Remote Production Playbook - Professional Essentials Guide

NCS | NEWSCASTSTUDIO.COM

CHOOSE

YOUR MODEL

ONSITE, REMI,

CLOUD OR

HYBRID?

SEPTEMBER 2026

NCS | NEWSCASTSTUDIO

PAGE

GAME PLAN

HOW SPORTS SCALES LIVE COVERAGE

PAGE

WHEN SYSTEMS FAIL

DESIGNING RECOVERY INTO EVERY SHOW

PAGE

AI IN PRODUCTION

WHERE JUDGMENT STILL MATTERS

PAGE

NCS | NEWSCASTSTUDIO.COM

MICHAEL P. HILL

FOUNDER AND PUBLISHER

NCS, the distinctive arrow mark, NewscastStudio,

Navigating the future of broadcasting, the color

red and related marks and trade dress are marks

of NCS | NewscastStudio.

The claims made in this publication are those

of the individual manufacturer or developer. This

publication does not guarantee the accuracy of

any claims made about a product or service and

any such use of a product or service is governed

by the business relationship between the end-

user and manufacturer or developer.

Statements and opinions made in this

publication

are

those

the

person

organization they are attributed to and do

not necessarily represent the opinions of the

publishers or other organizations featured in this

publication. This publication may contain content

that is provided and published as part of a paid

advertising relationship between the publisher

and the organizations featured.

Some product reviews may have been written

based on complimentary review units provided

by the manufacturer or developer. These units

may be retained by NCS.

Some photos in this publication may be file

or stock photography and not depict specific

products, individuals or companies.

NCS | NewscastStudio offers a wide

variety of partnerships for companies and

organizations to reach decision-makers

in television production roles around the

world, including advertising and advertorial

in future installments of this publication

Banner advertising on NewscastStudio.com

Email newsletter advertising

Sponsored partner content

Press release publication

For details, email our advertising sales team

at advertising@newscaststudio.com

DAK DILLON

EDITOR IN CHIEF

JACOB BILLINGSLEY

FEATURES EDITOR

ADVERTISING SALES

PROFESSIONAL ESSENTIALS

AN NCS PUBLICATION

Every live production is now an

architecture decision

The phrase “remote pro-

duction” used to mean one

thing: cameras at the ven-

ue, everything else running

from a hub. That’s not the

whole story anymore, and

this playbook is built around

the question that’s replaced

it — not whether an event is

produced onsite or remote-

ly, but which architecture

actually fits it.

The drivers behind that

shift are practical, not ideo-

logical. Broadcasters and production

companies are being asked to produce

more events, more outputs and more

platforms without a proportional in-

crease in budget or crew. Remote and

cloud workflows make that possible for

some productions and impractical for

others, and the difference usually comes

down to how the event itself behaves, not

what any single vendor promises.

We start with how to make that call

in the first place, then work through the

pieces that make any remote architec-

ture function: the network that carries

it, the control room that operates it, the

virtualization that makes it flexible and

the automation that keeps it manageable

as it grows. A close look at sports pro-

duction shows how those pieces come

together under the industry’s most de-

manding, most repeatable conditions,

and why sports has become the place

where broadcasters prove a workflow

before extending it elsewhere.

From there, we cover what it takes to

turn a single remote workflow into a plat-

form that can support several produc-

tions at once, and what reliability actual-

ly requires once the control

room is no longer down the

hall from the venue. Noth-

ing about this shift lowers

the stakes. A frozen camera

or a dropped feed still costs

the same whether the fault

started at the venue or three

time zones away, and several

of the articles here are really

about that fact — designing for

failure, not just for efficiency,

and knowing before air what

a production can still do when part of the

system stops behaving as planned.

We close with two practical tools: what

your crew needs to work this way, and a

checklist to run before your next produc-

tion leaves the building.

None of this argues for one model over

another. The people we spoke with for

this playbook run truck-based produc-

tions, REMI hubs, distributed teams and

cloud-native platforms — often more than

one in the same week. What they share

is a discipline about matching the archi-

tecture to the event, not the other way

around.

The sections build on each other, but

they don’t have to be read in order. If

you already know which architecture fits

your next event, skip ahead to the pieces

on connectivity, control rooms or reli-

ability. If you’re still deciding, start at the

beginning.

We hope this playbook gives you a

clearer way to make that call.

MICHAEL P. HILL

Founder and Publisher, NCS

HILL

WELCOME

GUIDE SPONSORS

NCS | NEWSCASTSTUDIO.COM

The right production model starts with

the event itself, not the technology stack.

A predictable weekly show with a known

venue and stable connectivity can sup-

port a very different architecture than a

fast-moving, RF-heavy production that

changes by the minute.

That distinction cuts through much of

the debate about remote production. The

question is no longer whether an event is

produced onsite or remotely. It is which

people, systems and decisions need to

stay at the venue, which can move to a cen-

tral facility and which can be distributed

across software and cloud infrastructure.

“It comes down to how predictable the

event is. If the venue’s known, the connec-

tivity’s known and the shape of the show

doesn’t change much from week to week,

remote works well. Where I still want ex-

perienced people physically there is any-

thing fast-moving, RF-heavy or creatively

ambitious, because those are the produc-

tions that change by the minute. The direc-

tor wants a camera somewhere awkward,

or the RF environment starts misbehaving,

and you need someone on the ground who

can just make a call and fix it, not someone

dialling in to find out what happened,” said

Alistair Horne, managing director at Hor-

nets Tech.

Traditional onsite, REMI, centralized,

distributed and cloud production are not

separate categories so much as points on

a spectrum. A production might acquire

pictures and manage RF at the venue,

switch and shade cameras from

a central hub, run graphics in the

cloud and bring in a replay oper-

ator from a third location — all for

the same show.

Choosing among those options depends

on the event’s predictability, creative

ambition, latency tolerance, connectivi-

ty, staffing needs, output count and risk

profile, and on whether the architecture

needs to support one production or be-

come a repeatable platform for many.

Start with the production, not the

location

The first question isn’t where the con-

trol room should be. It’s what the produc-

tion has to accomplish.

A recurring match from a fixed venue

is a different problem than a champion-

ship staged in a temporary environment.

A regional stream with a familiar format

doesn’t require the same operating model

as a global broadcast with specialty cam-

eras, complex RF, multiple languages and

dozens of outputs.

“Remote production is an excellent fit

for repeatable events with well-estab-

lished workflows, such as regular season

broadcasts where consistency and effi-

ciency are paramount. High-profile events

often introduce unique storytelling oppor-

tunities, specialized production elements

and a greater need for close collaboration,

making a larger on-site presence valuable.

For viewers, one element that still carries

significant value is having the talent onsite,

where they can better capture the energy

that makes live sports compelling,” said

Andrew Lahey, senior manager of solu-

tions services and project management at

Ross Video.

Predictability matters because remote

workflows reward standardization. Known

venues, stable connectivity, established

camera plans and reusable templates re-

duce the variables teams must solve in

real time, making recurring sports, region-

al competitions and studio shows strong

candidates for centralization. Productions

built around improvisation, intensive RF

or constant back-and-forth between the

director and the venue tend to keep a larg-

er onsite footprint — not because onsite is

the premium model and remote the econ-

omy option, but because each places peo-

ple and infrastructure differently based on

the work at hand.

