Hybrid event AV: the technical checklist most organisers miss

Most hybrid events fail in one of about six ways, and we’ve watched all six happen. The remote audience squints at a wide-shot of a stage, can’t hear panellists who turned away from the podium mic, and gets a Q&A that’s clearly designed for the people in the room. By minute 40 the chat has gone quiet and they’ve opened another tab.

None of this is inevitable. Hybrid done badly is “an in-room event with a Zoom link slapped on.” Hybrid done properly is two parallel productions sharing infrastructure but designed independently. Here’s the checklist we run through with every hybrid client.

1. The audio path is the make-or-break

If the remote audience can’t hear cleanly, nothing else matters. This is where most hybrids fall over.

The wrong setup: a single room mic feeding both the in-room PA and the stream. The remote audience hears PA bleed, audience coughs, and acoustic reflections. The in-room audience hears feedback because the broadcast mic is too close to the speakers.

The right setup: separate mix buses. The in-room PA mix is optimised for the people in the room (presenter mic loud and clear, room reverb feels natural). The broadcast mix is optimised for the people watching the stream (presenter mic loud, ambient mic at maybe -20 dB to give a sense of room presence, EQ tuned for laptop speakers and headphones, dynamics processing tighter so quiet asides are audible without spikes from loud moments).

That’s one console with two mixes, or a small split off the main FOH desk. Either way, your audio engineer should be making active choices about both mixes throughout the show, not setting up the broadcast feed and walking away.

Buyer question: “What’s your broadcast mix going to do differently from the in-room mix?” If the answer is “we send the main mix to the encoder,” ask follow-ups.

2. The remote audience needs a face, not a building

A single wide locked-off camera on the stage is what happens when streaming is an afterthought. From the remote viewer’s perspective, the speaker is a distant figure and the slides are unreadable.

What you actually need: presenter close-ups. Even one PTZ camera that can frame a chest-up shot of whoever’s speaking changes the energy of the stream entirely. Two cameras lets you cut between presenter and slide. Three or more lets you handle panels properly (each panellist framed, with a wide-shot reserved for context).

For broadcast graphics: presenter slides should be on the stream as a full-frame source, not a wide shot of the projection screen with the slide tiny in the corner. A clean output from the slide deck, switched in at the right moments, feels like watching a properly produced webinar instead of a pirated camcorder feed.

Buyer question: “What does the remote viewer see when slides are up vs when a presenter is talking vs when there’s a panel?” The answer should be three different shots.

3. Q&A has to work for both audiences

The classic failure: a presenter asks for questions, in-room people raise hands, the host throws to one of them, and the question is unintelligible on the stream because the audience mic never made it to that person.

Three fixes, pick at least one:

• Roving wireless mics for any in-room question, brought to the person by a runner. Works, slow.
• Roaming radio mics with a “mic runner” trained to position properly. Faster. Requires planning for crew.
• All questions submitted via Slido / Mentimeter / a moderated chat, with the in-room host reading questions aloud and the stream platform separately. Faster still, more inclusive of the remote audience, but requires moderation.

Option three is becoming the default for serious hybrid events because it makes the remote audience first-class. They aren’t tacked on; they’re a primary route into the conversation.

Buyer question: “How does a remote attendee ask a question that the in-room audience hears, and vice versa?“

4. Captions are non-negotiable

Captions help everyone: remote viewers in noisy environments, people with hearing impairments, attendees whose first language isn’t English, anyone watching with sound off on a phone.

Two paths:

• AI captioning (e.g. Zoom’s built-in, Otter, AWS Transcribe). Cheap to add. Quality varies with accent, technical terminology, and audio clarity. Suitable for internal events.
• Human captioner (CART service). Expensive but accurate. Default for public events, government work, healthcare, anything regulated. Roughly $400–$800 for a half-day session in Australia.

Whichever you pick, get them piped into the stream itself (burned-in or as a closed caption track), not just shown in the in-room screen. The remote audience is the audience that most benefits from captions.

