NOPE

I'm just getting started. But for people to literally insult me because of my first scene isn't on the level of a professional is ridiculous. I plan on making many, many scenes and probably some of my own assets in sketchup eventually "to cover up" the terrain. I know my way around sketchup pretty well, well enough to edit and make modifications to models on 3dwarehouse. I was just excited, that is all...

 

With your second post you directly compared your work to the Witcher 3 claiming superiority.

 

Long story short. - Just learn things better.. Often when I run into a limitation guess what - I figure out another way to go about doing it. That's why I love coding/game design in general. There's always more than one way to do something.

The only reason we/ a lot of us complained and practically insulted you is because you were trying to compare with AAA games. It's more so a warning not to do that or this will happen - unless it's genuinely certain that it's up-to par with it.

I've never heard of a 8 texture limit for the Terrain. I've had terrains with like 20+ textures once.
Unless this is from some special asset you bought. But then again, they should be able to make it where it can have as many textures as you want. Guess they just didn't implement that. Suppose it's harder than it appears. Because others have said, it's a general limitation to most engines...

Oh and the Shogun Total War terrains looked good with nothing covering them up. I swear I didn't see more than 4 textures for the entire terrains (Grass/Snow, Dirt Roads, and some Dirt Grass/Snow).

 

Once my terrain is decent enough I going to go into programming controllers and specialize in tool development and controllers and possibly multiplayer, since my skills are in OOP. Honestly just wanted a good back drop for me to develop controllers in. I want to focus on programming but I know that If I were to ever "get a job" in this industry my skill set would have to be beyond that of software. Level design + software is what i am thinking but not modeling... perhaps texturing and prepping models for the game engine ect. You have to be realist it is true because building a game does require a lot of specialty skill sets.

 

Well, I was going on 20 hours, a cappuccino and the fattest blunt you ever seen.

 

The "covering up" thing is a huge part of level design and game dev in general.

Unless you have a team of AAA artists (and likely, even then) - you're going to have tons of things that look like crap. Limitations or problems. Like half the tech involved in graphics and most of the technique revolve around trying to "cover problems" or hide stuff that looks like crap.

Things like corners, edges and spaces where different objects intersect is an obvious example. "where the wall and the floor connect looks horrible? How do we hide that?" - and SSAO is born.

I'd say that at the solo or indie level, hiding problems and working around crazy limitations might be the most important skill in terms of visual presentation.

A lot of the time, problems also get covered up in layers - like if you break down that amazing looking visual effect in game x - layer 1 looks like crap, layer 1+2 still looks like crap, but once you have all 5 layers of the effect working together, it starts to look amazing.

tl;dr - using grass or meshes to cover up problems isn't cheating or anything - it's something you always need to think about, and one of the skills we need to learn to master if we want to make great looking stuff.

 

An example of which would be that lighting breakdown that gets posted from time to time in engine discussions.

 

I'm sure most scenes will look amazing when the models like "cliffs" are designed to go on and wrap around specific parts of the terrain. That is what makes AAA game AAA.They arent just downloading an asset and plopping it, they are designing it to go in a specific place or to fit together with other elements so that it is seamless to begin with.

 

I think im going to look into making some of my own cliffs and rocks ect with autodesk and some noise algorithms. That may be a good start for me to start covering up the terrain in my island scene. The rocks and stuff on the unity asset store BLOW. I want circular shape cliff faces that can melt into mountain type shapes and long cliff faces ect.

 

Like was mentioned earlier Manufactura K4 is an excellent asset creator to check out. His packs include modular terrain frills designed for this very purpose.

https://www.assetstore.unity3d.com/en/#!/search/page=1/sortby=relevance/query=Manufactura&K4

 

and by the way there is no realistic grass or realistic anything in witcher go watch those videos again. its nice but nothing even close to having 500,000 individual blades of grass being rendered in real time.

 

K4 has some rock sets for free that are pretty good. I would definitely take a look at some of them (they're free, why not).

But you will probably still have problems in terms of intersections unless you get really fancy, RTP has some tools for blending objects into terrain procedurally (but it's a pain in the ass workflow).

Just keep in mind - especially in AAA - they're using every trick in the book to hide flaws and problems.


