Something interesting to consider that I hadn't thought of until just now, but it's worth noting that you can't necessarily just easily map the movements from an adult human to a character with proportions like Gollum in the LOTR... My guess is that you can still take all of the relative movements and apply them to each joint to get an animation to play, but I'm sure it took a lot of artistic effort and clean up to make everything look acceptable. Not really something I would have thought about without having seen the actor in the motion capture suit side-by-side with the character.
ncastaneda02
@bennyd87708 I think this is where topology becomes really important - as long as the actor and the character are topologically equivalent, it shouldn't be that difficult of a problem to smoothly map a movement in the actor to a movement of the character. You would likely need some touch ups by a CG artist, but I imagine that there is probably some smooth function mapping the mocap suit to any topologically equivalent character model
rheask8246
Lord of the Rings was the first movie to use real-time motion capture! The actor, Andy Serkis, performed with a motion capture suit as displayed on the slide, and animations of Gollum's face were animated manually.
sartk
@bennyd87708 Yes, I wonder if it will work anyway though. Especially if we consider body invariant information like joint angles, it's unclear how shared that is across different body types.
pcg108
I’m curious as to how modern motion capture technologies might use the actors own features and body shapes in the digital character and apply textures and transforms to them rather than build an entirely new object from scratch. I wonder if it would be computationally more efficient to do so using some of the techniques like the depth mapping from project 4
sZwX74
I'm not exactly sure how it works, but I think it will still work for different body types, as there are inside delves into films that show that these body suit motion captures even work to model humans with extra body parts, such as more arms or wings coming out of their shoulders. I think this is done in a similar way to how @ncastaneda02 described topologically.
Something interesting to consider that I hadn't thought of until just now, but it's worth noting that you can't necessarily just easily map the movements from an adult human to a character with proportions like Gollum in the LOTR... My guess is that you can still take all of the relative movements and apply them to each joint to get an animation to play, but I'm sure it took a lot of artistic effort and clean up to make everything look acceptable. Not really something I would have thought about without having seen the actor in the motion capture suit side-by-side with the character.
@bennyd87708 I think this is where topology becomes really important - as long as the actor and the character are topologically equivalent, it shouldn't be that difficult of a problem to smoothly map a movement in the actor to a movement of the character. You would likely need some touch ups by a CG artist, but I imagine that there is probably some smooth function mapping the mocap suit to any topologically equivalent character model
Lord of the Rings was the first movie to use real-time motion capture! The actor, Andy Serkis, performed with a motion capture suit as displayed on the slide, and animations of Gollum's face were animated manually.
@bennyd87708 Yes, I wonder if it will work anyway though. Especially if we consider body invariant information like joint angles, it's unclear how shared that is across different body types.
I’m curious as to how modern motion capture technologies might use the actors own features and body shapes in the digital character and apply textures and transforms to them rather than build an entirely new object from scratch. I wonder if it would be computationally more efficient to do so using some of the techniques like the depth mapping from project 4
I'm not exactly sure how it works, but I think it will still work for different body types, as there are inside delves into films that show that these body suit motion captures even work to model humans with extra body parts, such as more arms or wings coming out of their shoulders. I think this is done in a similar way to how @ncastaneda02 described topologically.