Traditional onsite production

keeps decisions close

In a traditional onsite production, most

of the crew and core systems travel to the

event — cameras, audio, switching, replay,

graphics and engineering managed from a

truck, flypack or temporary control room

PICKING THE RIGHT

PRODUCTION

MODEL

Continued on Page 4

REMOTE PRODUCTION

NCS | NEWSCASTSTUDIO.COM

at or near the venue. The model gives di-

rectors and technical managers direct,

fast access to the production environment

when a camera position changes or the

show departs from plan, and it depends

less on wide-area connectivity.

Those advantages carry a cost. Equip-

ment and specialists have to travel, mobile

units spend time moving between events,

and each venue requires setup and testing.

Infrastructure dedicated to one produc-

tion cannot easily support another at the

same time.

Onsite production makes the most

sense when local responsiveness and cre-

ative coordination outweigh the value of

sharing resources elsewhere. It can also

serve as a fallback layer within a remote

production, particularly when losing con-

nectivity would be costly enough to justify

keeping limited switching or transmission

capability at the venue.

REMI brings the control room back

to a hub

REMI separates acquisition from much

of the production operation. Cameras, mi-

crophones and a smaller technical team

stay at the venue while signals travel to a

central facility, where operators handle

switching, replay, graphics, audio mixing,

camera shading and monitoring. The cen-

tral hub is the defining feature: it lets an or-

ganization reuse the same control rooms

and specialist teams across multiple

events instead of rebuilding the complete

production environment at each venue.

“Broadcasters are coordinating major

events through centralized production

hubs connected by IP networks, remote

production (REMI) workflows and cloud-

based infrastructure. Rather than duplicat-

ing crews and equipment at every venue,

operators can manage switching, graph-

ics, camera shading, replay and monitor-

ing from shared facilities while local teams

focus on capturing the events,” said Martin

Lindsay, head of networked solutions at

Sony Electronics.

REMI works best when the production

calendar can keep the hub busy. A facili-

ty built for occasional use may simply re-

place travel costs with underused infra-

structure, while a hub supporting a steady

schedule of matches, news programs or

live events has more opportunities to

share rooms, systems and operators.

REMI is also not the same as cloud pro-

duction. A REMI workflow can run entirely

on dedicated hardware at a broadcaster’s

facility — its defining trait is separating the

venue from the main production location,

not the type of compute running the tools.

Distributed production loosens

the center

Distributed production takes the separa-

tion further, allowing operators, process-

ing and control to reside in several loca-

tions instead of one central control room.

A director might work from a production

hub, a graphics operator from home, a re-

play team from another city and cloud ser-

vices from one or more regions.

“Remote and distributed production

have evolved beyond remote access into a

true production architecture, where cam-

eras, operators, processing resources and

teams can be located across multiple sites

instead of being centralized in a single fa-

cility. IP-based technologies such as NDI,

SRT, VISCA over IP and FreeD connect

these distributed resources into a unified

production environment. This enables

organizations to share expertise and in-

frastructure across multiple projects, in-

creasing scalability and efficiency without

being limited by physical location,” said Ivy

Li, marketing director at Telycam.

The model gives organizations access to

specialists without requiring them to trav-

el or relocate, and it separates the scaling

of people from the scaling of facilities: add-

ing another operator does not necessarily

mean adding another seat in a fixed con-

trol room.

The tradeoff is coordination. Every add-

ed location brings another connection,

monitoring point and timing dependency,

so the workflow should not feel distributed

to the operator even when the resources

behind it are spread across several sites.

Distributed production makes the most

sense when geographic flexibility and ac-

cess to shared talent create measurable

value, and less sense when the effort to

connect and support participants exceeds

the constraints of putting them in one hub.

Cloud production makes

infrastructure temporary

Cloud production moves one or more

production functions onto virtualized in-

frastructure that can be deployed when

needed. Graphics, recording, transcoding,

monitoring and distribution are common

starting points. Switching, replay and cam-

era control are also becoming practical

in more workflows, although latency-sen-

sitive functions and the physical edge re-

main harder to move completely.

“The biggest challenges are at both ends:

at the edge, where cameras, microphones,

IFB, intercom and other talent-facing de-

vices still require physical connections;

and at the core, where the high-perfor-

mance compute and GPU resources need-

ed for virtualized production can be diffi-

cult to secure predictably in public clouds.

A hybrid approach, with predictable core

resources in a private or specialized cloud

and more elastic resources in the pub-

lic cloud, is likely to remain practical for

many live productions,” said Ulrich Voigt,

director of live production solutions at

Riedel Communications.

The economic case rests on elasticity

rather than the absence of hardware. Re-

sources can be activated for an event and

released afterward, useful for temporary

productions, seasonal peaks and services

that don’t justify permanent infrastructure.

But usage-based infrastructure does not

automatically cost less: compute, storage,

connectivity and data movement continue

for as long as the workflow operates, and

an always-on workload may be cheaper

on dedicated infrastructure. Cloud pro-

duction is strongest when its flexibility, de-

ployment speed or reach provides value

beyond a simple comparison of process-

ing costs.

Hybrid is an architecture, not a

halfway point

Most organizations do not need to

choose a pure model. Hybrid production

assigns each function to the environment

that best supports it: acquisition and RF

may stay at the venue, core switching and

audio may run at a central facility, graph-

ics or encoding may draw on cloud ca-

pacity during peak periods, and a remote

operator may control a dedicated system

elsewhere, with local fallback available

even when the primary show is produced

offsite.

“Broadcasters are centralizing orches-

tration, monitoring, shared processing and

specialist resources while retaining local

acquisition and other latency-sensitive

capabilities at the venue. Predictable, con-

tinuously used functions or critical recov-

ery paths may remain on-premises, while

cloud capacity supports temporary peaks

and geographically distributed events.

This hybrid model allows processing ca-

pacity, production tools and specialist

teams to be shared across events, without

making the entire production dependent

on a single facility, network or cloud envi-

ronment,” said Sarah Hackforth, interna-

tional sales director at Big Blue Marble.

Hybrid models can preserve existing in-

vestments while letting selected functions

Continued from Page 3

Continued on Page 5

NCS | NEWSCASTSTUDIO.COM

move when there is a clear operational

case, and they let organizations separate

steady workloads from temporary de-

mand. The risk is several disconnected

operating environments, each with its own

controls, monitoring and support require-

ments.

The aim isn’t to assemble one exam-

ple of every architecture. It’s to make the

boundaries between them manageable,

with consistent operator control, end-to-

end visibility for engineering teams, and

failure procedures that account for depen-

dencies crossing the venue, facility and

cloud.

Measure utilization, not just travel

Remote production is often justified

through smaller traveling crews, but travel

is only one part of the economic picture.

The broader question is how often people,

rooms, processing and network capacity

get used. A centralized control room im-

proves utilization when it supports a full

schedule; a distributed model can make a

scarce specialist available to several pro-

ductions; cloud infrastructure can absorb

short peaks without a permanent build. An

onsite truck can still be the efficient choice

when it arrives as a complete, tested envi-

ronment and the alternative would require

a complex temporary network and inte-

gration effort.

“Teams are expected to produce more

and more content across multiple plat-

forms with fewer resources. Investments

that can support multiple production

models and adapt to changing technology

standards provide greater long-term val-

ue. There is also a growing emphasis on

simplifying operations so teams can spend

less time managing infrastructure and

more time producing content,” said Bob

Caniglia, director of sales operations for

the Americas at Blackmagic Design.

The calculation should include contri-

bution, return feeds, monitoring, inter-

com, support, cloud consumption, facility

utilization, travel and engineering time

— and what the model makes possible.

Producing more events, adding alternate

feeds or reaching competitions that could

not support a traditional truck may create

more value than reducing the cost of an

existing show.