Look at the wall on the left - how much stuff did he cram in there to make that wall look 'natural' - he's got grass hiding the intersections, plants right next to the wall itself, a wheel for no apparent reason - and he stuck a plank on the wheel so the wheel feels more natural.

There are like 4-5 layers of stuff there used to make that tiny bit of wall 'feel' natural.

 

Billboard grass, when I originally posted this post, it was for that grass only. The directx11 Grass is by FAR better than anything in witcher 3, of course witcher couldnt use it because it may limit the "customer" base by requiring more hardware to run it. I get it. But why not build stuff that REQUIRES a high end graphics card, not because it isnt optimized but because you can and im sick and tired of AAA games not going all the way so someone on a card board computer can run it. I want to use my SLI to my fullest for more than just a nvidia tech demo or 2 or 3 games like crisis and far cry, mostly cryengine games.

 

It's your call.

But think about it like this: even on a high end machine there are limits, only so much stuff the gpu can crunch. Do you really want to spend that on the grass, or do you want to spend that on 20 animated characters, and buildings and props and trees and all the little details that make a scene really feel 'full' and 'rich'?

The sumono water looks great - but is it worth it? What do you need to give up to have that water effect? How much does the water add to the scene? Would the scene - as a whole - look better with those gpu cycles being spent elsewhere?

I'm not saying you should scrap the grass and water, but those are the questions that you as a level designer need to ask yourself.

I have stuck with Tenkoku sky forever, despite it being a bit of a resource hog (and trying to target mid-low end hardware). Why? Because even though it's expensive, I think it's worth it. There are tradeoffs though, and I had to scrap some other stuff in order to fit it into the gpu budget.

 

because most games are required to run on console too, and you dont want to do everything twice

and even if not, you don't want to limit your target audience to 2% of all gamers, because this would be a pretty retarded thing to do

 

This. The Witcher 3 targeted both of the major consoles as well as Windows.

Modern AAA titles require millions of dollars to develop and market. Targeting only people who buy the latest and greatest hardware would severely restrict your income. Just about anyone who goes down that path would have either (a) have financial problems, or (b) need to charge an exorbitant amount of money per copy.

The Witcher 3 cost about $80 million to develop and market.

 

What I have noticed so far is it isnt so much the Verts or Tris that consume CPU/GPU is different cameras rendering the transparency and reflections, any kind of real time lighting depending on the number of pixels its lighting in deferred rendering. The shadows seem to have no effect on my GPU at 250 units, which is plenty enough. What I do plan on having is ALOT of particle effects. Oh yes I want the water, lots of speed tree variation, lots of vegetation variation. I also plan on getting the asset to make Caves ect. I want to push my card to the limit "nearly" still need cpu time for the controllers I will be working on, and I intend on having a day night cycle which I am going to program myself, a nice interface, and it will be plugged into Suimono as well as my other assets like volumetric fog and mist, and it will have an API as well to plug into other assets.

 

here you go :

https://80.lv/articles/gpu-performance-for-game-artists/

 

Its a good thing my focus isn't on making money then For now this is just a hobby, so everything I create has nothing at all to do with what other people can "run" or "not run". Even if someday I did manage to build something with incredible graphic fidelity, If I built it by myself I wouldn't need to sell very many copies to support myself since I already made alot of money in SEM/Web Development/ Lead Gen and own everything I have including my house.

 

That's a good thing. About the only game I know of that attempts to push for high end visuals has had nothing but problems during its development despite having over $140 million in funding.

https://www.pcinvasion.com/report-reveals-problems-star-citizen-faces

 

Did you see my prototype of a seamless planetary landing system? I built it just to see if it was possible in Unity. Sure it isn't pretty but it demonstrates that it is possible with floating origin to have models at the scale of 1 Million Units or more.

What I need to know is how to proceduraly place mesh (infinite terrain) around a curve. That is a tough one to crack lol. No one has done it in unity, not to the scale of a planet size object. I have seen the planetary terrain systems on the asset store, I have all of them. They basically bend a cube into a sphere, not exactly what I need... What I need is much more complex to build but it IS possible in unity to "nearly" replicate what they have done in star citizen if you have enough time to build the grid system and figure out how to curve the grid.

