[mverta]

I rendered the same scene at 3 scales: 100% (Real-world scale), 1000%, and 1/1000th scale.

Other than the DOF, I don't see any difference. Even though I'm using Sun+Sky, shouldn't the bounce light energies be different, given that the reflected rays have much longer or shorter distances to travel?



I guess what I'm asking is: shouldn't the scale of this object be inherently obvious just by the different ways the light energies play on the surface, regardless of DOF effects? This scene was shot with identical lighting, and camera. Only the object's scale has been affected, which of course required moving the camera back or closer.

_Mike

[newmain]

the only thing must to know
that the small object will affect by light more then the big object in scale
intensity will be more in small objects

[mverta]

the only thing must to know
that the small object will affect by light more then the big object in scale
intensity will be more in small objects

But it isn't in these tests, that's precisely my point.

_Mike

[andronikos916]

I muself did a test like that with just a cube and a floor. I recommend to do these tests with a zoom camera view...

Even at starwars IV, V, VI (the old ones) the used scale models and look super realistic in some cases...especialy in exteriol dogfight scenes (like your example)

Try by placing your camera inside your model or very close to it.

...waiting for your results.
Andronikos

P.S. nice looking model by the way

[mverta]

Well that's what I mean... obviously, if I used an emitter, the scale would be immediately obvious, but I was interested in what I would get with the Sun+Sky, because my instinct tells me the shadow and bounce light qualities alone should tell a lot about the size of an object. But in this test, the size seems to have no effect on the quality of the shadows or indirect lighting... and the scale differences here are a lot greater than an ant to an elephant. More like an ant to the Empire State Building.

_Mike

P.S. What I really want to do is a Maxwell version of my R2-D2 (below, with mental ray), but the shaders we have aren't sophisticated enough yet.

[jeffg]

I've been working along the same lines and can't see much of a difference in my test either.

I placed a shoebox sideways on the ground and rendered. Then scales it up to the size of an amphitheater and rendered - no difference.

No difference in the length that the bonce light traveled inside the box.

Maybe someone can shed some light on this!

(bad pun)

[jeffg]

...ecause my instinct tells me the shadow and bounce light qualities alone should tell a lot about the size of an object. But in this test, the size seems to have no effect on the quality of the shadows or indirect lighting... and the scale differences here are a lot greater than an ant to an elephant. More like an ant to the Empire State Building.

_Mike

Have you thought of working backwards? Start with photo of a 747 or perhaps an aircraft carrier an grab a 3d model similar to either and work with Maxwell until it is giving you a good simulation of your photo. Then you will have your light/scale environment in which you can place your Corvette.

In other words, if you can trick your eye with something common - like a 747 and the Corvette still doesn't simulate well for you then the problem lies elswhere.

Just a thought.

[Micha]

Hmm, in the test the size of the environment stay the same (it is a Maxwell internal)? So, you compare a white air plane with a white "fly" in the same environment. I think, the same brightness is ok and if you use an emitter, than you will get different brighness.

[mverta]

I think you guys are missing my point... the ship here is not important; it could easily be a 747 (they're about the same size, and this ship is scaled correctly as it would be if it were real).

What's important is that I rendered this ship at 15,000 cm (real-world size), 15 cm, and 1,500,000 cm. The environment was all still Sun+Sky, only the object size changed. AND NONE OF THE RENDERS LOOK DIFFERENT - except for the DOF effects of the smallest one.

My question is: Shouldn't objects that are so vastly different in size have light affect them differently? Especially in the areas of bounce/indirect light? Because here, no matter whether the ship was 15cm long or 15,000 cm long, it looks identical.

I'm not worried about the look of this ship or the renders, I'm trying to understand why the light energy doesn't seem to be affected by the size of the object. This may be physically accurate, by my instincts say no. I'd like the NL guys to weigh in on this, if possible.

_Mike

[chrisvconley]

I would guess that the reason you don't see a difference is because sky+sun is "infinitely far away." You can't scale the object "relative" to the sky+sun -- it will always be infinitely far away...

Chris

[Mihai]

I think they are correct, in bright sunlight if you take a picture of a building and then a picture of an ant, the shadow intensity will be the same, right? Besides if you look at the 0.001% one, you will see indirect light doesn't look the same as the other ones, it looks more localized and harsh.

About that R2D2 model, is the grime part of the color map? If so I don't see why you can't render it in Maxwell, just use different plastic mats with different roughness settings. If the dirt isn't part of the color map just bake it in first with mentalray.

[mverta]

It's not the shadow intensity I'm wondering about, as much as it is the indirect/bounced light. If a ray hits a surface with X energy, then the further away the indirect surface is, the less light it will receive. A white piece of paper will cast a lot of bounce light on a surface 2 inches away, but not 2 miles away.

Similarly, I would think that the bounced light would be much weaker on a huge object, because the distance between surfaces is much greater, so the light energy would fall off more. But on a little 15 cm model, the bounced rays would still be strong when they hit the indirect surfaces.

The sunlight's intensity is the same, yes. The shadow quality may be the same, yes. But shouldn't the bounce light be showing greater falloff over greater distances?

_Mike

P.S., the grime on R2 would be relatively simple, but Maxwell can't even begin to touch the automotive paint shader I wrote for R2's blue, which is an extremely complex, multi-layered shader with metallic sparkles, and a realistic "flash" color change in the basecoat, as well as a separate realistic "flash" in the topcoat, and clearcoat. The shaders use lookups of the camera normal to compute a fresnel-based color change for the paint as its angle to camera increases (this is what happens with this paint in real life) and there's no way to do this is Maxwell yet. Maxwell can't do the necessary anisotropic reflections necessary to do the aluminum correctly, either.

[Mihai]

Have you tried putting in on a floor and test? Because now you are getting the same illumination from all sides, so that bounced light, might actually not be bounced light....

[mverta]

I considered doing that.. but shouldn't the adjacent surface bounces at least be affected? If you look at certain parts of the image, you can clearly see the bounce light being caused by adjacent surfaces, not just the "everywhere" light of the Sun+Sky. It's a good question, though...

_Mike

[Micha]

I see on effect - 1000% is the brightest image and 0.001% the darkest image, maybe, this is the effect you are searching.
I would try to render the scene with the space ship on a ground. It could be that the effect is better to see.