Standards & format conversion
Every live event starts with the same problem: contribution feeds arriving in different resolutions, frame rates, colour spaces, and HDR/SDR flavours — each needing normalisation before anything downstream can use them. Norsk handles standard feeds automatically. For suppliers whose feeds claim one format but behave differently, you write a per-source override rule once via API, and it holds at every subsequent event. No manual reconfiguration. No per-event setup grind.
How it works
A single workflow handles the full ingest-to-normalisation pipeline — with Spectrum conversion and optional TAMS integration at its core.
SRT, RTMP, UDP TS, and other contribution formats in production use today. Standard inputs require no per-source setup. Edge cases — suppliers whose feeds behave differently from what their metadata declares — get a per-source override rule configured once via API, not reconfigured manually before every event.
Spectrum converts resolution, frame rate, HDR/SDR, colour space, and aspect ratio to your defined target. Applied globally to standard feeds; explicitly overridden per source for the suppliers that need specific treatment. MQA quality metadata is written into the output at ingest, so quality telemetry travels with the content from this point onwards — no re-analysis required downstream.
Normalised output goes directly to a TAMS (time-addressable media store) — an open standard shaped by BBC R&D, built into Norsk by the engineers who helped define it. Content is accessible for clipping, editing, and replay the moment it is captured. No file transfer, no wait. In live news and sports, that is the window between leading the story and reacting to it.
About TAMS →Use cases
The operations that benefit most are those managing multiple simultaneous contribution feeds — where manual configuration per source simply does not scale.
A major sports broadcaster receives 30 or more contribution feeds per event from different suppliers, each arriving in different resolutions, frame rates, and colour spaces. The answer to 'can we take this contract?' is increasingly shaped by ingest headcount — how many technicians are available to configure feeds in time — rather than editorial or commercial judgement.
Standard feeds normalise automatically. For the suppliers requiring specific treatment, override rules are set once via API and held at every subsequent event — no per-event manual configuration. The pre-event setup grind shrinks to the genuine edge cases. Coverage decisions return to being editorial. Ingest headcount stops being the ceiling on what can be taken on.
Talk to us about your workflow →The gap between 'it was recorded' and 'the editor can use it' is a competitive window. While normalised recordings require a file transfer before anyone can touch them, a competitor working from the same live feed is already clipping the key moment.
TAMS integration means normalised output is written directly to a time-addressable store. Content is accessible for clipping, editing, and replay the instant it is captured — no transfer step, no wait. In live news and sports, that is the difference between leading the story and reacting to it.
Talk to us about your workflow →Dedicated standards conversion and mezzanine encoding hardware is expensive, peak-sized, and idle between events. But replacing it with a cloud alternative often just moves the per-source configuration burden to a new interface. The hardware is gone; the grind remains.
Norsk is cloud-native and scales per-feed on demand. Spectrum handles the normalisation that hardware used to — including precise per-source override for the edge cases that no generic 'automatic' mode reliably covers. Cost tracks actual usage rather than peak capacity. The hardware is gone. So is the grind.
Talk to us about your workflow →Capabilities
The full ingest-to-normalisation pipeline in one workflow — with quality annotation built in from the first frame.
Spectrum converts resolution, frame rate, HDR/SDR, colour space, and aspect ratio to a defined target — GPU-accelerated, applied globally to standard feeds. The normalisation that previously required a technician per source is handled automatically for feeds that behave as declared.
For suppliers whose feeds claim one format but need specific treatment, you write a per-source override rule once via API. Source matched, normalisation applied — regardless of what the metadata declares. The rule holds at every subsequent event without manual intervention. This is what separates normalisation you can trust from generic auto-detect that breaks on the sources that matter most.
SRT, RTMP, UDP TS, and other contribution formats in production use today. Standard inputs need no per-source setup. Edge cases — the supplier whose metadata is wrong, the feed that arrives in an unexpected flavour — get override rules, not manual reconfiguration at event time.
Write normalised output directly to a time-addressable media store. Content is accessible for clipping, editing, and replay the instant it is captured. TAMS is an open standard shaped by BBC R&D — and Norsk is built by the engineers who helped define it.
Quality metadata is written into the normalised output at the point of ingest. Quality telemetry travels with the content from the earliest point in the chain — readable by downstream monitoring including Norsk Probe, without re-analysis or separate measurement infrastructure.
Spin up instances per-feed for an event and scale down after. No hardware sized for peak, no idle capacity between events — cost tracks actual usage. Every configuration — normalisation targets, override rules, scaling parameters — is available via API so your orchestration drives Norsk the way it drives everything else in the stack.
If you manage multiple simultaneous contribution feeds and the per-source configuration tax is a problem worth solving, we want to understand your setup. We are talking to operations teams early to shape format support and override behaviour around real use cases.
Talk to us