Why does Unity look "flat"?

I have read through every response and it really does seem like you are doing comparisons of things without trying to keep high parity of settings between the comparisons etc.

What is the point of comparing what you have done when there are such clear differences in your setup? We have done multiple tests of arch vis setups for rendering between unreal and unity and never have we had issues like you mention. This will be down to the way you are setting up your scene, your settings used, and possibly the fact that it seems you are looking for problems; as in it seems like this is purely opinionated based on my own eyes.

You sure you have'nt been looking at this for too long?

I also want to point out you mention lots of times comparisons between unreal and unity and get a little defensive when some unity and rendering experts in here try to give you sound advice, but then later on say "its pointless comparing them" whenever a salient point is made that perhaps you have gone about this wrong.

So perhaps you are simply to "close" to this and need to take a step back before assuming everyone in here is "guessing" etc. Some extremely knowledgeable people have replied to you in here with some decent responses, and I would painstakingly try each one without making any assumptions or skipping any steps, no matter how wrong you think they are before anything else.

 

Even in the original shot that was modified, you can clearly tell the UE4 skybox is influencing the image. Unity shot doesn't have the same skybox (although it is similar in HDRP). When trying to get to the bottom of things, only the scientific method will do.

 

I've used Unity since 2011, so everything that has been mentioned here I have already tried.. More than once, I have gone to the lengths of doing exact like to like scene comparisons.. It's been a while sure so I can't remember in depth every single thing I did, I just know for the most part I have already done it en mass.. I went to the extremes of matching shaders, GGX implementations (you name it)..

On top of that, I've used every engine out there (studying implementations).. I was creating game frameworks with SDL + QT / XNA before Unity existed and I'm not a complete idiot..

I gotta ask thought how can you have parity between a photon mapping and radiance based lighting solution (that do react differently) hell lightmass reacts differently every time you hit the build button? Do I keep switching stuff off because the implementation is different until I get them to match? Shall I just do a comparison between two unlit scenes?

Where do we stop? What's the actual point?

There is no real point besides trying to analyse patterns beyond a simple scene stare and compare, I need contrast not similarity at this point.. If the like for like scenes would of given answers (as it's simple to do) then I wouldn't be asking. I see it in every other piece of "unity" work I've come across, so why would it be setting up "my" scene / or why would it be "my settings"?

Looks a little likely at this point doesn't it?

Because of things like this:

When things are taken out of context for their own personal amusement (again never said IT WAS the projection matrices), on a subject matter that does irk or they get their undies in a twist for potentially mentioning there's something wrong I just shut off to be honest.. I'd rather just find the issue my self, far less hassle..

On the other hand if something pops up I haven't thought of or a reasonable argument is made like many have already provided you can gurantee I will look into it.

 

@ShadowK how about making a few APK builds I've mentioned?

 

It would have to be for PC, AFAIK in Unreal they use a completely different rendering pipeline for mobile (Forward rendering).. I always use deferred and it's not necessarily the render pipeline that's the issue, all the post processing / reflections etc. differ for that platform. Also I'd have to re-install Unreal..

By default on PC we can both use ACES whether in Unity or Unreal (trying to bypass the tonemapper is an interesting concept in Unreal), I'm not really sure what their "mobile" tonemapper follows.. I never do mobile stuff so it's never been a concern..

Also what exactly are you looking for? Post processing on / off, lightmapping? RTDFS? DFAO vs. HBAO?

Light baking is obviously going to make things look more "complete" although what do you want out of that?
Gradient based environment lightmapping additions (w/ IBL?), IBL based lighting only enviro map off? AO included in lightmaps? Then what about matching the occlusion exponent?

Bent normal sky occlusion? (Stationary IBL lights use this be default).. Do I try to match the shadow bias? Or do you want me to keep it simple (turn off shadows etc. no IBL)..?

Your call..

