[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/3/ - 3DCG


View post   

File: 1.88 MB, 1244x532, Village1.png [View same] [iqdb] [saucenao] [google]
552675 No.552675 [Reply] [Original]

So I started to make my own game. I just wanted to get some opinions on my work so far.

Made my models in blender and am using unreal engine 4 for the game.

>> No.552677
File: 1.61 MB, 1253x536, Village2.png [View same] [iqdb] [saucenao] [google]
552677

>>552675
Another Pic of my first village.

>> No.552679
File: 1.75 MB, 1250x536, Mine_outside.png [View same] [iqdb] [saucenao] [google]
552679

A mine I got which connects to the first dungeon

>> No.552680
File: 821 KB, 1037x507, Mine2.png [View same] [iqdb] [saucenao] [google]
552680

the insides of my mine so far.

>> No.552681
File: 1.72 MB, 1259x542, mine4.png [View same] [iqdb] [saucenao] [google]
552681

>> No.552682
File: 1.60 MB, 1259x542, Mine_inside.png [View same] [iqdb] [saucenao] [google]
552682

>> No.552683

>>552682
These are not the final textures. Since I am making all textures myself I had to use the bark texture for the tunnels inside as I do not have a good looking mud or dirt texture.

>> No.552686

Lots of scope. What kind of game is it, anon?

>> No.552688

>>552686
Gonna be an action rpg. I know its an ambitious project. At the moment its just a hobby, as it gives me a reason to do lots of artsy stuff like creating models with clay and then scanning those. I started working on it in december.
Had the village done within the first few weeks. Then had to take a break cause of university. This week I made the mine and am now working on the first dungeon.

>> No.552689

>>552675
There's a lot of issues, however the one I feel the biggest need to mention is saturation. A lot of your textures are really, horridly saturated, even for a cartoon look. The grass, tree leaves, and roofs should all have their saturation turned down several notches.

https://docs.unrealengine.com/latest/INT/Engine/Rendering/LightingAndShadows/Lightmass/#gettingthebestqualitywithlightmass
This page sorta explains what the problem is. Basically, when all of your diffuse textures are super noisy, saturated, dark or contrasty, it kills your lighting and makes it hard to read (I really struggle to see where any of the shadows in your level are).

Tbf I can almost guarantee that's the main reason why your level looks uggo right now.

>> No.552691

why is it glowing? what is your skylight/directional light intensity?

>> No.552693

>>552689

That may be because I used a skylight so that I can see shit when Im working in the buildings. Gonna be working on my lighting when I have more time. But thanks for the tip.
I am focusing on making the models at the moment as I can do that in small chunks.
I will definitely look into changing the saturation.

>>552691
For my skylight its 1 and directional light is 10.

>> No.552695

>>552693
set directional light to something more like 3-5

>> No.552698
File: 176 KB, 1257x537, Villageout.jpg [View same] [iqdb] [saucenao] [google]
552698

>>552695
I decreased the skylight to 0.1 and the directionel light to 3

>> No.552700

that bloom is hella intense and doesn't look like it matches the low poly unrealistic style of everything else. can you just turn the bloom off entirely. also why is everything so grainy? jpeg compression? everything about these screenshots is about as unappealing as can be

>> No.552704
File: 126 KB, 1255x539, closeup.jpg [View same] [iqdb] [saucenao] [google]
552704

>>552700
The grainyness comes from the low Resolution I set the unreal engine to. So far my focus was on learning how to make 3d models and using the unreal engine.

My textures are all hd tga files.

I know I still have a lot to learn. Especially using the unreal engine.

Here is a close up of my textures.

>> No.552705

>>552693
Nah, your shadows read fine on the ground terrain cause its texture isn't blown the fuck out in saturation or contrast. I would honestly use it as a baseline because it's not too saturated, the trees and roofs seem to be at a literal 255 saturation in comparison.

>> No.552768

>>552675

Going to sound unintentionally harsh here, but:

It's very unappealing, the longer you keep saying "I'll fix it later" the more you'll have to fix and the less likely you'll ever fix it.

Even if you make a fundamentally good game, the visual aesthetic / art direction (or lack thereof) is so unappealing that even professionals, looking at this as a portfolio piece, will have a hard time looking past it. Get some concept art that you really like, settle on some kind of palette / color choices, pick some kind of definitive style, and develop a workflow (that is to say, a set of practices) that let you approximate that style as best you can on each of the new assets as you make them.

I think it's great that you've gotten this far and done this much, but so far this really is tremendously unappealing visually and if you don't fix that problem, you're going to end up with a really bad looking tech demo that people will just dismiss out of hand (the Unreal community is full of such examples).

>> No.552837

>>552768
Agree 100%

You should start working on your "final textures" NOW instead of waiting. If you're showing screenshots to people it looks extremely amateur. Maybe if you were posting amazing gameplay videos on YT it could make up for zero effort lighting and texturing. But not on a CG board.

