Fantastic timing Malcom. I just had a friend bring a ZED cam over to the studio this afternoon.
I was immediately able to see all of the image inputs from the drop down and could switch the resolutions and other parameters effectively.
The quality of the depth image i’m getting from my camera is pretty bad right now - objects end up having big blocky noisy shadows all around them - so i think i need to recalibrate it before I fully understand it’s capabilities.
Super helpful to have this top with so many parameters exposed. Will follow up once i’ve tested again after calibrating… hopefully that will tighten up the quality of the depth feed.
Currently the camera fps seems locked at 15fps (might upgrade my firmware though). No biggie for me actually for the near future it gets me where i need to go. If I use video device in top you can (even in the live release of TD) pick up the camera feed off the left in case anyone is interested.
Gotta say, Fill mode - wow! Really happy with those results. I was underwhelmed by the raw depth maps from the zed but this really kicks it up 10 notches
It won’t make it into the main 2018.20000 builds, but this build is based off of those, so as stable (except for any ZED related stuff). If there are features that you need in 2018.20000 that arn’t in the current ZED camera build just ask, and I’ll make a new one that is up to date with the latest 2018.20000 + ZED support.
Managed to test 3x zed depth cameras successfully at 30fps 720p with filtration on. Dedicated usb 3. ports required as well as running each zed in a seperate instance else you’ll get flickering. Also recommend for heavier gpu based scenes to assign your camera instances to a second quadro to prevent performance reductions.
We are compiled against the 2.8 version of their SDK. However you shouldn’t need to install anything to get it working. Maybe they changed their drivers or firmware though. I’ll check in the office soon.
You are using the v1 camera right? We don’t yet have a v2 in the office to test with.
I d like to ask if there is any TOUCH “TOP” planned (or already existing) for working with Spatial Object Detection and Large-Scale 3D Skeleton Tracking working with a Zedi2.
I d very much like to use it instead of a Kinect v2, as I would have availability to two Zedi2 and to one Zed Mini. And The positional Tracking on a FLOOR/PLANE seems to be very accurate with ZED
And does TOUCH support two Zedi2 on one PC? How do the TOPs react to that?
We are working on 3D skeleton tracking for Zed for our next experimental release, available in the coming months. We do not have Spatial Object Detection planned yet, but it could be added with Pro Support if required.
TouchDesigner supports multiple ZED cameras per process, though they do not sync or talk to each other in anyway, you can run multiple at the same time.
Pro support gives you direct one-on-one support with our team, custom builds, immediate bug support, and feature additions. You get Pro support with TouchDesigner Pro which includes 6hours to start with, you can use this time for anything. This includes custom feature requests for tasks we do not have immediately in our roadmap.
unfortunately I still cannot effort pro support.
I wanted to ask if spatial detection is maybe planned now or very soon?
as I asked our university if we could order two zedi2 cams for some projects… and I kinda should do something with them now.
and I am mainly interested in that feature (like the basketball game they promoted on their homepage, where each players position and movement is detected).
If anyone would be able to build a top and chop for it, I maybe could manage a single payment for this request, but pro support/pro account is to expensive for me atm.
is this ZEDi 3d skeleton tracking (and maybe also some other object detection - position of objects in a room) available in the experimental build 2023?