VideoConferencing with a Director
On October 15, 1951, television broadcast history was made when the first episode of "I Love Lucy" aired on CBS. The event is noteworthy because it was the first time multiple cameras were used to film a television show before a live audience. This was a task that many believed to be impossible as different lighting requirements were needed for various types of shots.
It is multiple cameras that make broadcast television more natural than video conferences. When I watch the news, the picture switches from talking head to talking head – it is instant and seamless thanks to the team behind the cameras.
This doesn’t happen in my group video conferences. When the active speaker changes, I either move my eyes on a panned-out picture of small talking heads or I remotely control the camera to the active speaker through a clumsy process that causes seasickness. Then when zoomed in, finding the next speaker becomes a frustrating game of hide-and-seek because all voices come from the same direction (the speakers). The lack of a crew behind the cameras is obvious.
In an era of self-driving cars, drones, and self-checkout stands – why can’t we automate what DesiLu did 62 years ago?
Evidently the folks at Polycom had the same thought, so solved the problem with the EagleEye Director. The product was introduced in 2011 and remains a clever solution to an industry problem. “Director” is the operative word as the unit determines where to aim the camera and the optimal zoom. It uses two cameras to make the transitions automagical.
Video conferencing technologies should be transparent. It should bring people together with an as-close-to an in-person experience as possible. Direct personal interaction is compromised when the view is of the entire table. This frequent default setting is sometimes referred to as the “bowling alley” effect referring to the long view down the table.
Despite the two cameras, the EagleEye Director provides one visual stream. It utilizes its built-in seven microphone array to triangulate the speaker’s general direction which then gets further pinpointed with face detection software. The microphones are only used to locate the speaker, so don’t interfere or complicate system audio inputs. When someone else speaks, the system initially switches cameras for a room orienting view until the other camera homes-in on the speaker.
Polycom changed audio conferencing with speaker detection almost two decades ago. It’s a similar, but more complex problem with video. Not only does the system need to determine the directional source of the speaker, but the optimal framing of the video as well.
The EagleEye Director is a set-top appliance. It has two pan, tilt, zoom HD cameras. It uses face detection software to ensure it doesn’t zoom-in on the back of a head. When it can’t find the speaker’s face, such as when the speaker isn’t facing the camera, it defaults to a room scene.
The software is integrated with multiple Polycom video solutions including the HDX 4500. At least for now the miniature film studio in-a-box is available exclusively to Polycom video customers. I would love to see Polycom release a version that could connect into a camera input of third-party video systems. It’s a very clever device that deserves a look, or two.