[3dtrialpractice]

hmm.. personally.. to me it looks like your closeUP NOT Maxwell SSS render is almost exactly looking (light/shadow on leaves) like your closeUP maxwel without SSS render-the reposted(2nd) w/o sss maxwell (APART from the over saturated green in the nonmaxwell sss AND the Maxwell closeup no SSS looks the best of all the leaf renders)

but I think i know what your wanting in thinSSS look from maxwell. i just dont think your nonmaxwell sss renders really show it either.. .. btu you want a glow and translcency to the leaves to relight the leaves under them right?

and if so.. i just dont see it in the CU non maxwell.. i barley see and transluceny in that nonmaxwell render OTHER than the shadowed areas of the leaves are brighter ..but it looks to me like a GLOBAL brighter shadowed area of leaves.. not a area translucent brightness..

I think shooting against the light with thinsss in maxwell from posted renders ive seen looks awsome.. but its when the light is behind camera that things look dark.. tough one to sort i guess..

[NathanDan]

Do it! I'd do it, but I'm really busy creating a Pharmacy sign to render

medication on the rise for maxwell users?

[big K]

3dtrialparactice, thanks for your comments.
i have to agree that the nonMaxwell render looks a bit oversaturated, but i just made it like this because i am nit able to get a fresh green with the maxwell system.
about the shadow part - i think this is what leaves look like (almost), they let a lot of light pass through to the other side with SSS and brighten the other leaves in the shade, BUT they are never kind of transparent like in the maxwell look. you are not able to see through one leave as you have the impression in the maxwell render and when the sun shines on them they get really bright (with this part i have the most difficulties in maxwell).

this is what i have right now after hours of testing.
still not quite satisfied, but it gets there. there seems to be an issue with the roughness in the material when used in the same BSDF as the thinSSS ! (maybe this is realted to the win64 roughness bug in SSS mats - i will try on win32 later)
when the roughness settings are high the material gets overbright (white) when lit from behind.
but you need this roughness otherwise the leaves get no light from frontal lights and therefore remain dark.
i was forced to use a second BSDF to get the front light with a rough surface.





it looks o.k. and i am sure there are trees that look like this, but still quite dark and the contrast in the sun should be higher. but this is due to the second BSDF as it only is "part" of the leaf.

i ahve seen the grass in the V2 announcements. this is the look i want. anybody of the tester is able to share this mat ? i would like to see how this is set up.
i am really not able to get it brighter.

[big K]

here is the problem with the roughness (i have tested under win32 and it looks the same)
this can´t be physically correct ?

[Tea_Bag]

Looks perfect to me!

[mashium123]

here is the problem with the roughness (i have tested under win32 and it looks the same)
this can´t be physically correct ?

yes.... this is exactly the problem that i have.
trying to create a leaf by follwoing the manual - as soon as i activate the roughness map (following the manual) it always crashes (except in the preview scene "leaf"... but hey, the thinsss is not made only for that one scene, is it )
when i use the roughness like you do in your scrshot (deactivate the the map for roughness), it doesn't crash... but doesn't look that good...

[big K]

mashium, besides the crash as you can see the thinSSS does not work with high roughness values. it gets overbright.
but you need the roughness because otherwise the surface of the mat is reflective and the sun does not light the material. (there is still color but this comes from transmittance and SSS)

[big K]

Looks perfect to me!

well, thanks, but it looks more like on a dull day without strong sun. (still looking good)

[big K]

i have done some more tests and i found out that the surfaces are actually brighter on the off side of the light source than on the side where the lights shines on.
i think this is the main problem which is either in my settings or in the thinSSS system.

for leaves this is of course wrong as they are at least as bright on the side in the direction of the light.
maybe this is the reason why it looks a bit off in my tests. (not to forget the roughness problem)

if you have a piece of paper for example it is always darker on the backside than it is on the side which is lit.

please some official tester have a look in here !

here you can see that the lit side is much darker than the back side

[3dtrialpractice]

wo bikK.. uv made soem BIG progress I must say.. yout newest tree examples are definatly the best for leaves so far .. but i feel your pain.. shouldnt have taken hours..
and im glad you posted screen shots and leaf examples of your explinations and tests.. it makes it way more understandable now..that we can see the examples..
i see what you mean.. the leaves form backlight amplify the light soo much and this isnt evident fromthe frontlight as the light behind the leaf is soo dark.

i noticed your asymetry was -.2.. maybe Boost this up to help "pusH" the light from the front lite.. maybe asymetry -.75 ot -.9? hell maybe a positve vaule try +.5 would fix the backlight problem..

OR maybe we just have to wait for the next build of engine that will address the other thinSSS bugs and will help the look of all thinsss renders.

I will next start to experiment w .thinsss myself.. but yove put some good work and thinking into it already thanks for the updates!

[brodie_geers]

I just ran some tests myself

thinSSS roughness: 0
view: towards sun


thinSSS roughness: 0
view: away from sun


NO thinSSS
view: away from sun


It seems like when you're facing the sun the leaves look very good. But when your back is to the sun all of the details are washed out and it looks far worse than without thinSSS. Maybe having the roughness turned up would solve that, but I wouldn't know since any roughness causes a crash.

-Brodie

[big K]

3dtrialpractical,
yes i was playing (and still am) with the asymmetry and have had bad results with the negative values (had started with them as they were set in the preset). they always look transparent and have the effect that the back is coming to the front.
i had better results with the positive values. the examples above are with positive asymmetry. (the one with the negative value only showed the white burn when using high roughness - but this effect is also with positive values - i had this only saved with a preview)

brodie_geers,
nice tests.
as you see the colors away from the sun are pretty dark. why should they be this darker than when seen towards the sun ?
it is not possible to get them brighter. (at least i have not succeeded) they don´t look like the sun is shining on them.
as you can see in your image without thinSSS the colors are saturated and also brighter. it should look more like this, but also with SSS and letting some light through.

[brodie_geers]

I found out from the bug forum that the roughness issue was only in ranges of like 1-3 so I cranked it up to 100 out of curiosity. The image with my back to the sun did look slightly better (you could see some of the details on the leaf however the shadows that one leaf should be casting on the leaves below it were all washed out so it was a blurry flat mess still), however, like the test above shows if you spin the camera around to look towards the sun the areas that let direct light through the leaf appear very white and completely unacceptable.

-Brodie

[big K]

if you want to continue testing you may try the version with two BSDFs. one with the refl0 and refl90 slots used with high roughness and one with the thinSSS + transmittance leaving refl0 and refl90 black.
you get quite reasonable results as seen above. but the SSS effect is weakend of course.

[big K]

i found out, that the scattering map isn´t used at all if the asymmetry value is negativ. is that as expected?

so the neagtiv values are for the transmittance map and the positiv for the scattering map.
but shouldn´t he influence of both maps go fluently from one to the other when the values are higher or respective lower than -1 / 1.
when using values above 0, 0.5 for example this seems to work. so there is still some transmittance influence and scattering gets stronger.