Build a model that can change

No production architecture removes

tradeoffs. Onsite workflows concentrate

people and equipment but make them

expensive to move. REMI centralizes re-

sources but depends on reliable connec-

tions to the venue. Distributed production

expands access to people and systems

but adds coordination points. Cloud pro-

duction provides elasticity but introduces

variable costs and dependencies. Hybrid

models offer choice but can multiply inte-

gration work.

The practical answer is often a model

that can shift by event: a regular-season

production using a small venue crew and

a central hub, a championship adding on-

site specialists and dedicated connectivity,

a lower-tier event leaning more heavily on

automation and cloud resources. The un-

derlying controls and operating practices

should remain familiar even as the place-

ment of resources changes.

The strongest production model is not

the one that moves the most functions

away from the venue. It is the one that puts

each function where it can be operated

reliably, used efficiently and changed with-

out rebuilding the entire workflow.

Continued from Page 4

///

NCS | NEWSCASTSTUDIO.COM

Moving production away from the venue

changes the role of the network. It’s no lon-

ger a link used mainly to deliver a finished

program feed — it becomes part of the

production system itself, carrying camera

feeds, audio, multiviewers, intercom, tally,

control data and return video among peo-

ple and systems that may be separated by

hundreds or thousands of miles.

That makes connectivity an architectur-

al decision rather than a procurement line

item. A remote workflow can have enough

nominal bandwidth and still fail operation-

ally because of variable delay, packet loss,

weak return paths or inadequate monitor-

ing. The question isn’t simply whether a

connection is fast enough. It’s whether the

network behaves predictably enough for

people to produce a live event through it.

“Production teams must manage more

sources, formats and destinations across

IP networks, remote facilities and cloud

platforms, often with smaller teams and no

tolerance for downtime. Addressing this

requires a common operating

model spanning venue-edge contribution,

centralized processing and cloud work-

flows, with high-density, ultra-low-latency

transport and built-in redundancy,” said

Ian Wagdin, vice president of technology

and innovation at Appear.

Bandwidth is capacity, not a plan

Bandwidth planning often starts with

the bitrate of each camera feed, multiplied

by the number of sources. That’s neces-

sary, but incomplete.

A production also needs capacity for au-

dio, intercom, return video, multiviewers,

file transfers, graphics data, telemetry and

control. Some of those services consume

relatively little bandwidth but become un-

usable when congestion introduces delay

or packet loss — a background transfer or

unexpected venue demand can affect a

connection that looked adequate during a

quiet test.

Engineers should model the production

at its busiest point, not its average: simul-

taneous feeds in both directions, protocol

overhead, redundancy and enough head-

room to absorb normal variation with-

out driving the path to its limit. That also

means understanding whether stated ca-

pacity is dedicated, shared or shaped by

the provider under certain conditions.

Not every signal needs the same treat-

ment. A camera feeding the primary

switch may justify a low-latency contri-

bution codec and a higher bitrate, a con-

fidence monitor for a remote producer

can often tolerate more compression, and

a high-quality return may be essential for

camera shading while a lightweight proxy

is sufficient for someone supervising sev-

eral productions. The goal isn’t to com-

The network is now part of the

production itself

CONNECTIVITY

Continued on Page 8

NCS | NEWSCASTSTUDIO.COM

www.lynx-usa.com

661-251-8600

www.yellobrik.com

NCS | NEWSCASTSTUDIO.COM

press every stream as aggressively as pos-

sible. It’s to assign bandwidth according to

what each feed does, how quickly an oper-

ator must react to it and what happens if its

quality declines.

Latency is an end-to-end budget

Latency is often discussed as if the net-

work contributes a single number. In prac-

tice, delay accumulates across camera

processing, encoding, network buffers,

routing, decoding, synchronization, pro-

duction processing and the return path. A

workflow that measures only the transport

leg can miss where operators are actually

losing time.

“One of the biggest lessons is to stop

treating latency as a single number at-

tached to a remote production. Different

buffers, processing stages and network

paths can all add delay, so you need to

measure at multiple points if you want to

know where delay is actually being intro-

duced. Otherwise, you risk compensat-

ing at the end without ever fixing the real

problem,” said Nicola Milburn, technical

sales manager at Hitomi Broadcast.

The acceptable total depends on the

task. A graphics operator may tolerate

more delay than a director calling camer-

as. Intercom and IFB are especially sensi-

tive, since even modest delay can disrupt

conversation and make talent hesitant,

and camera shading, replay control and

remote switching all depend on respon-

sive feedback.

Low latency shouldn’t come at the ex-

pense of stability. Removing too much

buffering makes a stream vulnerable to

jitter; adding too much produces a steady

picture that’s too delayed for the opera-

tor. The useful target is consistent latency

within the production’s operating budget,

not the lowest number achieved during an

ideal test.

Match the connection to the

consequence of failure

Remote productions can use dedicat-

ed fiber, managed IP services, the public

internet, private or public 5G, bonded

cellular and satellite connectivity. These

shouldn’t be arranged into a simple hierar-

chy from professional to unsuitable — each

presents a different combination of cost,

reach, service guarantees, setup time and

operational control.

“Broadcasters are using a mix of dedi-

cated networks, managed IP services, and

secure internet-based transport depend-

ing on the importance and budget of the

event. Redundant signal paths, monitor-

ing, and backup workflows are essential,

because no single connection should be

treated as failure-proof. Low-latency en-

coding and reliable transport protocols

are also making remote production practi-

cal in more locations,” said Abe Abt, senior

product consultant at AJA Video Systems.

A managed service can provide defined

performance, an SLA and an operations

team that understands the path, making it

appropriate for regular or high-value pro-

ductions. Public internet transport offers

broad availability and attractive econom-

ics, particularly when protocols compen-

sate for packet loss, jitter and changing

conditions, and it can be the practical

choice for venues without access to ded-

icated media networks.

A primary camera path for a major event

may require managed capacity and phys-

ically diverse routes. A remote guest or

supplemental feed may work well over

a tested internet connection. A smaller

event may accept more compression or a

simpler backup if that’s what makes cover-

age economically possible.

The correct choice follows the conse-

quence of failure: how long a signal can

be unavailable, whether degraded video

is preferable to no video, and which func-

tions must continue locally if the wide-ar-

ea connection disappears.

Redundancy requires real diversity

Two connections don’t automatically

create a resilient production. Circuits or-

dered from different providers may share

conduit, local exchange facilities, power

or upstream infrastructure. Two cellular

modems may ultimately rely on the same

carrier. A primary and backup encoder

may depend on the same switch, clock or

electrical circuit.

Path diversity starts with mapping those

shared dependencies. The primary and

backup should fail differently whenever

practical — separate last-mile routes, dif-

ferent carriers, independent edge devices,

redundant power and transport protocols

capable of protecting or reconstructing

the stream when one path degrades.

“The biggest lesson is that the public

internet is a fully viable option for prima-

ry sports contribution if you manage the

traffic right. Relying on a single network

provider is simply a single point of failure

during a live broadcast. Real operational

security comes from using transport pro-

tocols with network bonding to tie multi-

ple cellular and fixed networks together,

ensuring that the video stream remains

stable even if an individual connection

fails,” said Peder Boberg, product owner

at Intinor.

Bonded cellular and 5G can serve as a

primary path for some productions and

an independent backup for others. Their

value lies in combining available networks

and adapting when conditions change,

though they still require planning around

spectrum congestion, carrier diversity,

data limits, antenna placement and how

many people are competing for service at

the venue.

