Glass questions again

Announcements, requests and support regarding the Blender Indigo export script
scatman
Posts: 63
Joined: Thu Jun 11, 2009 12:12 pm
Location: North Carolina, USA

Glass questions again

Post by scatman » Wed Mar 30, 2011 6:42 am

Hello, all.
Please compare the renders below:
Image from Blender

Image from Indigo

How can I get the black areas out of the Indigo Render? Note: I've already tried lowering the Gain value in Blendigo, but it didn't fix the problem.

Any suggestions?

Thanks in advance.

User avatar
Meelis
Posts: 383
Joined: Sat Mar 14, 2009 11:29 pm
Location: Estonia

Re: Glass questions again

Post by Meelis » Wed Mar 30, 2011 7:15 am

Hi :)

Intersecting glass mediums need to have different % value - glass that's inside another glass needs to have % 9 or smaller than 10% if that glass in what it is has 10%

<scenedata>
<medium>
<name>GlassABC</name>
<precedence>9</precedence>
<basic>
<ior>1.50137</ior>
<cauchy_b_coeff>0.00809424</cauchy_b_coeff>
<absorption_coefficient_spectrum>
-----

All normals out (select all and ctrl + n) in blender edit mode.
No holes in solid mesh, all geometry has thickness (no single flat plane with 0 thickness)
----

Also u need HDR map for lighting if u don't want glass to be flat and boring.
http://www.openfootage.net/?cat=15

scatman
Posts: 63
Joined: Thu Jun 11, 2009 12:12 pm
Location: North Carolina, USA

Re: Glass questions again

Post by scatman » Wed Mar 30, 2011 12:14 pm

Thanks for responding, Meelis.

(wow 20-something views as of 8pm EST and only 1 reply...guess I gotta ask something more controversial?)

Anyway...I've tried changing the Gain and Precedence settings, but nothing is working. My normals are calculated outwards, too.
The rods are made of 2 uvspheres (cut in half and extruded) directly beside each other. Do I need to put 1 uvsphere slightly inside of another (kind of like a capsule pill)?

Is is ok if I attach the actual .blend file for you to tinker with?

Thanks,
scatman

latest attempt:
Image

User avatar
Headroom
Indigo 100
Posts: 1058
Joined: Wed Aug 08, 2007 1:07 pm
Location: Spartanburg, SC, USA
Contact:

Re: Glass questions again

Post by Headroom » Wed Mar 30, 2011 2:17 pm

It somewhat depends on how your object is created.

Zip your .blend before you post it.

scatman
Posts: 63
Joined: Thu Jun 11, 2009 12:12 pm
Location: North Carolina, USA

Re: Glass questions again

Post by scatman » Wed Mar 30, 2011 2:51 pm

There you go...the .blend file is attached.

Thanks for your help.

scatman
Attachments
DNA setup 4a Indigo - TESTdnaRna.zip
(1.8 MiB) Downloaded 274 times

User avatar
Meelis
Posts: 383
Joined: Sat Mar 14, 2009 11:29 pm
Location: Estonia

Re: Glass questions again

Post by Meelis » Wed Mar 30, 2011 5:58 pm

The capsel looking dna pairs (ACT) are made of 2 meshes, they can not be opened from center where they join - close them both. Also close the spiral's from both ends (total 4 circle rings to extrude (edge mode) - scale to center in median point mode). W remoe doubles and A and ctrl + N.

Image

User avatar
galinette
1st Place Winner
Posts: 923
Joined: Sat Jan 09, 2010 1:39 am
Location: Nantes, France
Contact:

Re: Glass questions again

Post by galinette » Wed Mar 30, 2011 7:56 pm

Hi,

What you want to do is tricky to do with Indigo I think.
If you want to make an object made of two kinds of glass, the interface between both cannot be defined directly. You need to have one part (for instance : the red one) slightly bigger than the yellow part, and overlapping. This means, for instance, that you need to have slightly different diameters for the two parts of the rod, to make the overlapping work correctly.

I would use a totally different approach for the rods. Make a single rod without split in the middle, and apply a material with an absorbance shader defining the two colors. I'm sure Indigo can do this, however, neither the exporter nor the material editor support it.

For the small rings inside the transparent helix, you should make them closed, and slightly smaller than the helix wire, with high precedence.

