The move towards CGI. What consitutes 'the look' and how close are we getting?

Well it's remarkable what we are achieving now on a $300 little box. In realtime - allegedly - compared to those CGI shots, produced by immensely powerful render farms at god know what rendering times. Sorry but I think the FFTSW is starting to look very, very dated even compared to today's tech.

Couple of points:

- I'm not entirely sure if the Drake renders are from the engine and not from Maya using Mental Ray. But they're certainly not what the engine can achieve in the actual game in real time. Dialing up the quality settings is exactly one of the major differences in what makes offline CG look better.

- Again, do not confuse technology with art. The tools used for creating Drake are vastly superior to what was available 15 years ago - basically, 1.0-2.0 versions of Maya and Photoshop. Today you can sculpt in real time using a tablet on a character with 5-15 million polygons in Zbrush, view and paint displacements and multiple 4-8K textures in real time in Mari, and the poly modeling tools have gone through a small revolution as well.
Then there's a huge amount of experience on best approaches and practices to create digital characters, with a far better understanding of how human skin behaves and so on.

- Also, CG technology was not that advanced in those days, it's less than 8 years after Jurassic Park and the first realistic CG characters. Back then, there was no subsurface scattering, raytraced reflections, or global illumination; and less obvious but probably even more important is just how much better the techniques for facial animation have become.
 
Those are some good teeth!

Have to mention though that those images aren't from FF TSW, but from an amazingly talented guy called Luc Bégin:
http://forums.cgsociety.org/showthread.php?f=43&t=1192739

The head model's worth a look as well ;)
file.php
 
I think there's the potential here for you and Laa-Yosh and others to learn how the world of graphics is perceived b the lay-person. If Joe Public can't tell the difference between one complicated paint shader and one simple one, you'll learn you can save time and effort going simple (although having to live with the ignominy and creating deliberately sub-par art!). ;) On the subjective case of how people perceive and distinguish between realtime and non-realtime renders, it has to be worthwhile to determine what exactly people think it is that elevates games to 'that look'. I know I'm curious on others' theories as to why games look CG to them.

We are actually looking at these issues all the time, for various different reasons.

There is of course the rule of diminishing returns and the gap between CG and games is closing. But - and there's no modest way to say this - most of the time if there's a question about our work, it's usually not if it's realtime or not, but if it's live action or CG. So what we actually have to think about during projects is not about what we could remove and get away with, it's about what else can we do to make it even better. We're not working on streamlining everything as much as we can, but on developing better tools and practices and making them easier to add. You can say that any of these small improvements is barely noticeable to the average audience - but still, in the overall picture, they will add up, people will notice them unconsciously and thus they're definitely worth the extra investment.

I am just a modeler / facial rigger and haven't really been involved with anything else, but even I could easily point out dozens of small issues on any image from The Order character screenshots, and if there was a way to do a before/after image, you'd probably agree about the importance of these details that you'd believe to be superficial and minute.


Not that we don't care about efficiency either - but it's more about getting even better at focusing our efforts and resources to what matters, and less about stuff like building the hero characters faster and such.
 
I am playing The Order now and I don't know what exactly but I prefer ND facial animation by far. Cutscene are realtime but the light on the face are much better than during gameplay. And the face look much better. Same thing on UC4, I saw a better light in the cutscene.m between the to brothers.

Some materials are not far from offline rendering.
 
Why does it still look not like a photo of a real person?

Well, what I've suggested to Luc was to add even more imperfections and asymmetry - even though he aims to create a "beautiful" woman, these things are still crucial to achieve realism.

I've talked about this with regards to working with scans of real people, things in nature can get pretty far out and yet as a whole they're somehow less noticeable as a whole. It's incredibly hard to find the right balance if you're working from scratch, which is what he prefers to do.
 
- Also, CG technology was not that advanced in those days, it's less than 8 years after Jurassic Park and the first realistic CG characters. Back then, there was no subsurface scattering, raytraced reflections, or global illumination; and less obvious but probably even more important is just how much better the techniques for facial animation have become.
Man, creating those CGI dinosaurs back in the 90's must have been super challenging. They still hold pretty well considering the technology they had back then.
 
Yeah, this is another thing: her skin looks to perfect, also the shape and silhouette of her jaw. The inside of her nose. The eyes seem to be so reflective to me. Btw, from an modeling aspect: aren't the eyes a tad to large? And there is something with the ears I can't put my finger on.


