UC4: Best looking gameplay? *SPOILS*

[Jupiter]

I see the biggest problem in the generalization. Some games may have an excellent effect in one place. But in the other places it has not the same quality of this effect. As for Crysis 2 I can say tessellation but I can say that this is consistently.

A ten point system is too complex. When points will be used, I would say Maximum five. 0 = effect is not there - 5 = best in category.

 

[Clukos]

In the topic of tessellation i still think it must go to Battlefront, in pictures it looks really good. But then there's the problem of this numbering system, Battlefront only does tessellation in very close proximity, this doesn't translate well into pictures (which can look absolutely mind-blowing in terms of geometry), and while this should deduct points off of BF you can't really judge from pictures unless you play the game yourself, and i'm sure not many people buy every game and have every platform.

Same thing with TAA, it may look great in motion but you can get ugly stills in almost every game that is using TAA. Should that deduct points? Even if it doesn't really affect the IQ when playing the game?

 

[Billy Idol]

To me it is clear that any grading, independent of how smart it is, can do game graphics justice.

However, it would be a start. And with the experts on this board, it would be better than any other known tech discussion.

B3D graphics grading+assessment threads for each category: +1!!!

 

[Clukos]

And then you have weird games like Dreams which go in a completely different direction both visually and tech wise, how can you score that game against other titles? There's so much that can go wrong with this. Like Jawed said i think it's better to examine each title individually and see what it tries to do, how consistent the output is, where it falters etc. Comparing different games in different genres, using different engines is pointless imo.

 

[Jawed]

In the topic of tessellation i still think it must go to Battlefront, in pictures it looks really good. But then there's the problem of this numbering system, Battlefront only does tessellation in very close proximity, this doesn't translate well into pictures (which can look absolutely mind-blowing in terms of geometry), and while this should deduct points off of BF you can't really judge from pictures unless you play the game yourself, and i'm sure not many people buy every game and have every platform.

Can you see the tessellation "popping" in that game as geometry translates from non- to tessellated?

Same thing with TAA, it may look great in motion but you can get ugly stills in almost every game that is using TAA. Should that deduct points? Even if it doesn't really affect the IQ when playing the game?

In World of Tanks TAA there are two obvious artefacts: 1. a shimmer is seen on regular, repeated, fine geometry, like an exhaust grille, even though it is static in the frame; 2. when trees fall down the leaves produce a peculiar blurry trail. The first won't be seen in a screenie.

GIFs can easily be supplied to prove/disprove temporal anomalies, whatever the technology under discussion.

 

[Clukos]

Can you see the tessellation "popping" in that game as geometry translates from non- to tessellated?

Yes it's quite obvious when playing the game (2-3 meters range).

 

[Clukos]

Here

 

[VFX_Veteran]

It's happening guys ..

Please don't share this webpage until it's complete. We have a lot of work to do and I don't suspect this Congress will come to an agreement on these features anytime this week..

 

[Clukos]

And to get back on topic a bit, with photomode not adding anything fancy on top of the image (like driveclub sampling from several frames and then combining the result) if this is representative of in-game quality (just one environment spoiler 40 minutes into the game, nothing major, not even characters on screen)

Spoiler

I think ND have done an incredible job with the AA. It basically looks like what you would call a bullshot.

 

[Jupiter]

Yes, sharp and aliasing free in 1080p. A lot of other games with TAA/1080p also have no aliasing but the imagine is much blurred.

In the topic of tessellation i still think it must go to Battlefront, in pictures it looks really good. But then there's the problem of this numbering system, Battlefront only does tessellation in very close proximity, this doesn't translate well into pictures (which can look absolutely mind-blowing in terms of geometry), and while this should deduct points off of BF you can't really judge from pictures unless you play the game yourself, and i'm sure not many people buy every game and have every platform.

That is correct. It is more easy to give the point in this area since tesselation is used very rare because of Sonys/Microsofts hardware choice.

EDIT:
But I have not seen clearly visible Pop ups in Crysis 2. The tesselation draw distance is high.

 

[pjbliverpool]

Here

I'm sorry but I have to say it, even with the Tessellation pop in (which only happened in a single instance that I could see), that looked better than the UC4 video's I've seen posted thus far in my opinion. UC4 looks awesome, no doubt about that, and it can be argued it's doing certain things better than BF and vice versa, but on balance, BF looks a lot more realistic to me.

I will grant though that the image quality (assuming there is no photo mode trickery going on) is absolutely incredible for 1080p. Probably the best AA I've ever seen in a console game and only comparable to very high end AA solutions on the PC (8xMSAA, 4xTXAA, SSAA etc...).