5. The internet uplink isn’t trustworthy by default

Venue wi-fi is fine for emails. It’s not fine for sustained outbound video. Even venue wired internet often shares bandwidth with everyone else at the conference (you, your speakers, the catering POS, the badge printers).

What we do for any hybrid event over Tier 1: bonded uplinks. Venue wired internet as primary, with a 4G or 5G bonded modem as backup. The encoder is configured to fail over automatically. If the venue’s connection drops, the stream stays up on cellular. The viewers see a brief bitrate dip and nothing else.

The wrong way to discover venue internet is unreliable: at minute 23 of your keynote.

Buyer question: “What’s the backup if the venue uplink drops?” “We’ve never had that happen” is not an answer.

6. Encoder redundancy for events that matter

For events where a stream outage would be expensive (sales kick-offs broadcast to your sales team across the country, AGMs viewable by remote shareholders, regulated events), running two encoders in parallel is cheap insurance.

Primary encoder + hot-spare encoder + manual failover at the switcher. If the primary glitches, the spare is already running and capable of being switched live. Total added cost is usually a few thousand dollars, and the conversation you don’t have to have with your CEO post-event makes it the easiest dollars you’ll ever spend.

7. The remote audience experience needs a producer

Someone needs to own the stream as a discrete production. In smaller crews this is the same person running the vision switcher; in larger ones it’s a dedicated stream director.

Their job: watching the actual stream the way a remote viewer would (on a phone, on a laptop), monitoring chat or Slido, signalling to the in-room host when remote viewers have questions or when something visual isn’t working, calling camera changes so the remote view isn’t a static wide for ten minutes at a time.

Without this person, the stream is whatever falls out of decisions made for the in-room audience. With them, the stream feels like its own production.

8. Recording the stream is separate from streaming the stream

Most encoders can record locally while streaming. Use this. The local recording is your fallback if the live stream drops (you can re-upload after the event), and it’s typically higher quality than the compressed live stream (suitable for editing post).

For better quality still: ISO recording. Every camera records independently to its own card, plus the program output, plus separate audio buses. This is the post-production master, not the broadcast. It lets you re-cut the show after the fact for editing, archive, or speaker-specific deliverables.

9. Pre-event rehearsal that includes the stream

In-person events do a tech run. Hybrid events should too, and it should include the stream end-to-end. Someone on the team watching the stream on the actual platform you’ll broadcast to, from a remote location, while speakers are running their content.

This catches: bad camera framing, audio that sounds fine in the room but muddy on laptop speakers, captions lagging by 8 seconds, lower-thirds with typos, presentations that look fine on the projector but reflect glare on the camera feed.

A 90-minute hybrid rehearsal the morning of the event prevents the failure modes that matter most.

10. The post-event matters too

Hybrid events have an unusually long tail. People watch back. Speakers want their segment for social. Sponsors want their on-screen moments.

Plan the post deliverables before the event, not after. What’s the highlight reel? What’s the per-speaker cut? Where does it live (your own LMS, YouTube, a sponsor’s page)? Who’s responsible for each deliverable? If post is included in the AV scope, the file delivery dates and formats should be in the proposal.

A summary checklist you can copy-paste

• Separate broadcast audio mix, not a feed of the in-room mix
• At least one presenter close-up camera (not just wide stage shot)
• Clean slide feed sent as broadcast source, not pointed at a screen
• Defined Q&A flow for both in-room and remote audiences
• Captioning (AI for internal, human for public)
• Bonded primary + cellular backup internet uplink
• Redundant encoders for events where downtime is expensive
• Dedicated stream producer watching the actual stream during the show
• ISO recording for post-production master
• Tech rehearsal that includes the stream, watched remotely
• Post-deliverables planned and quoted before the event

If you’re getting an AV proposal for a hybrid event and any of these aren’t mentioned, ask. The vendor’s answers will tell you who’s done this before.