Maxwell Render | Official public forum

Hello I just rendered out a image and it become like this.

60min on a amd64+ 1gbram



If im happy with it? Nah, not really, its no sharpness and I got no caustics even tho I checked all the parameters for it..

If someone could approach some settings and comments how to setup a studiolighted scene I would be glad.

this is the model






and what is the diffrens between the Engine and Render Engine tab?

Here is the settings used for the image above

http://img214.exs.cx/img214/8152/render62jn.jpg

(rendered with 1.1.23)

your reflection cards are being used with an emitter meterial? what setting is your emitter, also you might wanna uncheck physical sky.

Brett

Yup - physical sky is broken for caustics for alpha 1.1.23. (But worked for 1.1.10.)

Also when placing an emitter, be sure to check emitter's normal is pointing toward the objects to be illuminated.

-tak

Lower also your bounces to around 16. 40 is a big waste for this simple render.

thanks to all, Ive made some changes on the settings.

Lowered the bouces to 16 in both engine and render engine. Turned of the physical light and using a sky dome and emitters with normals towards the objects. But I get lowsy results as well, and no caustics.

The emitters was set to
Type: D65
Intensity: 1000
Units: Power

And I have tryed DOF to 5.6, 20, and 65

something tells me something big is wrong

this is the image



It must be something with the camera?

Regarding your initial post - well, maybe you can verify the caustics from the emitter working by without sky dome first.... intensity of 50 (default value for the dome) look too much compare to the emitter at 1000.

One thing to note is the caustics on the secondary ray (glass caustics casted on the floor, reflected in the chrome cap) may not show up (maybe a bug), but I tried emitter on the perfume bottle causing caustics casted onto the floor (also using sky dome at intensity of 1).

Btw - how are normals of the bottle's inner surface ? both normals should point outward to define the glass boundary to derive the thickness of it, or you will get its refraction all wrong.


-tak

two caustics happening on below image;

1) caustics from the bottle
2) second caustics from the chrome cap which acts like bounced light

Takumi Takahash

About the normals, the inner surface were not facing out, Its fixed now and I try without the sky dome.
Rendering right now, be back later with render-results

thanks

Ill give up:) I cant seam to get it right anyway...

It still not satisfied me :/

The one I liked the best is the bright one, if you just make a little less bright, I think it will be a really good image.

Mihai Iliuta wrote:The one I liked the best is the bright one, if you just make a little less bright, I think it will be a really good image.

Ok now im happy with the rendering





This is result I want, or the materials I would like to recreate. But its hard to find any good settings for harpness and correct color.

if you want to make it as is then set the scene as is:

i suggest you to do the following steps:

1- model just the base, texture it with a marble as seen and give it a bump (use plastic shader here)
2- set the lighting exactly as is and also try to catch the wattage, use a box emitter here and place it in a black diffuse cone to achive a spot.
3- make test spheres for testing further materials, like orange, metal and red glass
3- render 3 test spheres (with orange metal and red glass) at the same scene and be sure if they look same as here
4- finally invite your real model, maybe first just the cap and render it

PS: here there's no need of a stage setup... only a plane and 1 lightsource
with 16 bounces will be enough...

Hey Tom
How Do you gauge the bounces in the scene?
You say 1 light with 16 bounces, ive taken # of bounces to a precentage since 100 was the highest. so 16 is 16% of possibel bounces?
Can you help clear up what the number really means?
and how number of lights vs objects and shaders involved change how many bounces you need
thanks.
luke

Dear Luke,
Think of a ray, it launches from an emitter and hits to a surface,
then bounces from there to hit another surface and this goes on.
So it depends on the complexity of your geometry.
For such scenes (i mean studio setups) with low detailed dielectrics,
I experienced 16 bounces is enough... 32 would be slightly better,
but 50 or 100 would be absoutely useless.
Don't worry, just set it to between 16 and 32 depending on your patience
and the wait for at least 17 samples. That would be cool for sure...