forums.naturalpoint.com

hi

im new to all this mocap stuff, but i need some facial mocap (pointcloud data to use in cinema4d - no realtime) for my next projects.

i can create some pointclouds (30 markers) by using 3 x standart DV cameras (720px/576px 25fps) and some tracking software. it works, but the tracking prozess is quite time consuming and the DV cameras cause way to much motionblur. Searching for affordable cameras with > 60fps i finally stumbeled over this v120 slim camera ($369).

my question are:

is there any way to create a pointcloud (30 facial markers), by using only 3 cameras (v120 slim), 1 optihub and the tracking tools?

Do i need tracking tools do get 3d-pointcloud data, or can it be done using the OptiTrack SDK ?

Hello,

The best way to accomplish face motion capture is to use one of our Expression bundles. They combine the Expression software with V100:R2 cameras and have been designed for face animation. They will give you the simplest, least time consuming, most reliable and powerful experience.

Otherwise, it may be possible to achieve what you are interested in using the Tracking Tools software - however it has not been optimized for face tracking like Expression has. It lacks face marker template tracking and post-capture editing tools (gap filling, marker swap, etc). It does create 3D point cloud data, tracks rigid bodies and supports .csv and .c3d export. Note that its single marker tracking is not as robust as when tracking is performed with rigid bodies.

The OptiTrack SDK only offers 2D data for the markers' location on the sensor, so it won't suffice for what you need.

Regarding cameras, you're probably better off with the V100:R2, as the V120:SLIM does not have a built-in IR source. The R2 does, which will give you the ability to track passive face markers, which for most users is more ideal than active face markers. If you choose to use the V120:SLIM be sure to purchase the Sync breakout connector so that you can hook all cameras together, this will synchronize their exposure which greatly improves capture fidelity and 3D reconstruction. R2 cameras include sync support out of the box.

Hope this helps!

thanks for answer

this far, animation is only my hobby, so i cant afford $6000 for one of your expression packages.

For my first experiments ive made myself a active LED Marker setup already, and captured people will not be allowed to turn their head to much , so i hope it will work using three of your v120 slim starter packages for $369.

sorry...some more questions:

in the starterkit v120slim there is this lens:

Lens M12 4.5mm F#1.6 No-Filter, Low Distortion

is this the right lens for capture distance about 3feet ?

is the pointcloud tracking tools creates clear of marker swapping, or do i need arena for that ?

can i use tracking tools and three cameras without the optihub ?

We generally recommend 5.5mm lenses for our 6+ camera face volumes, because they give you better resolution/precision. But, because you only want to use 3 cameras, I would recommend the 4.5mm stock lenses. Their wider FOV might give you a better shot at decent capture for all 30 markers.

You should note that what you're trying to do will be tricky, and your results will not be the same as one of our officially recommended setups.

If the markers are not part of a defined rigid body, they will not be labeled, so you won't have the ability to address marker swapping within our application. (Expression's templates help address this nicely.)