Hey If anyone has used the Kinect attachment for the Xbox, you know that you are often driving a digital avatar in realtime. When will we see an iClone plugin so we can capture basic motions (and maybe audio capture too?)

Other people are already hacking the Kinect for other platforms and purposes. And Microsoft is not suing anyone over it.

This would essentially enable iCloners to make movies almost the same way that James Cameron did with live motion capture in Avatar.

Isn't it amazing... within the a year of the release of Avatar, comes the stereoscopic 3D output feature of iClone 4.3 and the general public availablity of the Kinectic motion tracking device.

I'm pretty sure it was a litle bit more complex than Kinnect.

Did you see the whole set of posts on ipiSoft. It's a motion capture software that can use a few webcames (cheap webcams) to do markerless motion capture. I'd rather use that as the cameras are cheaper... only thing with ipi is I don't really feel like dropping 1,000 us dollars for it.

Maybe if the software was like 100-300 then ok... maybe even 500... but not a thousand.

Also, you could only really get it from a few angles.

Now if you had multiple kinnects, then you could construct full 3D objects from real ones since Kinnect can create a virtual camear that you can move around your rom. That'd be pretty cool.

----------------------------------------------------
***BEST OF THE CORE**
:A music reel demonstrating the different types of music I produce for multimedia. Hope everyone enjoys.


Player not working? Use this link.

My other audio work.

Facebook page:

Creation begins at the core.

I´ve posted the same idea in the german forum some time ago, but I´ve never got any answers to it.

I totally agree, this would be more than just great.

In kinect microsoft is also using a bone system, and therefore lot´s of struggle wouldn´t happen, because it is allready implemented and ready to use.

Perhaps you wouldn´t even need a hacked driver for windows. RL could write a "game", which is simply capturing the motions (Microsoft will also provide support for kinect in the XNA in the future);
Afterwards the "game" could translate the kinect bonesystem in RL bonesystem, and export this to a attached USB stick for example.


I wouldn´t expect this to be as good as a 50.000$ motion capturing system of course.
But I do not expect iClone to be able to render a movie like avatar on the other hand either.


But, for home use, I think motion capturing with kinect would be more than enough

Wishlist. Definitely

________________________________________________________________________________________________________________________________________
Danke für Flex - Haare und Kleidung, die nicht durch den Avatar gehen !

ipiSoft is working on a mocap thing for Kinnect I think.

----------------------------------------------------
***BEST OF THE CORE**
:A music reel demonstrating the different types of music I produce for multimedia. Hope everyone enjoys.


Player not working? Use this link.

My other audio work.

Facebook page:

Creation begins at the core.

Having a pre-rigged iclone skeleton and export platform from kinect would be just sweet. A simple and cost effective way for truly anyone to capture motions for their iclone characters would be a wonderful thing

WOLF'S TRADING POST IS OPEN COME IN FOR A VISIT

"May you one day see what my eyes have seen...and be better for it" - Wolf
www.wolfzonedigital.com

"No. You get a lot more spacial data from multiple camera angles (i.e., 4 to 6 PS3 cameras placed in a circle or semi-circle); Kinect can only record data from one angle. Now don't get me wrong, Kinect is a great system for getting the most out of one camera angle because you at least have some z depth information, but this approach has its limitations, especially for full body capture. For example, it can't track body parts that are occluded by the body.

And, as explained here and elsewhere, you can't use multiple Kinects because it uses IR to record depth. (IR coming from multiple angles will corrupt the depth information.)"

Greenlaw, Senior Digital Artist - Rhythm & Hues Studios

General Picture - animating now to life itself(tm)
Member of Content Wizards

Agreed, however, a simple effective way to capture many rudimentary motions that one might not otherwise have in their library, i.e. genero sitting, eating, idle, common type animations from a multi purpose system that you use for other things like gaming, excercise etc...and without the space requirements or cah outlay of a multi camera setup would be a good tool to have imo. It's not going to get you cinematic mocap, but it would certainly get the job done for many simple and natural movements.

