Lighting Challenge #1: Fruit Bowl


Even though it’s a bit late I’ll give this a spin, there are some great looking renders here.


Here’s my version. I tried going for a flat paint look hence the absence of rim lights. I did some photoshop work but it’s all layers and colour grading. Thanks.


this is the sample image of what i 've done in the free time after work. hope u like it. there r lots to be done with the texture and lighting as well. . fine tuning the gi is still at work.
critics r most welcome.


thats a amazing use of light. . great image


Finally yestarday i got some time off to update the droplets. tell me what u think . i think they look a bit exagerated.


Better. Maybe the side of the pear shouldn’t be so bright, and if you lit it more like the other fruit then more subtle glistening drops could be have the brightest highlights in some of those areas.



yeah i guess that willl make more sense.


Well, here’s my go, a little late. Rendered in Cinema 4D 8.1.


This might not be in the spirit of the challenge, I decided to go x-ray.
No AO or IBL
No image-based textures added
Procedural shaders only
No postwork

This is/was a great challenge
Admire all the sharp minds at work
::::: Opera :::::


hi all . . .
this is the most recent image that i ve done. . i just thought about freezing it right here.
rendered with mental ray and post in photoshop. Render time about 5 minuts with gi on.
hope u like it


Whoah, here I go away for a few days and what do I miss? Jeremy using my shader and the lighter for Cars rendering in mr!? Okay, time to buy a lottery ticket or something! :wink:

Very nice image. The “metallic flakes” shader work nicely in the bowl!

To your question, albeit to late I guess:

I don’t quite understand what you mean by “focus” paramter… the car paint shader is actually a Phenomenon™, i.e. a mental ray collection of other component shaders that presents a set of subshaders as a new node and exposing a (sub)set of the parameters to the users.

You can see what/how this works by looking in the .mi declaration file for the paint shaders.

Reflections are handled by the mi_glossy_reflection shader, which has a whole bunch of parameters in itself, but only a subset is exposed in the car paint phenomena. The mi_glossy_reflection has parameters for distance-falloff and other groovy things to help keep render times in check and the image clean (i.e. the single sampling of environment, the shader assuming you have a pre-blurred environment map, to avoid noise in the environment reflections).

Alas, I ramble.

Very nice image, Jeremy!



Thanks Master Zap. I think I’ll leave the carpaint on the bowl just like that - although I have picked out a fabric for the background in the next test. I haven’t decided if I’ll also use your SSS on the grapes, I was going to but when frontally lit they look pretty good without it.

By offset I meant the issue of what gets reflected; does it reflect objects closer to you or farther away? Sortof like the index of refraction does for refraction, you need a slider for what objects appear at what location within a reflection too sometimes, right? Like where 1 represents a “physically correct” reflection, at 2 a sphere would reflect the environment around it including what’s all the way behind it, less than 1 would mean the sphere reflects an area more focused around the camera? I might be using the wrong word but this must be a concept that’s around somehow, right?

My first test didn’t include a reflection on the bowl of the banana, because I didn’t like how it got broken up into two parts. I’ll put a banana reflection in the next test but probably from a shorter banana. I wish I could adjust the offset on the plate’s reflection to fix a problem like this, but a separate reflection object will have to do for now.



Yeah I think a fabric may be more suitable there.

Well, as they say “try it you’ll like it” :wink:

Ah, gotcha, no, there is no support for that. Interesting thought, I never thought of that. We tend to think more “physically correct” over here (although I’m the resident “phenomenologist” that puts my finger to the wind and goes, “naah, this should be divided by… … .pi … squared… yeah, that’ll do”)

Actually, a normal sphere does that (except the very area covered by the sphere). It’s a common misconception that a sphere reflectson only the hemisphere facing the camera.

I keep thinking you could wire something like that up with the existing base shaders as a phenomena, but it may be tricky, actually… it’s of course trivial to code such a shader.

But I keep also thinking it’ll look wierd and distorted, at least on any flat-ish surface where the eye is able to “intuit” the reflection. (But yes, for curved surfaces our intuition for reflection go totally out the window and it could lend some extra artistic control… cool idea… but no, not in the shader today, m’fraid)

Actually, I found that kinda cool in my tests… also how the cherries get kinda streched in the reflection.

I’ll keep your idea in mind, though, for future shaders.


“pinch me” I must be dreaming :wink:


I haven’t had much time to play with this in recent weeks, but made some time the last few days. It’s not complete or completely satisfying, but I have made some progress.

All procedural textures (except for the bump on the leaf) and a simple key/fill/rim light rig with ambient occlusion.

Thanks for looking.


Here’s another test image. I smoothed out some of the geometry and put a little fog in there, too. I still have to do something to the end of the banana to make that more organic, maybe a transparency map to make a rough torn-off type edge?

Master Zap - I’ll see if I can do a SSS pass and see how that looks.

Regarding things being “Physically Correct,” that’s a great default for the first day’s dailies on most things. But, when the director asks you to change the perspective of a shadow or reflection, if you tried telling him that you don’t want to make his change because the computer’s output was physically correct on your first version, then you might not work too many days in the industry. :wink:



I´ve seen alot of renders these days of some Spheres with a fancy IOR material: " The render is small, grainy and I had to stop it after 90 hours, but hey!. It´s physically correct!" :wink:

Great render Jeremy. Some of the fruits are a bit glossy for my liking, but the light, texturing and cloth are spot on.
To to me it has this “overcast sky/daylight light coming in the window” feeling.
If that´s what you tried to do, I think you nailed it.

You could get a job in the industy some day… :wink:


Looking good! Maybe there’s a bit too much bump mapping on the apple, an apple is such a smooth surface.

Bump maps and diffuse light don’t really mix anyway. I use bump maps to break up highlights and reflections, but on diffuse shading bump mapping usually looks so fake that I’d do anything else before using it on my diffuse.


PS - I see lots of people have been visiting this thread, hope I get some feedback too…


Which it might not even then be…even the ones who claim physical correctness often still render in a vacuum without simulating polarized light, fluorescence, dispersion in the camera lens. I don’t think there is any renderer that is completely correct, it’s all just an approximation that comes more or less close.


HEllo everyone, Wow lately so many good stuff, i have some update on the pear and well im not really sure about it. anywyas Jeremy i think u mean this.

About your image is great im not reallly sure if u told us ur mood u were looking for , to me looks great somehow i loook at it alot of time. i like your those fruits behind the apple . but im not sure about the color of the orange grapes. i guees ur mood is like old fruits. anywyas have a good one.

i didnt get how to put an image so i can see here.

anyone know if theres another name for the sss subsurface scatering in MAYA 7.0

#300 and, but I’m sure there’s others around too. Correct me if I’m wrong though.

Anyhow, I guess it’s more about physical simulation than beauty, although this terms might not exclude each other.

Grüße :wink: