Realism in for CG Characters, How
I couldn't find any thread about this, i am not sure how to search, so if it is duplicated, i am sorry.
I would like to know, how does realism is achieved in CG Characters, especially in VFX movies in Cinematic game trailers? LIke Blur, Blizzard, ILM Does in their projects.
It's looks so stunning, i want to do myself.
I would like to know, how does realism is achieved in CG Characters, especially in VFX movies in Cinematic game trailers? LIke Blur, Blizzard, ILM Does in their projects.
It's looks so stunning, i want to do myself.
Replies
[ame="
If only they put that hard work in to Diablo 3.
That's probably the best mental logic when doing realism. As realism in my experience happens to only fail when you start over analyzing/stare at your work. Analyzing critically about your work, I notice only works well when you do fictional types of work (cartoon, stylized, etc.).
In the end, after work is usually made, its really all down to your shader and light setup + renderer.
But like its always said... Study your surface/length anatomy of your subjects. Having exercising good mental memory of imagery you see and replicate it on your surfaces well. In the end will do you all the justice you need.
Character guys may say I'm wrong on this. But don't go nuts studying your inner anatomy. Because it really has no relevance to the subject of art, unless you are a technical artist. Whereas knowing your edge-loops is really the most technical you should get with it. Because you will have already studied length and proportion, with surface information while constructing your final character.
How to learn these?
normaly to speed the process up you have to render with pass so you can tweak the settings in compositing.
How do you learn anything in life? You try, you fail miserably, you figure out why you failed, you fix it, and you repeat the experience.
How to find this?
I think what people are getting at is you need to research a lot of stuff and practice yourself, there's no sure fire way to making incredible characters from the start.
Rendering for cinematics requires using multiple layers and passes that take a lot of time and care. You need to use a compositing package (Nuke, or Composite by Autodesk) for the final animation sequence [they are node-based and much more powerful than layer-based aps]. If you want to produce a cinematic by yourself you will have to learn many skills (concepting, storyboarding, modeling, anatlomy, sculpting, texturing, animation-ready-final model, rigging, skinning, animating, advanced lighting, rendering with MRay, VRay, Renderman, etc., advanced shaders (such as SSS shaders), displacement mapping, effects (particle effects, physics, cloth dynamics, fluid effects), look development, composition and camera techniques, and possibly matchmoving (if using video shots). I am only aware of two people that know all those skills (that have produced a cinematic). Look up Alex Alvarez of Gnomon (founder as well as instructor) for inspiration if you want to head down this path. That guy is amazing (he taught a Master Class at Gnomon that covered all of the above). Just start and practice, practice, practice.
it was very high budget approach back then but nowdays it much more common.
this is one of the best ( and affordable way to do your own)
http://www.ir-ltd.net/
facebook page
https://www.facebook.com/InfiniteRealities?fref=ts
but still need human eye of detail and creative touch to make it perfect and believeable,
this thread is an example if everyone given the same 3d head scan and textures. and challenged to render it as real as possible.
http://www.cgfeedback.com/cgfeedback/showthread.php?t=751
But i am not sure.
I am not really good for coding, at last i am not that smart for programming and coding.
Is it possible to know everything in CG and to make own cinematic characters?
No, it's not true. For one, how do you scan someone who doesn't exist? Most CG characters are fictional, idealized people.
What you need is a very talented and experienced artist to sculpt the head and paint the skin texture. Fortunately we have one, Karoly Porkolab, and he can do this:
Then what our shading/lighting/comp guys do is to take a lot of very high quality reference photographs and tweak the shaders and then the final comp made from many many layers to resemble the photos. They use specific lighting and they also took a lot of extra images of side-lit and back-lit heads to study the amount of SSS and so on.
I've seen traditional software users like Max or Maya capture peoples likeness more vertice by vertice then 3D-Scans did in majority of the games.
However, most of the scans are also heavily processed. And the voice talent and the scan talent are usually different people, too.
Mass Effect 1 may have used scans as starting points for some of the human(oid) characters, too - Shepard, Liara, Kaiden and Ashley were all based on real people.
