Mocap solution for C4D coming soon?

Become a member of the CGSociety

Connect, Share, and Learn with our Large Growing CG Art Community. It's Free!

REPLY TO THREAD
 
Thread Tools Search this Thread Display Modes
  12 December 2013
Originally Posted by Crix Dorius: How about finger motion tracking?


Hi,
this is a possible option for next product iterations, not for the initial release.
The system is flexible enough to add such options later if requests are coming in.
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  12 December 2013
Originally Posted by EricM: It would be ace to have a working out-of-the-box FK to IK rig in C4D similar to HumanIK, Ikinema or nevronmotion to deal with retargeting all thosenew mocap files...


I can have a look at this. Should be possible, as everything can be accessed through Python, C.O.F.F.E.E. and XPresso even while recording or just playing back the incoming data.
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  12 December 2013
Originally Posted by saigam: Any hint about the price?
If the price is right for me, I'll be the first in the line as soon as it comes out, this is very exciting.


I cannot give you a final price yet, as the production costs of the cases are still to be negotiated. The price of an optimal full body set might be too expensive just for hobby/fun use, but will be much cheaper than already available Pro Systems. Of course it depends also on the amount of data you need. For example a doctor just might be interested in arm or leg movements of his patients. You are always free to start with just a few sensors and add more later. We wanted to make sure to keep the production costs as low as possible without making compromises about the features. I will take care to make the prices available as soon as I can.
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  12 December 2013
Originally Posted by Andy1010: Does this differ from the Xsens suits or is this the same thing just more customized for C4D? Also can you do 6dof camera moves with this, or us it limited to body captue only. More info would be great. Keep up the good work


Hi,
yes, XSens has a comparable system, but our price point will be different. Also we differ in some features like external sensors and tried to keep the hardware and software requirements as low as possible. Our systems runs with MacOS and Windows and most of time a simple laptop will do.
As you use C4D to read out the sensors, you can apply their data to any object, parameter channel or Expression setup you like. The MoCap can also be about a horse, a machine or a car, driving down a street. Doesn't matter. Just absolute position data cannot be captured by sensors. That is a general limitation of all sensor based systems. As long as you don't need exact object positions in space, sensors are the most flexible option for MoCap, as you can use them on set/on location with no additional infrastructure around (cameras, DOF sensors).
Thanks for your interest! I will try to give more infos as soon as I have something new to show. Hopefully in early January.
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  12 December 2013
if this is working good count me in...looks very promising. do you have any stuff to show already?
 
  12 December 2013
Originally Posted by MAN0: if this is working good count me in...looks very promising. do you have any stuff to show already?


Thank you. I will record some videos with the early prototypes I have in front of me, to give you all an idea about the implementation.
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  12 December 2013
Quote: I can have a look at this. Should be possible, as everything can be accessed through Python, C.O.F.F.E.E. and XPresso even while recording or just playing back the incoming data.


Thanks for considering it. Because I actually learned in my brief homemade mocap experience that there's more to it than just capture in a successful mocap pipeline.

It's obviously the first step, but if you don't have the right tools to edit/apply/retarget/layer and sequence the raw capture, it's pretty hard to do anything useful with it. And C4D really lacks some key elements for it :
- A standardized human figure with FK/IK controls embedded: Daz/Motionbuilder/HumanIK/Nevronmotion for Lightwave they all have it and that's why it's easy to use them. It could maybe be done as a new template from the character builder, but it's beyond me.
- A better retargeting tag (see Cactus Dan's) with rotation offsets and drag and drop lists.
- Euler filtering (thanks to per anders for providing a scipt for that)

For now, I'm usually doing the mocap retargeting/layering/editing in an app like Daz or Iclone 5 (it's got Autodesk Human IK included at a fraction of the price of Motionbuilder) and only export the final result in C4D for rendering.
 
  12 December 2013
Originally Posted by EricM: ...
For now, I'm usually doing the mocap retargeting/layering/editing in an app like Daz or Iclone 5 (it's got Autodesk Human IK included at a fraction of the price of Motionbuilder) and only export the final result in C4D for rendering.


Totally agreed, its so much harder without a good standardised Full Body IK rig to work with. The ikinema system is very nice.
__________________
Cinema R15, MacOS X 10.9, iMac & MacBook Pro
www.sector3.com.au
www.3sprockets.com
www.trickstargames.com
 
  12 December 2013
This all sounds very cool. I was going to ask about using it for quads and I see that horses were mentioned. So, can this system be applied to a dog and a cat? Can we do a person riding a horse and get both objects working together in C4D obejects? Can we do a scene with two or more people and a dog, for example?
 
  12 December 2013
Originally Posted by oliveoyle: This all sounds very cool. I was going to ask about using it for quads and I see that horses were mentioned. So, can this system be applied to a dog and a cat? Can we do a person riding a horse and get both objects working together in C4D obejects? Can we do a scene with two or more people and a dog, for example?


