Page 1 of 1
Diamonds
Posted: Sun Dec 10, 2006 2:28 am
by w i l l
I need to render some diamonds. I've found some links on the forum to DXF files but I'm wondering if there is a magic way to model them. I think a lot of the linked files are modelled in Rhino.
I'm using Solidworks... the geometry of diamonds is quite hard to figure out when looking at an on screen image, due to the transparency of them and i'm too poor to buy bread let alone diamonds.
Does anyone have any tips on how to model them and are they regarded as easy or difficult to model just out of interest.
Posted: Sun Dec 10, 2006 2:32 am
by vinys
Maybe this is of help:
http://www.gemcad.com/
Posted: Sun Dec 10, 2006 2:43 am
by ricardo
Hi Will,
There is a plugin for rhino called techgems that delivers many sample cuts, most of what you see should come from there.
Have a look here for more info on the subject:
http://en.wikipedia.org/wiki/Diamond_cut
If you just want a round cut please pm an email adrees and I'l send you and iges file.
Ricardo
Posted: Sun Dec 10, 2006 2:48 am
by w i l l
Thanks very much for those links!
ricardo - I have no idea what I want. I need to model a crown with a few diamonds.... I'm not even sure how the diamonds are fixed to the crown, I'm assuming they're crimped in place. I'll need to look at a few images online.
Has anyone used Maxwell to render an image that has 100's of diamonds. What is likely to happen on a standard PC... 1000000 hours render time?
Posted: Sun Dec 10, 2006 4:18 am
by glypticmax
Hi Will,
A couple of tips.......
Boolean Difference the gems from the setting. Otherwise, the metal model overlapping the gem material could *deaden* it.
As for time, that's a good question. Delete any of the gems not visible and do not affect your scene.
The most I've done at one time is 24 diamonds with two multi-lights and dispersion on. Smallish stones, 4mm or so diameter. Dual core, 4 gigs of RAM. 20+ hours to SL 20. I would stop at SL 16-18 and save a lot of time.
Try IOR cubic carbon (diamond).
Keep your scale real.
Good luck.
Edit: Time mentioned abvoe should have been 11+ hours for the dual core render. Softbox HDR's on two single poly emitters.
Posted: Sun Dec 10, 2006 1:20 pm
by deadalvs
http://www.3dlapidary.com/HTML/Models.htm
these models would just do it concerning diamonds. why model more when these are perfect ?
* * *
i am right now rendering a scene with 8 diamonds, an the mentioned IOR file with dispersion, caustics and one light from the side. (rendering the same image on my four different computers)
the image is still full of noise and colorful dots and i am not sure at which SL this will vanish. i'll render some more and post an image.
* * *
it could be that DOF causes a lot lesser benchmark since it's more complex to render the image. it that true?
* * *
deadalvs
Posted: Sun Dec 10, 2006 1:47 pm
by vinys
Cool, thank you for this link

Posted: Sun Dec 10, 2006 1:57 pm
by deadalvs
vinys wrote:
Cool, thank you for this link

don't thank me, thank Maximus3D in this thread:
http://www.maxwellrender.com/forum/view ... hp?t=19352
* * *
deadalvs
Posted: Sun Dec 10, 2006 2:01 pm
by glypticmax
Its sometimes a good idea to optimize, recalculate normals or in some way clean up the DXF files from 3dlapidary.
I take mine into C4D for a check up before using them in Maxwell.
Posted: Fri Dec 15, 2006 12:16 pm
by deadalvs
well: damned. i messed up my rendering. i saved it after rendering instead of setting a path to render to.
if i had known this i could have posted the merged image of four computers that working 22 hours each...
* * *
well, when i have time, i'll re-render...
* * *
deadalvs
Posted: Fri Dec 15, 2006 12:28 pm
by w i l l
Can't you save a jpg if you don't set the path?.... I thought it was just an mxi that you cant save after.
Posted: Fri Dec 15, 2006 12:31 pm
by deadalvs
the .jpg yes, but not a .mxi for later merging...
see here:
http://www.maxwellrender.com/forum/view ... hp?t=20030
* * *
deadalvs
Posted: Fri Dec 15, 2006 12:33 pm
by w i l l
Edit - sorry I just noticed that you need to merge them.
Posted: Fri Dec 15, 2006 12:50 pm
by deadalvs
w i l l wrote:Edit - sorry I just noticed that you need to merge them.
since maxwell works near-perfect everyone is so decent here ...
no problem...
* * *
deadalvs