Videoconferencing Finally Establishes Its Shot

October 29, 2025 by David Danto

As part of my long history in our industry, I’ve spent decades joining video meetings with people I’d never met in facilities I’d never seen. More times than I can count, I’d connect to a call and find myself staring at a picture of people’s knees or feet – because the camera in the room was pointed wrong. I’d have to patiently explain how to find a touch panel or a handheld remote, then talk someone through tilting the camera up, panning it left or right, and zooming correctly – all until I could actually see the humans I was supposed to speak with. It was a ridiculous ritual, and it highlighted one of the big flaws of early video conferencing: you needed to be a video-conferencing hardware expert (or have one nearby) just to get a decent shot. For years, that was simply how it was.

Telepresence exacerbated the problems; Hollywood had a solution

Then came the TelePresence era – sometimes called Immersive TelePresence. The promise of those systems was that the cameras were invisible, perfectly placed, and didn’t need manual adjustment. The rooms themselves were mirror images, designed so every table, chair, and wall color matched, creating the illusion that participants were sitting across the same table in one seamless space. It was revolutionary for its time and eliminated the “camera-aimed-at-the-floor” problem. But it also introduced a new limitation: while you could now clearly see the faces of those speaking, you lost all sense of where they were. You couldn’t tell how the room was arranged or who else might be off-camera and listening – with a smile, a frown, or a raised eyebrow. It was a clinical, floating-head experience – better, but oddly detached.

That brings us to something Hollywood has understood for more than a century: the importance of the establishing shot. Just about every film scene begins by showing you the space before it shows you the people. Think of the opening of a courtroom drama – first the wide view of the judge, the jury, and the gallery – then the lawyer at the podium. Or before Julie Andrews starts singing about hills, we see said hills. These wide shots orient the viewer, provide exposition, and set the tone. Without them, you’d see faces but have no sense of the relationships between them. The same logic applies to video meetings. If all you see is a cropped-in speaker, you lose the context that defines the conversation. You don’t know who’s in the room, how they’re arranged, or how their body language might shape the dialogue.

In filmmaking, the establishing shot builds spatial awareness; in meetings, it builds social awareness. We humans rely on that context. We’re constantly reading posture, eye contact, and subtle cues from others in the room. That’s why, when you’re attending in person, you don’t just stare at the presenter. You glance around the table to see how people are reacting. If a senior vice president is speaking, you might actually be more interested in the CEO’s reaction than in the SVP’s slides. That’s the dynamic that traditional video conferencing – and even immersive telepresence – missed: the ability to “read the room.”

During that same telepresence era, the industry first began to realize what was missing. Tandberg was one of the first to introduce a “filmstrip” of room views beneath the main image – a set of smaller shots showing everyone else at the table. That small addition gave remote participants a sense of presence again. You could finally see how others were positioned, who was leaning in, who had crossed arms. It restored part of what we’d lost when conferencing went from chaotic camera swings to tightly framed close-ups. Just as cinema learned to cut between the wide shot and the close-up, video meetings were learning the same rhythm.

The death of the remote control / touch panel

As AI and machine learning entered videoconferencing, we finally reached a plateau where hand-held remotes and touch-panel controllers could be thrown away. Cameras became smart enough to make the correct shots themselves. Modes like “speaker tracking” and “group framing” became automatic. Users could walk into a room, connect, and start speaking – the cameras were smart enough to find them and frame them properly. These capabilities brought TelePresence-like quality to much less expensive systems. But they still didn’t solve the problem of context – where were the speakers in the room, and how were the in-room participants reacting to what was being said?

We’re now entering an era where systems are finally smart enough to solve that problem. One of the earliest products to approach it was the Meeting Owl, a center-of-table BYOM device widely used in SMB and education settings. It produced a composited image of the last three speakers layered below a panoramic strip of the whole table. While early Owl Labs models had questionable camera quality, the company deserves credit for throwing down the gauntlet around compositing. It democratized the idea that a single, intelligent image could tell multiple stories at once – the speaker, the context, and the audience reaction – without human control.

Zoom followed this, introducing its platform-based Intelligent Director. Its approach used multiple cameras to capture different angles of the room and then sliced and composited those feeds so each participant appeared in their own square. It solved for visibility – you could finally see everyone’s face – but it didn’t yet restore spatial awareness. Still, it was a pivotal moment: for the first time, the platform itself, not the room hardware, was making creative decisions about how the meeting looked. Since that time, we’ve experienced a continuous balancing act between the two – platform intelligence and edge intelligence.

Crestron, after acquiring 1 Beyond, expanded the concept dramatically with its Dynamic Composition Capabilities, turning multi-camera systems like Automate VX into real-time directors capable of showing multiple viewpoints simultaneously. In practice, the screen might display a wide establishing shot, a close-up of the speaker, an audience reaction view, and the presentation content – all blended seamlessly. Integrators can set rules: a training room might prioritize presenter-on-content mode, while a conference room might emphasize two-up discussion framing. The result feels less like security camera footage and more like a professionally produced broadcast – executed entirely by the system itself. Remote participants experience a sense of being in the room, not just watching it.

Cisco, at its recent Webex One conference, took this concept further with its new AI-based Dynamic Camera Mode coming in RoomOS 26. In Cisco’s solution, users and administrators no longer need to select “speaker tracking,” “group framing,” or any other mode. The system’s device intelligence continually evaluates the room and its occupants, composing the best possible image on the fly. It merges an establishing view and close-ups into a single, cohesive scene – always presenting both content and context. More impressively, it adapts as the meeting unfolds, filtering out reflections, background movement, or people walking past glass walls. Cisco’s approach represents the logical endpoint of two decades of camera automation: no more manual control, no more mode selection. It’s not just AI automating the camera; it’s AI directing the meeting.

AudioCodes has also embraced this evolution with its Composite AI mode on the RXV200 platform. It blends feeds from multiple cameras into a unified image – one lens capturing the whole room, others zooming in on active speakers. Its “Focus Speakers” feature enlarges up to three people currently talking, while “Presenter-on-Content” overlays the presenter on shared material. These are the kinds of choices a live video producer would make instinctively – now handled entirely by the system.

All these innovations – and surely more are coming – share a common thread: they’re restoring context. They’re bringing back the establishing shot, not as a single static angle but as a living, responsive composition – always evolving to match the conversation.

The next time you’re in a hybrid meeting and notice that you can see both the speaker and the subtle reactions around the room, that’s not an accident. That’s the technology quietly doing what good filmmaking has always done – helping you understand not just what’s being said, but where and how it’s landing. After all these years, the technology is finally catching up to how humans naturally see and interact. This videoconferencing concept may just have legs…