The sensors are already quite compact, but not small enough for smaller animals like cats. You might be able to attach some sensors on larger dogs, but this all has to be tested.
Larger animals (monkeys, horses, cows, elephants etc.) shouldn't be a problem if they are kind enough to carry the sensors without eating them for lunch :-)
You can capture different humans or animals and humans at the same time. It just depends on the amount of sensors you own. If the amount of sensors gets too heavy for one WiFi network, you can even use additional PCs and networks, because the data is timestamped. The data will be in sync again when you merge it on one PC. Capturing just one or two humans isn't a problem with just a single network though.
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  12 December 2013
Originally Posted by ArndtvonKoenigsmarck: Larger animals (monkeys, horses, cows, elephants etc.) shouldn't be a problem if they are kind enough to carry the sensors without eating them for lunch :-)

If the amount of sensors gets too heavy for one WiFi network, you can even use additional PCs and networks, because the data is timestamped.

The data will be in sync again when you merge it on one PC. Capturing just one or two humans isn't a problem with just a single network though.

What would be an estimate on the number of sensors per wifi? Could maybe 8 people, such as a conference room scene around a table, work? I am curious and hopefully gaining a better understanding.

How big are the sensors and how do they attach? Is each sensor also a transmitter making it all a wireless system? If so, I would assume that you the have a different channel assigned to each sensor and a computer app maps these channels to each joint of the object skeleton? Or, am I way off thinking this way?
 
  12 December 2013
Originally Posted by oliveoyle: What would be an estimate on the number of sensors per wifi? Could maybe 8 people, such as a conference room scene around a table, work? I am curious and hopefully gaining a better understanding.

How big are the sensors and how do they attach? Is each sensor also a transmitter making it all a wireless system? If so, I would assume that you the have a different channel assigned to each sensor and a computer app maps these channels to each joint of the object skeleton? Or, am I way off thinking this way?


The number of supported sensors per router/access points depends on the available bandwidth and type of router/access point. Typical numbers are about 50 to 60 sensors supported by a recent access point. That should be more than enough for common MoCap needs. The main sensors will be about the size of a cigarette pack. We also have plans for slave sensors, that are much smaller in size and can be cabled to the main sensors, but basically the setup could be cable free. Each sensor has a unique IP. You can bind the sensor data to any parameter or object in CINEMA 4D by special C4D Tags. By this you can also use the data of one sensor to drive multiple Objects or parameters at the same time.
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  12 December 2013
Originally Posted by ArndtvonKoenigsmarck: The number of supported sensors per router/access points depends on the available bandwidth and type of router/access point. Typical numbers are about 50 to 60 sensors supported by a recent access point. That should be more than enough for common MoCap needs. The main sensors will be about the size of a cigarette pack. We also have plans for slave sensors, that are much smaller in size and can be cabled to the main sensors, but basically the setup could be cable free. Each sensor has a unique IP. You can bind the sensor data to any parameter or object in CINEMA 4D by special C4D Tags. By this you can also use the data of one sensor to drive multiple Objects or parameters at the same time.
Thank you Arndt for this information. I look forward to this happening.

Besides full body Mocap, will there be any ability to do full facial and head Mocap? Eye direction and mouth for lip syncing? Fingers and thumbs on hands? I am guessing the sensors are too big for those types of placement.

Last edited by oliveoyle : 12 December 2013 at 01:50 PM.
 
  01 January 2014
Just to give you some new infos, here is the final case design.
Facial or finger tracking is not included yet. Regarding facial tracking, this would require an optical/camera solution. There are already products available for this, so combinations might be possible. Finger tracking is not an easy task, at least if we talk about precision. We are already thinking about special gloves, but not for MoCap.
Attached Images
File Type: jpg Preview_small.jpg (68.2 KB, 58 views)
__________________
Kind regards
Arndt von Koenigsmarck

private: http://www.vonkoenigsmarck.de
company/webshop: http://www.rodenburg-verlag.de
 
  01 January 2014
if the mocap data would be close to xsens data, in terms of clean/noise, this will be a game changer. I cant wait to get my hands on it. It would be cool to know the ballpark price, to get prepared for it . There are a couple of wireless mocap options out there @ 4k price point, but data is too noisy/jerky

It would be possible to have two full body packs to interact with each other? (fight, etc)
__________________
My link

Last edited by stoiQa : 01 January 2014 at 08:31 PM.
 
reply share thread



Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

vB code is On
Smilies are On
[IMG] code is On
HTML code is Off
CGSociety
Society of Digital Artists
www.cgsociety.org

Powered by vBulletin
Copyright 2000 - 2006,
Jelsoft Enterprises Ltd.
Minimize Ads
Forum Jump
Miscellaneous

All times are GMT. The time now is 11:35 AM.


Powered by vBulletin
Copyright ©2000 - 2017, Jelsoft Enterprises Ltd.