3D Coat - PBR and more

rcallicotte · October 28, 2014, 05:04:52 PM

Have you heard? Andrew is introducing the use of PBR within 3D coat. Here is some more information I found on The Foundry websites about other things coming someday -

He supposedly hired 3 additional programmers (not counting the guy who handles the MAC and Linux builds, and Raul, who handles the LiveClay tools) earlier in the year, and one of those is supposed to be working on expanding the Retopo room to be a more full-fledged Poly modeler, and they plan to implement NURBS mainly for export to Manufactuing standard file formats (ie., STEP, IGES, etc).

After PBR, I think Andrew plans to add true Sculpting layers (per object) in the "Sculpt" workspace. You can already do that with the image based sculpting toolset in the Paint Workspace, but most people are going to do most or all of their sculpting in the Sculpt room/workspace. Then he will likely handle the poly modeling tools. So, I think that is the progression.

Upon Infinity · October 29, 2014, 10:30:16 AM

I haven't heard, nor do I know what PBR is or does. However, I do know they have wanted to move towards more poly modelling in the retopo room.

If it's usable, it'll be great, as 3D coat is pretty terrible to use for geometric objects right now.

rcallicotte · October 29, 2014, 01:23:09 PM

Some useful explanations of PBR -

http://www.marmoset.co/toolbag/learn/pbr-theory
http://www.garagegames.com/community/forums/viewthread/136389

physics and beyond me - http://www.cs.rpi.edu/~cutler/classes/advancedgraphics/S13/papers/lafortune_94.pdf

It's becoming more popular in many texturing applications, like Allegorithmic's line and Quixel.

jaf · October 29, 2014, 02:59:02 PM

I'm watching this closely since I have 3D Coat and the Allegorithmic's tools. Substance Painter looks really interesting though I haven't tested it much (yet!)

Upon Infinity · October 29, 2014, 03:10:39 PM

I think I understand the basics, but correct me if I'm wrong, but is this to mean that it will only improve the rendering in the render room of 3D Coat and not exported objects? If so, I think I'm more interested in the new modelling tools in retopo.

Tangled-Universe · October 29, 2014, 03:53:24 PM

Quote from: Upon Infinity on October 29, 2014, 03:10:39 PM
I think I understand the basics, but correct me if I'm wrong, but is this to mean that it will only improve the rendering in the render room of 3D Coat and not exported objects? If so, I think I'm more interested in the new modelling tools in retopo.

Yes exactly!

I don't believe in apps which "can do it all" but I rather believe in highly specialized/niche apps.
PBR is nice for turntable presentation of your creation, but it is completely lost as soon as it goes outside of 3D Coat.
Unless the whole industry starts to standardize their shading/texturing methods I see little use for a modeller to have a PBR-er.

rcallicotte · October 29, 2014, 04:45:35 PM

No, PBR is useful outside of 3D Coat, which is the only reason to do it. The Allegorithmic set of tools, including Painter, can all use it. The maps that can be created using PBR can be used in any application and have already been used in some games. It's taking off, because it's a different way (and seem a better way) to do materials and texturing. Quixel uses PBR. Marmoset uses PBR. Unreal Engine uses PBR.

Allegorithmic's Painter is pretty amazing and so much less money than other similar painting programs. The Allegorithmic staff is very helpful and I think you would find their texturing and materials through any of their applications are top notch.

An example for UE4 - https://www.unrealengine.com/blog/physically-based-shading-in-ue4

Tangled-Universe · October 29, 2014, 04:58:06 PM

They are all doing the *same* PBR?

jaf · October 29, 2014, 07:30:50 PM

I think the biggest concern right now is from what I've read, PBR doesn't have a well defined spec. So I'm not sure about Andrew's implementation, but I'm sure he's privy to more information than I can get with general searches on PBR.

But yes, TU has a good point -- apps like 3D Coat or Zbrush are usually not the end of the creation chain, so it needs to be implemented in a way other apps can take advantage of it.

