PDA

View Full Version : Physically Plausible Shading - what is it?


DreaminSoInsane
10-02-2012, 04:03 PM
Physically Plausible Shading seems to be the IT words at the moment. But I cant find much information on what exactly it is other than the obvious details the name suggests.

How does it work? and what are the pros and cons?

LucentDreams
10-02-2012, 05:29 PM
A good readon the subject http://diglib.eg.org/EG/CGF/Volume13/Issue2/v13i2pp109-120_abstract.pdf

ShawnDriscoll
10-03-2012, 01:05 AM
Physically Plausible Shading seems to be the IT words at the moment. But I cant find much information on what exactly it is other than the obvious details the name suggests.

Who is using those words other than Renderman developers?

LucentDreams
10-03-2012, 01:50 AM
Well technically it was robert Lewis' Paper that I posted the first page form that coined this term all the way back in 1993. It's been used in many papers since, but yes in terms of recent popularity among non paper and technical discussions any TD or artist I know talking about it is usually in reference to Pixar's specifically, so not sure where it has suddenly become the IT word. perhaps a school using PrMan and all the students are raving about it without knowing what it actually is. See that all the time.

DreaminSoInsane
10-03-2012, 12:07 PM
Ok, so maybe people are using the lingo wrong.

I read the linke above and an interview in fxguide with Pixar on its Renderman Pro Server

fxg: Could you explain what physical plausible shading is?

Ford: Physically plausible shading is a term originally coined by Robert Lewis in a paper entitled “Making shaders more physically plausible” from the Fourth Eurographics Workshop on Rendering, Jun. 1993, pp. 47-62. The basic idea is that shader developers need to pay more attention to the physical laws of energy conservation from which production benefits can be derived analogous to those produced by “linear workflows”. RPS 16.0 provides a set of tools for shader developers who wish to produce physically plausible shaders that run with efficiency. This is where the new integrators and a standard radiosity sample enter the picture. We have implemented MIS (a standard optimization for physically based renderers) so that shader developers do not need to. We have also provided a basic (and growing) collection of building blocks in the form of RSL shaders.

is Physically plausible shading an actual workflow capable of being implemented like linear workflow or is it more of an idea which is just an aim for shader developers?

Surely physically plausible shading eliminates control of spec which although might not be actually exist in the physical real world but in the demands of the film industry is very useful given most directors don't care what is physically real or not they just want it to look cool!

These may be stupid questions but I just see shader and lighters TD talking about how physically plausible shading is the only way to go

playmesumch00ns
10-03-2012, 01:14 PM
Physically plausible shading basically just means using and enforcing principles from physically accurate rendering in production pipelines. This can range from quite complex light transport solutions in shaders to just using area lights and BRDFs that enforce energy conservation.

The pros are that you get more believable lighting and shading. The only cons are that you generally need to rely more heavily on raytracing to get the most out of it, which depending on your scenes and the spec of your render farm, may or may not be an issue. Lighting artists also need a bit of re-education to get the most out of the pipelines.

It's become a buzz-word recently because most of the big facilities started upgrading their lighting and shading pipelines a few years ago to work in a more physically correct manner and this work is really coming to fruition now. This upgrade was mainly due to an increased awareness of what physically correct shading is, as well as raytracing in PRMan finally getting to the point where it was practical to use it in production.

The first time it really became known about, to my knowledge, was around the time of the first Iron Man movie and ILM's subsequent promotion of their work on that. Nowadays it seems that everybody and their mother is pushing their physically plausible shading pipelines in conference sessions.

Moreover, Pixar themselves have spent a lot of time implementing a physically plausible shading library that ships with PRMan, as well as continually optimising the raytracing functionality that's needed to support it. Most people imagine this is because of the long-overdue emergence of a renderer that looks like it might actually challenge PRMan's dominance in the high-end market (Arnold). Whatever the reason, it's a good thing for everybody involved in the business of making photoreal VFX.

DreaminSoInsane
10-03-2012, 03:14 PM
ahh thank you. Thats a much clearer explanation. So basically there are multiple ways of implementing this workflow and basically it is to do with recreating the energy conservation of light in order to achieve more realistic fall off.

If you are using more physically plausible shaders does this not limit the artist with how much they can adjust parameters to achieve certain looks requested? People talk that this should be the only practice, which I initially felt is a bit naive since you wouldn't tell anyone to limit themselves only to one renderer but worried that maybe I just didn't understand the wonders of it, seeing as shader writers in particular seem very very passionate about it.

playmesumch00ns
10-03-2012, 05:17 PM
ahh thank you. Thats a much clearer explanation. So basically there are multiple ways of implementing this workflow and basically it is to do with recreating the energy conservation of light in order to achieve more realistic fall off.


To be done properly you should be doing all those things, i.e.
- Using a well-tested and robust monte-carlo integration including multiple importance sampling for efficiency
- Using only area lights (with correct falloff), not traditional "delta" lights
- Using energy-conserving BRDFs and taking care to enforce energy conservation when artists are building shaders

How much any of those things are done in any particular company, show or shot depends on the circumstances...


If you are using more physically plausible shaders does this not limit the artist with how much they can adjust parameters to achieve certain looks requested? People talk that this should be the only practice, which I initially felt is a bit naive since you wouldn't tell anyone to limit themselves only to one renderer but worried that maybe I just didn't understand the wonders of it, seeing as shader writers in particular seem very very passionate about it.

It's not limiting at all really. Depends on the implementation, but normally it's perfectly possible for artists to put insane, or broken values into shaders and lights. All that's happening is the mathematical foundation for it all is a bit more rigorous than it used to be.

In an ideal world, no artist would ever put in a silly value, or all those sorts of parameters would be locked down to "plausible" values. But production doesn't work like that :)

CGTalk Moderation
10-03-2012, 05:17 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.