Failover behavior matters as much as

the backup itself. The team should know

whether switching is hitless, automatic

with a visible interruption or manual, who

has authority to make the change and how

operators confirm the backup is healthy

before it’s needed.

Synchronization holds the

production together

A remote control room may receive

feeds that traveled over different paths,

used different codecs or encountered dif-

ferent network conditions. Without align-

ment, cameras can arrive several frames

apart, audio can drift from video and

switching between sources can expose

discontinuities.

Timing design starts at acquisition and

continues through transport, processing

and monitoring. At the venue, sources may

share reference and time information;

across the wide-area network, gateways,

buffers and frame synchronization bring

feeds into a common production timing

domain, keeping audio, video, metadata

and control associated even when they

take different paths.

“Broadcasters have learned that reliable

remote production depends on consisten-

cy more than chasing the lowest possible

latency. Stable timing and accurate syn-

chronization give production teams the

confidence to work across different loca-

Continued on Page 9

Continued from Page 6

“Reliable remote

production depends

on consistency more

than chasing the lowest

possible latency.”

NCS | NEWSCASTSTUDIO.COM

tions without affecting the viewer experi-

ence,” said Andy Rayner, chief technology

officer at Appear.

Return paths are part of that timing

problem too. Directors need a multiview

that reflects the sources they’re switching,

talent needs program audio and IFB that

support natural interaction, and camera

operators need tally and instructions that

match what’s happening on air. A contribu-

tion path can’t be evaluated independently

from the feedback loops that let people

use it.

Monitor the path and the

production

A decoder producing a picture doesn’t

prove the workflow is healthy. Operators

need to see packet loss, jitter, bitrate, buf-

fer behavior, path switching, timing status

and device health alongside the media it-

self, with enough context to know whether

a problem affects one feed, one venue, one

cloud region or the entire production.

Monitoring should follow the signal

from acquisition through production and

delivery, correlating network telemetry

with media checks and production meta-

data rather than leaving each vendor’s

dashboard as a separate island. Alerts

should identify the affected service and

likely point of failure, not simply report

that a threshold has been crossed.

“Every additional version creates an-

other output that must be protected and

monitored. Any fault affecting one output

should be isolated before it reaches the

main programme or another destination.

Alerts also need to show which services

are affected and where the issue began.

This allows operators to focus on the rel-

evant part of the workflow and activate

recovery quickly, which is essential when

even a short interruption can affect live

coverage,” said Colin Moran, vice presi-

dent of production products at LTN.

Monitoring also has to be usable. A wall

of green indicators can hide a developing

problem, while a flood of undifferentiated

alarms slows the response. Dashboards

should be organized around the decisions

each role must make, with engineering

able to reach deeper data when diagnosis

is required.

Design for the degraded state

The final test of a remote production

network isn’t how it performs when every

path is healthy. It’s what the production

can still do when capacity drops, a route

fails or a venue loses contact with the hub.

A resilient design may lower a stream’s

bitrate, reduce the number of returns,

move to a backup path or preserve only

the feeds required to keep the program

on air. Some productions may retain local

switching or recording; others may priori-

tize contribution over remote monitoring

until full capacity returns. These choices

should be made before the event and writ-

ten into clear operating procedures.

Testing must include realistic failure,

not merely confirmation that the backup

connects. Teams should interrupt the pri-

mary path, load the network, force a de-

vice failure and rehearse communication

when normal tools are unavailable, and

know how the production returns to its

preferred state once the fault has passed.

Connectivity becomes the production

backbone when it’s engineered with the

same discipline once applied to the truck

or control room. Capacity matters, but so

do consistency, timing, visibility and re-

covery. The network is ready when oper-

ators can trust it during routine moments

and know exactly what will happen when

the routine breaks.

Continued from Page 7

///

10 NCS | NEWSCASTSTUDIO.COM

The control room was once easy to lo-

cate. It was the room with the switcher,

audio console, replay systems, graphics

engines, multiviewers and the people op-

erating them. Its physical boundaries also

defined much of the production workflow.

Those boundaries are becoming less

useful. Acquisition may remain at the ven-

ue while switching runs from a central fa-

cility, graphics operate in software and a

replay operator works from another city.

The equipment and crew no longer need to

share one room, but they still need to be-

have as if they do.

That is the central challenge of the dis-

tributed control room. Moving individual

functions is technically possible. Turning

them into one coherent operating environ-

ment is harder.

The room becomes a control layer

A distributed control room is not simply

a traditional gallery accessed remotely. It

separates the operator’s interface from the

location of the processing behind it.

The switcher panel may control software

running in a production hub. A browser in-

terface may configure a graphics system in

the cloud. A hardware panel at one facility

may be reassigned from one production to

another without changing the underlying

room. What matters is not where each ap-

pliance sits, but whether the operator can

reach the required resources with predict-

able control and feedback.

“We are seeing a move towards switching

systems becoming more software-driven

and location-agnostic, allowing production

teams to access resources wherever they

are needed and scale as required. Rather

than being tied to a specific control room,

switching capabilities can now be distribut-

ed across centralized facilities, remote pro-

duction centers, and cloud environments.

This gives operators and program makers

greater flexibility while helping organiza-

tions make more efficient use of technical

resources,” said Narinder Ball, director of

technology at Gravity Media.

This allows control rooms, processing

and operator positions to be assembled for

each production rather than permanently

wired together. The same room can sup-

port different events as software assigns

the required resources.

The operational model, however, must

remain legible. If every production requires

engineers to rebuild signal paths, permis-

sions and interfaces, physical flexibility has

simply been exchanged for configuration

work.

Distribute resources, not

complexity

The case for distributed control is stron-

gest when it makes scarce or expensive re-

sources easier to share. One replay system

may support several productions. A graph-

ics operator may work across regions. A

central facility may provide switching and

monitoring for multiple venues without du-

plicating a complete control room at each

one.

That efficiency depends on abstraction.

Operators should not need to understand

which server, network path or facility is

providing every function during normal

production. They need a clear view of the

CONTROL ROOM

Continued on Page 12

Inside the distributed control room

NCS | NEWSCASTSTUDIO.COM

Drop a remote-production photo: OB truck, venue uplink, or remote gallery

Prob a bly

S E E T H E W H O L E C H A I N

tagvs.com

The most expensive word in remote production

Your show travels a link you don't own, out of a venue you're not standing in.

“Probably fine” is a bet on infrastructure with no SLA.

TAG watches your feed leave the venue and watches it arrive. If something went

wrong in between, you know before your audience does.

Deploy anywhere, see the whole chain.

12 NCS | NEWSCASTSTUDIO.COM

sources, controls and system state relevant

to their role.

“Control rooms have accumulated more

systems, but the operator still has the same

two hands and the same pressure when

something changes on air. The priority

should be bringing switching, graphics, au-

dio, routing and monitoring into workflows

that make sense to the person using them,

so operators can focus on the production

rather than another layer of controls,” said

Ole-Andreas Løvland, chief executive of

Bitfocus.

Abstraction cannot mean removing the

ability to diagnose or intervene. When

something fails, engineering teams need

to see the underlying route, device, appli-

cation and timing state. The interface must

hide routine complexity without conceal-

ing the information required to recover the

production.

A unified workflow gives each operator

the right level of control while preserv-

ing deeper technical visibility. A simplified

dashboard may only appear easy until the

first unexpected condition occurs.

Software does not eliminate the

panel

Browser interfaces and software-defined

controls make distributed production prac-

tical. They can be deployed quickly, updat-

ed centrally and tailored to a specific role.

