NOPE

Check out my latest creation, I seem to be getting better and better at this . Should I become a game developer? In this scene I am trying to achieve the same level of graphics fidelity as games like Witcher 3, created by big budget studios. You can download and run this scene if you have a mid grade graphics card and windows PC.
Does anyone have any suggestions for my next scene? I am thinking of recreating the yellow brick road TOUR in VR from the Wizard of Oz.
https://drive.google.com/open?id=0B5gBU0VVYsRIalRSQ3h5bzVsMTQ

 

I meant directx 11 grass in title !

 

This scene was just to see what I could run and still get away with a decent FPS. The grass is a must have for me, not even witcher 3 has grass this good . Did quite of bit of optimizing to get it to run well, I am using unity terrain along with the direct X grass so that I can use unity billboard grass as well. Unfortunately with this setup I have to plant each individual tree and have them on their own layer to prevent the water cams from duplicating the render.

 

Recommendation: stay away from mentioning games like Witcher 3.

Exploring, learning and playing is great. It's a neat scene. But we are firmly in the amateur work category, and this is very much hobbiest level output.

Keep going dude, keep learning, keep making cooler stuff.

But really look around, look at what other people have done, and be realistic about the level of output. Have fun with it, don't start talking about stuff like Witcher 3, just don't even open that subject. Compare your work to other hobbiests and see if you can start to push to the absolute top tier of that category first.

Again, I hope I am not demotivating you - keep making better stuff and keep learning. Just don't be one of those guys who's like "now I can build skyrim!"

 

Why? I was looking at witcher 3 and comparing it, witcher 3 does not have realistic grass or volumetric clouds.

The only thing that is really nice in that game as far as the environment is the fact they have a lot of variation.

I wanted to see If it is possible to at least match them on fidelity in unity and in my view it can be even better than witcher 3 with unity 5 as far as environment.

I made this in 2 days so... Imagine if I spent a few weeks on a scene what I could accomplish.

Im going to buy map magic or world generator and learn to use it. This scene was made with basic terrain.

 

Yah I need about 200$ in terrain assets to get to that level

 

This is more of a modded Oblivion than a Witcher 3. Still, not bad for a push button environment. It would be a good place to start making a more realistic environment.

 

I could of spent more time on it before sharing but I want to move on to a bigger scene, using terrain tools and proceedural generation. My next scene will have more speedtree variation, the directX grass with more variation and a much larger terrain. I'm thinking 4000x4000 with a yellow brick road using EasyRoads.

 

Can you show me the best unity environment with DirectX 11 grass? Or infini Grass.

Or the best "proffessional" level environment in general?

 

Just saying, look around at what other hobbiests have done in similar time frames.

The "I dropped some assets into a scene, look what I did in 2 days" thing is deceptive for a lot of people. The learning curve starts getting very sharp and never lets up.

Try to be honest with yourself and try to look at things critically. If you can't see the difference yourself between this scene and almost anything in Witcher 3, then I certainly won't be able to convince you otherwise.

PS - a lot of people feel the same way - its absolutely amazing how easy it is to get a small isolated scene set up and look great. But this experience is so crazy far from real game work or a real professional quality environment it's kinda crazy. Have fun, relax, make some cool environments. Just don't start to think that you are gonna make a skyrim or something.

 

This scene is using FREE speed trees, Free nature assets.

Obviously Witcher 3 has better trees and models + terrain (My terrain was just made quickly).

Matching witcher 3 or even oblivion terrain inst a problem obviously if you are using world generator + map magic (which I don't have currently).

The scene is only to demonstrate the grass running with a complex water system like suimono.

Before I start work on my next scene which will be to the scale of an "open world", I needed to make sure that the assets will run on a mid grade card, GTX 550+.

My next scene I intend to spend at least a month on. And I will be making it with map magic most likely.

I will have lots of vegetation variation, optimized LOD ect.

 

Last time I looked at infini grass I passed, it was so insanely slow that I considered it not usable. I think you found the same thing with the "mid level card" requirement for a very simple scene. I've seen richer scenes that can run on mobile over a year ago.

Here's a thread showing some examples that a guy made with various random generators + a few post effects:
https://forum.unity3d.com/threads/procedural-worlds.398863/

Those environments are just like a few clicks and a few assets. Many are really pretty.

 

And no I did'nt say I "learned" everything I need to know to make this in a day.

First I had to get a good understanding of using multiple cameras with depth buffer, lighting, LOD ect.

