Thick Glass

[simmsimaging]

For those edges you can just make them a different material id and use the VrayMultisubobject Map and make the edges more glossy, that will give the right general effect (or bump map it - either way).

I agree that getting too hung up on 'physically correct' can be a hassle, but the more cheats and hacks you add to a scene the less intuitive it becomes to light and work with it, and the less your assets become useful for other scenes. I use plenty of cheats too, but I do try to keep things "realistic" simply because it tends to make everything simpler and less error prone.

YMMV.

b

[glorybound]

The science behind render engines, like V-Ray, are to be physically correct. We don't have to use the tool, the way it was made to be used, but it does make life easier if you are going for photo-real.

[Neilg]

it does make life easier if you are going for photo-real.

I really don't agree with this. Vray is built around cheats and biased methods to get quick renders. I can guarantee that a half day in vray setting up shortcuts and half a day in photoshop gets a better result that 3 days in vray meticulously following the laws of nature. Will render 5x faster too.

[mario74]

and half a day in photoshop gets a better result that 3 days in vray meticulously following the laws of nature.

i even dont agree with this...

the "perfect thing" out of any render-engine will never be able to compare with photoshop, especially (EXACTLY stills) interior / exterior scenes!

FYI: http://www.youtube.com/watch?v=-6zLayk-W4E

that kind of art will take you a hell of setup and rendertime also... but only 30 minutes in ps!

physicaly based - no, take a workshop for photographers

[glorybound]

VRay Sun/Sky, Vray Camera, material IOR... If this isn't physical based rendering, I don't know what is.

[trixian]

I prefer to think of this as mathematically sound as opposed to physically correct. Its all about getting all the numbers to add up inside the confines of the "physically correct" mathematical model.
But having specular/reflection separate from diffuse, and caustics, GI separate from direct lights is in no way Physically correct in regards to our physical world. They are approximations. But they all are designed to play along in a way that does not break the given rules for this specific (VRay) approximation.

[samuel_bubat]

Sorry...
Wall of text incoming


I ´m not sure if anything expect IOR and physical correct light distribution is physically correct at all. Maybe Vlado can give some answers.
Just some things I thought about recently. A Vray physical sun at noon has a multiplier of ~70. So now imagine you would have a monitor
that can display values as bright as the real sun. In this case a physically correct sun would have a muliplier of 1. So all these values
in my opinion highly depend on the output device wich varies a lot. Be it a monitor a printer or whatever. the next point are the textures.
At wich lighting condition using wich camera at wich exposure settings do you photograph a texture to have a "physical correct" material later in your scene ?
(Remember those posts.. "I made a photo of my garden, recreated it in 3D with correct camera settings but it looks completely different)
I have not found an answer, yet and I think... well they are as correct as they are incorrect and there is no real answer to this question as again
it would highly depend on the output device just like in my first example. I´m thinking a lot about this and for my part I came to the conclusion
skipping all these "handcuffs" makes a 3d artists live a lot easier. In detail I mean I do not use a physical sun, no physical camera, no gamma 2.2.

Also a common misunderstood term is "photorealism". Photorealism and realism can be two completely different Things.
What your eye can see, what a device can output and what a camera captures usually varies a lot.
You can make images that look damn realistic but are not photorealistic at all. A physical camera is an extremely limited device.
Well that´s just the nature of an analoge Device, i guess. A Vray physical camera simulates these limitations.
I found a way to visualize the color curve within a scene. And a Vray physical camera does the following.
Imagine a vertical line the bottom is black and the top is like the maximum amount of brightness before your eyes explode in a linear fashion. A Physical
camera would cut out a part of this line wich is the lighting you´ll get in your rendering. everything above or below is just beeing cut (clamped).
If you change the exposure the position of said cut out part will shift along the line. Now, because light has an inverse square falloff while a physical camera
is linear you´ll always run into the well know and often discussed problem that you´ll have overbright regions around the windows and to dark areas in the back
of a room. Some people thought they found the holy grail around this by using gamma 2.2. Well actually it´s correct a monitor usually doesn´t have a linear
color curve. So indeed it helps a little but in more extreme situations it´s simply not enough.