They also allow an operator to reach sys-

tems that are not physically present.

Dedicated surfaces still have value. Di-

rectors, audio mixers, replay operators and

technical directors rely on speed, tactile

feedback and muscle memory when deci-

sions must be made immediately. A gener-

al-purpose interface may be appropriate

for setup, supervision or occasional control

while a purpose-built panel remains the

better tool for continuous operation.

The distributed control room therefore

does not require every function to become

a window on a laptop. It requires the con-

trol surface and the processing to be sep-

arable. A physical panel can remain famil-

iar even as the resource behind it changes

from one production to the next.

“The driver is density: more sources,

more formats (HDR, HFR, multiple resolu-

tions), and more simultaneous outputs for

different distribution platforms. Hardware

that used to be fixed-function is being re-

placed by processing platforms that can be

reconfigured in software as requirements

change. At the same time, the human inter-

face has to get simpler even as the underly-

ing system gets more complex. The control

room of today needs to abstract that com-

plexity without hiding it. Operators need to

trust the system while retaining the ability

to intervene precisely and quickly when

something goes wrong,” said Ian Wagdin,

vice president of technology and innova-

tion at Appear.

The design question is not hardware or

software. It is which interface lets a person

perform a task accurately under live condi-

tions, and whether that interface can follow

the production when resources move.

One production needs one shared

state

Once the control room spans several lo-

cations, every participant must work from

the same understanding of what is happen-

ing. Source names, tally, routing, clocks,

multiviewers, permissions and production

status cannot diverge by location.

A director calling a camera from a de-

layed multiview, an audio operator hearing

a different return or an engineer looking at

stale routing data is not part of the same

control room in any meaningful sense. Con-

sistency across interfaces matters as much

as access to the underlying functions.

“The biggest challenge is that signal flow

is no longer confined to a single room or

a single signal type. Once you are dealing

with mixed environments across IP, legacy

infrastructure, virtual clients, and third-par-

ty systems, complexity tends to show up at

the boundaries: format translation, control

visibility, network behavior, and making

sure the operator still experiences the sys-

tem as something simple and predictable,”

said Dave MacKinnon, vice president of

product management at Clear-Com.

This is why orchestration and monitoring

become part of control-room design rath-

er than supporting systems around it. The

production needs a common control layer

that understands which resources are ac-

tive, how they are connected and what each

operator is allowed to change. It also needs

monitoring that shows whether commands

took effect and whether every location is

receiving the expected result.

The distributed control room succeeds

when its distribution becomes operation-

ally unremarkable. Operators should know

what they control, see the same production

state and respond with the same confidence

they would have in one physical room. The

technology can live almost anywhere. The

workflow still has to feel like one place.

Continued from Page 10

///

NCS | NEWSCASTSTUDIO.COM

Virtualization only helps if it

reduces the work

VIRTUALIZATION

Virtualization promises to let production

resources move, scale and change without

rebuilding the underlying facility. A function

once tied to a dedicated appliance can run

as software on shared infrastructure, in a

private data center, in the public cloud or

across several environments.

That freedom is useful only if the work-

flow becomes easier to operate.

Moving a familiar production function

into software does not remove its configu-

ration, monitoring, security or recovery re-

quirements. It can instead scatter those re-

sponsibilities across applications, compute

platforms and vendors.

The central question is therefore not how

much of the production can be virtualized.

It is whether virtualization reduces the

work required to launch, operate and sup-

port it.

Virtualization is separation, not a

destination

Software-defined production separates a

function from the hardware box that once

defined it. Switching, graphics, monitoring,

recording or transcoding can be deployed

on common compute rather than pur-

chased as fixed-capacity appliances.

That doesn’t mean every application must

run in the public cloud. The same software

may operate on dedicated servers at a facili-

ty, shared compute in a private environment

or infrastructure rented for a specific event.

The value lies in being able to choose and

change that placement without redesigning

the workflow.

“Organizations are increasingly achiev-

ing long-term flexibility through architec-

tures that separate workflows from specific

hardware. They are adopting software-de-

fined, orchestrated environments that allow

resources to be deployed where they make

the most sense operationally and econom-

ically, whether that’s on-premises, in the

cloud, or across both. This approach allows

them to continue leveraging existing infra-

structure, while gaining the ability to scale,

add new services, improve resilience, and

adapt to future requirements without dis-

ruptive rebuilds,” said John Mailhot, senior

vice president of product management at

Imagine Communications.

Cloud-hosted and cloud-native systems

also shouldn’t be treated as interchange-

able. An application moved from a dedicat-

ed server to a virtual machine may gain a

new location without gaining elasticity — a

cloud-native system is designed to allocate

Continued on Page 14

14 NCS | NEWSCASTSTUDIO.COM

resources dynamically, recover compo-

nents independently and update without

treating the application as one large, fixed

installation.

Simply relocating software can preserve

the same operational constraints while add-

ing cloud billing, data movement and anoth-

er infrastructure layer.

Move the workload, not the old

assumptions

The first functions virtualized are often

those already driven by software and those

with variable demand. Graphics, record-

ing, transcoding, monitoring and distribu-

tion can be easier to deploy temporarily or

scale for additional outputs. Functions tied

closely to physical devices, deterministic

performance or immediate operator feed-

back tend to remain on dedicated resourc-

es longer.

That dividing line isn’t permanent — it

shifts as software and infrastructure im-

prove. But each workload should be evalu-

ated on what virtualization changes beyond

its location. Can it be launched from a test-

ed template? Can capacity increase without

a manual rebuild? Can the same configu-

ration run in more than one environment?

Can the application recover without taking

down the complete production? Can it be

removed cleanly when the event ends?

If the answer to those questions is no,

the system may be virtualized in a technical

sense while remaining fixed in practice.

The lifecycle is the difficult part

Moving media among software applica-

tions has become more practical. Coordi-

nating the applications through an entire

production remains harder.

“In cloud environments, moving signals

between different vendors is no longer

the biggest challenge, with protocols such

as SRT, NDI and JPEG XS having made in-

teroperability relatively straightforward.

The harder problem is coordinating the

lifecycle of the multi-vendor infrastructure

itself: provisioning and deprovisioning ser-

vices, assigning resources, applying con-

figurations and monitoring dependencies

across vendors. Containerization helps, but

not every broadcast application is efficient-

ly containerized, so the industry still needs

pragmatic orchestration approaches for

managing mixed hardware, software and

cloud resources,” said Ulrich Voigt, director

of live production solutions at Riedel Com-

munications.

A production may depend on several ap-

plications starting in the correct order, re-

ceiving the right permissions, discovering

the expected sources and reporting their

status to a common monitoring system.

One configuration change can affect timing,

routing or downstream capacity, and turn-

ing the workflow off can be as important as

turning it on, particularly when unused re-

sources continue generating cost or retain-

ing sensitive material.

Orchestration is what turns a group of vir-

tualized products into an operating model.

It should apply known configurations, val-

idate dependencies, expose failures and

return resources when they are no longer

needed. Without it, flexibility becomes a se-

ries of engineering tickets.

Scaling compute doesn’t scale the

operation

The ability to create another process-

ing instance does not mean a production

can support another output without effort.

Each new feed may require graphics, audio,

metadata, monitoring, quality control and

someone responsible for its performance.

Software-defined infrastructure delivers

value when those requirements can be

repeated through templates and shared

services rather than rebuilt for every out-

put — scaling compute without scaling the

surrounding operation simply moves the

bottleneck.

