Forum: Poser - OFFICIAL


Subject: GC question

meltz opened this issue on Dec 14, 2011 · 16 posts


meltz posted Wed, 14 December 2011 at 7:21 PM

Ok so should i always use GC in poser pro 2012? I did a few renders on something im working on and i liked how the ones without GC came out. Should i re render with GC on and tweek the lighting more or is it really not that big a deal?


kobaltkween posted Wed, 14 December 2011 at 7:54 PM

You should always work to your own goals.  There's no benefit in taking extra time to make results you don't like passable.  I use linear workflow because it's the quickest way for me to get the results I want.  I do optimize my lights and materials for it, but since I tailor my lights and materials to my scenes and only use preset lights for testing, it's not such a big deal.  If you find not using GC gives you the results you want, then stick with it.  TBH, I haven't found the changes in lighting so consistent to find even the experience of changing lighting helpful.  I find it easiest to just build my lighting from scratch when I change a major rendering parameter.

Again, to be honest, I think this is the big problem people have with using GC.  They take a scene entirely built for one workflow then try to apply another.  And lots of people have all this advice for how to change everything, but that takes a huge amount of effort and you probably won't match the vision you built using the other workflow.  If you want to use a new workflow, start with a new empty scene and add your elements one by one, adjusting as you go.  It's much quicker and easier.



hborre posted Wed, 14 December 2011 at 7:56 PM Online Now!

Without any images posted it would be hard to judge.  However, incorporating Gc in your PP2012 render settings will require not only tweaking your lights but re-evaluating how your lights are set up in the scene.  You must realize there are other options light IDL and self light emitting objects which help enhance Gc effects.  If you are rendering your scene set up as you would in previous Poser version, then the scene will look similar in PP2012 without Gc.  But does this scene contain many lights to compensate for the overall darkness and do those lights have light intensities equal to 100%?  If so, applying gamma correction will only wash out those exact unmodified scenes.


bagginsbill posted Wed, 14 December 2011 at 8:25 PM

Don't do anything new. But then why do you have Poser Pro 2012?

You want to use scatter? You want it to look like real skin? Then you want to use GC. 

Or do you want to continue to use old content that doesn't use scatter? Then don't use GC.


Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)


Winterclaw posted Wed, 14 December 2011 at 8:51 PM

Built in GC is one of the main benefits of Poser Pro.  So for a render with realism in mind, I would say yes.  If you want a certain style, then it depends.  The only time that GC will really be a negitive is if you've loaded a character with a shader developed before GC (I'll call those pre-GC shaders from now on) and you haven't corrected the shader out of laziness.  Or you like the overall looks of the old shader for some reason and everything else is GCed so then you'd have to tweak the old shader for anti-GC. 

 

WARK!

Thus Spoketh Winterclaw: a blog about a Winterclaw who speaks from time to time.

 

(using Poser Pro 2014 SR3, on 64 bit Win 7, poser units are inches.)


bagginsbill posted Wed, 14 December 2011 at 10:20 PM

Attached Link: Why do I need gamma?

Read the linked thread.

Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)


meatSim posted Wed, 14 December 2011 at 11:44 PM

Sorry.. this may have been asked and answered many times already as GC isn't new anymore.. but how do I fix an earlier shader? Diffuse no higher than .8, I thought I read reflection shouldn't be higher than .4..  anything else?  I'm not really skilled enough to re-build the entire shader trees myself.  Is there a guide(s) that helps figure out appropriate set ups for different types of clothing materials in the GC/IDL/Scatter era?

Quote - Built in GC is one of the main benefits of Poser Pro.  So for a render with realism in mind, I would say yes.  If you want a certain style, then it depends.  The only time that GC will really be a negitive is if you've loaded a character with a shader developed before GC (I'll call those pre-GC shaders from now on) and you haven't corrected the shader out of laziness.  Or you like the overall looks of the old shader for some reason and everything else is GCed so then you'd have to tweak the old shader for anti-GC. 

 


Miss Nancy posted Thu, 15 December 2011 at 12:26 AM

diffuse + specular + reflection + refraction <=1.

diffuse=.8 is o.k. if everything else is zero IMVHO, but one should note that the diffuse channel is not used in some new shaders.



aRtBee posted Thu, 15 December 2011 at 1:58 AM

@BB

As I started the mentioned thread, but never really finished it, I can do that here.

I started V4 scenes and GC made her look sick. It turned out that this was caused by the cyannic color offset used in her skin textures. After I removed those, enabling GC is the best thing one can do for his images. Not because linear workflows, but simply because a GC image matches the way the human eye and mind deal with dark areas, and because nob-GC images have a very hard time dealing with details in the dark.

In the thread mentioned BB gave an excellent example (the black water surface reflector). With a little experiment you will find out that there is really no way of making a non-GC image of this scene and correcting it in post. Not even when handling HDR-formats.

So, IMHO, render GC should be on in all cases. Material GC should be set to render GC for color maps in order to avoid artifacts, while material GC for value maps should be set to 1. Only color maps created on older Macs might be set to gamma=1.8.
To my experience, if the image does not look good, the cause is in the materials, or in the lighting. Not in render-GC.

