UC4: Best looking gameplay? *SPOILS*

[Clukos]

Spoiler

Video title.

Yeah...

Anyway, i got some capture in similar positions (or at least tried to, camera ain't the same across games) from other AAA third person shooter main characters to compare, Jupiter do you mind posting some from The Division from similar position/lighting?

The Order - Galahad

Rise of the Tomb Raider - Lara

Quantum Break - Jack Joyce

These are gameplay all models from the closest camera position available in-game (wall hugging basically), RotTR and QB are the respective PC versions maxed out(in the case of QB i set upscaling to off to reduce blurry IQ for these images, the game is pretty much unplayable with these settings on my PC ). The only TPS i don't have to compare is The Division which from what I've seen has some really nice models up close

 

[VFX_Veteran]

It's like VFX dude is gathering a whole bunch of multiplatform games including Ryse, Tomb Raider, Batman, AssCreed Unity and then selectively choosing the best features of each respective titles and putting them all together against one single UC4, a real fair battle isn't it? ND must feel humbled indeed.

No. It's more like rating UC4's graphics for each area as just normal (save the animation -- which is undeniably astounding) compared to other games that have pushed the tech further in 1 category or another.

 

[VFX_Veteran]

In screenshots? Maybe. No game with TAA looks perfect in stills during motion, The Division, QB, ME:C, U4, FO4, Battlefront and others. In motion (when you are playing the game) it looks sublime(depending on the implementation).


That's non-interactive water, for all we know it's all pre-calculated(because it doesn't need to do anything else, they can spend resources more efficiently elsewhere).

What makes you think driving over an ocean of water in UC4 isn't pre-calculated besides the point-light source flares flying overhead and lighting up a radial area on the water? We are looking at what's on-screen. The shader looks better in Batman, IMO.

 

[Shifty Geezer]

No. It's more like rating UC4's graphics for each area as just normal (save the animation -- which is undeniably astounding) compared to other games that have pushed the tech further in 1 category or another.

Can people start prividing numbers in that case? For a given effect, fidn the best in class and set that as a ten, and then (arbitrarily) give UC4 (and other games) a score in that feature.

At least then we'd have the feintest pretence at a technical, objective discussion which is what B3D should be about.

 

[VFX_Veteran]

Feel free to refute the facts... and i didn't say that the game is always as sharp...

But yeah, a 1080p image can look pristine.



Yes

This is a subjective comment isn't it? 1080p has limited number of pixels to represent a continuous vision (i.e. artwork) regardless of AA technique. You'll never get that infinite frequency of the art -- even with 16k.

 

[Clukos]

This is a subjective comment isn't it? 1080p has limited number of pixels to represent a continuous vision (i.e. artwork) regardless of AA technique. You'll never get that infinite frequency of the art -- even with 16k.

You don't know how temporal aa works don't you? I suggest reading this: https://de45xmedrsdbp.cloudfront.net/Resources/files/TemporalAA_small-59732822.pdf

Same case with QB which manages to have beyond 720p IQ even if the base geometry buffer resolution is 720p. There's so much you don't understand in real-time rendering and make bold claims about.

Edit: Alternatively you can read some of the great posts from Sebbi here on the forum...

 

[VFX_Veteran]

Can people start prividing numbers in that case? For a given effect, fidn the best in class and set that as a ten, and then (arbitrarily) give UC4 (and other games) a score in that feature.

At least then we'd have the feintest pretence at a technical, objective discussion which is what B3D should be about.

I am going to do just that this week. I've started to prep my article which will include actually GRADING the graphics features of UC4.

Here is my list:

Anti-aliasing
Physics
Geometry
Tessellation
Normal/POM
Textures
Shadows/Dynamic light sources
GI
AO
DOF
Animation
PBR
Foilage rendering
Hair Shading
Skin Shading
Reflections
Volume Smoke/FX

And yes, the grades will compare to any game that has the best implementation (IMO) and go down from there.

EDIT: IN FACT, I have a GREAT idea! Let us all hash out the scores and give screenshots as to why we scored it this way and then post it as an article to be read so that people don't just have NX Gamer and Eurogamer as sources for technical reviews..

Thoughts?

 

[VFX_Veteran]

You don't know how temporal aa works don't you? I suggest reading this: https://de45xmedrsdbp.cloudfront.net/Resources/files/TemporalAA_small-59732822.pdf

Same case with QB which manages to have beyond 720p IQ even if the base geometry buffer resolution is 720p. There's so much you don't understand in real-time rendering and make bold claims about.

