Cut Crystal Tests

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 2:12 am

Here's version 5 with Sellmeier calculated IORs. This time around I have include the Office 2007 version of the sheet for those that have Office 2007 along with the Office 95-2003 version as well. I have no idea how the chart will show up in the .xls version...

EDIT: Links removed. Final version for proof of concept further down below somewhere

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 2:17 am

Render update:
~56h 30m, ~3760 samples/pixel

Kram1032 · Post by **Kram1032** » Mon Oct 08, 2007 2:26 am

you did a great job

why don't you show the updated result? xD

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 2:31 am

Kram1032 wrote:...if you directly compare the results, is there any difference at all?
Hu, if I take a really close look, I think, sellmeier is slightly lower than the measured results Am I right?

Code: Select all

Measured
	1.79026	1.79539	1.80187	1.81185	1.81335
	1.82038	1.83102	1.83650	1.83808	1.83957
	1.84636	1.84666	1.85504	1.87204	1.87425
	1.89393	1.91366

Cauchy
	1.79621	1.79756	1.80072	1.80933	1.81087
	1.81840	1.83005	1.83598	1.83766	1.83924
	1.84636	1.84668	1.85518	1.87152	1.87355
	1.89074	1.90642

Sellmeier
	1.79026	1.79539	1.80187	1.81185	1.81335
	1.82038	1.83102	1.83650	1.83808	1.83957
	1.84636	1.84666	1.85504	1.87204	1.87425
	1.89393	1.91366

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 2:33 am

Kram1032 wrote:why don't you show the updated result? xD

If you mean the render update? Because changes are so small from time to time now that it would be better just to wait till it's cooked

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 2:47 am

Well... So much for me modelling my crystal plates this weekend *sigh*

I'm off to bed got work in the morning

Ok Kram just for you...

Render @ 57 hours (JPEG compressed) 50% scale: Click for full res PNG

Wedge · Post by **Wedge** » Mon Oct 08, 2007 10:59 am

Very nice render. Especially liked seeing it converge in the earlier pages.

Hand made glasses...Just awesome!

CoolColJ · Post by **CoolColJ** » Mon Oct 08, 2007 11:13 am

That looks nice, definitely has a photo-real look

why the gaps between the tiles? shouldn't there be some stuff in there

Post by **OnoSendai** » Mon Oct 08, 2007 11:14 am

WytRaven:
Looking at your Schott SF57 IOR vs wavelength graph, with the linear wavelength scale, it looks like you could get a better fit for the cauchy curve - it's diverging quite a lot from the measured data near the edges.

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 12:37 pm

Well here's the problem. I still don't know what you are using for the A coefficient in your Cauchy calculations, we can only provide a B coefficient so you must be calculating A from something. I am using nD in mine and calculating the B coefficient as the gradient of a line between nF and nC. Without knowing how your coming up with your A it makes it hard to fine tune my B values.

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 2:59 pm

This evening I will setup that chart so that I can select a glass from the list and it will display the selected data. That will facilitate easier adjustment of the B coefficients because I will be able to see the changes as I make them.

If only there were a way to derive Cauchy Coefficients from Sellmeier coefficients

BbB · Post by **BbB** » Mon Oct 08, 2007 6:30 pm

I wish I could keep up with the technical stuff here. But it's a great image in any case!

WytRaven · Post by **WytRaven** » Mon Oct 08, 2007 6:36 pm

Lol. Well the image doesn't directly relate to the technical stuff rather the technical stuff was inspired by the image.

The 'crystal' material in the image is hacked from an IOR of 2 and a Cauchy B calculated from Heavy Flint glass IORs at the F and C wavelengths so it's not accurate at all. But it does look mighty pretty doesn't it

WytRaven · Post by **WytRaven** » Tue Oct 09, 2007 4:31 am

Good news. We now have n/ɑ pairs from 780nm - 350nm in 10nm steps.

Index of refraction values calculated via Sellmeier's equations.

Absorption coefficients were calculated by applying a spline function to the meassured values and performing intercepts on the curve at the desired locations.

I have setup a live charting system whereby I can adjust the B coefficients gradually and see the results visually instantly. I will tune them as best I can later.

Ono is there any particular reason why you just seem to refuse point blank to answer the question of how you calculate your Cauchy A values internally?

Render is now @ ~83h and ~5550 samples/pixel.

Kram1032 · Post by **Kram1032** » Tue Oct 09, 2007 6:32 am

great

I'm not sure, if he absolutely refuses, but sometimes, it's really hard to get an answer, out of him, especially, if it's indigo-internal-techniques related xD

Might be related to something like wanting to have free time, when looking through the forums, and not needing to think about Indigo's calcs, for some mo's

Cut Crystal Tests

Who is online