 

(offtopic, sorry)

Have you thought about doing the RPG top down like the escapists using sprites? Isn't the story the most important bit?
The story is what haunts you, not this hopping from engine to engine forever chasing that glimmer of ambition.

Would love to see a 2D story by you, I think most people would too. Visuals? matters less the more hyper realistic everything becomes, because if that's the new normal then stylised begins to be what people take notice of.

I've seen threads from you that dodge about this same rendering topic forever now, and I'm getting worried it's making you feel stressed and isn't beneficial.

I truly believe your stories, in the simplest form like 2D art, would bring the attention to the story not the visuals, just something to ponder.

 

Sound like ShadowK really need a competant artist on his side, to point at problem and find solution. No technique is ever perfect, and artist deal with limitation all the time. Think about it, the only renderer they have is their brain, they don't rely on fancy gi algorithm and hand place all colors. SO fixing less than stellar rendering by painting over is what they already do, in this case painting over might be suggesting artistic solution to technical limitation (like can I have some light on the side of the tree). That's how it happen in AAA anyway, the artist ask for things and the technical guy find a solution. And they have been doing great job since the fixed function pipeline.

 

Ha, you do cut to the matter at heart.. Yes I am obsessive compulsive when it comes to graphics, you do have to wonder when you start picking apart the book of the dead don't ya?

It's only a matter of time before I walk outside the house and say "the graphics suck"..

Some times it's just a matter of putting things into perspective, I might be chasing unicorns that don't exist. It's the game that matters, not how shiny the reflections are (or so I keep telling myself).. Anyway I'll crack on and shutup from now on.

 

You should get an Oculus Go, coming out of VR the first few times gave me a whole new appreciation of the fidelity our reality has. But all jokes aside, reality does mostly look pretty crappy compared to the highly art-directed stuff we see in movies and games.

I can relate. Maybe it'd be best to make a clear choice between embracing that and going for something fancy looking but with a scope of e.g. 1 room only for the whole game and less than ~2 dozen art assets, or doing what hippo suggests and staying away from all the graphics distractions and focusing on the game. The middleground where people want both seems to lead to an awful lot of abandoned projects (mine included).

 

Well originally I wanted to make something like Dragon Age (the first one) and that's far from a pretty game even when it got released. I made an R-Type game in less than a week, I also made a top down in about three months (funnily enough I had zero complaints when making these, not in art direction or anything it was all very straight forward)..

Not enough of a challenge.

It's about finding that "middle" ground, it's easy to jump from linear RPG to openworld behemoth AAA graphic insanity.. More people have done it than they care to admit (just check some of the other game engine forums) ..

Also when these beta testers on the original RPG I made said it looked "flat", I didn't know what they meant neither did I actually notice it.. I did get rattled by it and this one side of game development we don't talk about often, especially being too close to a project. How is it affecting you?

As hippo said to me at some point you just cut the bull and live with what you've got, it'll never be perfect.. Hell no matter what your budget is, no matter what the game is it's never perfect.

I need to move on, I'm sure 99.9% of people won't notice or care and it's them I'm making the game for.

 

Expecting stuff before conducting experiments is not very scientific. I will just look at the whole graphics pipeline and see how they differ.

As for "Unreal only uses forward rendering on mobile" - it may be true. I have quickly skimmed through resources and talks and it seems that this is the case. However, a lot of in-game screenshots of separate renderers don't have much difference, they look mostly the same, especially on static images. I bet it won't be a problem in that simple two-spheres case

So this is what I propose:
Make two builds from your original post (with two spheres on a plane) with that same settings and see how it looks on mobile. Make sure to use OpenGL ES 3.1 in both cases so it has the least of limitations for both platforms. Then see how it looks on mobile, make a few screenshots and share them here. If you don't have a capable enough device, just send those builds and I will take a look.

 

Locked this as the OP has clearly stated it's just really unhealthy chasing these things, and they don't matter - at all, and it's just lighting differences, which can be overcome with suitable direction.