“Buyers should assess whether an ar-

chitecture delivers redundancy, end-to-

end visibility and operational continuity

across cloud, hybrid and on-premises en-

vironments, while allowing workloads to

run where they make the most operational

and economic sense. The real value of soft-

ware-defined infrastructure is its ability to

help broadcasters launch new channels,

regional variants and live-event workflows

without a proportional increase in dedi-

cated hardware, engineering effort or op-

erational complexity,” said Srinivasan KA,

co-founder and president of global business

at Amagi.

Visibility has to follow the workload too.

Teams need to compare resource use, per-

formance and failures across environments

using consistent measures, or moving an

application simply creates a new monitor-

ing island.

Flexibility should remove

exceptional work

The most useful measure of virtualization

is not how many applications run in soft-

ware. It is how often the production team

can use a proven process instead of engi-

neering an exception.

A successful platform can deploy the

same workflow for a larger event, move a

workload when requirements change and

restore service without forcing operators to

learn a different production model. It makes

infrastructure placement an implementa-

tion decision rather than the organizing

principle of the show.

Dedicated hardware will continue to

make sense where it offers the clearest

combination of performance, reliability and

simplicity. Virtualized systems will make

sense where portability, reuse or temporary

capacity reduces operational effort. The

strongest architecture can use both without

requiring teams to manage them as unrelat-

ed systems.

Virtualization creates flexibility by sepa-

rating software from place. It creates value

only when the resulting production has few-

er special cases, clearer visibility and less

work between an idea and a reliable live

service.

Continued from Page 13

///

NCS | NEWSCASTSTUDIO.COM

Automation should handle the

routine, not the judgment

Live production has always depended on

automation. Macros, rundowns and preset

configurations have long helped operators

execute sequences that would be slow or

error-prone by hand.

What has changed is the reach of those

systems. Automation can now configure

workflows, route signals, generate metada-

ta, create clips, check quality and prepare

multiple versions of the same production.

AI can identify moments, detect anomalies

and surface material for an operator to re-

view.

The question now is where automation

reduces repetitive work without making a

production harder to understand or recov-

er when it departs from the plan.

Automate what should happen the

same way

The strongest automation candidates are

tasks with known inputs, defined rules and

repeatable outcomes: a production tem-

plate that establishes routes, loads graph-

ics and assigns processing; metadata that

applies the correct language, advertising

or distribution rules to each output; quality

checks that catch missing audio, frozen vid-

eo or configuration errors before air.

These functions don’t remove the oper-

ator. They reduce the routine actions com-

peting for the operator’s attention.

“Automation is reducing the amount of

manual intervention required to manage

increasingly complex workflows. Rather

than performing repetitive operational

tasks, teams are spending more time over-

seeing workflow performance, exception

management, and service quality. Orches-

tration platforms help operators manage

larger environments while maintaining

consistency and operational control,” said

Heather Mellish, vice president of global

sales at Zixi.

Automation is less useful when a task de-

pends on context that can’t be expressed

reliably as a rule. A template can prepare a

show, but it can’t anticipate every way a live

event will depart from plan. The objective

is consistent execution where it helps, not

removing judgment from situations that re-

quire it.

Exceptions become the operator’s

work

As systems handle more routine actions,

the operator’s role shifts toward supervi-

sion — an improvement only if abnormal

conditions stay easy to recognize.

Poor automation can conceal a chain of

decisions behind a single status indicator,

leaving an operator who knows something

failed without knowing which step failed,

what the automation already changed or

whether repeating the command will make

it worse.

Useful automation reports progress,

exposes dependencies and identifies the

service affected by a fault, with alerts orga-

nized by consequence rather than an undif-

ferentiated list of technical events — what

requires action now, what can wait and

what the system has already attempted.

This is management by exception in

practice: automation handles the expected,

people decide on the rest.

Editorial control begins with the

trigger

Automation doesn’t eliminate editorial

control just because one action triggers

several outcomes. Control rests in who de-

fines the trigger, which actions follow and

whether a person can stop or change the

sequence.

“The point many people miss is that edi-

torial control sits in how the automation is

triggered. Intelligent automation can fire

multiple processes and devices from a sin-

gle operator click, or automatically from an

event, for instance a game moment trigger-

ing not just on-screen graphics but lighting

changes and visual effects in the venue

and direct social messaging to fans. Far

from limiting editorial control, this opens

up creative possibilities that simply were

not practical when every action had to be

found and fired by hand,” said Ole-Andreas

Løvland, chief executive and co-founder of

Bitfocus.

A producer may decide a verified scoring

event should update graphics, mark a re-

play and create a social clip — that doesn’t

mean every detected moment should pub-

lish automatically, since different outputs

carry different editorial, commercial and

rights implications.

AI should surface options, not

hide decisions

AI expands the range of tasks that can be

assisted because it can work with unstruc-

tured material. It can suggest highlights,

generate descriptive metadata, locate con-

tent and identify quality problems across

large numbers of feeds.

“AI is becoming a powerful tool for im-

proving efficiency, whether through au-

tomated clipping, metadata generation,

content discovery, or quality monitoring.

However, editorial decisions still require

human judgment, especially in live sports

where context and storytelling are critical.

The most successful deployments use AI to

augment production teams rather than re-

place them,” said Yang Cai, chief executive

and president of VisualOn.

The useful distinction is between rec-

ommendation and authority: a system can

identify a highlight or a warning, but the

workflow still needs clear rules for when

a person approves, when the system acts

automatically and how that action is traced

afterward.

Preserve the ability to intervene

Every automated workflow should an-

swer a few basic questions for the opera-

tor: What is happening? Why did it happen?

What will occur next? Can it be stopped?

What happens if it fails?

Those answers require visible status,

manual override, audit history and a fall-

back independent of the failed automation.

Teams also need to test exception paths,

not merely demonstrate that the intended

sequence works.

The most effective automation is rare-

ly the most conspicuous part of the pro-

duction. It removes repeated steps, keeps

outputs consistent and brings exceptions

forward. People remain responsible for the

choices that shape the program; automa-

tion gives them more time and better infor-

mation to make them.

AUTOMATION

16 NCS | NEWSCASTSTUDIO.COM

Sports turned remote production

into a repeatable platform

Remote production did not become es-

tablished through one decisive event. It

advanced through schedules.

Sports gave broadcasters the same ven-

ues, competition formats and production

requirements week after week. That repe-

tition created a practical place to test con-

tribution paths, centralize individual func-

tions and refine workflows over hundreds

of events. At the same time, live competi-

tion left little tolerance for delay, missing

pictures or confusion between the venue

and control room.

That combination made sports an un-

usually effective laboratory. The industry

could prove remote production on recur-

ring events, measure whether it improved

resource use and then apply the model to

more complex coverage. The lesson was

not simply that a show could be produced

somewhere else. It was that a repeatable

production system could support events

with very different budgets, audiences and

editorial expectations.

The first opportunity was more

coverage

Remote production is often described

as a way to reduce travel and onsite staff-

ing. In sports, its more consequential ef-

fect has been to change which events can

justify live coverage at all.

A traditional mobile production carries

a substantial cost before the first camera

is positioned. The truck, crew, travel and

setup may be reasonable for a major rights

package but difficult to support for a re-

gional competition, lower division, Olym-

pic qualifier or niche sport. A centralized

workflow changes that calculation by al-

lowing production infrastructure and spe-

cialist operators to be used across a larger

schedule.

“Remote production is now well suited

to many league matches, regional compe-

titions and secondary events where cost

efficiency and scalability are priorities.

