IBC 2026 Observations: The Second Half of Broadcast Cameras Hinges on Scene Understanding

A decade ago, broadcast cameras competed on who captured the clearest image. Today, the competition has shifted to a different axis.

At IBC 2026, one takeaway is unmistakable: pure spec competition is receding, while AI capability has taken center stage. The IBC Conference’s designation of “AI content production” as a top annual agenda item is well founded—as resolution, sensor, and interface metrics become increasingly homogenized, a camera’s ability to “read” the frame and “understand” the scene has become the new dividing line.

The VX90Pro demo zone at VHD’s booth offers a clear microcosm.

Live Experience: AI Tracking Stability Proven in Real Time
On opening day, many attendees walked into the VX90Pro capture zone to stress-test it—pacing side to side, turning abruptly, crouching and standing up. The camera’s AI tracking box remained locked on target, following smoothly without stutter or loss of lock.

The adjacent K70 control keyboard demo was even busier. Directors used the 4D joystick to control the PTZ head and, with AI visualization shortcut keys, executed camera positioning and function switching seamlessly. Trajectory recording and one-touch replay drew technical leads from production companies, who stopped to examine the capability: fixed-route shots can be configured once and recalled repeatedly, eliminating the need to re-block camera movement each time.

These interactions appear simple, but they entail substantial technical barriers:

·Accurate recognition: moving objects must not be misclassified as people, and recognition must not fail when a person changes clothing color.

·Stable tracking: camera movement must be smooth, not jerky, to avoid viewer discomfort.

·Reacquisition after loss: if a person is briefly occluded, the camera must reacquire immediately rather than stalling after target loss.

In-House ReID + FaceID Dual Algorithm: Fast Reacquisition After Occlusion
VHD’s VX90Pro, VX90, and VX720LG all support multi-mode AI tracking—auto-framing, presenter tracking, and zone tracking—covering diverse use cases.

The true differentiator is the re-identification capability of the in-house ReID + FaceID dual algorithm.

Many AI-tracking cameras lose lock or misidentify a person after occlusion—for example, when someone crosses in front of the lens. VHD’s dual-algorithm approach identifies targets using both facial features and body/clothing features: when the face is unavailable, clothing and posture provide recognition; when clothing changes, facial data serves as fallback. This two-layer redundancy significantly improves re-identification speed and accuracy after occlusion.

This is not a technical showcase gimmick; it is a capability that genuinely reduces cost.

One Camera Operator, Ten Cameras?
Broadcast production faces a practical cost problem: more camera positions require more operators. A mid-sized studio or live event with five or six positions typically requires five or six camera operators, making labor the dominant cost.

This is where AI tracking delivers value—auto-framing, auto-tracking, and auto-switching enable cameras to handle most routine scenarios, leaving one operator at the control console for overall scheduling and live adjustment.

From one operator per camera to one operator for ten cameras—this is not far-fetched; it is already underway.、

This echoes the industry signal from IBC 2026: AI’s value has shifted from feasibility to cost savings. This is a highly practical value proposition and the core rationale behind VHD’s continued investment in AI.

Final Thoughts
In the second half of broadcast cameras, the contest is not only about larger sensors or higher zoom ratios, but about whose AI better understands scenes, deploys in practice, and creates real value.

When a camera can truly “understand” what it captures—how many people are in frame, who is speaking, and whom to frame—many production and broadcast workflow steps can be redefined. That is precisely what makes the “new-gen camera” most worth anticipating.