I'm joining the waiting team... I hope it's early September. Also I wonder if I should second kinect v2, or stay and wait to see if with one the results will be okay for animation...

Also a question towards ipisoft, now when the kinect can track up to 5 people, your software will be capable of doing the same?

Best wishes...

AFAIK, up to 6 people. But that does not matter.
Our software does not use the skeletal tracking provided by Kinect SDK, as it has pretty low quality for general mocap purposes (because its main purpose is gesture recognition):

• Works reasonably well only when a person is facing the camera (no turns)
• Very limited joint freedom (for example, practically no spine bends)
• Does not combine data from multiple sensors for analysis (to overcome occlusions)

So, there is no any connection between the number of persons supported by Kinect SDK and our software.
Besides, there is no much sense to do mocap for 5 actors in a limited working area of Kinect. They would have practically no space for motions.

What will be the maximum area of capture, with one or more Kinect v2?

Can I capture area 14x15x3 meters?

The new sensor itself has no difference in its working range - it's about 0.8 to 5 meters. Thus, the capture area is the same - about 2x2 meters, maybe a bit more depending on conditions.

Are you planning to announce a release date, or to just release the software when it's ready? I've been waiting patiently to buy your software until I know for a fact that I can upgrade for free to v3

According to Brekel, K2 is a way more precise than K1. Are you guys already working on support for K2 in iPiSoft ?

We will officially announce the release of V3 in approximately a couple of weeks before the date.
According to current development situation, we estimate the announce to be in September and release in the very end of September or early October.

Thanks for the update! I'm eagerly looking forward to this release. :)

G.

_________________
Greenlaw
Artist/Partner - Little Green Dog | My Demo Reels (2013,) (2015,) (2017,) and (2019)

Watch a one minute excerpt on Vimeo now!

What exactly does his "precise" mean?
The main improvement for us is far less noise level, especially when using multiple sensors. This does not add any precision, but reduces the number of tracking errors.
We also see that the real depth resolution of Kinect 2 is about twice better (in linear size). This helps with small details like feet and practically has no value for tracking of large body parts.
Another improvement is skeleton tracking in Kinect SDK. 6 persons instead of 2, better head rotations, limited hips bends, rough hands tracking. However, our software does not use it at all, we have our own tracking algorithm (in contrast, Brekel relies on tracker of Kinect SDK). And still, there are limitations: expects an actor is facing the sensor (practically no turns), no spine bends, poor feet tracking.

So having 2 K2 is not as good as having 4 PS Eyes ?

What about 2 K1 vs 2 K2 ?