Subsurface scattering: my first work

Kram1032 · Post by **Kram1032** » Mon Jul 21, 2008 5:52 am

nice renders

Better than the photo from before!
jk

ave · Post by **ave** » Mon Jul 21, 2008 10:45 am

... yeah, I am afraid of you guys!
If you say that photos looks "promising"... !

Anyway, i have tried in these days to emulate the light condition of the two photos. These are the results.
Still not 100% happy but...
BTW, I am having problems with the henyey-greenstein coefficients.
In these two renderers I put it to 0 for RGB, while using the values tabulated by Jensen i get a dirty gray substance! Also, keeping the value but reversing the sign (I thought that maybe the sign convention was different) I get a bluish stuff. Any ideas?

ave · Post by **ave** » Mon Jul 21, 2008 10:49 am

... while these are with h-g coefficients from jensen paper.

BbB · Post by **BbB** » Mon Jul 21, 2008 7:47 pm

Those look nice but still too opaque for me, especially compared with your photo. Perhaps you should use some artistic licence and make the milk a notch more transparent than the data suggests.

Have a look at this Maxwell milk glass here, (scroll down the page).

Post by **OnoSendai** » Mon Jul 21, 2008 11:55 pm

Hi and welcome Ave.
The Henyey-Greenstein phase function parameter 'g', should definitely be positive for forwards scattering.

Kram1032 · Post by **Kram1032** » Tue Jul 22, 2008 8:14 am

I like the milk as it is

If I remember correctly, WytRaven found some problems with h-g and also a way to correctly convert the values....

ave · Post by **ave** » Mon Aug 18, 2008 8:49 am

I played again a bit with this milk glass. Basically I just changed the lighting: now it is an env map from evermotion which i converted to exr. I am pretty satisfied, but still I dont understand why using the h-g coefficients from Jensen I get something wrong... here I just put g uniform and = 0.

CTZn · Post by **CTZn** » Mon Aug 18, 2008 2:39 pm

Maybe you are already used with that, but did you check in the exported igs if the value written are the ones expected ? I don't use c4d, it's just a suggestion... maybe there is a known mismatch needing data conversion as kram mentioned ?

Post by **fused** » Mon Aug 18, 2008 6:39 pm

i can assure, that if you enter the data right in cindigo, it will be exported as expected. for rgb its a bit tricky to get the right values, tho.

heres a quick howto:
1. set RGB slider to 0...100% (0% represents 0.0 and 100% 1.0)
2. give your values in a valid range (0-1; eg 85% for 0.85, 12% for 0.12, 0.045% for 0.0045, etc.)
3. tweak the "gain" parameter to get the right values, its just a multipier (in case of the 0.045%, if the desired value was 45 it has to be set to 10000)

matsta · Post by **matsta** » Thu Aug 28, 2008 11:38 am

beautiful renders man. very nice scene setup! for some reason i can still see VERy slight subdevision on the last image. is it just me?

greetz
mat

Borgleader · Post by **Borgleader** » Thu Aug 28, 2008 11:47 am

matsta wrote:beautiful renders man. very nice scene setup! for some reason i can still see VERy slight subdevision on the last image. is it just me?

greetz
mat

Looks good enough for me....can i drink it now? *begs*

ave · Post by **ave** » Thu Aug 28, 2008 5:46 pm

Thanks for your comments!

2 matsta: you are right, there are still subdivisions visible

but mybe some postwork would be enough to get rid of them?

PureSpider · Post by **PureSpider** » Thu Aug 28, 2008 11:26 pm

About that grey milk... are you sure the scene scale is set right?

Subsurface scattering: my first work

...

my last milk

Who is online