 

[zed]

Actually I think it'd be a decent discussion if people approached it seriously and came to consensus. There could be a scale (the Abinger-MacMillan graphics quality scale) and games could be compared. Those comparisons would be subjective, I'm sure, but at least it'd progress these very common discussions in a meaningful, productive way, and give us something new instead of the endless subjective screenshots and hperbolic one-liners.

You've changed your tune from when I proposed this a few years ago ;-)

 

[Scott_Arm]

I see the idea of objective reviews will not die. May I interest any of you in my perpetual motion machine?

 

[HTupolev]

That way we lose the subjectivity (as much as we can) regards the 'this game does everything so is the best in graphics.'

Not really. The basic difficulty in these discussions isn't that people disagree about what's objectively going on; that aspect of the discussion is usually perfectly healthy. The difficulty is that different people weigh things differently. What the graphics are doing at a given moment is objectively quantifiable, but broad "graphical quality" is a subjective determination. The output of a rubric that weighs objective factors into a few "quality" numbers is essentially an opinion, and fundamentally not much different from asking a person what they think. All you accomplish with a rubric is that people will criticize the rubric, which isn't really a change from what happens right now.

And what about constructing that rubric? There are so many qualitative variations, and so many things whose significance depends on other things. Even something as basic as size and expressiveness of dynamic lights is an insanely complex realm to judge. Most lights in Halo 1 are enormous and can cast (white) specular; Halo Reach can handle far more lights and supports specular color, and the specular is more expressive; but most of these lights are smaller and many diffuse-only, so in some ways Halo 1 often sprawls more interesting dynamic light across the screen. How do we rate these two very different cases against each other in a way that is both useful and can be evaluated in a purely objective manner? If we can't evaluate it in an objective manner, then the rubric is worthless, since different people will come up with different numbers, which violates the whole purpose of having this "objective" rubric. And if we simply cast aside these kinds of complexities and leave them out of the rubric, the rubric is worthless, because there are hugely significant aspects of a game's graphics that are being totally ignored.

And what do we do with the rubric? Humor the conversation with it as a starting point? If someone makes a dissenting opinion, the rubric wouldn't have any kind of special (intellectually honest) leverage against that judgement, since it's ultimately an opinion as well.

The reasonable thing is for people to stop panicking over whether a game's graphics are the best. Comparative discussion is great, and it's entirely possible to have a healthy discussion where people explain why something is their favorite. The issue is that people turn it into a battleground where the goal is to win. Creating a rubric which can be evaluated from objective data won't change that.

 

[Jawed]

As I alluded to earlier, development studios themselves have to make decisions about what techniques to use in their games and what research they might underwrite in order to create results that fit their agenda. They are making decisions about quality on each and every technique they deploy.

If they choose to go with an off the shelf engine, then there's not much else to say.

If development studios can make those decisions (and they clearly do, otherwise there'd be no graphics) then we can talk about them. We can compare techniques till the cows come home. We can talk about stencil shadows versus cascaded shadow maps or lightmaps versus lightprobes. We can talk about the apparent resolutions of various render passes. We can observe inconsistencies across the game. We can observe common failure cases. We can track the evolution of the state of the art.

e.g. I think a screenshot war on the subject of ambient occlusion, in a thread that's all about ambient occlusion, is a pretty healthy start. I happen to think NVidia's VXGI technique in RotTR (on PC) is the first of the next wave of AO techniques. In a few years' time, prolly next console gen, all the games with high end graphics will be doing that.

 

[Recop]

Why do people don't consider the complexity of action scenes ? I mean, it's much easier to get good graphics in unambitious gameplay sequences.

It's like the amazing chase scene of Uncharted 4 had never existed...

 

[London Geezer]

Is this a serious thread now?

Boring.

 

[pjbliverpool]

Why do people don't consider the complexity of action scenes ? I mean, it's much easier to get good graphics in unambitious gameplay sequences.

It's like the amazing chase scene of Uncharted 4 had never existed...

So it doesn't have the best graphics, but when you account for the 'complexity of action scenes', it does?

 

[Recop]

So it doesn't have the best graphics, but when you account for the 'complexity of action scenes', it does?

To me, Uncharted 4 is by far the best looking console game and it's even more impressive when you you know that ND was able to maintain these graphics with very ambitious action scenes.

But to be fair, not all studios work in the same conditions. I mean, you can't really compare ND to RD for obvious reasons (budget, size, etc.)

 

[Clukos]

Few more gameplay shots from Uncharted 4 (areas we've seen in trailers)

Spoiler