 

On the topic of seamless planetary landing systems... Kerbal Space Program was made in Unity.

http://store.steampowered.com/app/220200/

 

I was not impressed with kerbal. Its very unimpressive. I can build something better in a week if I kept working on my prototype, I have actually terrain on a SCALE planet in my prototype.

Kerbal is not to scale and it is NOT terrain on the planets surface. It is very basic and graphically unappealing, the orbital simulation is nice but doesn't interest me.

 

What i did learn from kerbal is too look heavily into floating origin and now that I know how to build it, I can make a scene that is nearly infinite in scale basically. The only limitation is the unity editor itself, it gets a little skidish with objects larger than 1 million units. You cant move things manually you have to do it with math.

 

I have a theory for a grid system to replicate star citizen and elite dangerous planetary grid system. with a scaling Grid with grid spaces near the player in high detail and far away becomes a single mesh. Making it LOD seamlessly though that is the tricky part for sure. I will probably go back to it after I get this island done and a boat simulation controller.

Before I get into mesh programming it would be nice to have a decent understanding of "how" to build a nice looking scene to begin with.

 

I actually thought that demo was kinda promising really. It showed you were actually doing work and not just dropping a few assets into a scene.

Do you man, have fun, learn a bunch, maybe down the road, you will make something really sick.

Just keep learning. Don't Dunning Kruger yourself into thinking you're Rembrandt when you're still learning to mix paint. You just end up holding yourself back and slowing your own improvement.

 

Prove it.

 

Oh, I'm learning every day .... soon, very soon I'm going to go into procedural mesh generation programming and specialize in it. In order to build the "planetary scale" terrain system in unity I'm going to literally become a specialist in mesh vert programming, and try out my theories. I learned a lot already from the assets on the store.

Mesh programming as opposed to modeling is much more appealing to me, it will also enable to make assets to sell on the store that "no one" else has made because there are few in the field as far as I can tell.

 

Haha, in due time... yall irked my girken now I have to build a bad ass island somehow with at least "top tier" hobbyist graphics quality before I go back to it or I won't be satisfied.

 

Here is an example:
https://www.assetstore.unity3d.com/en/#!/content/75041

Why is there only 1 "rock generator" on the asset store and WTF why is it low poly seriosly? Where is the cliff and boulder generators? Or how about planes made to work with tessellation? Why isn't there more model generation systems?

 

If you want to jump into the subject now:
http://jayelinda.com/modelling-by-numbers-part-1a/

That's the best intro I've ever seen on the subject. Really well written and very clear.

 

The reason I put it aside is I realized it would probably take a minimum of a month spending every extra moment I have to even get to a prototype and probably an ugly one, but a working one. Another month after that to perfect a LOD system, not to mention shaders. I understand how I would program the mesh and LOD systems but I am not a shader artist. I am going to pick up shader forge soon and hopefully that will be enough to suit my needs.

 

I may need some help on the project, a shader artist and specialist would be nice, once I have the prototype. Its a daunting task to learn so many things so quickly and retain all of the knowledge.

 

It took me a while to figure out how to walk around inside a moving space ship, I attempted to do it when I first started playing with unity lol. I had to get a good understanding of cameras first. I went with a mirrored camera / ghost ship system avoiding the physics system, but you cannot tell the difference at all, it is seamless. And as far as landing on a planet surface, planet gravity is easy and so is the fps controller. In the end I leaned it is completely possible to replicate star citizen in unity as far as "seamlessly" getting into a space ship and flying it off into space, exiting the controls while the ship is moving in any orientation as well as exiting the ship into space into an FPS "space sim controller" and landing and getting out of the ship "anywhere" seamlessly. That is what I set out to find out and I did it.

 

yeah man - wait until youre actually trying to build out gameplay and integrate all the different elements together into some kind of entertaining experience, while maintaining an aesthetic, real game mechanics and at least some sense of narrative, coupled with bug fixing, ux design, and the total lack of interest anyone is going to show in your early progress.

this is deep water bro.

 

its a video:
https://drive.google.com/file/d/0B5gBU0VVYsRIUFhHUkMzQjZmWEU/view?usp=sharing

This is the first thing I programmed in unity, my "first scene" ok. Now don't make fun of me! I made this in my first week of ever playing with a game engine.

