This makes AVATAR look like a cheap video arcade game

Status
Not open for further replies.
Postmaster said:
Have you ever tried to make a 3D model of a human behave and move like a believable human by hand Rodney?

Compared to this, the task of tying a 3D model (human or navi or whatever) to motion capture data is pretty straight forward and simple. Once the model is rigged you just throw the data at it.

Heck it´s almost automatic. Check it out: http://www.xsens.com/en/general/mvn

Frank
Click to expand...

I suppose thats one way to devalue all the work that went into the animation... :thumbdown:
 
noob1nw8.jpg


hey, you asked for it :D
 
Loki said:
I suppose thats one way to devalue all the work that went into the animation... :thumbdown:
Click to expand...

But you are realizing that the NaVi are NOT animated right?

There is no animator doing the moves (except the tail) or face expressions - they just copy-paste the capture data performed by real life actors on the 3D characters. It´s a mathematical process or workflow, not animation like in Toy Story.

Frank
 
But to me, humans aren't even the discussion here. One of the most disappointing aspects of Avatar is how bad the machines looked, compared to THESE machines.

Maybe it was the sheer amount of work in Avatar that forced them to make some choices about what to focus on and what not to. I think there were what, 800 digital artists on this movie?
 
At this point I think it's the lighting more than anything. The lighting and texture of the image in this guy's work is what sells it the most. Also, as others have mentioned, flat mechanical shapes are about the easiest thing for CGI to do well. James Cameron said in an interview that the lighting of the CGI to him was the most challenging aspect and I don't think they cracked it. They still went with the tendency of CGI artists to render everything too cleanly. People compare this to District 9 but that's not entirely fair because there you had CGI in real environments, so they had to match the CGI lighting to the rougher textures and imperfections of the captured footage. Avatar was often 100 percent CGI so they had no reference. On top of that, you had an alien planet with different natural light sources than what you would find on earth. Still, they could have been a little bolder about introducing imperfections into the lighting and film characteristics to give it a more realistic feel.
 
Batutta said:
At this point I think it's the lighting more than anything. The lighting and texture of the image in this guy's work is what sells it the most. Also, as others have mentioned, flat mechanical shapes are about the easiest thing for CGI to do well. James Cameron said in an interview that the lighting of the CGI to him was the most challenging aspect and I don't think they cracked it. They still went with the tendency of CGI artists to render everything too cleanly. People compare this to District 9 but that's not entirely fair because there you had CGI in real environments, so they had to match the CGI lighting to the rougher textures and imperfections of the captured footage. Avatar was often 100 percent CGI so they had no reference. On top of that, you had an alien planet with different natural light sources than what you would find on earth. Still, they could have been a little bolder about introducing imperfections into the lighting and film characteristics to give it a more realistic feel.
Click to expand...

This is why they hired Roger Deakins for Wall-E. And the first half stuff with wall-E alone on Earth was really really stunning and beautiful. Did Avatar have a traditional DP on the CG stuff?

What I think would be interesting would be to give some approximation of the lighting on those empty warehouse sets. Like some gobos to represent tree leaves, etc, a single huge source to represent sunlight. Not sure how much this would help or hinder the capture process though...
 
Postmaster said:
But you are realizing that the NaVi are NOT animated right?

There is no animator doing the moves (except the tail) or face expressions - they just copy-paste the capture data performed by real life actors on the 3D characters. It´s a mathematical process or workflow, not animation like in Toy Story.

Frank
Click to expand...

I think that might be more true for the body movements than the facial expressions. I read an article where they said they were able to use the facial captures as a reference but couldn't map them automatically because of the fisheye distortion that was caused by the cameras being so close to their faces.

I generally like animation done by hand than by motion capture as well. Motion capture probably makes more sense if you're going for ultra-realism though.
 
Postmaster said:
they just copy-paste the capture data performed by real life actors on the 3D characters.k
Click to expand...

Not even copy-paste ... they where supposedly doing it in REAL TIME, on set. Cameron's "Viewfinder" had (lower-res) real time renders of the characters including facial expressions as he looked "through it".

At that point, it's just a question of processing power before you could have a full quality live interview with an avatar of an actor. (Na'vi or other character).

Only some of the Na'vi you see in the movie are "avatars" as characters in the movie (Neytiri not being one for example); however they are ALL actual digital avatars of the actors.

Avatar isn't just the name of what's going on in the movie; it's the technology used to make it.



And yes, the video posted here is very good work. Leaves you feeling in awe at the work that has gone into it. Fantastic stuff.
Though, it's not 100% if you are using photographed material for textures (including sky).

- Mikko
 
Status
Not open for further replies.
Back
Top