- - - - - 

Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.

visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though


meatSim posted Thu, 15 December 2011 at 2:55 AM

Thanks,

I've noticed that about the diffuse not being used as much.  I'm in the process of converting a bunch of my old v4 and v3 stuff for use with Antonia~WM.  A lot of those shaders would have been built before we had GC to play with.  I was wondering if it was just nostalgia that had me thinking they used to look a lot better than they do now.. quite likely the GC is a bigger part of that than I had thought

 

Quote - diffuse + specular + reflection + refraction <=1.

diffuse=.8 is o.k. if everything else is zero IMVHO, but one should note that the diffuse channel is not used in some new shaders.


meatSim posted Thu, 15 December 2011 at 3:05 AM

So does this apply to the alt specular and alt diffuse channels that dont have values attached to them?

Quote - diffuse + specular + reflection + refraction <=1.

diffuse=.8 is o.k. if everything else is zero IMVHO, but one should note that the diffuse channel is not used in some new shaders.


aRtBee posted Thu, 15 December 2011 at 3:40 AM

Quote: quite likely the GC is a bigger part of that than I had thought

Well,

  1. render GC could be switched off to check (then back on again)

  2. (non-PPro?) materials are loaded with "set to render GC" on all imagemaps which is fine except for the value maps. No work required.

  3. there is some default script setting the GC (to 1) for the value maps. One click only (I'm not behind my Poser machine right no, so you've got to take a look yourself. sorry).

  4. render GC is not effecting the handling of materials as such, it mainly effects the results of shadowing (aka: lighting). With flat, omnipresent lighting GC will have no effect at all. So if the image looks different from the original, my quess it's on the lighting. Usually, older scenes have lights etc added to compensate for the lack of GC (and the lack of IDL). Therefore, when enabling those, most lighting have to be adjusted (=reduced).

- - - - - 

Usually I'm wrong. But to be effective and efficient, I don't need to be correct or accurate.

visit www.aRtBeeWeb.nl (works) or Missing Manuals (tutorials & reviews) - both need an update though


Miss Nancy posted Thu, 15 December 2011 at 5:45 PM

Attached Link: http://www.runtimedna.com/forum/showthread.php?63356-SSS-Tester&highlight=jade+shader

> Quote - So does this apply to the alt specular and alt diffuse channels that dont have values attached to them?

in attached link was posted bill's attached shader with diffuse and specular channels = 0.  whilst jade and other amorphous silicates can exhibit refraction, it's difficult to measure jade's refractive index AFAIK.  however, renders of objects using this shader don't emit light nor do they seem to glow or bounce off more diffuse than they receive. that may imply the shader has diffuse + specular + transparent + reflect + refract <=1.  I don't know how to show using math that this shader obeys that. 

it may be that alt_diff + alt_spec <=1 in that the fresnel blend node somehow fixes it that way.  e.g. if one dropped the colours of alt_diff and alt_spec to gray, it would look way too dark, but if one altered the hsv node or tried to add a refract node, it might cause the object to exhibit diffuse independent of lighting.



bagginsbill posted Thu, 15 December 2011 at 6:21 PM

Index of refraction is not measured by measuring refraction. It is measured by measuring reflection.

The IOR affects reflection and refraction. Refraction, however, can be suppressed by the material, particularly if the IOR is complex. (The imaginary part is nonzero in electrical conductors. It is zero in dialectrics, by definition. A dialectric = a material with a real IOR and no imaginary part.)

So - if you measure reflections you can solve for IOR. And this is certainly possible for any material.


Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)


bagginsbill posted Thu, 15 December 2011 at 6:26 PM

The principle whereby refract + diffuse + reflect <= 1 is called Conservation of Energy, or CoE. You can't have more energy coming from an object than the energy that reached the object, unless the object is emitting energy itself instead of just reflecting or refracting it.

In my shaders, CoE is always obeyed. Not so in most Poser shaders.

However, the "value" property of Poser specular nodes turns out not be normalized in the way we understand applies to Diffuse_Value and Reflection_Value. Therefore, you will see shaders from me where the sum appears to be greater than 1. If you normalize all the numbers, then they do not exceed 1, but the denormalized numbers don't have to (and can't) obey that rule. It's a non-rule actually, as you will see me break it all the time. Look in my latest shaders and see Specular_Value = 6 or more.

Meanwhile, the Fresnel_Blend, as seen above in my Jade shader, is blending between two normalized inputs and the blend is never greater than 1. The reason should be obvious as I've explained the general "Blend" function a billion times.

Blend(a, b, f) = (1-f)a + fb

If the quantities a, b, and f are all between 0 and 1 inclusive, then this equation is also always between 0 and 1.


Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)


bagginsbill posted Thu, 15 December 2011 at 6:29 PM

Suppose you measure the reflection value, at normal incidence, of a dialectric surface. Let's call the value x.

Then the index of refraction (IOR) of this material is easily calculated from:

(1 + sqrt(x)) / (1 - sqrt(x))

This is easy to derive from the inverse of the Fresnel equation with theta = 0.

There is no need to measure the angle of refraction or amount of refraction. The reflection value, refraction value, and refraction angle are all driven by the same equations. Solve for one, and you've solved for all.


Renderosity forum reply notifications are wonky. If I read a follow-up in a thread, but I don't myself reply, then notifications no longer happen AT ALL on that thread. So if I seem to be ignoring a question, that's why. (Updated September 23, 2019)