It isnt my "first scene", I suppose, I have made many others learning, uncluding one where you can fly from an earth scaled object to space seamlessly:

I got kind of bored when I realized how much time it may take to make it look "really" good. But I did mater floating origin and optimizing it eliminating ALL stutters or frame rate drops from translating the world and particles. I also wrote the controller for the space ship which is still a work in progress.

The scene uses multiple cameras as well as mirrored cameras inside the ship to allow you to seamless enter and exit flying the ship to walking around in its interior in any orientation.

 

I don't care if my scenes run on mobile devices or integrated video cards ...

My grass even in it current state blows the grass away in that link you just posted.

 

Critiques on the waterfall scene:

• Avoid using the hyper-green color for grass. Looks fake.
• Avoid using Depth-of-Field when you have such a small scene.
• Water - Use PlayWay
• Get some Global Illumination going

 

What advantages does playway have over suimono?

Have you used both? Suimono seams kind of heavy on the hardware, out of everything in this scene it is the water that uses most FPS, not the grass or trees.

And I do have the water Cams layered properly and not double rendering.

 

You misunderstand this scene. I can RIGHT now expand it nearly indefinitely. The grass is optimized. The scene is small sure but I could make it 10,000 times bigger and it won't make a difference FPS wise with the optimizations I am working on.

 

I'm about to buy Gaia, any of you have experience with it. It seems like a good terrain editing extension for the money.

https://forum.unity3d.com/threads/s...texturing-planting-and-scene-creation.327342/

 

I just got Gaia, I'll be back in a few weeks to post my next Scene . One much larger.

 

Damn you guys, now I had to buy PlayWay just to put it next to Suimono and see what the difference is. :/

 

The only thing I don't like about suimono really is that, when you are adjusting the ocean surface it does not save your settings while running in play mode.

It is super easy to code buoyancy systems.. and the simulation is very nice.

 

Any conclusions?

 

No offense but I see nothing even closely resembling Witcher 3.. I mean it's not bad, but really? Witcher 3? If you're going to try and compare, then show results that are actually in comparison with the same fidelity or better.

Again it's not really that it's "bad", just nothing even closely resembles the fidelity of Witcher 3. Time doesn't matter, a day, two days a year or a month, if your textures/shaders aren't up to par or better then there's no comparrison. This scene looked fake from the second I looked at it.

You can have all the Volume clouds, sumino or whatever else but if one massive thing stands out it kills the rest of the scene instantly.

Again - it's not really that bad, it's good in terms of things I see people showing all the time, but it's no AAA quality either.
Just keep at it and perhaps one day you will get the desired result. But this isn't it however. If I were you I would have waited until I did that full month you mentioned, at least then you put in enough time to actually do things properly rather than just throw things together and asking an opinion.

Well you got my opinion at least, and many others. Just don't give up, keep on trying.

 

Are you referring to the massively overused bloom? Or the neon green grass? Or the water that magically goes from a solid blue color at a distance to perfectly clear up close as if it didn't have a single particle of dirt in it? For a first time effort in Unity it's likely far better than most of us were able to create but I wouldn't label it realistic in the slightest.

No. Simply buying assets won't make you a better game developer. You need knowledge and experience. Some of the best games out there were created in a time when we didn't have game engines with powerful programming languages and fancy content creators.

 

So you got yourself a typewriter and think you have written the next Macbeth....

Seriously though, while the scene looks nice, its nowhere NEAR AAA level of fidelity. A technical system like a DX11 grass shader, or a water system does not make a goodlooking scene. Even there: does the character interact with the grass? With the water? Is the grass physically shaded? See where I am going? You found some of the good thirdparty assets in the store and actually went the extra mile to build a scene from it, nice. But lets not get overconfident shall we?

What really makes or breaks an AAA level is the skill of the leveldesigner in using the assets at hand in a creative way to
a) set a mood: The level does not only have to look nice. It has to fit the mood of a story, both the overarching of the game and the one of the level.
b) integrate into gameplay: you usually don't create game levels just so your character has an empty sandbox to run around in. Each level should fill a purpose in the game, give the player a nice difficulty progression and so on. Even more basic, there SHOULD be gameplay in a level. Once you have to get into how your level should play, you will have to cut many of the niceness, because gameplay usually comes first, and might demand cutting some of the best looking elements in the scene.
c) look artistically balanced: Color balance, modifying bought assets to fit well together, background elements fading into the background compared to gameplay relevant elements, the art itself giving players hints to story, gameplay, even the right way to go.


