[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / r / s / t / u / v / vg / vr / w / wg] [i / ic] [r9k] [s4s] [vip] [cm / hm / lgbt / y] [3 / aco / adv / an / asp / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / qst / sci / soc / sp / tg / toy / trv / tv / vp / wsg / wsr / x] [Settings] [Home]
Board
Settings Home
/3/ - 3DCG


Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
File
  • Please read the Rules and FAQ before posting.
  • There are 4 posters in this thread.

10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
06/20/16New 4chan Banner Contest with a chance to win a 4chan Pass! See the contest page for details.
05/08/16Janitor acceptance emails will be sent out over the coming weeks. Make sure to check your spam box!
[Hide] [Show All]



I'm a bit puzzled by this, and a bit curious.

Let's say I bake normals on the mesh presented above. To save time, I unwrapped the bottom part of the grenade cylinder as one island, and the cap at the bottom as one island (as can be seen from the SGs in the third image - not literally all of it is one island, but the parts with the hexagonal indents are.

I know I gotta make each UV island into its own smoothing group, if I don't want seams on my bake. However, it seems that if I do here, the bake gets wavy as shit. Weirdly enough, this is something that some applications have no problem displaying (like substance, which displays the normals perfectly fine even if they were baked in xNormal), but it starts to break into other places like Marmoset and Sketchfab.

My questions go as follow:

1. What are the exact guidelines for separating hard-edged faces into their own UV islands? The little beveled grips on my mesh (on top of the cog, around each side of the grenade's spoon) are mostly one UV/SG, and they came out just fine.)

2. What exactly makes it so that certain programs can display these "wavy" normal maps just fine, while others will break the shading in arbitrary ways?
>>
File: 1482096950712.png (53 KB, 188x226)
53 KB
53 KB PNG
bake gets wavy as shit because that's how your lowpoly mesh is.

put some support loops and bake again. but oh that's right you triangulated your model. next time only triangulate for export
>>
File: uwott.jpg (503 KB, 1680x540)
503 KB
503 KB JPG
Some more documentation. I can sorta understand the artefacts in sketchfab, but those in marmoset, or why certain engines can display the model fine while others don't, is beyond me.

>>546496
I did triangulate for export, I'll look at your suggestion.
>>
>>546497
>I did triangulate for export
no you fucking didn't. because you can't edit your model now.

look at my pic, start packing your bags.
>>
>>546498
Adding support loops only seems to compress the distortion, rather than solving it (but again, that's mostly what support loops do).
>>
Well yeah, what the hell is going on with the smoothing groups on your low poly? When applying smoothing groups, you have to think where the hard edges would be in real life.
Did you seriously assign one smoothing group to the whole bottom part?
>>
>>546501
That information would be carried on the highpoly, and onto the normal bake. While I evidently shouldn't have created this big of an UV island in this case, I believe you don't know what you're talking about either.
>>
>>546500
im only saying that as what you can do right now. you basically smoothed a fully triangulated model without marking any smoothing groups. because those dark area's that you see actually represent what happens to a complex model when it is smoothed right away. you should go over smoothing groups and re-bake
>>
>>546504
Did you even read the OP?

I know how I can fix this problem. I can redo the SGs and get tiny seams, or redo the UVs and get a clean bake. I made this thread because I'm wondering:

1. What makes it so you can sometimes group hard edges together in one UV island for a bake, and sometimes not?

2. What makes it so that unsynced normals sometimes break in programs, and sometimes not?

2.
>>
>>546503
Normal baking takes the smoothing groups of your low poly into consideration too man.
It's trying to smooth the indents with the round cylinder, you can see that the waviness in your normal map corresponds with the waviness of your smoothed low poly. I't tring too smooth the indents with the big cylinder.

In other words, don't be lazy; fix your smoothing groups, split the UV's and your problem is going to be fixed.
>>
>>546507
you didn't read the questions in the OP either, did you
>>
>>546506
>1. What makes it so you can sometimes group hard edges together in one UV island for a bake, and sometimes not?
what does that mean?

>What makes it so that unsynced normals sometimes break in programs, and sometimes not?
increase the lightning and zoom in, you would find that it does.
>>
>>546509
>what does that mean?
Take the pic in >>546497 for example. The object on the right is composed of several hard edges, yet its "top" surfaces were all unwrapped as one island. While you can see the gradients from the single smoothing group coming onto the normal map, it doesn't seem to have any problems displaying on any apps I've checked.

Compared to the SGs that do break the normal bake, one reason I can guess why this one is fine is because it's a lot simpler, and has less hard edges at extreme angles (nothing above 30, while the grenade bottom has some 60 degree angles). However, I'm wondering if anyone has figured some sort of established guidelines for what hard edges not to group in UV islands, cause that'd be a lot easier than guessing every time and having to fix your UVs because you were too optimistic.

>increase the lightning and zoom in, you would find that it does.
That's absolutely not true, a bake from substance is going to display just fine inside substance, even with these SGs. The problem comes when exporting to other apps, because they all seem to handle normals in a slightly different way.
>>
>>546511
>The problem comes when exporting to other apps
i know what you mean. in SP the color seem more defined, the normal intensity is much better in SP viewport. in online browsers it seem weak/dim/faded. so yeah its just how they code the renderer.

>The object on the right is composed of several hard edges, yet its "top" surfaces were all unwrapped as one island. While you can see the gradients from the single smoothing group coming onto the normal map, it doesn't seem to have any problems displaying on any apps I've checked.

problems with seams happen for numerous reasons.
a pole might interrupted the shading, maybe the triangels were a little too dense or inconsistent with eachother.
or maybe you had inner faces that you forgot to delete, and maybe you some verticies needed to weld because some edge was open and loose. shit happens. you will get better with practice
>>
>>546497
I'm assuming the smoothing is calculated differently across programs or the shaders calculate the normals differently or something?
I know your issue and I've had them too but I'm not sure why it's the case.




Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.