it together is a rejection of the assumption that better monitoring means more data flowing to more people. The case both speakers make, from opposite ends of the delivery pipeline, is that smarter monitoring is monitoring you can trust to filter — automating the decisions a system can already make correctly and reserving human attention for the moments that genuinely require judgment.
The full webinar covers each of these threads in far more depth, including the technical mechanics of how quality data gets embedded in a stream, and a live demonstration of failover monitoring in action.
If you’re going to be at IBC, you can find out more at the Norsk stand (5.A53) and the Touchstream stand (5.H54). Or get in touch at info@norsk.video.