 

This is very doable, and has been done. Just because you don't see it on the asset store doesn't mean folks haven't done it. I've built something that was pretty much exactly that (just the display was done differently). Technically a spherical world wouldn't be infinite though, it would just loop around. It can also be done very easily with a traditional tileable world set up, using shaders to curve the terrain. It's important to remember, that you only have to achieve the appearance of something, the actual mechanics may differ. There is virtually nothing that you can't do in Unity (or any other modern 3d engine), from a practical standpoint.

 

Oh I already have an idea of what I want to do keeping it simple and fun, arcade style and doable solo. I have played every space sim known to man and I know what I can and can't achieve in a reasonable time in a space sim. The controllers arent hard, the story and any kind of complex "ship" damage system would be time consuming, but not an exploration / arcade shooter with a decent story .

 

Sorry, but calling KSP very basic and unimpressive shows some lack of knowledge on how to handle large worlds/planets/solar systems in games. Even worse when you say you can do better in a week.

The scale in KSP is real. Just like every game there's a lot of trickery to make it efficient, but it's still real. Plus, the planets actually have motion and orbit the star, and the planets also have angular rotation.

You might want to watch the talk they did at Unite, describing a few systems they did:

 

I don't want it to be a "single" or spherical object. I want the terrain to curve around only until it is out of the view fulstrom but not connected to any pieces behind the view culling. So in a sense its like an infinite grid but only withing the view of the camera.

 

Can't you tell by looking at it that it is not to scale... it is made to apear large but I put 100$ that planet is no where near 1 mill units in scale.

 

I was speaking only about the graphics fidelity. And the fact its just a third person view, you are not inside the space ship looking down. There is no high quality atmospheric effects. It is not immersive in the sense of feeling like you are outside of a planet look at my planet man. My plantets graphical fidelity blows kerbal away man.

 

See, you already know about trickery and hiding flaws, look how you walk away from the window at 4:16 to avoid showing off the awful transition!

I'm making fun, but it's true - smoke and mirrors - games aren't perfect flawless works of art. There are always problems and ugly bits, and we need to do the best we can to make sure the gamer doesn't see those. In video like this - walking away from the window for a couple seconds is a good way to do it. In terms of working with terrain or other limitations, there are other tricks.

 

Skip to 2 minutes and look at my planet from space, I don't even have high enough resolution textures for it yet because the planet is so massive in scale. The ship is literally a "tiny spec" in comparison.

 

Its because I'm no where near done with it yet. I didn't want to work on the atmospherics until I get a terrain system going. But after I do I fully plan on adding volumetric cloud effects as well as a planetary exit and re-entry simulation with tons of particle effects as well as storms in the atmosphere that will interact with the ship causing turbulance, the "burn" effect on the front of the ship ect. I want it to be more of an immersive experience than a "scientific" or "orbital" simulation. It will be more of an arcade style cinematic experience.

Oh I have plans and im getting closer and closer. Each scene I create teaches me a little more and brings me closer to my dream of creating a specific kind of planetary landing system, NOT like in star citizen or elite dangerous, more focused on the re-entry itself and planetary atmospherics and interaction. I want storms that can cause you to crash your ship and die for example.

 

By the way the back door of the ship probably wont be open in the release version so sure I may hide some things and use some trickery if that is what I must do to make my vision become reality.

 

And you cant really compare kerbal to what I want to do. Kerbal is an orbital simulation which has no graphic fidelity to speak of, whereas mine will not have orbital physics most likely, at least not in the beginning. It will be the complete opposite focused on graphic fidelity and re-entry animations and effects not only from a third person view of the ship but from the cockpit OR while walking around on the ship but leaving the controls during re-entry may be bad for your health unless you planning on jumping, I will be making a skydiving controller too... which will be fun.

I could also very easily have a vehicle inside the ship for roaming the planet surface. The ship in that video is just something I got off sketchup, I need a modeler for a rigged ship made for the game. Its all but impossible to find any good game ready ship with an interior especially large enough to hold a buggy or something.

This game will be single player as well and I have no initial plans to make it multiplayer... I am only one person after all.

 

Stop talking and get to work.

 

if I get a working prototype 80% there will you help me find a ship modeler? Are you a modeler? I hate to do all that work and get stuck with a crappy ship.

 

Nobody will help you.

Get to work.