Sunday, February 28, 2010

HOW TO CAPTURE MOTION

The idea of capturing motion like a butterfly in a net is appealing to me, however, what’s captured from the actors in black spandex MOCAP suits is actually a file with a list of numbers indicating the positions of specific points in 3D space. This computer data is then manipulated by motion capture (mocap) specialists to produce films like BENJAMIN BUTTON and AVATAR, among others.  

James Cameron says AVATAR is not animation and, since the action is created by actors in mocap suits with built-in and data points, I have to agree with him. Mocap is not CG animation because no animators are involved in creating it.  It is a different medium which may soon have its own Oscar category.

After questioning several mocap specialists, here’s what I learned:  
While the  data is being captured from actors in mocap suits, computers can render it almost in real-time, with some fairly realistic lighting, texturing and physics simulation, creating characters which can then be used directly on screen. These software-generated characters are stiff and wooden and mostly very creepy.

For the big movies, mocap data is generated by actors and seen simultaneously on computer monitors as low-resolution models. You can see this happening here (just skip the blather). The data can be further processed by CG animators who animate extra bits not on the captured motion like clothes, tails and hair, but these animators don't create the action, they just refine it.

And this brings us to performance capture or PERFCAP, where data points include fingers, face and facial cameras to capture actors’ expressionsAccording to one mocap specialist, “The cutting edge of performance capture is still what they've done on BENJAMIN BUTTON…” Take a look at this.  Then there's WETA's work on AVATAR close-ups. AVATAR is a good effort, but the characters are still numb-faced and creepy no matter how big and translucent their eyes. The lights may be on but there’s nobody home. See the uncanny valley effect.

Frankly, since we already have excellent motion picture cinematography and excellent animation (CG and traditional), I just don’t see the need to capture motion and torture it into digital graphics. 

Thanks to all who provided information for this mocap overview and who chose to remain anonymous and to Eloi Champagne, who didn't.
*
See also:  An interview on MOCAP with ELOI CHAMPAGNE, Motion Designer, 2D/3D animator and President of STUDIOCRONOS[4] located in Montreal, Canada.

2 comments:

Rebecca Sutherland said...

Very Interesting. Haven't seen Avatar yet, but I've seen the CGI effects you've featured here. Understand your point about the numb faces. Benjamin Button's face was much better, but I still spent the whole movie wondering (and worrying) how they did it, rather than concentrating on the action/dialogue. I guess that might be me having to suspend my belief a little more.
Seems like they have to make the technical process more complicated before they make it simpler.Or are we all too knowledgeable about the smoke and mirrors? I think not!

Nora Lumiere said...

I just don’t see the point of MOCAP, really; it’s not as if animators can’t create the exact same actions as the actors.
I think it's easier to suspend disbelief when the characters aren’t so lifelike, as in traditional animation for example.
This was a hard post to write and I'm glad you found it interesting.