Yes, he posted again and confirmed it was indeed running in real time on a Pro.
https://twitter.com/HIDEO_KOJIMA_EN/status/804542406959853568
Accurate human rendering in game [2014-2016]
- Thread starter babcat
- Start date
- Status
https://twitter.com/HIDEO_KOJIMA_EN/status/804542406959853568
While everybody's looking at the detail in the skin and the blood, can I just say "Dat fabric!"
Yibing Jiang's been busy? https://www.artstation.com/artwork/6kEmV
I really like how convincing Ellie's eyes look, you can tell the anger and determination from them not just some random specularity sprinkled on, it's even better in motion too. The lighting also feels like it has more depth and skin tone looks more natural in comparison to UC4.
Not sure if there are separate "cinematic" and "gameplay" models for the hero characters, it's just not worth the effort to build and rig two assets.
Shaders and lighting are a lot easier to scale back (requires practically no artist work), and they probably have a much greater effect on frame rate anyway.
Shaders and lighting are a lot easier to scale back (requires practically no artist work), and they probably have a much greater effect on frame rate anyway.
I was talking about this. I want less of those reductions in future games.
That is a backwards approach - features are not removed for gameplay, but added for cinematics.
Locking the scene's content and the camera's movement allows the developers full control over the rendering performance. As cutscenes are usually about character moments, they usually only feature a few actors and a closed environment, so the load on the renderer is much lower than what it has to deal with in an open space with many characters and lots of action. This frees up a lot of resources, and it is a logical move to utilize it for better rendering quality. Cutscenes will always work better with more directed lighting, and better shading enhances that even further - it can direct the viewer's attention, add to the mood of the scene and so on.
So what you actually want is for developers to leave the CPU/GPU resources unused and settle for cutscenes that look worse, and have less dramatic effect. It makes little sense, at least to me.
Locking the scene's content and the camera's movement allows the developers full control over the rendering performance. As cutscenes are usually about character moments, they usually only feature a few actors and a closed environment, so the load on the renderer is much lower than what it has to deal with in an open space with many characters and lots of action. This frees up a lot of resources, and it is a logical move to utilize it for better rendering quality. Cutscenes will always work better with more directed lighting, and better shading enhances that even further - it can direct the viewer's attention, add to the mood of the scene and so on.
So what you actually want is for developers to leave the CPU/GPU resources unused and settle for cutscenes that look worse, and have less dramatic effect. It makes little sense, at least to me.
It's kind of the reverse in movies, at least with stunt doubles - they take over when there's a lot of action and it's harder to see who's doing the things. Also, nowadays, we have digital stunt doubles and head replacements and such to make the illusion even more convincing 
Body doubles, now, that's a completely different thing...
Body doubles, now, that's a completely different thing...
While it would be foolish to not use the extra resource on cutscenes it is still very annoying going from one extreme to the other. In tomb raider there is a huge difference with cutscene graphics even when completely maxed out 4k on pc. In uncharted 4 the difference is even more noticeable:
That's why while i am still impressed with the last of us 2 teaser i am careful not to get too hyped because the gameplay will never end up looking that good. I personally would enjoy it better if the cutscenes look exactly like the gameplay so there are no discrepancies other than hero lighting. Correct me if I'm wrong but the phantom pain only adds hero lighting to the cutscenes with no extra shaders and they still look great, every effect is still present in gameplay.
That's why while i am still impressed with the last of us 2 teaser i am careful not to get too hyped because the gameplay will never end up looking that good. I personally would enjoy it better if the cutscenes look exactly like the gameplay so there are no discrepancies other than hero lighting. Correct me if I'm wrong but the phantom pain only adds hero lighting to the cutscenes with no extra shaders and they still look great, every effect is still present in gameplay.
Last edited:
Orion
Regular
What I'm wondering is what the heck is going on over at hollywood. Even human characters that appear for like 2 seconds on screen you can still easily tell they're cg, cough, rogue one, cough. For a few moments parts of the Death Stranding trailer looked more impressive, and far more convincing, yet obviously a fraction of the budget and resources were used.
And yet, most people only noticed small hints that Paul Walker's face was CGI in the closing moments of the movie when it was clear film-makers were saying goodbye to him, even though his face was CGI across large chunk of a film.
IMO, there is a strong chance Disney did not want to pay [or they did not have enough time] to implement those Rogue One characters as best as they could. Film was fast tracked, and massive reshoots changed the flow of entire 2nd half of the movie.
IMO, there is a strong chance Disney did not want to pay [or they did not have enough time] to implement those Rogue One characters as best as they could. Film was fast tracked, and massive reshoots changed the flow of entire 2nd half of the movie.
Cody Walker looks a lot like Paul Walker. Going from there is plausible a bit more easy. Nevertheless the work there was of course masterly.
I found the faces in Rogue One obviously much better than in any video game. Probably it was this Uncanny Valley effect for many people. Which is a subjective effect. Some years ago today's faces in games would have been also called Uncanny Valley for many people. It may take a time to settle.
As far as I know there were also several human CIG faces in Rogue One. Not just
.
I found the faces in Rogue One obviously much better than in any video game. Probably it was this Uncanny Valley effect for many people. Which is a subjective effect. Some years ago today's faces in games would have been also called Uncanny Valley for many people. It may take a time to settle.
As far as I know there were also several human CIG faces in Rogue One. Not just
Leia and Tarkin
Last edited:
ILM spent a year and a half on that CG character. It wasn't a question of money, people, or effort.
Don't get me wrong, I didn't think it worked either. It's just such an incredibly complex challenge.
You're mistaken.
Red Leader and Gold Leader were brought back using unused footage from A New Hope, cleaned up and inserted into CG cockpits. Which is why they work flawlessly...
- Status
Similar threads
- Locked
