VRay Sun, Sky and Physical Camera video tut + Bonus Script!

[studioDIM]

lol, some rushing after each other :P
At times like these i miss a shoutbox with rooms :P

Lele

[teabag]

bloody hell u know guys how long it took me to understand LWF...and now this....lol

[flipside]

Hi Lele,

I understand that the outdoors should be blown out in an interior shot. But it seems a bit extreme here.

First one is rendered with 0.255:



This one with 0.8



And with 1.0



This is an image without gamma correction or anything else. I tried this because the look of the 0.255 image reminded me of interior images when I didn't use lwf (gamma correction)



Now my question, do you think the first image looks like it would look in real life?

I'm like teabag, first lwf and now this

[aaron-cds]

Does anyone have an interior image rendered with exactly this technique?

Not exactly, I used 1.0 instead of .255, which blew out the exterior and window surrounds too much. I'm waiting for the interior video tutorial. Patiently that is. No pressure Lele.

[teabag]

this is definitely going to be a massive thread....

...you know what u have just started lele?

[studioDIM]

...you know what u have just started lele?

hardly...

Lele

[studioDIM]

Ok a few examples.
I am using the kitchen probe developed as latitude-longitude, its HDR characteristics left intact.

-4 stops, display gamma 2.2


-4 stops, display gamma 1.0 (no gamma)


-1 stop, display gamma 2.2


-4 stops, display gamma 1.0 (no gamma)


-2.5 stops, display gamma 2.2


This should answer your concerns.
If it looks overblown, use fog in the glasses to cut down on direct lighting, compensate with exposure.
If it doesn't , bring it on, i'm all curious to see where this leads...

Lele

[Micha]

@aaron-cds: your image looks phantastic. A photo?
I like the burned out winodws.

Did you try the Reinhard mapping with a lower burn value? I think, if you like to get less burned out windows, than you have three choices without strong fakes:
- burn value lower
- post colormapping with various tools
- simualtion of an analog film gradiation curve (vray framebuffer)

But from my view, the 0.255 method maybe good for exteriors only, because the known side effects. The 0.255 increase the intensity of the background and decrease the GI bounce intensity dramatically.

[aaron-cds]

Thanks Micha, I'll look into using the Reinhard mapping.

[studioDIM]

But from my view, the 0.255 method maybe good for exteriors only, because the known side effects. The 0.255 increase the intensity of the background and decrease the GI bounce intensity dramatically.

From the kitchen images above, i am not sure the .255 goes terribly wrong.
My personal guess is that being able to be descriptive, and faking at will, lead to conditionally overexposed images (ie. darkened skies but less overall contrast).
They have a more powerful descriptive value, as they show the surroundings better, but they do part from the photographic look.
As i said, i'm curious to see experiments done.
For instance, i see nothing wrong at all with AAron's image, which makes me wonder.

Lele

[Micha]

Lele, I don't understand the HDRI example. Could be nice to see a example of the MR_scene from you. You mean, the 0.255 method works for interiors too?

For instance, i see nothing wrong at all with AAron's image, which makes me wonder.

I totaly agree, one of the best images I ever have seen. Only the repeating wall pattern show me, that it is rendered (it's no easy to find).

[aaron-cds]

Only the repeating wall pattern show me, that it is rendered (it's no easy to find).

Son of...

stupid Arroway materials

[Micha]

[flipside]

I also see nothing wrong with the image, but he doesn't use the 0.255 multiplier...

Here are some more test with my appartment scene:

Original photo:


If I take a pic that looks like this, I usually adjust levels, like this:



Rendering with mutliplier=1.0:


Rendering with multiplier=0.255



Exterior scene, lit by domelight with kitchen hdr texture, materials 1.0


And with materials set to 0.255




In the appertment scene, I think that the 0.255 image actually looks more photoreal (don't look at the splotches of the GI). But like with the real photo, I would probably adjust levels in post to brighten up the dark areas.

The exterior one with 0.255 also definatly looks better than the 1.0...

For me it seems that this method has the gamma 2.2 advantages, but keeps the contrast as we were used to before the gamma 2.2 workflow. If a photographer would come and take a pic of this appartment, he would probably also use extra light to brighten up the interior so that there wouldn't be this much contrast in lighting. I guess that's what needed in the 3d scene too. But then you would again come closer to the image with the 1.0 materials...

Not sure which way to go with all this. The fact that you need to put vraycolor maps everywhere is a bit of a hassle (even with the script).

lele what were you talking about with brdf that vray doesn't have? I have no clue what you're talking about here Would there be a way to implement something in the vray material so that the color map isn't needed anymore?

Thanks,

wouter

[studioDIM]

I agree on most counts, Wouter.
First on the fact that what makes me wonder is precisely the fact that Aaron did not use the .255 method.
Second, i tried and tested it a bit on one particular scene which looked good to start with, coming from a production the studioDIM did a while back.
And i have to agree with you, it's a LOT quicker for interiors to adjust the dark tones in post (or through a vfb curve).
The idea of adjusting after the gamma is applied though defies a bit the intentions of the method
I'd say 0.5 has an ideal "grain" for interior shots, all in all.
Bear in mind all the tutorial is about is finding an easier way about, some sort of finer-dented cog to treat such strong light intensities.
.255 worked a treat in outdoors with little or no shadowed areas.
This, however, is an early take on the aforementioned scene, at 0.255.
Notice we're against the strongest possible light, and that the ratio between the walls and sky in the Kitchen hdr was 1:15 (indirectly lit wall : blue of the sky), whereas in this shot it never reached a ratio of 1:7.
So math versus real world lighting modeling tells me we're still a tad above the ideal for diffuse vaues.
Here's the test:


Lele