Road to somewhere: Learning thread for me!

luge

Ok Polycounters,

I haven't been doing so great with my high-poly modeling, and in my other recent thread for a project I am working on, it was brought to my attention that I was approaching things not necessarily wrong, but definitely not right either. I was making my edges too real.

So, now I am on the road to working out how to do high poly better, and to texture better and make my materials better. my current project is put on hold for the time being, while I practice my workflow and get the idea down first before tackling that project again.

So, without further wait, here is what I am practicing on currently:

Reference:


High-Poly:


I currently have two different renders for the piece, one in marmoset, and one in maya viewport that I cleaned up. In marmoset, for some reason, I can not get my textures to render properly right now. which I am working on.

Marmoset Render:


Maya Viewport:


Textures:


The material definition is not where I want it to be right now. but I'm still working on it and trying my best to learn more and get it where it needs to be.

Here is a link to the other project thread, that is currently on hold:
http://www.polycount.com/forum/showthread.php?t=132573

Joost

Marmoset tb2 uses a different kind of rendering, "physically based rendering". It requires different maps but produces better results. Here's a tutorial and the theory behind it. I suggest you read both thoroughly. :
http://www.marmoset.co/toolbag/learn/pbr-theory
http://www.marmoset.co/toolbag/learn/pbr-practice

Your gloss (old spec) map is going to need more detail. And your diff/albedo needs less detail.

Not saying this is perfect but it might help you:

Click on the image for full res.

luge

I did glance through the articles before. been reading around trying to understand it more. but I only skimmed the articles. I'll take a deeper look into it and figure this stuff out.

by the looks of the spec/diffuse/ref I'm guessing glos looks more like old spec maps (without color) , ref is your spec (atleast in marmoset). alright. got it. well, I'll still be reading the articles to get it more ingrained and learn more.

thanks for the info though.

luge

so... late night again. didn't sleep. decided to work instead. I read through both of the articles you linked a few times over to try to get things to click. I think I got it for the most part. some of it I had to play around with in order to actually understand it and how it affected the texture and materials more. and this is the result.





And the textures:


I'll probably look up a tutorial about setting up for a more proper presentation. but otherwise, I'd feel that this knife is about done. maybe a tweak here or there, but its about time to move onto the next object to practice on. something a tad harder. Thanks again for the help.

MrOneTwo

[ame="http://www.youtube.com/watch?v=LNwMJeWFr0U"]Physically Based Rendering for Artists - YouTube[/ame]

Here you have great stuff about pbr. Your model looks good for a learning experience.

luge

thanks for the video. and, i'm not really new to modeling per say. I've been doing it for awhile. this was more for practice with high-poly modeling, since I've been doing things wrong for so long, which I think I've finally gotten an understanding about how to do... which is in part due to people on here giving me advice and such. but i'm really happy with how things baked down after I figured things out and it has been fun.

RobeOmega

komaokc wrote: »

Marmoset tb2 uses a different kind of rendering, "physically based rendering". It requires different maps but produces better results. Here's a tutorial and the theory behind it. I suggest you read both thoroughly. :
http://www.marmoset.co/toolbag/learn/pbr-theory
http://www.marmoset.co/toolbag/learn/pbr-practice

Your gloss (old spec) map is going to need more detail. And your diff/albedo needs less detail.

Not saying this is perfect but it might help you:

Click on the image for full res.

Is there anyway to disable PBR for marmo?

luge

from what i can tell, PBR is what marmoset is all about. so I doubt it. at least I never saw the ability to. since it is a PBR based renderer. there are other options though.

Joost

There are ways of getting the original marmoset, which I think is still pretty useful if you're texturing for non-pbr engines.
@luge: Try disabling the sky background for your renders, it makes the knife stand out more. The textures are looking better, but I'd try to add some oily smudges to break up the even noise.

luge

yeah, i'll try adding some fignerprints or something to break up the noise. also, i'll try out disabling the sky background. hopefully it does help with presentation and such. i'll put up a new shot later today.

luge

Ok, I think I can call this knife done now. I'm pretty happy with how it turned out. Obviously there could be more tweaks to it. but I'm gonna call this done. I'll still take C&C on it obviously.



And thank you to everyone.

Ave

Nice one

luge

Thank you very much

luge

welp, started the next one earlier today. was originally just gonna do a quick blockout... then took it a step forward cause why not?

Concept (Not mine):


Low-Poly:



High-Poly:




The low-poly model needs some slight tweaking in certain areas. and can lose some tris in others. I did miss a piece, but I dunno, I like it the way it is. if I want to, I can always add it afterwards, of course I'd have to redo some parts, but thats what saves and copies are for right? ^_^

Joost

It looks good, but be mindful of your edge thickness. There's some very tight edges in your high poly.

luge

ah, i see where you're talking about. thought i made sure they wouldn't be so tight oh well. I'll fix those spots up.

Joost

It's definitely a big improvement over the knife. There's a few areas like the screws and the bottom of the bullet that are really sharp and probably won't bake well. And then there's some areas that could be softened up a little bit, like the rail/sight at the front.

luge

Well, I did do a bake of it to see how well it would bake. some parts were fine... others not so much. I also ran into some major issues in the normals for some pieces that made no sense whatsoever. I fixed up the high-poly, and rebaked. still had some issues, even the same one from before... but i decided to just clean up the texture and so how well it goes. I may rework some edges again though in the high-poly.