So instead of a physical camera i use a standard camera (or sometimes a vraycam without exposure). This preserves the whole intensity range of our
imaginary line when rendering, resulting in an overbright image. Than I use Reinhard color Mapping and have completely control of what to do
with this whole range. I can squeeze it, I can make custom curves depending on the architecture of a room (Big or small windows, floor, wall ceiling materials
are the biggest factors) or the wishes of my customer and I can control all of this with just the 3 parameters of reinhard color mapping.
It makes gamma 2.2 obsolete and you can theoretically light a factory hall with a light through a tiny window without any burned out or to dark areas.
Of course you can do this while also having gamma 2.2 active. But for my part I don´t like to turn the same knob in two different places if it ´s not neccessary.

Why I don´t use a physical Sun/Sky. Well first it´s the perfect solution if you have to animate a sunset for example. I wouldn´t take anything else.
But for archviz , just like a physical camera, it´s a limitation. A physical sun is nothing but a direct light as said before with a multiplier of ~70 at noon
that changes it ´s color when you change the azimuth. Well why not replace it with a standard direct light with a multiplier of 70 ? You preserve full control
of the lighting. For example a customer sees a preview rendering and would like to have a little bit of direct light on a chair that is more in the back of a room
If you have a physical sun the lighting color will change dramatically than you´ll have to white balance it again
wich also means you´ll change the colors of your textures wich in the worst case have to be corrected as well. With a standard light you just change the light
position and that´s it. Who cares if a sun at a certain position would maybe give a more yellowish light in reallity ?
Same counts for the environment. I usually use a simple gradient instead of a physical sky. I use the RGB Multiplier as intensity control. So I can precisely control
the colors as well as the intensity compared to the Direct Light. Usually my direct has a value of 50 and the env. gradient and rgb level of 25. But if I want less
or more contrast (speaking of direct vs. env lighting) I increase the env multiplier or turn it down, just as needed. So you have complete control about every aspect of your lighting. It´s fun, try it ! Also this makes any use of white balance obsolete. I mean imagine you want a clean white lighting. With a physical
sun you´ll first add a yellow light and a blue environment just to color correct it with whitebalance. Just skip those steps and make a white light with a white
environment, problem solved.

I ´m preaching this for years . Actually I shouldn´t really care how others light their scenes. But sometimes it makes my blood boil when I see
how complicated things get for some people only because they are trying to stay "physically correct".. because dozens of terrible tutorials out their
tell you this is the only way.
And I blame this all on maxwell It all really started with this horrible renderings of photo flashlight images that no customer ever on this planet
would accept as an architectural rendering But from that point onwards everyone wanted a physical camera and a physical sun and what not
and thought this is the only way to get a realistic image.
Oh and now that I´m allready ranting, let me tell you how terrible I find all these bidirectional, unified sampling etc. nonsense. I´m so glad Vray offers
such a wide range of different algorithms for different phenomena. It makes life so much easier (even if it requires some knowledge).

Oh and one last rant for today. My absolute favourite... inverse .. is.... using a physical sun in conjunction with a HDR for an interior lighting.
This is wrong on so many levels that I don´t even know where to start... but i think it´s enough for today.. maybe another time

cheers !

[Neilg]

using a physical sun in conjunction with a HDR for an interior lighting.
This is wrong on so many levels that I don´t even know where to start... but i think it´s enough for today.. maybe another time

this is how I light morning/late evening shots. How can something that looks good be called wrong?

[voltron7]

interesting debate from bobby asking for a way to make his glass have green edges...interesting to hear different opinions on the matter.

While not feeling particularly argumentative about this stuff...(trying to remain neutral, lol.)
I will just say that for still image production in arch-viz, not a single client I have had, so far, could give a rat's ass about physically correct, or proper, or how the scene is lit.
Bottom line is schedule/deadlines and end product that looks good, not necessarily correct.

Colors right? Too bright? Too dark? Too grainy? Larger resolution? Camera view okay?
Those are what seem most important for 100% of my jobs.

I cringe to even try to discuss these technical ideas with them, when they are usually still struggling to
save attachments to their hard drive, or to use Winzip/Winrar.

Still, this has been a good thread to read, so thanks to folks for putting in their 2 cents.

[trixian]

Well, to be honest, there is one rather interesting aspect of having the "physical camera" and sun\sky models with appropriate intensities (and ies lights to a degree), and that is for periodical sanity checks and "calibration".
With out this virtual base point in our knowledge, materials would be a hell of a lot more finicky to create, and would behave very differently under differing conditions (like certain oldschool game shaders that only work in a specific lighting setup).