Now, I know we live in a world were every small john doe can pick up Unity and produce an FPS asset flip in days, and can even then go on an try to sell it on steam with a fair chance that steam will put his asset flip on their store front. These low quality games will msot probably have NONE of the aspects of good level design in them.
How would they, because level design cost the most valuable resource these asset flipper will never spend: time!

But as you seem to be ready spend some time designing your levels, have lofty aspirations (Witcher 3 as a competition, really?), and seem to be eyeing a position in the game industry, maybe have another look at your scene and ask yourself these questions:
a) are my colors balanced? Are they pleasing to the eye, and picked from a consistent palette of colors? (My opinion: colors are WAY oversaturated, and often clash with each other. I would modify the textures, or at least use a LUT filter posteffect to get the colors under control)
b) does my level have any gameplay? If not, is it built around the idea for gameplay? (My opinion: your scene currently looks like a scene built to look nice, kind of a tech demo. That is not how game levels work. Look at the scene and think what kind of game it should be for. Alter the scene so the gameplay would work in it. Make it CLEAR to the viewer how the scene would play (for example using assets to show the right way to go, making cover and stuff like that obvious))
c) Does my level have a story? If yes, is the story showing through? (My opinion: at the moment this is a generic outside grassland tech demo scene. Nothing too exciting about it once you look beyond the tech driving it. Go the extra mile. Build a story into it with assets, and the placement of color, lighting and terrain/trees. So much things that would instantly make the viewer care more: a skeleton laying at the bottom of a cliff with some decayed pieces of gear around it... A tall tower in the distance. A smoking volcano... a flying caste in the distance... the ruins of ancient cultures)
d) extra points: does my scene not only play well IN GENERAL on mid range PCs... can I also make it play well on low end PCs with minimal drops in fidelity? Are there any hotspots where framedrops occur?


I would guess you would be surprised how much work went into the Witcher 3 levels. And you would be surprised how badly your scene matches in this comparison when you ask 100 people. The Witcher 3 artists where not idle, and are most probably more skilled than you are.
But I am not saying your level looks bad, or that you shouldn't have lofty goals... maybe be a little bit more modest here in the forums, because you are kinda unfair to Witcher 3... after all, Witcher 3 is about 100 bigger than your tech demo, and actually has to support gameplay and not just a camera fly through.

 

All of this. (especially the point that a scene != game).
Additionally, it is important to understand that effort to result isn't a linear time scale. 2 hours to build a passable looking forest scene doesn't mean that in 4 hours it will look twice as good. The last 10% takes 90% of your time.

It's not horrible for assembling a bunch of assets, and I have certainly seen worse, but really take a look at your aspirational targets, and look at the details that set them apart.

 

Not bad for a first environment, but nowhere near witcher 3 quality.

This is one of the best-looking scenes I've seen made out of fairly vanilla unity, although some $300 was spent on the asset store for effects:



It uses some art from the blacksmith demo, and megascans materials. Finished in 24 hours. Not a bad result at all, if you ask me, but I wish it hadn't been so heavily colorized. You can also download the project I think, from the link above.

Disclaimer: I think that Unity 5 lighting shotcomings will hold back graphics quality slightly in comparison to other engines, but don't worry, I'm not planning to go on about that here!

So if I were you I would aim for something at least that quality ^^ if you want to be the best graphically in Unity 5.

 

PFfftt...

You have to cover up the terrain to even make unity look good. I bought Gaia spent all night learning how to use it, got me and island made, starting texturing it and ran into issues at 8 textures.

Seems Unity terrain is so limited, it is like we are lving in the 1990's. I mean seriously, 4-8 textures give me a break.

Having to break out a modeling program to make some terrain that looks decent is ridiculous. I mean sure its "ok" with such a limitation but you should be able to have unlimited textures on the terrain.... Cmon....

 

For no apparent reason the occlusion culling broke on my scene, it was working perfectly and I ddnt even touch it and it just randomly broke.. Unity is really starting to get on my nerves.

 

Hate to break it to you but this limitation is not unique to Unity. Other engines have similar limitations. This is why terrain systems are often ignored in favor of using meshes.

 

Its a ridiculous limitation, anyway, I am not a modeler but I am willing to check out any mesh terrain tools if you know of any. The thing is, how am I suppose to use Gaia, or any procedural terrain tool in unity with a "mesh" ?

 

You can export the terrain heightmap data and import it into a modeler like Blender. From there you can generate a mesh from it, texture it as necessary, and export it to an FBX to be re-imported back into Unity.

 

Nah, it was always really easy to make a good looking screenshot or a nice scene drag and drop.

