Lathing a transparent solid

Luke,
It's one of the newer 3Ghz dual-Quad core (8-cores) with 16GB RAM (although since C3D is only 32-Bit, it doesn't SEE all that RAM).

I have all my settings very high for HighResoultion print-quality renderings. In addition to radiosity, I have 3 (three) area lights with roughly 80-100 samples each.
See screenshots.

------

My point wasn't that it takes longer for Transparency... That is OBVIOUS that it needs to calculate more. My comment was how MUCH longer it is. In LightWave (I know, I know... this ISN'T LightWave)... BUT... in THAT application with high settings... turning transparency on only changes the render time by a small percentage of the total time. In C3D, it multiplies the render time by 3x, 4x... on some 5x.

That's all.

Not ALL my renders need to be that High-res and high-quality.

-------

So... you're saying if I "FLIP" the normals... (and the surface looks BLACK in the layout window) that it will render as I am looking for...? I will have to try that when I get back to work. Thanks. I ignored other people's suggestion because I thought that THEY thought I had lathed incorrectly and that I had my normals "wrong"... BUT, you're saying I need to INTENTIONALLY make them wrong for it to render as I am seeking. Am I correct...?

CORRECTION:

I basically put in the values here on my home machine to take a screenshot, but I forgot to lower the "Threshold" for the camera settings: On the scene I was using at work, the camera settings were:

edge
8x8
8x8
0.020 (much lower)
...and the rest is close to identical.

Dear Scott

I'm saying that you flip the inside ( like the black area in the wire herewith)

So only half the model needs to be flipped.

Also looking at your settings ) I'm not the expert here) but do look overkill for what you are trying to render.

Like the x8 min sample for camera and the 80 samples on the area light seems a lot to me, for a render that size I get away with x1 min and about 32 samples on an area light...but see what the others here think.

Regards

Luke

Well... I'll be honest, it was simplified evolution that got me that far.

I did all my test lighting and settings at small resolutions, like 800x600 and the renders were very fast & acceptable... for me to bump the lights from 15, to 30, to 100 samples had VERY little render-time effect at small resolutions, so I just kept experimenting and got a "look" that I liked.

I really haven't experimented with other types of lights and/or Ambient Occlusion for radiosity.

Maybe that's my next goal... to back-off all those settings and bring them up a little bit at a time. You're right, it's probably overkill. I just didn't have time to re-think my "scene" that I "perfected" at small scale.

Thanks - and thanks for the tip about flipping the interior normals. Will try it.

Just a side note ´bout

: a thing that might gonna get out of focus:



((by the way: the grey square in the upper left corner is the same color as at the lower right corner))

Dear Scott

doing some more quick tests.

I quickly lathed a bottle based on your sample pic added one area light and a .hdr ran the pic at 3600 x 2400 at the settings I mentioned on my previous email and the render took 260 secs on a quad 2.6 Intel mac which means, you should get a result in less than half the time with all the detail.

Martin's default settings are always very good and all the info for tweaking is very clear on the help files.

All that needs to be done is mess around with the liquid, glass and flip normal settings until you get the kind of feel you want.

Hope this helps and look forward to seeing the next renders


Regards
Luke

Frank - Thank you

Luke - Did you use some sort of reflection map...? I'd love to see the actual file to see your light(s) settings, position(s), etc.

Admittedly, lighting is my weak-spot. I love the reflections you have in the glass in your scene. Looks very "natural".

-sf

Dear Scott

File to big to post.

So. lighting as per setup herewith, simple backdrop, 1 area light 16 or 32 samples, raytrace+trans on. AO radiosity and .hdr from the examples file in cheetah will do (use the old town one).

Bottle outer in brown glass inner skin (blue) use an air material i.e. totally transparent and 0 refraction. and then the liquid (middle mesh) like glass but set to 1.33 refraction an no specular.

Then it's a matter of playing around with the area light position and camera angle.

camera x1 min and x4max no camera light and that's basically it.

I hope this makes sense and helps

Regards

Luke

ps there is this good tutorial which I have taken bits of and adapted for cheetah which might also help: http://www.freewebs.com/adriantiba2/glass.htm

Putting a center that's explicitly made of "air" is a nice hack which works around the limitations of the material system. Unfortunately, it won't work for open-topped vessels.

Martin,
Will there be volumetric rendering of transparent objects in v5...?

I have yet to achieve good results on my own using the "tricks" here.