Chaos Corona Forum
Chaos Corona for 3ds Max => [Max] I need help! => Topic started by: DarcTheo on 2016-05-13, 11:34:46
-
trying to work on metal materials with Corona. Finding that ours tend to look quite flat. I've tested with premade metals from other users and it still seems to happen, so something is going wrong somewhere.
I've attached two examples, one from Vray and one from Corona. Using a basic scene and just using sun & sky for both versions. Vray metals tend to show a gradient from bottom to top that make the panels look less flat. Corona seems to light the whole face the same, making it look flatter.
What can I do to achieve a look closer to the Vray example? I've tried various IOR values.
people in the studio are preferring the Vray metal over the Corona metal.
thanks
-
Umm.. Have you tried increasing reflection glossiness? It is clearly higher in the Vray version. Also, the background sky is much darker in the Vray one.
-
Umm.. Have you tried increasing reflection glossiness? It is clearly higher in the Vray version. Also, the background sky is much darker in the Vray one.
i will try playing with gloss. I tried to leave sun and sky at the same settings in both and same physical camera settings, I guess it does not translate perfectly between the two renderers but i hoped it would make it close enough to be a fair comparison.
-
And the plane-to-reflect (look into the mirrorball) is strangely bright, seems like "Cast Shadows" is disabled on the object.
Good Luck
-
And the plane-to-reflect (look into the mirrorball) is strangely bright, seems like "Cast Shadows" is disabled on the object.
Good Luck
Ah yeah, i turned that off to see if it would help some how, i'll turn that back on.
thanks
edit - It casts the foreground in shadow with shadows turned back on, and it was turned off in both the corona and vray renders to stop that, it was only there to help reflect something. Seems Vray handles it a bit differently?
-
Yes, it´s different, different technology. To achieve the same with Corona you have to pull out the swiss army knife: Corona Rayswitch Material.
See
https://coronarenderer.freshdesk.com/support/solutions/articles/5000516723-how-to-disable-shadow-casting-of-an-object-
This is exactly what happens here.
Good Luck!
Edit: It´s now projected to wrap it automatically at rendertime:
https://forum.corona-renderer.com/index.php/topic,11965.msg77118.html#msg77118
I suggested this about 1 year ago after discussing those object property options and what users would expect, gave up and was told that I may write my own maxscript material using rayswitch to achieve this ;) Glad to see this very important step of a perfect max integration now in the first 1.5 daily
-
Yes, it´s different, different technology. To achieve the same with Corona you have to pull out the swiss army knife: Corona Rayswitch Material.
See
https://coronarenderer.freshdesk.com/support/solutions/articles/5000516723-how-to-disable-shadow-casting-of-an-object-
This is exactly what happens here.
Good Luck!
thanks Frood. Went ahead and did that for consistancy in the backplate. It now looks closer, just missing the highlights. alot less light etc hitting it. Maybe this isnt such a fair comparison, i'll try tweaking the sun now with gloss
Would anyone want to try with the scene and adjust the metal as well?
-
could you share a scene for testing? or show settings?
have you disabled diffuse on metals?
there are some differences in settings, though should be simple to set as visually same
-
could you share a scene for testing? or show settings?
have you disabled diffuse on metals?
there are some differences in settings, though should be simple to set as visually same
Attached is the material, it was converted from VRay and then I have been trying to tweak it to look how we want it. IOR doesnt seem to be making a big difference at the moment, i'm thinking maybe the fall off curve does not work so well from Vray to Corona.
As above it just seems much darker than the Vray one. I set the scene up for corona then just hit convert scene to make it Vray, so the sun position was the same, sky default. only thing I wanted to be different was the renderer. but ofc the material came the other way round from a Vray material first.
-
Hi. It seems like all your troubles are coming from using Vray "tricks" in Corona. That was the case with "don't cast shadows" and IMHO it's also the case with "fake" IOR in the materials. I would disable the fresnel falloff in reflection map (use a color instead) and set IOR to something between 25 and 40. IMHO using the Corona IOR gives a better result than this Vray method.
I might be wrong tho... :)
-
Did some tests myself :- )
It's not 100perc (ok, not even 50perc. but it will do)...I dedicated (wasted lol) 30 minutes this late evening to this whole experiment ,and 2x10 minutes from that are render times.
Random notes:
-Does D65 mean different temperature in white balance between Vray and Corona ? Corona is colder, but perhaps I am stupid.
-But outside of this, I did not use any clamping or tonemapping. Both are linear outputs, rendered at final exposure (no tweaks in VFB or PS, so correct MIS is computed in both renderers), with default MIS 20 in both Vray and Corona.
-I use slew of Dubcat's tricks. In Vray, to use flat reflection, I simply uncheck fresnel. In Corona, 999 is not enough, CoronaColor with 9999.. value in HDR is used in FresnelIOR slot.
-But I also use converter, which divides the anisotropy in half. But it still looks different nonetheless. I can't judge honestly, I am not sure what is correct or better.
-What looks like GI/Direct light noise in grey area in Corona render, are reflective (GI) caustics. I unfortunately forgot to turn them ON in Vray. MIS20 still allows a faint of them. My mistake, like I said, 11AM, rushed it...
The lack of GI caustics in Vray perhaps contributes to slightly brighter Corona render....damn, should have thought of that.
Anyway, to keep as much consistency, you can see I only use fallofs to drive mixing, not values, and Siger's ComplexFresnel for metalic speculars, and Siger's ThinFilm for coating.