Heres what it looks like now with a little color. might be hard to see, but i'm leaving it like this for tonight and will get back to it tomorrow.

luge

Decided to rework the hi-poly a tad, and rework some parts of the low-poly so it would bake better, did get a better bake with less cleaning and fixed up some issues. I like it so far, time to actually get into the texture.

Joost

The edge thickness looks better, but there seem to be a lot of shading errors? Do you have smoothing group splits at uv seams? And what did you bake in? Maybe your green channel needs to be flipped.
I Suggest you use handplane if you're not already.

luge

I think i'll look into handplane. just read up on it since you suggested it, seems neat. I have been uving where the difference in smoothed edges are, but I'm using maya. maya does not have smoothing groups per say, only smoothed/hardened normals. I'm also baking through maya, which works fine for the most part, but I've been baking everything together. I may try a different approach of taking it all apart and baking different parts. I tried to bake in xnormal, but... yeah, didn't go so well, not exactly sure why but it was an unfixable mess...

Also, what I'm assuming you are referring to as shading errors would be the jaggedy lines? For some reason, those only appear on seams when baking out through maya. I tried to add seam fixing, but it appears on the seams no matter what. still trying to find another work around. the only thing I can think of is to just clean it up, because it does appear in the normal and ao.

luge

So, I found a work around for the problem. it was the baking process in Maya, fixed it up and am currently working on the texture itself. heres what i got so far (still needs a bunch of work.)





Still need to define the materials more, add some more wear and tear and the like. still a little ways to go. although, i feel like i should have finished this piece already, had i not been fighting with the bake.

Essentially, the the problem was that I was baking the low as one object, so I broke it up into multiple objects, setup the transfer maps for it, and then fixed up my envelopes to cover what needed to be covered. there were some errors still, minor ones but it worked out.

luge

Could fix it up some more, but I'm calling it done... now off to the next project. Think I'm gonna start back on environments. every now and again i'll post some practice for critique

BARDLER

If you are doing the metalness workflow you only have full black or full white, for not metal and metal respectively. When something is a metal it will use your information in your albedo as the specular. None of the materials on the gun are making sense and I can not tell what type of metal you are trying to create, and I have no clue what the red material is supposed to be. Really try to focus on the materials and the recreating the reference you are using.

luge

ok, I get what you mean... I was confused when I first read your post. I actually felt like I got what I wanted. but I see what you mean.
I guess I wont call it done just yet.

luge

since I guess i'm not done with it, i'm reworking the diffuse. sucks that with metalness you lose the ability to control your non-metal reflectivity. oh well.

Still trying to get the material right for most of it. the material in the concept looks like a painted metal for the bottom part, and a non-painted for the top, but I'm having issue getting that kind of look right now. this is what i've got:



right now it just looks dirty, worn away, or like someone has been touching it a lot, and hasn't cleaned the grease off.

Joost

It looks pretty good, but everything looks quite blurry. What resolution are your textures?

You can achieve a lot with just glossiness. With PBR more glossiness= more reflectivity (smaller highlight) If you cheat by using the metalness map you're basically creating unrealistic materials.

The metal needs more scratches. Right now it only has soft details like smudges. Something like this could work http://www.cgtextures.com/texview.php?id=57525&PHPSESSID=e52512hki2j5tmhu1fjd4safk0

You need to decide what material you want the red part to be, look up reference pictures and then base the texture on that.

luge

its actually at 1024x1024. not quite sure why it blurs, but, i do have some scratches, they just aren't very visible because I toned them down quite a bit. I'll fix that up at some point. right now i am fighting with my computer >_>

luge

Update:




diff lighting:

BARDLER

So painted metal is kind of a funny material because if the paint has any opacity to it than you get a metal effect and non metal effect so the metalness workflow kind of breaks down. If you want to stick to the metalness workflow than you could try to hack it using a middle gray value for the paint and see what you get. I think the better option would be to do a spec map so you can get better control over the material. On a side note though I have no idea why a gun would have a painted metal grip like that, it seams kind of out of place.

Here is a quick little mock up to help you out. Nothing to crazy just setting my base values down to get my materials to look decent and then adding in some simple gloss detail. Obviously on a model you want the scratches to follow the geometry and usage, but this shows the basic info.

I used a hires picture of a gun and a painted wall to get the materials to read correctly and then I blended between the two with a mask. I can send you the photoshop doc if you PM me your email.

luge

i had actually already thought of doing a 50% gray for the "painted metal" then decided that I didn't like the idea of it being painted metal... felt, weird. so i dropped that idea. so i went a different route instead. but thank you, I wouldn't mind seeing what you did though in the photoshop doc. i'll pm you my email.

oh, and, after you said something about how the metalness works, i looked more into it to understand exactly how things work with it. and I'd agree that using a specular map would be better for the control, but in this case, it allows me to learn more about the limitations and how to work around things like this, or where I can change design ideas to adjust for workflow or whatever the case is.

luge

Still working this plastic material...

What I'm aiming for:


What I've got:



I feel like its getting there.

BARDLER

Plastic is not a metal so the specular should be low, generally anything that is not a metal is about .25 or lower. The plastic on the Wacom that you circled is a dark diffuse, a dark spec, and a bright almost white gloss.

luge

I'm still using the metalness workflow, so I'm trying to work the gloss to get the desired look I want.

luge

Well, I decided to work a Spec instead of the metalness. I got the concept of working around metalness maps, so I set up a specular.