Major international tournaments and pre-

mium live events still benefit from larger

on-site teams because of the complexity

of production, the volume of feeds and the

need for close coordination around edito-

rial decisions. Increasingly, though, we’re

seeing hybrid models that combine cen-

tralized production with targeted on-site

expertise,” said Srividhya Srinivasan, chief

technology officer at Amagi.

The lower tiers were not merely a re-

SPORTS PRODUCTION

Continued on Page 17

NCS | NEWSCASTSTUDIO.COM

duced version of premium television. They

gave production teams a high volume of

events on which to standardize camera

plans, test remote communications and

learn how much support a venue actually

needed.

World Archery offers one example. The

organization brought production into a

central facility and adopted an IP-based

workflow, giving it greater consistency

and more control over how events were

produced and distributed. The important

result was not simply a smaller venue foot-

print. A federation with a global schedule

gained a production system it could reuse.

That is where the economics of remote

sports production diverge from a cost-cut-

ting exercise. The same infrastructure can

make additional events viable, extend a

season’s coverage and give a league or

federation more control over events that

previously depended on outside produc-

tion capacity.

A schedule becomes a production

platform

The model becomes more powerful

when several events overlap. College

conferences, regional leagues and early

rounds of tournaments may have multiple

games happening at once. Building a com-

plete production island for each venue du-

plicates switching, replay, graphics, moni-

toring and specialist talent at the moment

demand is highest.

A central platform can allocate those re-

sources across the schedule. Some capac-

ity remains assigned to a game from start

to finish, while other systems and staff can

move between events or support several

at once. The value comes from utilization:

a replay room or graphics team that would

otherwise travel and sit idle between pro-

ductions can remain active across a full

day.

Sky Sports+ is one concrete example

of that relationship between shared infra-

structure and expanded coverage.

“The clearest lesson is that cloud and

remote production can meet the demands

of major live sport, but only when broad-

casters treat them as a new operating

model rather than a direct replacement

for existing equipment. Resilience has to

be designed in, workflows need to be test-

ed under real pressure and staff need time

to become confident in a different way of

working. The proof is now public and at

scale: Sky Sports+ launched in summer

2024, delivering more than 50 percent

more live sport at no extra cost to viewers,

made possible by cloud production,” said

Tim Jackson, senior director of sales and

business development for North America

at Techex.

The significant figure isn’t a reduction in

crew or infrastructure. It’s the increase in

live inventory. Remote and cloud produc-

tion allowed the service to offer substan-

tially more events without treating every

additional game as another full traditional

build.

That shift requires planning at the port-

folio level. Control rooms, operators, li-

censes and processing capacity must be

scheduled across events. A delay in one

game can collide with the next booking,

and a workflow that works for one produc-

tion may fail when several venues request

the same shared resource. Sports forced

broadcasters to solve those problems be-

cause the schedule made them impossible

to treat as occasional exceptions.

Specialists can serve the event,

not the location

Centralization

also

changes

how

high-value production functions are as-

signed. Replay, graphics, camera shading

and commentary can require experienced

operators or specialized systems that are

difficult to duplicate at every venue. Re-

mote access allows those resources to fol-

low the event without physically traveling

with it.

One public-cloud replay deployment

shows how specific the opportunity can

become.

“Replay is becoming more accessible

through simple, affordable software-based

systems that deliver high-quality creative

output without requiring extensive, pur-

pose-built on-premises equipment. We’re

seeing this in practice. For example, a

major U.S. broadcaster is using Riedel’s

SimplyLive in the public cloud for super

slow-motion (SSM) replays, with a venue

gateway on the OB trucks sending the SSM

phases into the cloud. This allows one SSM

backend and one operator to cover multi-

ple games across the country in a single

day, while keeping the technical footprint

in the OB truck to a minimum,” said Ulrich

Voigt, director of live production solutions

at Riedel Communications.

This is more than moving a replay oper-

ator out of a truck. It separates a special-

ized production resource from a single

event and turns it into a service that can be

scheduled across games. The same prin-

ciple can apply to graphics, off-tube com-

mentary, quality control and other func-

tions with uneven demand.

It also creates a new constraint. Shared

specialists cannot be scheduled as if every

event will follow its clock perfectly. Over-

time, weather and technical delays must

be built into resource plans. Remote pro-

duction makes expertise more accessible,

but it does not make that expertise unlim-

ited.

One event now feeds many

products

Sports also expanded the definition of

production output. A game may need a pri-

mary broadcast, alternate commentary, lo-

calized graphics, streaming feeds, vertical

video, instant highlights and social clips

while the event is still in progress.

These are not separate editorial worlds.

They draw from the same live pictures, au-

dio, metadata and moments. Remote and

cloud workflows can make those assets

available to teams and automated pro-

cesses without building a separate chain

for every destination.

NBCUniversal and ViewLift illustrate the

shift.

Continued from Page 16

Continued on Page 18

18 NCS | NEWSCASTSTUDIO.COM

“Broadcasters like NBCUniversal are

using AWS Elemental Inference to deliv-

er vertical video on Peacock in real-time

during major live events, reaching fans on

mobile with just 6 to 10 seconds of latency.

ViewLift is doing the same, transforming

what used to be manual, labor-intensive

clip generation into automated workflows

that deliver results in minutes during live

broadcasts. The through-line is the same:

get content to audiences on social and mo-

bile while the moment is still relevant, not

hours later,” said Steph Lone, global leader

of solutions architecture for media and en-

tertainment at Amazon Web Services.

The remote-production lesson is that

scale is no longer measured only in the

number of games. It also includes the

number of useful products created from

each game. A centralized workflow can

keep the primary production intact while

making its sources and metadata available

for additional outputs.

That changes the business case. Rights

owners can extend one event across more

platforms, sponsors and audience seg-

ments. The production investment sup-

ports a wider content operation instead of

ending when the final whistle sounds.

Marquee events test the model

differently

The growth of remote production at

lower tiers did not confine it there. Premi-

um sports have adopted centralized cam-

era control, replay, graphics, commentary

and other remote functions, often as part

of a larger hybrid operation.

During the 2026 global soccer tourna-

ment, FOX Sports used a standards-based

REMI architecture across 104 matches

and multiple venues. NBC Sports also

made remote production a central part

of its Winter Olympics coverage. These

events tested whether reusable systems

could adapt to venues with different tech-

nical demands while maintaining local ac-

quisition and operational control.

They also demonstrated why the sports

laboratory does not produce one univer-

sal model. A championship carries more

feeds, specialty cameras, talent, security

requirements and editorial coordination

than a routine league match. The value

of remote production is not measured

by how few people remain onsite. It is

measured by whether each role is placed

where it can contribute most effectively.

Camera operators, field producers, ven-

ue engineers and talent often benefit from

proximity to the action. Central teams

can handle functions that gain more from

shared infrastructure, consistent facilities

or access to specialists. The balance may

change by event, venue and even produc-

tion day.

This is a more mature use of remote

production than simply replacing the

truck. The architecture becomes a set of

choices that can expand or contract with

the event while preserving a familiar oper-

ating model.

What the laboratory proved

Sports showed that remote production

works best when it is designed as a repeat-

able platform rather than a cheaper copy

of an onsite show.

The schedule rewards standardization,

but live competition guarantees excep-

tions. Shared resources improve utili-

zation, but only when they are planned

across overlapping events. Centralized

specialists can improve access to exper-

tise, but delays and peak demand still re-

quire capacity. Additional outputs create

new value, but each one still needs edito-

rial ownership and monitoring.

Those lessons extend beyond sports.