>>552704
These textures look like you downloaded a picture of tree bark and applied a single photoshop filter. You can buy simple textures from the Unreal store if you need placeholders. But this is unacceptable.

>> No.552838
File: 125 KB, 630x418, rothenburg-ob-der-tauber.jpg [View same] [iqdb] [saucenao] [google]
552838

OP, I really hope this is either your first game or you are just testing things, cause this sucks. If you are trying to be artistic with this you need to go back to the drawing board. Everything about this screams uninspired RPG. Take a look at real world reference and do some sketches first.

>> No.552842
File: 67 KB, 600x928, CIYYRCFVAAEyglU.jpg [View same] [iqdb] [saucenao] [google]
552842

I wonder if people were this critical when Notch first posted Minecraft here

>> No.552844

>>552842
It's 2017.

>> No.552845

>>552837
>You should start working on your "final textures" NOW instead of waiting.
>work on your final textures with 0 gameplay, 0 programming
wut. Have you ever made a game

>> No.552875

>>552845

"I'll fix it later" is the mantra of the nodev. Stop trying to pull OP into your world of failure.

OP if you're still in this thread, I'm not trying to discourage you, nor am I accusing you of "low effort", you've obviously worked very hard but you've got a poor foundation and it's resulting in poor results despite a ton of work.

>>552842
Minecraft succeeds despite its "programmer art", not because of it.

>> No.552876

>>552842
he posted it to /v/, not /3/.
You're going to get completely different criticisms depends on what board you're posting to.

>> No.552878

Looks better than anything I've done in Blender and Unreal. Good work, OP.

>> No.552881

>>552768
You are right. I will start to learn how to make good lighting right after my exams. At the moment I only have time for modeling. Next months I have 2 months to really invest time into learning how to make good lighting and animations and modeling characters (which will be done by scanning clay models as Im way better in doing those than making them in blender). These will be my next steps. But I personally see no point in learning how to make lighting when you dont even have any models. So thats what I started with. First I learned how to do the basics in 3d modeling. Walls, grass, trees etc. After that I went on with learning how to use the unreal engine.

>>552838
I actually started to learn how to model with blender at the end of december. Paused for a bit at the end of january to have some time to study for exams. Now I started picking it up again.


>>552842
I really appreciate the honesty here. Telling me what really "sucks" about my project is the only way for me to get better.

>>552837
I did these textures by taking pictures of stuff and using some gimp filters. I still have to learn how to use all the programs better. So at the moment I'm playing around a bit to see what works and what doesnt. And as a "poor" student I can't afford better programs (piracy or educational licenses are no options for me)

>> No.552882

>>552875
Everyone needs to start small. For me this is a hobby. I dont expect to make any money of this game ever. I try to take step by step but at the same time not to rush anything. My world is seperated into parts and I started modeling the smallest one right now. which consist of a river, a mine, a village and a small dungeon. When Im finished with modeling these I can go on to make those more beautiful. After that I can worry about the right workflow.

>>552876
/v/ is toxic as fuck. I would not trust those guys in giving me advice on anything. On this board I can expect to get some good and honest tips.

>> No.552883

>>552875
>"I'll fix it later" is the mantra of the nodev. Stop trying to pull OP into your world of failure.
is this some kind of 2017 troll? OP has 0 code, 0 gameplay and you want him to make textures that are FINAL ?

>> No.552907

>>552675
Where's the cute girls?

>> No.552938

>>552883
He is on an forum for 3dimensional aesthetics.
Showing placeholder art is like showing no art.
I can totally relate to his problem, but if he hasn't anything to show he simply shouldn't asking for advice here, or the result is like this.

>> No.552940

>>552883
>>552938
Calm down people. Its no placeholder art. Its the best I could do at the time. But Im happy about all criticism. Especially when all I get when asking friends and family is: It looks great. That does not really help at all. Just keep it civil please.

>> No.552945

I am quite a misfit here as i am more of a programmer, but when using unreal what do you program in? Is it possible to use C++ which is by far my most preferred, as many of my algorithms work in this.

>> No.552946

>>552945
Oh, and I actually thinks it looks quite nice, although it appears more as a village sim, than an rpg with all the Bird View shots.

>> No.552947

>>552940
couple of points
>work on your buildings. look at how buildings are actually made. you have wooden columns terminating straight into stone, or down into nothing, looks bad. the combination of wood and stone doesn't look plausbile, especially on the towers
> work on the mine, its too regular inside. You can still have straight tunnels, but make the wall geometry a bit rougher. This is a working environment, not a temple or tomb. Add litle projections of stone and dents in the wall.

>> No.552958

>>552945
Not OP but when working in UE it is by far preferred to use Blueprints for the majority of game logic and reserve C++ for performance-critical code. (Writing C++ for UE also involves entering their nasty template-heavy compile environment.)

