Zoom Rooms Intelligent Director: what it does and what it actually requires

Lightly edited for readability.

When we do video conferencing, one of the goals should be to make it more like in-person, right? And one of the best ways I’ve seen to do that is with Zoom Rooms Intelligent Director. Have you seen it? Let’s check it out.

[Watching Zoom’s demo video] This is cool. Now, I think these cameras are a little high, but let me pause this. Look - this is it. You get a video of the entire room, but you also get individual video feeds of each person in the room. And I don’t mean like a mosaic. It’s as if all these people in the room had a laptop right in front of them with Zoom open, sending that video feed down to the far end. That means the remote users can double-click on a face and make it full screen, because it’s its own individual feed - or they can rearrange tiles, things like that. That’s really powerful.

Why are these cameras so high? That’s crazy - I think they should go lower. Now, why do you need three cameras? You can do this with one camera, and I have done that. This is the AVer CAM - looks like the 570 or the 550. That’s a really nice camera: good wide-angle lens, great low-light features. But you just get the one view. When you have cameras in the corners and on the edges, Zoom sees the same person from multiple angles and chooses the best one. Instead of using the camera that sees the back of someone’s head, it’ll use the one that gets a front-facing view.

So let’s go see what kind of hardware is necessary for this. Zoom Rooms license - yep. Zoom Rooms controller - fine. Two or three Zoom-certified cameras that support presets, installed on the same wall. Zoom recommends a camera capable of 4K capture. Support for up to 24 people across nine streams using three cameras.

There are codec requirements for two-camera systems, which I would never recommend - why do two? You should at least do three. And the compute: the Mac Minis are a screamer. You can’t get the M1s anymore, and Windows loves to get in your way - sign up for this, sign up for that. Mac just gets out of your way, and these things will run and run and run. The requirements say M1 and above, so an M4 chip with 16 gigabytes is going to be just fine, smallest storage, gigabit ethernet - about $800. That’s a killer deal.

Next, the certified cameras. Three cameras go in the center, left, and right positions - or two cameras as front-of-room plus center-of-room. Rally Bar for front of room… Logitech’s come a long way. My biggest problem with Logitech is - I can’t prove this - but they seem less reliable than anything else, and it seems to have something to do with their USB stack and firmware. If you’ve ever had a Logitech room that’s mostly good but sometimes fails, you know what I’m on about. Their optics and their newer audio are pretty good. It’s just something about reliability I can’t quite put my finger on.

Other cameras I really like: the CAM520, the CAM550. Notice the asterisks - the left/right camera model can be used as a center camera, so you can run the 550 across all three positions. That’s great. The AVer VB350, though - the wide angle on this is massive. The dual lens delivers up to a 113-degree field of view. That’s overly wide. It’s going to put blank space on both sides, so you’re wasting pixels - unless your front of room is really close to the first row of seats.

The Huddly L1 - this one’s a bad mamma jamma, because it has an extra-good sensor: a 6K image sensor. I like that. But it only uses digital zoom. One of the values of optical zoom cameras like the AVer CAM550 is you can zoom in on the far-end participants. If you’re doing 25 people and maxing the room out, you’ve got about 12 on each side of a long table - that’s a long room, and you don’t want digital zoom cameras there. The L1 has great optics and good low light, but I’d use it for a shorter room.

How it handles data: it analyzes data from the multiple camera angles to select and display the best angle of each participant, and the video streams are temporarily stored and processed on the Zoom Room equipment. So it stays local. That’s great.

Here’s how you enable it: go to the Zoom web portal and look through the thousands of options. It’s starting to get a little crazy, but one of the best things about Zoom Rooms is how many configuration switches there are - pages and pages. I’ve been doing this since there were only a couple of pages, and it’s fun to watch how many features Zoom keeps adding. Hard to keep up, though.

There’s some calibration, and then - best of all - make it the default camera mode for all meetings, so it’s just on. Your users will never turn it on themselves. They’ll hit Start and go sit down, never adjust the cameras, and leave everything at default. The tyranny of the default, as Steve Gibson says - you want that default to be Intelligent Director.

One thing I love about Zoom Rooms is the notifications. You can send an email - I think a webhook too - if there’s trouble, and there’s a notification for when Intelligent Director calibration is lost. If you make it the default, definitely enable those notifications.

So: Intelligent Director. An excellent way to let your remote participants feel more in-person and more included - more meeting equality. Remote people are at a disadvantage: they’re not in the room, they can’t feel the energy, and often they just see the wide-angle shot. Being able to see the non-verbals - the slight furl of a mouth, those tiny things - can be so impactful. Each of these cameras can cost a couple thousand dollars. But what’s the cost of miscommunication? It could be millions, depending on what you’re working on. Seems like a good ROI to me.

Hope this helps. Thanks for watching.