The original demo scene is literally Tenkoku Sky + Sumono water + a few speed trees and dx grass. He didn't even change the grass color from the horrible default green.

Demo scenes in asset packs like manufactura and big environment had some really stunning results you could copy or tweek.

Probably the first really nice scene I ever made was a process of taking one of K4's demo scenes and step by step, rock by rock recreating a small area. Going through that process gave me a real idea of what went into building a good looking environment.

It really takes a lot of work, and the details are what counts.

 

My island is coming along nicely with the exception of the ugly terrain with limited textures "bs".. and im going to straight up call it that. I don't care what unity's excuse is. It seems obvious it is like that for a reason, to prevent solo developers from more easily creating creative works on par with the bigger studios. You can't tell me that if you could have 46 + textures on a terrain it would be night and day..... blueray vs 8-trac player.Instead I have to find a way to Hide the terrain with models, and add more resource use to the hardware.. that sounds "smart".

 

it's a poor craftsman that blames his tools. Your examples are lackluster because you're a noob, not because Unity is holding you back.

 

I suppose I could convert my terrain to a mesh and use a mesh painter, but what about the grass and trees ect...

 

I have yet to see a single good terrain that isn't cloaked with models. Show me one that looks good with 4 textures lmao..... Do it now. I worked really hard on that for hours on end learning Gaia. Sure Just like you I will probably end up downloading every terrain tool under the sun... map magic, world generator ect. I'm sure there is some trick to it. How about offering some advice instead of being a jerk.

 

Constructive Advice:

• Stop thinking you should be able to create Witcher 3 by downloading a couple asset packs. Accept that there is a lot to learn. That you need to really invest tons of time into getting better: months, years.

• Start really learning: download great examples, like the Blacksmith demo the Viking Village, etc. Look at how guys like K4 did it, he went on to work on Witcher 3, you can see his stamp all over that game. If you pick up his assets you can see how he builds and crafts an environment.

• Accept that these things take time. You may be talented, but you will not suddenly be able to match top quality work over night.

• Really look at your work critically. Compare it to other, "why does my scene look like crap and his scene look amazing?" What's different?

 

Large/good studios have developers and artists that are deeply skilled, and know how to work within the limitations that come with game development. They also have the ability to create tools and processes to achieve the end goals they want. There are always limitations when building games, regardless of platform. There are amazing games done with Unity, not all developers are equal, and not all face the same limitations that others do. A pencil or paintbrush is going have very different results depending on who is wielding it and it has nothing to do with the tool itself.

 

And FYI, I write 6 programming languages and 2 database languages, including HTML/CSS and Javascript.

I am a programmer, not a modeler. If I was a modeler, oh it would be easy to create the "assets" to cover up the terrain now wouldn't it.

 

That's cool man. I am a programmer too. It takes time to learn a new kind of skill. You don't suck personally, maybe you're a real smart guy, but you still need to give yourself time to learn a new and different kind of skill, especially if it's very different from your previous background.

Don't make excuses, don't give up because of limitations... git gud.

 

For sure! I've created (and thrown away) dozens and dozens of terrains... using RTP, Terrain Composer and Gaia. I'm doing it for the sake of learning, and I still suck at it.

It also appears that Suimono hasn't been even tweaked... the original demo scene didn't even have underwater caustics enabled. I also own Suimono and it can create some pretty sick looking water but it's not a magic bullet. Still takes learning and practice!

 

I think the thing to learn from this thread is: don't barge into a place like you own it, bragging about what you've done when you've only done 10% of what is actually used to make a "game."

You may get some pushback.

 

10% - no. This is a non measurable quantity unless we're talking about asset flip level work. This is an early step in a learning process.

Useful thread lessons:

• Be critical and realistic about your work.
• Accept that you need to learn.
• Don't make excuses.

 

By the way searching for information related to the limitations of Skyrim and Fallout 4 (six textures per quad by the way), I stumbled upon a GDC 2014 entry for The Witcher 3's landscaping. It's about an hour long video.

http://www.gdcvault.com/play/1020394/Landscape-Creation-and-Rendering-in

 

All I see is a bunch of models covering up the terrain which makes up less than 20% of the space on those images.

 

If you need better screenshots you're free to look for them but I figured simply showing you the pictures would let you know that two extremely popular open world games are working with very much the same limitations as you.

By the way having meshes covering a good portion of your terrain is to be expected when you are working with heightmap based landscapes because a heightmap cannot represent complex features (eg overhangs).

 

Yeah, and?

You don't make a game with terrain alone.