Upgrading a Mocap system at my University

[J0hn]

Hi guys,

The 3d program at my university is upgrading their motion capture system because we will have a bigger dedicated space of 30ft X 60ft. I am doing some research to write a proposal for a new setup and I have a few questions about your system.

I saw that the cameras on the site had "large volume mocap" and "medium volume mocap". I tried looking to see how big the capture volumes were and wasn't able to find the information anywhere.

The old system has 20+ cameras right now and our capture volume right now is 11ft x 19ft. We can only capture body right now and would like to expand with facial animation. The software we use are Iq and Motion Builder. We use the setup for 3d animations and video game animations for classes.

Questions:
1. I would like to know how to determine the capture volume for a 30ft X 60ft space?

2. Which cameras would serve that purpose and also the lenses?

3. If we wanted to can we mocap small animals like a dog or a cat?

4. How many actors can be captured in the space?

Thanks for the help,
John

[Seth Steiling]

Hi John,

Thanks for your interest in our mocap tech. We can definitely set you up with a system that will perform beautifully in that space, with better performance than you're currently enjoying.

Your space would qualify for our large volume camera--the Prime 41. We don't have projected volume sizes up on the site yet, but we will soon. In the meantime, I'll give you an approximation of what we'd recommend and what you'll get out of it.

Note that capture volume size depends on many variables, so there's not a simple "X volume in Y setup" answer. But I'll paint a picture for you based on what would be an optimal setup and camera count in that space (with optimal camera count being the number at which camera-to-camera FOV overlap and camera-to-floor convergence are the same):

Setup: 30' W x 60' L x 9' H
Optimal camera count: 30 (4 on each short side, 9 on each long side, 4 in the corner)
Resulting capture volume: 15' W x 45' L x 9' H

Reducing the camera count below 30 cameras in your setup will decrease the size of the capture volume. Increasing the camera count above 30 will provide better marker continuity (even better data), but won't substantially increase capture volume size. For these reasons, we'd consider 30 a sweet spot for your setup.

Regarding your other questions:

• The Prime 41 only comes with one 12mm lens option, but it will be ideal in this setup. It offers a generous 51� FOV to go along with 100' camera-to-marker range (the best in the industry by 20'-40', and plenty for your setup area).
• Small animals shouldn't be a problem. Tools for creating custom non-bipedal skeleton hierarchies are still in active development, but in the interim, you should be able to use rigid bodies to track animal body parts and constrain them in MotionBuilder.
• The maximum number of actors we've tracked so far in this type of setup is 4, though 5-6 should be possible in real time. We will be doing some additional testing in the coming weeks. With offline processing, the number should be considerably higher still.

[J0hn]

Hi Seth,

Thanks for responding. I really hope that the Professor and the University will pick your system. The information that you have shared with me will be really helpful for the proposal. I am excited for the program as they will be able to offer more.

I have talked to two local businesses here in Dallas, Texas on what systems they are currently using and it seems like Vicon is predominantly used here. I'd like to diversify the systems here in Dallas.

One more question how close would the cameras be placed to the capture subject to record facial mocap? It would be easy to set one up with tripods in the same room.

Thanks again,
John

[Seth Steiling]

Sounds awesome, John. With the recent launch of the Prime 41, now's an ideal time to add an OptiTrack system into the mix in Dallas. You'll be able to install the best-performing tech on the planet at a substantially lower cost than any of those Vicon systems.

It might be helpful in your pitch to your professor to know that the Prime 41 is besting Vicon's T40s and even T160s in several key performance metrics--in some cases by nearly 2x. (Far better range, more precise image processing, tuned for IR illumination, etc.)

Additionally, the cameras include several cool features that are unique to our tech and really simplify the mocap pipeline. Some of these include:

- On-camera Aim Assist button activates a camera for aiming with one button press on the back of the camera, with no need for a software operator scrolling through cameras during aiming. For anyone that's been up and down ladders hundreds of times, this is a HUGE time saver.

- Convert any Prime 41 into a sync'd and calibrated reference camera (with options for back-projected 3D marker, skeleton, and mesh overlay) with one mouse click.

- On-camera visual alert system provides visual feedback about camera health and activity using a ring of RGB LEDs on the front of the camera. As camera status changes, so does the indicator ring. Hard to communicate in text, but this is an awesomely useful feature.

- No proprietary cables! Vicon systems use expensive proprietary cabling. Our systems use standard Ethernet cables--inexpensive to purchase, inexpensive to replace.

Hopefully some of that will help you persuade your professor. If you need any more info, just let us know.

[Seth Steiling]

Oh, I almost forgot--face volume! The distance from performer to cameras can range from around 2-3 feet up to maybe 10-15 feet. It depends on the camera used, the number of markers, the size of markers, and to an extent what camera settings you use. There's a pretty good sized "safe zone" though. Tripods in the same room shouldn't be an issue at all.