Specular map thread link?

Dave Jr

Hi,

there was recently a post on one of polycounts forums on how to quickly and easily achieve a decent looking specular map. It was a short video on vimeo and whilst I believe I bookmarked it... apparently haven't

If anyone can link me; itd be greatly appreciated.

EDIT: Found it... if anyone else wants it: http://vimeo.com/90507404

Dave Jr

perna wrote: »

Always be sceptical of any suggestion of "quick and easy" specular maps. The reflective properties is a material are THE most important data, and consequently require the most amount of work.

Also keep in mind that this video is using a mostly outdated, non-PBR method, picking values mostly at random, not using gloss, treating specular as secondary to diffuse, using a lot of old-fashioned non physically based "tricks" and not synchronizing spec and diffuse correctly. The tutorial is also using clipped specular ranges, which is a huge sin, and breaking a whole host of other rules.

So just a warning.

Oooo...

Thanks for the heads up Perna, I always learn so much from you... probably more then my degree

(you can probably quote that for the education for games thread) - itll be worthwhile hahaha

Flight

On that subject, has anyone got a good PBR texturing tutorial? Been looking around but haven't found anything useful.

Acid_Wire

Flight wrote: »

On that subject, has anyone got a good PBR texturing tutorial? Been looking around but haven't found anything useful.

+1

There are some useful threads here, as well as the guide on the Marmoset site, but some start-to-finish, in-depth videos or something would be awesome.

Xoliul

No disrespect for the guy doing this, but this is the complete opposite of what I'd do these days. It's about as far from 'PBR correct' as possible, fudging the value all over the place, using the sliders to tweak things, no care put into value ranges, etc..
Though i have no idea if he claims it to be PBR-correct, he does have the PBR checkbox on inside Marmoset.
On top of that, once he gets to the tiles and so on, he starts to manually paint stuff out. This is just my opinion, but doing that stuff in Substance Designer allow you to tackle it in much smarter, less destructive ways, allowing you to tweak things better and reuse logic for other textures. Manually painting things out like that is pretty insane imo...

tynew

Acid_Wire wrote: »

+1

There are some useful threads here, as well as the guide on the Marmoset site, but some start-to-finish, in-depth videos or something would be awesome.

I agree, the only one that went into the "most" detail was probably Andrew Maximov's? tutorial with his Cerberus pistol model. Even then, it was just a quick video breakdown. I haven't actually seen someone complete a tutorial with a start to finish, multi material PBR asset. That would be awesome.

One thing I would be interested in is how someone can sample their data for materials when charts/values that don't exist at the moment for that material type. Guesstimating stuff off the charts is generally only correct in one lighting condition afaik, and can look wrong in some "extreme" scenarios where the environment is radically different.

So many professionals out there, not enough tutorials/workflow examples

SuperFranky

tynew wrote: »

I agree, the only one that went into the "most" detail was probably Andrew Maximov's? tutorial with his Cerberus pistol model. Even then, it was just a quick video breakdown. I haven't actually seen someone complete a tutorial with a start to finish, multi material PBR asset. That would be awesome.

I'd love to make some kind of tutorial once I'm fully comfortable with PBR workflow. Right now I'm dedicating all my time for this stuff, it's really really interesting.

Acid_Wire

perna wrote: »

<education>

Thanks for the explanation
I've been wondering about this kind of thing since I'm just starting to find my feet with the changes from UDK to UE4.

tynew

Cool insight perna. I'm confused a bit with your example. I thought metals were so dense that they can't absorb light. Which iterates the point they are black without reflection/IBL as your first example and technically would make that "physically correct" right? In what circumstance would you not have metals not lit with IBL like your first image?

tynew

Okay awesome thanks perna

Torch

Xoliul wrote: »

Though i have no idea if he claims it to be PBR-correct, he does have the PBR checkbox on inside Marmoset.

Gonna sound a bit noobish but where is the checkbook in toolbag? Haven't seen it but would be good to know.

tynew

I think that is the conserve energy check box under the specular parameters.

Torch

Thanks Tynew, yeah I toggled it on and off but didn't seem to make any difference. Will have to read up more on documentation

D4V1DC

^I think if you turn it off i read somewhere things are no longer pbr?
I think it works with metalness, idk don't quote me. ;p

Xoliul

It only makes a difference for the glossiness and the specular map. I'm not a fan of their setup in that regard, you go into this weird hybrid shading model then, and it invites so much more misunderstanding. Just make sure you really, really read through their documentation and get what they are saying, before you start messing around in any way like that linked video here.