Etienne
Eclat-Digital Research
http://www.eclat-digital.com

User avatar
Headroom
Indigo 100
Posts: 1058
Joined: Wed Aug 08, 2007 1:07 pm
Location: Spartanburg, SC, USA
Contact:

Re: Glass questions again

Post by Headroom » Thu Mar 31, 2011 3:20 am

While you're at it also the helix in the .blend file is pretty large (several meters).
One Blender unit is one meter, so the default cube is already 2^3m in size.

If you scale it down to say 20-30 cm in diameter it will yield much more interesting results.

Some HDRI with more point-like light sources would potentially generate some nice caustics if you place the helix on a background.

scatman
Posts: 63
Joined: Thu Jun 11, 2009 12:12 pm
Location: North Carolina, USA

Re: Glass questions again

Post by scatman » Thu Mar 31, 2011 7:22 am

Thanks to all for your comments.

I tried your suggestion, Meelis; the result is below
Image

As you can see, it changed the ends of the spirals, but Extrude+Scale to center+Remove doubles+Ctrl N didn't quite fix the rod.
Next I'll try resizing one of the rods, per galinette's suggestion.

scatman

User avatar
dag
Posts: 189
Joined: Tue Feb 02, 2010 2:28 am

Re: Glass questions again

Post by dag » Thu Mar 31, 2011 8:25 am

Set up for blendigo 2.6.1.4
Smaller model = more transparent glass
Attachments
27.jpg
27cm diameter
dna_hdri.jpg
looks groovy with envmap
dna.zip
blend
(2.41 MiB) Downloaded 253 times
Last edited by dag on Thu Mar 31, 2011 12:41 pm, edited 1 time in total.

User avatar
Headroom
Indigo 100
Posts: 1058
Joined: Wed Aug 08, 2007 1:07 pm
Location: Spartanburg, SC, USA
Contact:

Re: Glass questions again

Post by Headroom » Thu Mar 31, 2011 9:58 am

Nice DAG.

Yields even better results if thelead cystall glass from the matDB is used. http://www.indigorenderer.com/materials/materials/385

I also used a lightprobe from the St. Peters Basilica in Rome.from here
http://ict.debevec.org/~debevec/Probes/

Set the mapping to spherical. You may have to turn your scene upside down because otherwise the lightprobe stands on its head.

You als need to convert the .hdr file into an .exr using for example Picturenaut:
http://www.hdrlabs.com/picturenaut/

scatman
Posts: 63
Joined: Thu Jun 11, 2009 12:12 pm
Location: North Carolina, USA

Re: Glass questions again

Post by scatman » Thu Apr 07, 2011 10:18 am

Thanks to everyone that responded. After much trial and error, I now know what I was doing wrong.
1. I wasn't closing both ends of my rods with Extrude+Scale to center+Remove doubles
2. I wasn't using the Edge Split modifier (I didn't even know it existed)
3. I had the 2 'rod halves' touching each other; there should be a tiny gap between the 2
So, if anyone else wants glass beside glass in Indigo (I like caustics), please do the 3 things above, and don't forget to RECALUCLATE NORMALS OUTSIDE with Ctrl+N.

These are more like what I had in mind
Image

Image

Thanks a billion to you all.
scatman

Silmä
2nd Place Winner
Posts: 147
Joined: Fri Jun 26, 2009 7:46 am

Re: Glass questions again

Post by Silmä » Mon Apr 11, 2011 5:30 am

scatman wrote:3. I had the 2 'rod halves' touching each other; there should be a tiny gap between the 2
There really shouldn't be a gap, but the rod halves should have different precedence values.
Attachments
precedence.jpg

User avatar
galinette
1st Place Winner
Posts: 923
Joined: Sat Jan 09, 2010 1:39 am
Location: Nantes, France
Contact:

Re: Glass questions again

Post by galinette » Mon Apr 11, 2011 7:19 am

In fact, there should be either a gap, or an overlap.

A gap will give not 100% correct results, as it will simulate a thin layer of air between the two glass materials. It will render without issues, but will not be exactly what we want.

An overlap will give what we want. However, it's hard to model in this case, and requires a half-rod to have a slightly bigger diameter around the overlap.

No gap, no overlap will cause render issues (black areas generally)

Etienne
Eclat-Digital Research
http://www.eclat-digital.com

User avatar
dag
Posts: 189
Joined: Tue Feb 02, 2010 2:28 am

Re: Glass questions again

Post by dag » Mon Apr 11, 2011 9:24 am

Ooooor just use a absorption layer transmittance texture.
Attachments
alt.jpg

Post Reply
24 posts

Who is online

Users browsing this forum: Google [Bot] and 23 guests