Every modern video game is written with some sort of internal scripting language, from Quake's scripting language QuakeC, and earlier.

>> No.553160

>>552675

Nature has more plants than one type of trees and green grass.

You should place few birches (white bark) and pine trees, at least 5 types of flowers, bushes with flowers or berries depending on season, white/green/blue mold inside mines, moss covered rocks, where's all loose stones? did people collect all of them?

Can't see any farms, what do people eat? do they only hunt? no animal would come near that town to eat green grass.

What's the point of that fence? There's much more steeper and higher drop down nearby.

And, trees grow longer branches on the south side of the tree or on the side that has most light coming from. Now trees look like growing every direction, I don't feel immersed.

Also, those mountains look like piles of rubble.

>> No.553419
File: 101 KB, 1069x487, netter lights.jpg [View same] [iqdb] [saucenao] [google]
553419

Just for those interested. I worked a bit on my lighting. Increased saturation, made some colour grading and changed the lightcolour a bit. I will go on to work on more Types of trees and other types of natur foliage.
Also either within those walls or outside of them I will put some farms and maybe even a mill.
Those mountains arent final either since Im still beginning.

Thanks for the criticism guys. I really appreciate how much you gave me to think about.

>> No.553425

Stop manually painting the landscape OP, waste of time.

>> No.553427

>>553425
You are right. I should put an auto-material tool.

>> No.553432
File: 142 KB, 898x557, maya_2017-02-28_19-33-29.png [View same] [iqdb] [saucenao] [google]
553432

Since this thread is about a game in UE4 I thought I could ask (hope OP doesn't mind) some simple questions that don't warrant their own thread, about the importing pipeline between Maya and Unreal:

1 - I have this small environment, however it's all split into parts right now (walls, sliding doors, chairs, table, fence, etc.). Is it necessary to convert everything into a single object or is it fine to export like this?

2 - Send to Unreal or Export Selection/All... command? Does Send to Unreal do anything special besides creating an import folder in your UE project?

3 - UE supports Lambert, Blinn, Phong/Phong-E and Anisotropic material imports. However, none of these have familiar terms in the fields when plugging in textures like Metallic, Roughness, Ao and so on. Is it better to use these materials for something simpler that would need, at most, an Albedo and AO map, or is it better to import the raw mesh into UE and fuck around with the shaders in there?

>> No.553441
File: 3.87 MB, 1685x1035, _AutoLandscape.png [View same] [iqdb] [saucenao] [google]
553441

>>553427

Pic related is the landscape I just finished. It's all auto based on slope and height, sand around edges not shown (my game will be on an island). I can also paint layers if I need to, and adding materials to landscape is as simple as popping in some textures.

The grass is also auto generated which saves so much time

>>553432

1. It depends, do you want the in-game object to be just one mesh or separate? This question doesn't have an answer, it's up to you.

2. Dunno. I just export my meshes as fbx and import. Sometimes I'll do several and choose not to combine them during import.

3. Dunno wtf you're talking about, but Roughness/Metallic/etc are pretty easy to comprehend, just get a nice mask texture and mess with sliders using a lerp node to see what they do

>> No.553442

>>553441
I meant the material fields as in, Blinn and related has stuff like Ambient Color, Incandescence, Translucence, Eccentricity and alikes, which seem more geared towards something like CG animations? Whereas something like Stingray PBS has more familiar terms like Color, Normal, Metallic, Roughness, Emissive and AO Maps.

The thing is that Unreal doesn't import Stingray PBS natively but it does for Blinn. Since I have all these maps, is it better to try and use Blinn anyway or to go to Unreal for something *like* Stingray PBS?

Sorry if this isn't very clear, I'm a virgin when it comes to shaders and shader import pipelines between modelling softwares and engines.

>> No.553443

>>553442

Sorry if you're a virgin then I'm a retard in this matter. I've never even made materials in software such as maya, I just model, texture in substance/ndo, and export textures to unreal. substance will even pack rough/metal/ao into a single texture for ease of use in ue4

>> No.553444

>>553443
This is probably just an indication that I should stop being stupid and download the student version of Substance Painter.

People say it's fucking amazing for easy texturing. Either way, thanks anon.

>> No.553445

>>553444

Or just pirate it! I spend like a year struggling with texturing/modeling using GIMP/Blender, then I realized I can just pirate everything. Now I use zbrush (amazing), max, maya, substance painter/designer, b2m, knald, I even pirate all the unreal marketplace shit to use for learning...

OMG! I would have spent 10s of thousands if I had to purchase all this shit

>> No.553517
File: 1.61 MB, 2816x2112, dscf9114.jpg [View same] [iqdb] [saucenao] [google]
553517

>>553441
You know grass doesn't grow on this level of steepness only near the sea, right? And even there it's a rare sight. You need an angle of over 80degrees in order for grass to not grow naturally and then most likely in a mound like this the soil will rot out and will create a 90 degree angle drop, I'll post an example in the other picture.

