Analysis: Live production’s problems have moved up the stack

The live production industry has largely stopped debating whether production should be distributed. The more useful question now is where distribution actually works.

Across technologies and deployment models, a surprisingly consistent boundary has emerged between workloads that virtualize cleanly and those that remain anchored to physical locations, people and deterministic infrastructure. At the same time, many of the problems that once dominated live production engineering are shifting upward, from moving signals to coordinating systems, resources and operators.

That line, not the cloud-migration narrative most coverage still defaults to, is the more interesting story.

Those patterns emerged across responses gathered for a recent Industry Insights roundtable on the state of live production, which included vendors spanning production, monitoring, transport, cloud and infrastructure.

The interoperability problem has moved up the stack

For years, interoperability meant getting a signal from one vendor’s gear into another vendor’s system without a conversion box in between.

That problem is far more mature than it once was. SRT, NDI, JPEG XS, SMPTE ST 2110 and NMOS have standardized much of the plumbing.

But calling interoperability “solved” misses where the friction has gone.

Ian Wagdin, vice president of technology and innovation at Appear, pointed to the remaining problems at system boundaries, where control models, protocols, timing implementations and media formats still have to behave predictably together. Standards such as NMOS and ST 2110 help, but orchestration and end-to-end testing remain part of making a multi-vendor environment function as one system.

Advertisement

Ulrich Voigt, director of live production solutions at Riedel Communications, pushed the problem another level higher.

“The harder problem is coordinating the lifecycle of the multi-vendor infrastructure itself: provisioning and deprovisioning services, assigning resources, applying configurations and monitoring dependencies across vendors,” Voigt said.

Containerization helps, but Voigt noted that not every broadcast application containerizes efficiently yet.

The industry has not eliminated integration.

It has moved from a hardware-integration problem it understood toward a software-orchestration problem it is still learning to manage.

The elastic-versus-sticky divide is structural

The same boundary appears repeatedly in how companies describe cloud migration.

Transcoding, packaging, graphics, overlays and monitoring move relatively easily because they are software-native and their demand is elastic. Camera shading, audio mixing, intercom, IFB and tally are more resistant because they remain tied to latency, physical presence or both.

Wagdin described the same split in infrastructure terms. Rather than treating hardware and virtualization as competing end states, he said organizations increasingly judge each workload on performance, cost and adaptability. Dedicated hardware remains the stronger fit where latency and resilience matter most, while software and cloud capacity provide elasticity for new services, regional productions and temporary peaks.

That makes hybrid production look less like an intermediate phase and more like the likely end state.

Big Blue Marble’s account of ORF’s 2026 Eurovision Song Contest coverage in Vienna makes the point concretely: signal contribution, dual-region cloud distribution and delivery to more than 50 public-viewing locations ran through cloud infrastructure, while camera acquisition, intercom, high-frame-rate replay and latency-critical switching stayed on the ground.

Riedel described the same split more generally.

Even after large-scale cloud deployments proved the quality case, the connection to talent on site, cameras, microphones, tally and IFB can still undermine either the production or the business case.

Advertisement

The industry is not confused about what belongs in the cloud. It largely knows. The challenge is bridging the elastic and sticky parts without creating two operational worlds.

Scale is no longer just a capacity problem

The cloud has also changed what scaling a live production means.

Colin Moran, vice president of production products at LTN, described a production model in which one core event feeds television, streaming, FAST and social platforms, with language, graphics, advertising and presentation varying by destination.

Technically, spinning up additional outputs is getting easier.

Operationally, it is not.

Moran noted that each language feed, graphics package or regional service still has to be managed throughout the event. Three to five versions can now be part of the plan from the beginning, making reusable configurations and shared resources essential.

That exposes a more important bottleneck. Compute can scale quickly. Human attention cannot.

The technical ability to create another feed is becoming less meaningful than the ability to monitor it, understand it and intervene when something goes wrong.

The efficiency story is partly a labor story

Blackmagic Design, Magewell and Telycam all framed tighter integration between switching, graphics and replay as an efficiency gain, allowing smaller crews to run more sophisticated productions without adding equivalent complexity.

That is true, but it is only half the story.

Advertisement

“One of the biggest operational challenges is staffing live events: the number of incoming professionals in today’s workforce willing to work irregular hours and travel extensively for live productions is limited, even as productions grow and become more distributed and complex,” Voigt said.

Moran described the same pressure from the investment side, noting that operational demands are increasing faster than many organizations can expand their teams.

That reframes the consolidation trend.

Packing switching, graphics, replay and other functions into fewer systems is not purely an engineering achievement. It is also a response to an operating model that cannot always assume the traditional number of people will be available to staff it.

Telycam’s observation that automation is raising the skill floor rather than lowering it runs counter to the assumption that fewer operators means lower-skilled work. Operators increasingly need broadcast fundamentals and IP-networking knowledge at the same time.

The industry is asking fewer people to know more, then building software to make that sustainable.

Telemetry has become the substitute for a shared room

The least flashy theme may be the most structurally important.

Boland, Big Blue Marble and Lawo each emphasized monitoring and observability not as dashboard features but as the mechanism that allows distributed teams to trust that they are looking at the same reality.

“Color-accurate, dependable monitoring gives teams a trusted visual reference, helping ensure that everyone is making decisions based on the same picture regardless of where they are working,” said Gary Litwin, national sales director at Boland Monitors.

When a production team shared one control room, “the same picture” was largely a given.

Once the workflow spans venues, remote hubs and cloud regions, that shared reference has to be manufactured deliberately.

Lawo described dashboards using tools such as Prometheus and Grafana to surface only what each operator needs. Big Blue Marble described end-to-end quality telemetry closing the loop between playback and delivery.

Moran adds another layer: as productions create more customized outputs, operators must know not just whether video is present, but whether the correct commentary, graphics, advertising and platform format are attached to each version and where a fault began.

Observability is therefore becoming less about monitoring machines and more about making distributed production understandable.

The actual state of live production

Strip away the individual product pitches and the industry’s problems are moving.

Signal transport is increasingly standardized. The harder work is shifting toward orchestration, workload placement and operational visibility.

Cloud and software can scale quickly, but many things around them cannot. Cameras and talent remain physical. Human attention remains finite. Skilled operators remain difficult to replace. Multi-vendor systems still have to be coordinated as one production environment.

That makes the central question less about what can move to the cloud.

The more useful question is what still becomes difficult after it does. And increasingly, the answer is not bandwidth or compute.

It is making the entire production behave as one system, and giving the people running it enough visibility, context and control to trust what is happening across distance.