WELL, but as you can see....nothing drastic right :- ) ? Looks more than close. As we say in mountain biking, choose your renderer (wheel size...) and be d*ck about it :- ).
-
do you have the reflective caustics enabled in vray?
-
do you have the reflective caustics enabled in vray?
I edited my post but maybe too late. Oh, don't want to re-render... Actually I can't, I was lazy and didn't save before conversion :- D
I unfortunately forgot to turn them ON in Vray
-
Hi. It seems like all your troubles are coming from using Vray "tricks" in Corona. That was the case with "don't cast shadows" and IMHO it's also the case with "fake" IOR in the materials. I would disable the fresnel falloff in reflection map (use a color instead) and set IOR to something between 25 and 40. IMHO using the Corona IOR gives a better result than this Vray method.
I might be wrong tho... :)
I recognize this setup, it's from VMPP/CMPP by Siger :- ). The Brass is setup bit over-complicated, and not as full-scale metal. But it definitely doesn't have "fake IOR", it's actually way more correct, even if setup with fallof like this. It's not Vray method either, it works fine in any renderer. But using Color plus simple IOR artistically work just fine, I still hold it's good method :- ) But nothing that would make the 'hard-core' way 'fake' or incorrect.
But because that particular setup (this brass above) is so complicated, I think it didn't translated correctly by converter script.. more so that it is such a multi-blend with AO as well.
-
Yes, I can see my wording was inaccurate. By "fake" I meant: "not Corona/build in" - not "physically incorrect". I haven't the knowledge to judge physically incorrectness on this level :)
Second, I'm an old mental ray user, not Vray, but I've seen many Vray materials build like this, so I assumed it was a common Vray-thing :)
But it's always good to learn something new :)
Regards
-
I wanted to keep conversion errors to a minimum, this is what I did.
- Used the same HDRi in every application.
- Made a quick material in "Substance Painter 2", keeping everything PBR.
- Used the same "Physical Camera" with same "Physical Camera DOF" settings (I wanted to test DOF too).
- Kept all the renders Linear.
The V-Ray render is 0.3 RGB darker.
Some V-Ray people actually lower the GI Contrast to around 0.75 to mimic Corona/Octane.
Everything is controlled with maps, so there was no reason to show the material settings.
(http://i.imgur.com/byDgzm4.jpg)(http://i.imgur.com/tO8cLke.jpg)
(http://i.imgur.com/cAFe9i5.jpg)(http://i.imgur.com/qjNDvUY.jpg)
(http://i.imgur.com/5nRQpnG.jpg)
-
I wonder what's causing those artefacts in Corona?
-
The bump is done with procedural noise. But yes, Vray doesn't have it. Hmm, no idea. Could be some form of terminator ? (despite having no diffuse). It's super contrasty HDRi (Chelsea stairs from HDR-labs)
-
There's a checkbox in Arnold called smooth subdivision tangents that deals with this faceting effect that occurs with anisotropy. Not sure if that gives a steer towards a solution in Corona?
-
Lowering bump value might help. to sth like .05-.03
-
i'm getting the same artifact even with no bump, didnt find a walkaround still
i noticed that is somehow related to the anisotropy of a material and it occures both in uvw and in local normal orientation, if it's set to 0 the facets disappear
-
Thanks for the replies and thanks dubcat & juraj for showing your examples.
I will try setting up a new test scene and try again using your inputs. You both use HDRIs, does this make a big difference on metals? My studio tend to use just the sun&sky as the images get a style in photoshop, so its not aiming for photorealism. I did try use an HDRI for reflection override but the scene was being lit with sun & sky still.
It's clearly possible to get very similar results, just frustrating that i'm doing something wrong. perhaps I need to convince the use of HDRIs in our work flow.
I also noticed those artifacts appearing on my corona shader, not sure what caused it, will try lower bump values again.
-
i'm getting the same artifact even with no bump, didnt find a walkaround still
i noticed that is somehow related to the anisotropy of a material and it occures both in uvw and in local normal orientation, if it's set to 0 the facets disappear
Correct. I did another test and it's because of ano not bump :- ).
-
Thanks for the replies and thanks dubcat & juraj for showing your examples.
I will try setting up a new test scene and try again using your inputs. You both use HDRIs, does this make a big difference on metals? My studio tend to use just the sun&sky as the images get a style in photoshop, so its not aiming for photorealism. I did try use an HDRI for reflection override but the scene was being lit with sun & sky still.
It's clearly possible to get very similar results, just frustrating that i'm doing something wrong. perhaps I need to convince the use of HDRIs in our work flow.
I also noticed those artifacts appearing on my corona shader, not sure what caused it, will try lower bump values again.
Well, in this small example, I needed contrasty environment for metals to truly look like metal :- ).
HDRi or Sun/Sky is just type of light, it doesn't dictate reflections but with Sun/Sky, you need geometry to complete the environment, while HDRi can already have on.
I use Sun/Sky 90perc. of time :- )
-
Well, in this small example, I needed contrasty environment for metals to truly look like metal :- ).
HDRi or Sun/Sky is just type of light, it doesn't dictate reflections but with Sun/Sky, you need geometry to complete the environment, while HDRi can already have on.
I use Sun/Sky 90perc. of time :- )
Good to know its not sun & sky then, I quite like this simple way of setting up the base lighting. Haven't had time to try any tests again yet but hopefully I can get similar results soon.
-
did anybody find a workaround or a way to fix the facets appearing when using anisotropy?