>> No.553519
File: 3.22 MB, 3072x2304, 14150876.jpg [View same] [iqdb] [saucenao] [google]
553519

>>553517
Here's a reference photo for a hill in my city, look at how complex the terrain is. I assume the right hill in your picture is something like that, but the smalles hill in my city is around 160m in diameter, yours makes no physical sense. Get more references and do it properly.
Also, don't forget that without trees/tree system, a hill won't have the necessary soil structure to maintain it's shape and it will most likely tumble down, so you can't have these steep hills without trees on them..

>> No.553520
File: 528 KB, 1600x899, DSC09061.jpg [View same] [iqdb] [saucenao] [google]
553520

>>553517
here's the rotted hill I promised.

>> No.553523

I think it's good looking. I wouldn't be a pirate though. Original, hand crafted textures will be respected more and will be more original/ Jus sayin

>> No.553524

>>553523
no one gives a fuck if you made your textures in a pirated copy of SP or in mspaint, pixel by pixel

>> No.553525

>not final textures
I seriously hope so.

>> No.553636

>>553517
>>553519
>>553520


If there's one thing I know about games is no one cares about whether the grass is correctly placed or not. And while we're on the subject, the huge majority of players don't notice something that a graphics artist would. They hardly notice repeating textures, or bad seems, low poly models, etc..

Just look at some of the biggest games of our time. They all LOOK like garbage, but have good gameplay. Spend like 30 seconds standing still in FO4 as an example, and you'll see just how awful everything looks.

>> No.553649

>>553636
Currently playing The Witcher 3, cannot confirm

>> No.553655

>>553636
>Spend like 30 seconds standing still in FO4 as an example, and you'll see just how awful everything looks.
That's because FO4 has objectively poor quality graphics compared to modern games.

But you're right that the average "gamer" can't tell the difference between good graphics and shit graphics.
They think they can, but they really have no fucking idea.

>> No.553669

>>553636

People do notice those things, they just can't articulate why the scene "looks like fucking garbage and the artist should kill himself".

Likewise, if you do it right people still won't be able to articulate why it looks right but at least they'll say "The artists did a good job on this one, let's buy their game."

>> No.553680

>>552700
>>552704
>bloom is hella intense
seconded, plus contrast is a bit harsh, regardless of final resolution. i like the general idea and what you're going for though

>> No.553854

>>552675
Looks cool. It's good you're making low poly and focusing on quantity. Very large project thought. If it doesn't seem to be going in the direction you want, don't be afraid to start over, or an entirely new project all together.

>> No.553886

Should've gone with unity if you're just starting, it's easier than baking eggs.

>> No.554314

>>553419
since nobody commented OP I thought I'd just mention your lights look 10x better in this scene.

You need to work on adding objects to the landscape though, like boulders and plants and things to break up the green/brown thing you have at the moment. And work on making the geometry at the edge of those canyons more rocky and broken - placing large rock assets over the landscape to cover the smoothly ounded bits will make it a lot more believable.

With regards to the little town it looks like a good start but you should spend some time on placing the houses in an interesting way. Try to break up sightlines a bit and create alleyways and town squares, giving more variety to the spaces the player can go in. Maybe go play Witcher 3 and run around Oxenfurt or Novigrad a bit for some ideas. Try placing upper level bridges or stone supports between houses, add a clock tower or a church or a castle keep, add a market square with a bandstand, that sort of thing. Lots of assets that aren't just houses or castle too - things like fences, barrels, tables, workbenches, flagpoles, animal pens, food stalls, etc. Also work on the bridge, there should be some more interest as you go over and see the gatehouse ahead, it's not just a bridge and then a hole in the wall.

Keep it up though, looks like you're making some good progress.

Oh and final thought, historically these sort of towns were surrounded by smaller settlements and farmlands that are necessary to support the town's food needs, as well as any industries that were too space-hungry or unpleasant ot house inside the walls (like leather tanneries, kilns, etc). Might be an interesting way to break up the surrounding landscape - have some small buildings by the bridges, some guardhouses, then further out have farm buildings surrounded by fields of barley. That'd look nice.

>> No.554324

>>553419
Second >>554314
The saturation of your roofs is f u c k e d but the rest is a substantial improvement over your initial post.

>> No.555760

>>552675
All I want to say is be careful when relying on a flat plane for terrain. Especially with steep slopes try adding very large boulders where the big fall offs are to add more detail so mountain sides don't look uncanny. Also use many many many reference pictures. Just look at a shitton of photos

>> No.555846

>>552675
Decent first attempt. Keep at it mate. 5 - 10 years you'll be a pro