storylinki
← Blog
Streaming Tools

The Encoders, Graphics Engines and Monitors Behind Our Streams

A look at the tools Priya's streaming crews rely on: encoding, graphics, signal monitoring and backup paths, and why each one earns its place on a live online production.

Priya Nandra

6 min read

The Encoders, Graphics Engines and Monitors Behind Our Streams

Start with the delivery path, not the camera

Every streaming job I plan begins at the viewer's end and works backwards. Before anyone talks about lenses or lighting, we ask where the audience will watch, on what kind of connection, and how long the programme runs. Those answers decide the bitrate we aim for, the resolution we send and the number of backup routes we need. The tools follow from that path. A venue with a strong wired connection lets us send a clean, high quality feed with modest compression. A site with a weaker link needs a more conservative encode and a second route ready to take over. We write this down in a single page that the whole crew can read.

It sounds dull, but this page prevents most of the arguments that used to happen on the day. When the producer, the engineer and the graphics operator all work from the same description of the delivery path, each of them can make small decisions without stopping to ask. That keeps the gallery quiet, and a quiet gallery is where good streaming gets made.

  • Audience location and device mix
  • Venue upload capacity and a second route
  • Target resolution and bitrate
  • Programme length and any planned breaks

Encoding: hardware for the main feed, software for the spare

For the main programme feed we prefer a dedicated hardware encoder. It does one job, it has no background updates to worry about, and its behaviour is predictable over a three hour show. We set the keyframe interval and bitrate during rehearsal, then leave them alone once the show is live. A second encoder, usually software running on a separate machine, sits ready as the spare. It takes the same programme signal and sends it to a backup destination. If the main unit stumbles, we do not rebuild anything under pressure. We switch the destination and carry on, and the audience sees a short rebuffer at worst.

The choice between the two is rarely about raw quality. Both can produce a good picture when set up properly. We choose hardware for the path that must not fail, and software for the path that must be flexible, because the software one can be adjusted quickly if the venue network behaves differently from what we tested.

  • Hardware encoder for the primary feed
  • Separate software encoder as the live spare
  • Settings fixed at rehearsal, not during the show

Graphics that the audience can actually read

Online viewers often watch on a phone held at arm's length, so our graphics are built for that screen first. Names, scores, session titles and sponsor lines use larger type and more generous margins than a television-only show would need. We test every lower third on a small display before the doors open, not just on the gallery monitor wall. The graphics operator works from a prepared running order, with each title loaded in advance and numbered to match the cue sheet. Late changes happen, so we keep a few blank templates ready for names that arrive at the last minute. Typing a correction into a template takes seconds, and it avoids the temptation to improvise a layout while the show is on air.

We also keep the graphics machine on its own clean signal path. If a title misbehaves, it can be taken off air without touching the camera feed underneath. That separation is a small design choice, but it has saved more than one programme from a visible glitch.

  • Large type and safe margins for phone screens
  • Titles numbered to match the cue sheet
  • Blank templates for last-minute names

Monitoring: the screens nobody sees on camera

A streaming gallery has more monitors than most people expect, and only some of them show pictures. We keep one screen on the outgoing stream exactly as the audience receives it, delayed by the platform's normal lag. We keep another on audio levels, and another on encoder health, showing bitrate, dropped frames and connection status. The monitor that matters most is the one showing the real public stream. The programme monitor in the gallery shows what we intend to send, but only the public stream shows what arrived. More than once, that screen has shown us a problem the gallery could not see, such as an audio drift or a frozen graphic on the platform side.

Someone in the room is always responsible for watching those screens and nothing else. On smaller jobs that person is me. On larger ones it is a monitoring operator whose only task is to say, calmly and early, when something looks wrong.

  • Public stream view as the audience receives it
  • Audio level metering on the programme and the return
  • Encoder status: bitrate, dropped frames, link health

Comms, cables and the unglamorous kit

Much of what keeps a stream steady is plain equipment. Talkback headsets with reliable batteries, labelled cables, spare adapters and a tested power distribution unit matter as much as any encoder. We label every cable at both ends and keep a printed signal diagram taped near the desk. Our engineer, Callum, builds the signal routing so that any single piece of equipment can be replaced without redoing the whole chain. Each connection has a known replacement, and the spares live in a labelled case rather than a general toolbox. When something fails, the fix is a swap, not a diagnosis.

We also bring a small wired network switch of our own. Relying on a venue's shared network for the control room's internal traffic invites trouble, so our machines talk to each other over a closed local network and only the encoders reach outwards. It is a simple habit, and it removes a whole family of surprises.

  • Labelled cables and a printed signal diagram
  • Spare adapters and power kept in a marked case
  • Closed local network for the control room

Rehearsal as a tool in its own right

We treat the rehearsal as part of the toolkit rather than a formality. A full run with live graphics, a real encode and the backup path switched in and out tells us more than any amount of checking in isolation. We deliberately trigger the failover once, so everyone has seen it work. The rehearsal also gives us a reference recording of the public stream. If something sounds or looks different on the day, we can compare it with the rehearsal and find the cause quickly, whether that is a changed setting, a different microphone or a busier network.

In these shorter autumn weeks, with many events moving indoors and schedules tight, that spare hour of rehearsal is often the first thing organisers offer to cut. I would ask them to keep it. Everything else we use depends on having practised it together once.

The tools behind a good stream are mostly ordinary ones, chosen for predictability and arranged so that one failure never becomes two. A hardware encoder with a ready spare, graphics built for small screens, monitors that show what the audience sees and a rehearsal that tests the failover: none of it is exciting, and that is the point. A calm gallery usually means the kit was settled long before the show began.

You may also like

  1. What a Broadcast Control Room Costs and What Moves the Price

    A plain look at how control room design is priced, from desk layout and monitor walls to signal flow and comms, and which choices make a budget grow or shrink.

  2. Stand-by, Cue and Go: Gallery Calls Explained in Plain Words

    A producer's three-word rhythm of stand-by, cue and go keeps crews, cameras and graphics in step. Here is what each call asks of you and why the order matters.

  3. Bitrate, Keyframes and Latency: Streaming Terms Explained Plainly

    A producer's plain-words guide to the numbers on a streaming encoder: what bitrate, keyframe interval and latency actually mean, and why each one changes what your audience sees.

Get in touch

Write to us: hello@storylinki.info