What we rely on to run a live broadcast gallery

When people picture a live gallery they picture the vision mixer, the panel with the rows of buttons. In practice the mixer is the least interesting part until the inputs feeding it are sorted. Before any event we write an input list: every camera, playback machine, graphics output and remote feed, each with a fixed button position and a plain label. That sheet is taped above the panel and mirrored on the multiviewer.

We favour a mixer with enough physical buttons to cover every source without paging through banks. Paging is where mistakes happen at speed. If a show needs more sources than the surface offers, we reduce the number of sources before we accept menu diving.

Keeping the layout identical from one job to the next matters more than any single feature. A vision mixer operator who knows where camera two lives will cut cleanly without looking down.

• One written input list per event
• Fixed button positions, plainly labelled
• Avoid paged banks on live cuts

Behind the gallery sits a router, and a good one is boring. We map every source and destination on paper first, then build the cabling to match the map. Colour-coded, numbered cables at both ends mean a fault can be traced in seconds rather than minutes, which is the difference between a glitch and a lost segment.

We keep routing decisions conservative. Where a source must reach several places, such as the mixer, a record machine and a confidence monitor, we split the signal properly with distribution amplifiers instead of daisy-chaining outputs. A chain looks tidy until one device fails and takes the others with it.

Reference sync is handled deliberately too. A single sync source feeding every device stops the slow drift and occasional frame slips that otherwise appear an hour into a long event.

• Paper map before any cabling
• Numbered cables at both ends
• Distribution amplifiers rather than daisy chains

Lower thirds, scoreboards, name straps and full-screen cards all come from a dedicated graphics machine, not from the mixer's built-in keyer alone. We agree the look with the producer in advance and build the templates before the doors open, so changing a name on the day means typing text, not redesigning.

Spelling is the quiet hazard. We ask for the running order and speaker list as a single document and check every name against it, because a wrong title on air is remembered longer than a good camera move.

A second operator on graphics is not always affordable. When it is not, we keep the graphics page short and the cue sequence numbered, so the person cutting the show can also fire the right card without losing their place.

• Templates finished before rehearsal
• Names checked against one speaker list
• Numbered graphics cues

A gallery works because people can hear each other. We run a clear talkback system with separate channels for the director, the camera operators and the audio desk, and we test every headset before the first rehearsal. A dead beltpack discovered during a live cue is entirely avoidable.

On the picture side, the multiviewer shows every source with its label and a clear program and preview indication. We also keep a plain, uncluttered clean-feed monitor so the director sees exactly what viewers see, without overlays confusing the judgement.

Audio matters here as well. I mix from the same gallery, so I check that programme sound and the crew's talkback never leak into each other, and that the level the director hears matches what leaves the building.

• Separate talkback channels per role
• Headset check at every rehearsal
• Clean-feed monitor for the director

Live work needs a way back when something fails. We carry a spare for the items most likely to stop a show: a second power supply path where the equipment allows it, spare cables of each type, and a backup record of the programme feed running throughout.

We do not duplicate everything, because doubling every device adds complexity that itself causes faults. Instead we decide beforehand what a failure of each part would cost the show and plan a specific response, written down and shared with the crew.

The simplest backup is a rehearsed one. We walk the team through a lost camera, a frozen graphic and a dropped feed during setup, so that during the event the response is already familiar.

• Spare power, cables and a backup record
• A written response for each likely failure
• Rehearse the failures, not just the show

When we compare kit for a job, brochure features rank low. We ask how the device behaves when it is pushed, how fast an operator can recover from a wrong press, and whether the crew can understand it after a short briefing. Familiar equipment that the team knows well usually beats a newer unit with extra functions nobody has practised.

That is also why we resist adding tools mid-project. A new piece of equipment introduced on the morning of an event brings new menus and new failure modes at the worst moment.

For clients, the practical effect is a calmer gallery and fewer surprises. The tools are chosen to be understood quickly, labelled clearly and recovered from easily.

• Judge recovery speed, not feature lists
• Prefer kit the crew already knows
• No new equipment on the day