Tangled-Universe · October 29, 2014, 08:30:35 PM

As long as 3D Coat's PBR isn't the same PBR as anyone's PBR (did I just say...PBR?) then it isn't adding much, because you would still need to convert stuff to make it work in other parts of your pipeline.
As long as such, among many others things in CG, isn't standardized, I see little use and desirability for such functions.
Frankly I still don't see the fuzz about this?

goldfarb · October 30, 2014, 01:15:11 PM

PBR may not be standardized (but then again nothing in CG is) but EVERY studio (ILM, Weta, DD, MPC etc etc etc) are moving in this direction as are ALL the major rendering tools (PRman, Mantra, v-ray etc etc)...
and there are some pretty standardized BRDFs (Bidirectional Reflectance Distribution Functions) - these define how light interacts with different surfaces.

basically PBR removes all (most/many) the hacks that renderers have had to use to approximate how light interacts with a surface - Phong is a hack, GI is a hack etc etc.

for truly photorealistic imagery you have to replicate how real light interacts with real materials (note I say 'interact' and not 'reflect' or 'bounces' - very different things).

check this out:
http://en.wikipedia.org/wiki/Unbiased_rendering

Tangled-Universe · October 30, 2014, 04:02:03 PM

So if PBR is not standardized then why would the big players move to PBR? What's their advantage? Just so that assets come in 90% OK from another studio?
I see no big advantage if it won't be standardized eventually.

Anything which does not compute the render equation by Kajiya brute force is basically/essentially biased rendering, right?

goldfarb · October 31, 2014, 10:04:27 AM

it's not about transferring assets it's about the efficient (read cheap and easy) way of rendering an object with real world light and real materials - true photrealism.
the difference is the same as rendering human skin before subsurface scattering and after - no one would think to render human skin these days without sss...so soon PDR rendering will become the default method - you'll always have the opportunity to render in the traditional ways, and in some cases this might be better (faster or more suited to the style you're after) but anyone doing VFX and rendering elements that need to integrate with live action footage will need to use PBR.
and when I say PBR isn't 'standardized' that really just means that there isn't one single method/implementation, it's a young technique and is improving all the time...PRman 19 will mark a big change in the way most studios render, Mantra will have an unbiased option from what I hear, and Arnold is getting more popular.
as for Terragen, there isn't any reason to feel that it isn't 'keeping up with the industry' ot that it needs to change anytime soon - but a physically based render out of Terragen would be pretty cool, given all of Terragen's other abilities.

Tangled-Universe · October 31, 2014, 12:44:05 PM

Yes I see your point Michael, but why oh why would 3DCoat need a PBR and what is the fuzz about it?

I remember some guys asking for camera animation and PBR-like rendering for World-Machine. I only agree with such proposals/reqeusts if the software itself is very mature in regard to its functionality and possibilities.
So to get back to 3D Coat, I agree with Upon Infinity then who says it would be more interesting then to have new retopo tech.

Ghehe sorry for derailing this topic, but I just can't wrap my brain around pushing software outside of its own niche/strengths.

(Better standardize surfaces (OpenSubdiv) volumes (OpenVDB) etc. and everybody really use that to THEN create niche applications with highly specialized functionality and allow the standards to allow for easy exchange and production.
The CG industry is so counterproductive and constantly re-inventing the wheel.
Given the huge recent discussion on CGSociety it seems that this won't change anytime soon, because they actually seem to like it.)

goldfarb · October 31, 2014, 12:56:06 PM

ah well now we're really talking about a company adding a trendy feature just to say they have it...
I agree that developers should really stick to fixing bugs and solidifying their feature set, implementing things that are becoming more 'standard' (alembic, openSubD, VDB, bullet etc) than adding things that don't really contribute to the core of their application...
I can see the use of having PBR in 3DCoat if you were using it for look dev - but frankly NO ONE is using it for that - this isn't a criticism of 3DCoat, just a fact of the industry - but it may have been easy for them to implement and may not have taken any developer time away from their main features...