WOLF'S TRADING POST IS OPEN COME IN FOR A VISIT

"May you one day see what my eyes have seen...and be better for it" - Wolf
www.wolfzonedigital.com

Well, I thought IPISoft was trying to tell us to rely on their Playstation Eye 3 approach, but it is now clear that it isn't true anymore. So to everyone who has been advocating the use of Kinect for mocap, I apologize.


If you are willing to wait for it, you'll probably see Kinect mocap from IPISoft next year.

Let's hope that it will be easier to calibrate than four PSEye3's.

General Picture - animating now to life itself(tm)
Member of Content Wizards

I would agree with jasonjbrown that the Kinect sensor makes Avatar-style motion capture possible and I would include the possibility of facial and hand-gesture capture. Reallusion should consider the possibilities.

Dreamcube017 notes that the Avatar mocap was more complex than might appear and this is true - they had a vast network of hi-res cameras to collect the body movements of many actors simultaneously and each actor had a muzzle type camera in front of their faces to capture expressions and speech.

But looking at the ipisoft option is, I also agree, barred by expense, although ipisoft is offering until 15th jan 2011 a 30% discount bringing the price down to $662. Not only is the expense considerable, but calibration is extraordinarily difficult and even the small net of cameras demands a 20' X 20' space just to allow you a couple of paces either way in the central area. Others have cast doubts on the idea of multiple Kinects arguing that their infra-red beams would interfere giving chaotic depth measurements in the z dimension.

AndreasM notes that there is already a bone system in the Kinect Avatar driver system but my information is that this is confined at the moment to 20 joints for some reason - perhaps because it is purposed to allow more than one person simultaneously. Now your ten fingers have thirty joints which puts a constraint on hand gestures.

I certainly agree with Wolfzone who argues that a pre-rigged iClone skeleton and an export platform would be just sweet.

My own point is that I find library motions dull and pointless. People asked to carry out an action rarely do it with any emotional coloration. But this coloration is precisely what drama demands.

Body language is language. We can read the emotions of animals by their movements and they can read ours by our movements. people do not simply walk: they walk purposefully or aimlessly, angrily or calmly, easily or laboriously. It is the same with all other movements. We have a seven-million-year-old language in movements and facial expressions. Libraries of dutiful docile actions don't cut it.

I applaud jasonjbrown's suggestion and recommend it to the company.

For multiple Kinnects, I think it COULD work... but there would have to be a program or driver written to tell the cameras how to handle the depth cue. They would have to communicate with each other and know where they were.

Multiple Kinnects would allow you to put your entire room... or person) into a 3D object. So instead of just motion capture... you could capture the entire person because all the Kinnects would allow for full vew of the person at all times.

And who says that the depth cues have to be read at all. They can simply be ignored and you wouldn't have a problem..

Motion capture libraries do have things like angry walk and tired run... but the main issue is finding one that fits exactly.

There are some good ones, but a lot of times, the issue is just finding a really good one quickly. CMU has a great collection... but unless you like looking through an enormous text file and going through folders, it won't do you much good. Animeeple has some great ones, but it still takes a little bit of time I think.

Finding animations for basic things like walking, running, jumping, and kicking is pretty easy in most cases. But what if you want someone to rock back and forth on their knees or what if you want someone to be sitting and leaning back ... in one of the chairs that don't lean back.

Or really any suttle animations for that matter? You could go and look for them, but motino capturing them would be quicker.

It's kind of like sound design. Sometimes I'm looking for a specific sound effect like a shower door opening or someone banging on a screen door. I could spend an hour looking for somehng that sounds kinda like that... or I could just go and record it myself.

I think it's possible to use smaller spaces if you're using smaller scale animations... but I do wish the price was a little les than 1K. As a 3D hobbiest, 1K is still a bit much to drop for mocap. I know it's on sale... but it won't always be on sale and I'm not ready to purchase itat this point in time.


Oh, one last thing... as good as mocapping stuff is, I've found that keying animations with a mixture of some dynamic motion, you can get some very nice looking movements. You can always take a video camera, act out your scenes (with help from friends) and then use the video as a referance. That's what some comanies do and it turns out rather well.

----------------------------------------------------
***BEST OF THE CORE**
:A music reel demonstrating the different types of music I produce for multimedia. Hope everyone enjoys.


Player not working? Use this link.

My other audio work.

Facebook page:

Creation begins at the core.