Much to the concern of the other animators on my team, we've managed to secure a motion capture suit for our studio! We're not going to use it to replace anyone; I only have three artists on my team right now and it's hard enough making 3D animations as it is. This will be a great way for us to increase our rate of production.
Here's what we're working with:
Xsens MVN Awinda
https://www.xsens.com/products/xsens-mvn/
Before purchasing the MVN Awinda, I requested some sample data, which Xsens was quite happy to provide. I used it to run a bunch of tests. Below is the first attempt at linking the data to some of our characters.
Test 01:
I've done some light cleanup in Motionbuilder to improve the foot contacts and adjusted the hands to keep them from intersecting with hips and heads. This is very near to the raw data. I should probably mention that we already animate all our characters with an HiK rig, either in Motionbuilder or Maya, so implementing mocap is theoretically very easy for us.
Test 02:
The first test went well so I decided to do a more detailed test. I applied the same data to Wanna-Know-Wally, a character that I recently remodeled/rigged to get ready for a short he's soon to star in. To save time, I also used this render to test some linear lighting setups that I want to use in the mentioned upcoming short:
Here I've kept the minimal cleanup, but added some very rough lipsync and facial animation. This level of quality represents about an hour of animation time, total. It's a pretty low bar, but all the non-artists I showed it to loved it.
Test 03:
To provide a wider range demonstration, I pushed the next test further:
In this test, I've spent a whole day cleaning up the animation: adding new animation, embellishing, making better lip-sync, and adding prop interactions. This process is definitely slower than just accepting the rough & dirty data, but spending one day of labor on 10 seconds of animation is still a great rate of production for us. Although this is not likely to win any awards, it doesn't look too bad, either. Our target audience of young children won't likely mind.
That last test locked in the sale for us. Our VP was on board and we ordered our Awinda. Xsens got it to us in about six weeks.
Test 04:
Here's my first test with data captured in-house:
This was a really basic test, mostly just to show that it works. We have a request to do an interview with one of our characters in the real world, so I used this as a chance to show what our characters would look like in that environment. I spent a few hours on this total... but a few days doing handstands, running around, testing the range, and enjoying this suit like a happy artist should.
I'll mention here that the range on this suit is awesome. Full demo coming soon! We're wrapping up production on a very quick Club UYN project today, it'll be our first published piece that uses the suit.
Test 02:
The first test went well so I decided to do a more detailed test. I applied the same data to Wanna-Know-Wally, a character that I recently remodeled/rigged to get ready for a short he's soon to star in. To save time, I also used this render to test some linear lighting setups that I want to use in the mentioned upcoming short:
Here I've kept the minimal cleanup, but added some very rough lipsync and facial animation. This level of quality represents about an hour of animation time, total. It's a pretty low bar, but all the non-artists I showed it to loved it.
Test 03:
To provide a wider range demonstration, I pushed the next test further:
In this test, I've spent a whole day cleaning up the animation: adding new animation, embellishing, making better lip-sync, and adding prop interactions. This process is definitely slower than just accepting the rough & dirty data, but spending one day of labor on 10 seconds of animation is still a great rate of production for us. Although this is not likely to win any awards, it doesn't look too bad, either. Our target audience of young children won't likely mind.
That last test locked in the sale for us. Our VP was on board and we ordered our Awinda. Xsens got it to us in about six weeks.
Test 04:
Here's my first test with data captured in-house:
This was a really basic test, mostly just to show that it works. We have a request to do an interview with one of our characters in the real world, so I used this as a chance to show what our characters would look like in that environment. I spent a few hours on this total... but a few days doing handstands, running around, testing the range, and enjoying this suit like a happy artist should.
I'll mention here that the range on this suit is awesome. Full demo coming soon! We're wrapping up production on a very quick Club UYN project today, it'll be our first published piece that uses the suit.