I must admit, as an absolute laymen, the Nathan Drake UC4 shoots look way more natural to me.
 
At first glance it looks real, but then the eyes and mouth start to not look quite right. It's a lot more obvious when viewed shrunk to fit the screen than seeing it full size and viewing it piecemeal.

Searched the web for something similar. http://view.stern.de/de/rubriken/me...le-light-color-christin-original-2685048.html

I think the skin shader/lighting is off. I wonder if the neutrality of the pose also makes it veer off real?

Image1.jpg
 
Last edited:
I am playing The Order now and I don't know what exactly but I prefer ND facial animation by far.

Drake's face has a vastly higher number of bones and a much more detailed model, all of which allow for more nuanced and evenly distributed skin deformations.

The-Order-1886_models-1.jpg


vs

44jb.png


Drake's face is about as detailed as some of our previous work like Blackbeard:
blackbeard_wireframe.jpg


The one major difference is that we always model the face of the hero characters with a polygon layout that follows the individual facial wrinkle patterns, so we don't have to rely on animated displacement/normal maps and can sculpt those details into the model instead. We're using blendshapes and they have to be created from scratch anyway, so the only extra work we do is a few days spent on building the model.

But ND is using a different method: painting bone weights is incredibly time consuming, especially with such a big number of bones - so they're using the same mesh and bone layout for all characters in order to reuse data and effort. They just refit them to the shape of the different faces and then they only have to rework the elemental facial expressions, instead of starting from scratch.
 
Searched the web for something similar. http://view.stern.de/de/rubriken/me...le-light-color-christin-original-2685048.html

I think the skin shader/lighting is off.

This is a pretty complex question, because that photograph has very carefully tweaked lighting and make-up on the model and could also have been re-touched.

But this also shows just how important references are, it's a lot easier to create something realistic if you have something to look at and compare with.
 
Well it's remarkable what we are achieving now on a $300 little box. In realtime - allegedly - compared to those CGI shots, produced by immensely powerful render farms at god know what rendering times. Sorry but I think the FFTSW is starting to look very, very dated even compared to today's tech.

frank-tzeng-sdbvs.jpg


Making-of-Uncharted-4-Nathan-Drake-22.jpg


Uncharted-4-Nathan-Drake-Character-Model-Gets-Gorgeous-Hi-Res-Renders-468065-3.jpg

I am still amazed by this shots. Especially the first one in b&w: it hides the eyes, which are super difficult imo and being b&w helps also a lot because the difficulty of getting realistic colors are gone...

Also the last shots in this bright light look astounding to me.

Amazing what is possible today.
 
I must admit, as an absolute laymen, the Nathan Drake UC4 shoots look way more natural to me.

Near perfect females are always a lot harder IMHO, compared to a ruggedly handsome kind of male. You just sort of have less to work with, the eyes have a lesser number of points of interest to register, and so you have to make them a lot more accurate to sell the image.
 
Yeah, I was just thinking about the same: could it be that for a men, it is more difficult to create a convincing woman than another man? And if so, does the same hold the other way around: is it simpler to create CGI men for a female audience?
 
Well, I'm certain that there are some psychological elements there, although in my opinion it's a bit more about just how vastly different our (men's and women's) idea of female beauty can be. There's a huge range just for a start: the girl next door, fashion models, actresses both contemporary and classical, all kinds of traditional art (sculptures and paintings), celebrities, even porn stars. There's an incredible range of hair styles, make-up, cultural inheritance, and so on. There's a LOT of serious research as well, with things like the 'beauty mask', averages of hundreds or thousands of photographs, theories... Compared to this, men are really simple already ;) Although, maybe London boy could provide another interesting perspective here.
However, Drake is an interesting example because the team behind the character had both men and women in it, so I'm not sure if there's anything like this could be involved.

But I still think that the visual aspects I've mentioned are probably more important. Things like stubble, wrinkles, scars, more complex forms and such provide a lot more detail for the eye, or in other words the visual elements of the brain, and so it's a little easier to get away with less then completely realistic execution, the minute flaws are less noticeable because they're hidden under all the "noise". And of course Drake is still a stylized character, even if he's far more realistic now.
So males are probably easier to create no matter your gender.

Which makes me even more curious to see female characters from UC4. I'm quite sure that they're actually a much harder task for the team and this is at least in part the reason why we haven't seen any glances of them yet.
 
Back
Top