Any organization producing recurring live

events can benefit from reusable configu-

rations, shared production resources and

infrastructure sized around a portfolio

rather than one show. What sports adds

is evidence at both ends of the market: re-

mote workflows can make smaller events

possible and support some of the largest

productions in the world.

The lasting change is not that the control

room moved away from the stadium. It is

that sports production became a platform

capable of covering more events, creating

more versions and assigning resources ac-

cording to the needs of each competition.

The laboratory is now an operating model.

Continued from Page 17

///

NCS | NEWSCASTSTUDIO.COM

Scaling remote production means

designing for concurrency

A remote production can be designed

around one event. A production platform

has to work when several events need the

same rooms, operators, processing and

network capacity at the same time.

That is the shift organizations face after

REMI becomes routine. The engineering

question is no longer whether signals can

travel from a venue to a hub. It is whether

the same infrastructure can support the

next event without another custom build,

then continue operating when schedules

overlap, requirements change or one pro-

duction runs long.

Scaling therefore depends less on mak-

ing any single workflow larger than on mak-

ing resources reusable, schedulable and

isolated across a portfolio.

Design for concurrency

The first remote productions were often

treated as individual projects. Equipment

was assigned, connections were tested and

a team was assembled for a known show.

That approach can succeed repeatedly

without becoming scalable.

At the platform level, demand must be

considered across the full schedule. Two

events may require the same replay capac-

ity, control room, graphics operator, codec

license or support engineer. A production

that enters overtime may collide with the

next reservation. A regional feed added late

may consume capacity held for another

event.

“Remote production is redefining the

economics and scalability of live production

as broadcasters are asked to produce more

events with smaller crews, tighter budgets,

and faster turnaround times. We’ve seen

production companies reduce setup times,

increase the number of simultaneous pro-

ductions they can support in a day, and low-

er operational costs by adopting a remote

production workflow. As these workflows

mature, the focus is shifting from proving

remote production works to making it eas-

ier to scale, manage, and repeat across ev-

ery production,” said Francesco Scartozzi,

vice president of sales and business devel-

opment at Matrox Video.

Concurrency planning has to include

more than compute and bandwidth. Con-

trol surfaces, monitoring positions, inter-

com ports, user permissions, specialist

staff and engineering support can all be-

come limiting resources. The system needs

enough reserve to absorb delays and fail-

ures without turning every busy day into

manual triage.

Turn venues into profiles

Scaling slows when each venue is treat-

ed as a new integration — the camera count

may change, but the production shouldn’t

require teams to rediscover how contribu-

tion, timing, return feeds, control and mon-

itoring fit together every time.

A venue profile can define connection

types, source names, codec settings, laten-

cy expectations, communications, security

rules and fallback procedures. Production

templates can then assign that venue to an

established workflow instead of rebuilding

the signal path around it.

Continued on Page 21

SCALING PRODUCTION

20 NCS | NEWSCASTSTUDIO.COM

Reliability depends on recovery,

not just infrastructure

Remote production changes where a

program is made, but it does not change

the consequences when the program fails.

The difference is that the cause may now

sit far from the operator who first sees the

problem.

A frozen camera could begin with a ven-

ue encoder, a contribution path, a cloud

service or a control-layer error. Several

vendors may meet their individual service

commitments while the production re-

mains off air.

Reliability is therefore an operating mod-

el, not an infrastructure specification — rec-

ognizing failure, limiting its effect and re-

storing service. The questions are whether

duplicated systems can fail independently,

who has authority to act and what the audi-

ence receives during recovery.

Redundancy is not resilience

Duplicating equipment or connectivity is

useful, but two instances don’t necessarily

provide two independent paths. Circuits

from different providers may share a con-

duit. Applications may share a cloud region,

identity service or control plane.

Resilience begins by mapping failure

domains across the production. A shared

dependency in power, timing, authenti-

cation, orchestration or communications

can disable otherwise separate systems. A

backup is credible only if it can be reached,

controlled and observed when the primary

is unavailable.

“The main operational challenge is main-

taining control of a workflow spread across

venues, remote production hubs, cloud re-

gions and multiple delivery networks. Resil-

ience depends on independently recover-

able layers, clearly defined failover triggers,

shared end-to-end telemetry and clear op-

erational ownership across vendors. These

mechanisms must be rehearsed together

under realistic load, as backup paths and

escalation procedures that work in isola-

tion may still fail during a live event,” said

Sarah Hackforth, international sales direc-

tor at Big Blue Marble.

The test is not whether every layer has

a backup. It is whether the production can

continue when a complete failure domain

is removed.

Define the degraded production

Not every capability needs the same pro-

tection, and few productions can afford to

duplicate everything. The practical alter-

native is to define the minimum viable pro-

gram before the event begins.

That decision should be editorial as well

as technical. A sports production might pre-

serve the main camera cut, program audio,

score and a clean transmission path while

losing isolated feeds, enhanced graphics

or some replay functions. Secondary lan-

guage versions or digital outputs may have

a different recovery order from the primary

service.

The purpose is to prevent a partial fault

from becoming a complete outage because

the team has not agreed what can be shed.

Operators need thresholds for switching

paths or dropping a nonessential output.

Producers need to know what those chang-

es mean on air.

A degraded mode also needs a route

back. Signals may need to be resynchro-

nized, states reconciled and delayed con-

tent checked before returning to the pri-

mary system. A plan that ends at failover

covers only half the incident.

One person must own the incident

Distributed production divides responsi-

bility among venue teams, network provid-

ers, facility operators, cloud platforms and

application vendors. That specialization

can improve support, but it can also cre-

ate an ownership gap at the moment when

speed matters most.

Every critical production needs an inci-

dent lead with authority to coordinate the

response. The role is to establish a common

view, assign actions, invoke the degraded

mode when necessary and keep technical

work aligned with the program.

“A good integration is one that works re-

liably in real operations, not just in a proof

of concept or project plan. It must be resil-

ient, observable, scalable and supportable,

with clear ownership of handoffs, monitor-

ing, escalation and service performance.

Projects still fall short when they focus

too much on connecting systems and not

enough on day-to-day operational readi-

ness, including redundancy, metadata in-

tegrity, security, support and cost control,”

said G Morgan, executive vice president of

sales at Globecast Americas.

Contact details, decision rights and ven-

dor boundaries belong in the production

plan, not in a message thread assembled af-

ter a fault. Suppliers can own their services,

but one person must own the effect on the

show.

Rehearse failure and recovery

A successful launch proves that a work-

flow can operate normally. It says little

about how the same workflow behaves

when dependencies disappear or several

events compete for the same resources.

Tests should interrupt real functions

under representative load. Teams can dis-

connect a primary contribution route, stop

a processing instance or simulate the loss

of return audio. They should confirm that

alarms reach the right people, communi-

cations remain available and the producer

understands the on-air effect.

The exercise should continue through

restoration. That reveals stale configu-

rations, incomplete state replication and

procedures that depend on a particular en-

gineer. It also gives operators practice mak-

ing decisions from incomplete information.

Runbooks should identify the symptom,

the evidence needed to isolate it, the per-

son empowered to act and the expected

effect of each recovery step. A list of com-

mands is not enough.

Security is part of availability

As production systems become reach-

able across facilities and public cloud en-

vironments, security controls become part

of the live signal path. Identity services, cer-

tificates, remote-access tools and software

update policies can stop a production as ef-

fectively as a failed encoder.

Security design has to preserve con-

trolled access during an incident. Teams

need governed emergency accounts, clear

RELIABILITY

Continued on Page 21