Not sure if they scanned the talent for Miranda, Samara and whoever else was based on real people in the sequels, though. Kolby Jukes may be able to tell you as he did the Miranda character. Although he has other celebrity sculpts in his portfolio for the Dark Knight figures and I think those were 100% hand sculpted.
Also, I've seen our sculpt guy do a totally real Samuel Jackson head when we were working on the (canceled) Avengers game intro.
So scanning does happen for games, and of course for all digital doubles in movie VFX. However, when you're working on a CG character that you want to show up close, it's usually something that doesn't exist - like the old Brad Pitt in Benjamin Button - because otherwise it'd be a lot cheaper to just film that person.
Although... well they did have a bunch of CG actors in Beowulf, but they all looked creepy...
Whoa crazy.
To add to the discussions, seems like the final elements for a realistic character are also the camera, and the character's fidelity of animation.
The camera for the subtle artifacts of realistic perception, and the same for animation... minor movements in the skin, or articulation can be enough to sell something as real.
I almost would suggest that life like qualities have more about how things move and act than how they look. You can have decent cg, but you can't have just decent animation with realistic renderings.
So after you sculpt the characters, you need nice texture and shaders on it, followed with nice lighting, render and comp to get some realism on characters?
I know there so much tutorials on modeling, sculpting, but, how about making shaders, texturing, lighting, render and comp?
Are you guys gonna recommend me some tutorials about cinematic characters? Or should i take Gnomon course, that it's from Alvarez himself, Shading and texturing characters?
There is no golden answer as to how to make amazing shaders, etc, like its already been said it is a matter of practice and dedication, trial and error. If you need to check out some tutorials, by all means check out Gnomon, even Digital Tutors, etc. but I don't really think you needed any of us to tell you that.
just a shame they can't give the same level of credit to people that actually created in-game assets for them properly i guess.
"yeah let's give credits to the guy who stood on a plate for 20 minutes, and ignore the guy that gave us an entire levels worth of assets and weeks of his life to do it".
[ame="
yes thats one term thats used.
However, it should be noted that when you are looking at the higher end of CG character work in movies, that it will be more than one person working on a shot like that.
Someone will model the character (and also texture), sometime another person will do the texturing. Then a diiferent person will do the lighting and rendering, plus there can be someone else just writing and tweaking the shaders. Then after being rendered out, someone will composite the shot and then at the end, everything will be graded, again by a different person.
And I haven't even mentioned stuff like dynamics like hair, cloth and clothing, which again will be done by different people.
Basically, overall you can have somewhere between 5-10 people working on a character and shot.
Yeah, a single person can do, of course. What Bellsey said is that in a production pipeline, with deadlines, work will be usually divided, to make things easier and faster.
I am not sure to ask here, or to register and post at cgtalk. Is there any sources of creating a high end character for movie or cinematic?
http://www.3dtotal.com/index_tutorial.php
The credits for the game are nearly 10 minutes long, and they list a lot of people under just a company name (for example Digic is nearly a hundred people) - if all the contractors' names were to be included, the credits would never end.
I personally have no problem with this...
Do what you want to do. If you enjoy doing it you will be better at it. You might like sculpting or modeling or texturing or animating or whatever else you can think of.
Honestly, just stop asking questions and go do it. Google is your friend. Grab some 3D applications and just start following tutorials. It sounds like you're just looking for some magic tutorial which will teach you everything in a couple hours. Truth is, these artists are amazing because they've spent years refining their work.
Just create
Another thing to consider, is quality. By spreading the work to specialists, you can get the very best at every stage.
It's very hard to have one person literally good at everything. I'm sure there are people who will say, 'not me, I'm good at most things', but in reality I'm sure there's some aspects to their skillset that (despite being good) perhaps aren't quite as strong as others. Besides, I'm very much of the opinion that no matter how good you think you are, there's always room for improvement and new things to learn.
would like to jump in.
What render engine should i learn/use, Vray, Mental ray or renderman?