What does that have to do with having a final pixel count of 1920x1080? I'll read your article just to save face.. *sigh*

 

[Clukos]

And yes, the grades will compare to any game that has the best implementation (IMO) and go down from there.

So this is a comparison between a single game and many individual games? What is the point of that?

 

[Shifty Geezer]

What does that have to do with having a final pixel count of 1920x1080? I'll read your article just to save face.. *sigh*

The IQ is exceptional. Where the 1080p comes in...1080p with 32x super sampled AA is going to have better IQ than 4k with zero AA, even is 4K is downsampled. Except maybe on a 4K display which is firstly deabateble and secondly irrelevant for consoles.

 

[Shifty Geezer]

So this is a comparison between a single game and many individual games? What is the point of that?

It's the only sane way to attempt to discuss this and other games. Other games can also be measured against the reference points. That way we lose the subjectivity (as much as we can) regards the 'this game does everything so is the best in graphics.' That is, a game that scores 7,7,7,7,7... can be measured against a game that scores 5,10,5,5,10... with the same averages and we see two products with two different takes - one with more range in results but both equally good. Then we'd be left with individuals saying whether they prefer a uniform higher average or more best-in-class tech even if accomopanied with bigger compromises elsewhere.

There'll probably be a lot of arguing regards the best-in-class reference points though.

 

[Jawed]

We need a new thread for all the things we can grade in-game graphics on. Let's add: weather, time of day, caustics. What else?

 

[Clukos]

We need a new thread for all the things we can grade in-game graphics on. Let's add: weather, time of day, caustics. What else?

Real-time planet trajectory calculation

 

[Billy Idol]

I like the idea of grading! Bring it on! We should make a B3D grading thread with a list of the actual leaders in each category and an overall leader (largest sum of grades).

The individual grades could be generated by voting and then averaging...if we want a B3D overall grade.

 

[VFX_Veteran]

Yeah...

Anyway, i got some capture in similar positions (or at least tried to, camera ain't the same across games) from other AAA third person shooter main characters to compare, Jupiter do you mind posting some from The Division from similar position/lighting?

The Order - Galahad

Rise of the Tomb Raider - Lara

Quantum Break - Jack Joyce

These are gameplay all models from the closest camera position available in-game (wall hugging basically), RotTR and QB are the respective PC versions maxed out(in the case of QB i set upscaling to off to reduce blurry IQ for these images, the game is pretty much unplayable with these settings on my PC ). The only TPS i don't have to compare is The Division which from what I've seen has some really nice models up close

The Order seems to be the best in that scene. Issues I have:

UC4: excellent skin and texture detail but the lighting is way too flat. Not enough self-occlusion anywhere. Hair doesn't have self-occlusion either and seems like ambient is the dominant contribution here.
RoTR: her skin is much less impressive, also no self-occlusion. Hair shading is also very flat with no shadowing.
QB: It has the self-occlusion but they messed up the damn eyelids. Because of this, he appears to have some light inside of his head. Hair looks the best so far (probably pre-baked).
The Order: also has light leaks if you rotate the camera to the front of him while a light source is behind his head. It's hidden most of the time though.

 

[Shifty Geezer]

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.

 

[VFX_Veteran]

The IQ is exceptional. Where the 1080p comes in...1080p with 32x super sampled AA is going to have better IQ than 4k with zero AA, even is 4K is downsampled. Except maybe on a 4K display which is firstly deabateble and secondly irrelevant for consoles.

I get that.. but is that number an exaggeration? Are their games going up 32x from 1080p?

 

[VFX_Veteran]

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.

Ok.. then, who is in? Let's all come to peace -- avoid biases and come up with a kick-ass article!

 

[Clukos]

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.

There are too many variables and too many different games targeting different genres/techniques/art-tech direction to make sense out of this. How can you compare Star Citizen to The Division for example, both have strong points and negatives but there's not much common between the two because of the different style, gameplay, scope etc. I'm at least interested in this but i don't know if the outcome of the thread will be actually useful.

 

[Jawed]

Real-time planet trajectory calculation

Operation Flashpoint, from 2001, has real time sun, moon and star positions that are accurate for the date and time of gameplay (including the earth's precession according to orbital position).

This came in useful when the player was forced to navigate with neither a map nor compass.