Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - selene

Pages: [1] 2 3
1
Vray2sidedMtl not converting, nor vray lights.. was fine on corona 6

Sorry to hear that INVIZ. Please could you send us the scene which manifests this problem?

2
Thank you!
RC7- now the shader tree looks much simpler and reasonable. No more unnecessary Mix maps - that's good :)! (Converter7_vray-corona_mat-structure_RC7.jpg) and rendered result is visually the same as it was with RC5 and close to original look of the shaders (Converter7_vray_original-converted_mat-RC5-7.jpg)
This time material structure is much cleaner for further work with it which is great, now I understand how almost all maps were converted in order to use all aspects of Vray 2sided material within single Corona Physical material using FrontBack maps (Converter7_vray-corona_mat-structure_RC7-1.jpg), but:
1. could you explain, please, the meaning of use maps in Vol. adsorption and Vol. scattering of leaf material and why the map which was used in Vray material to map translucency (Map #11) is used here, not in translucency slot of Corona Physical material (Converter7_vray-corona_mat-structure_RC7-3.jpg)?
2. Why IOR parameter of the trunk material is mapped with some Mix map? Is it the way to imitate mapped reflection parameter of Vray material (Converter7_vray-corona_mat-structure_RC7-4.jpg)? And would be interesting to know about the usage of CoronaColor map with applied Solid HDR color higher than 1 in IOR mapping, like IoR map from this example has.

Also would be interesting to hear what do you think about the way of conversion of Vray 2 sided material which I used in my experience. I used two Corona materials plugged in to Corona Layer material with Corona FrontBack map used as mask. Here is a very quick example of this kind of conversion (Converter7_converted_mat_RC7-layerd_2sided_mat_strc.jpg). Material structure is pretty simple, render results you can see also here (Converter7_vray_original-converted_mat-RC7-layered_mat.jpg and Converter7_converted_mat_RC7-layerd_2sided_mat.jpg)

Hello Bormax!

Answer to question 1:

The same phenomenon which is defined by Translucency Fraction parameter in CoronaPhysicalMtl is defined by Translucency color in VRay2SidedMtl (not by VRayMtls' parameters connected to the VRay2SidedMtl).

The Map #11 is not used in Translucency color slot of CoronaPhysicalMtl, because VRay2SidedMtl automatically uses base color of the opposite side as a translucency color (that's why we pick Map #10 and Map #16).

CoronaPhysicalMtl ignores both Volumetric Absorption and Volumetric Scattering parameters when its Thin-shell parameter is enabled, but Corona Converter always propagates all maps from the original material, even when they do not contribute to look of the converted material. If VRayMtl uses some other Translucency modes - e.g. SSS, we plug the corresponding maps to these slots so we decided to do the same in this case to be consistent.

Answer to question 2:

Because the original VRayMtl uses reflection amount parameter / reflection amount map, however this parameter is not physically plausible (that's why it's not present in CoronaPhysicalMtl). Since the only CoronaPhysicalMtl's parameter which may affect its reflectivity is IoR (if the material is not metal), we try to create a new IoR value by decreasing the original material's IoR using the reflection amount parameter / reflection amount map in order to preserve look of the original material. However note that this is a "hack" created to preserve look of the original materials. Alternative (more strict) method would be to ignore / drop such parameters / maps, but this could lead to many look-breaking changes during the conversion.

Answer to question 3:

To be honest I have never tried this before, but it looks like a good idea :) May I ask you whether the results look identical? And which approach leads to better render times?

3
The RC 6 version of the converter constantly badly detects metal / non-metal materials when converting, Very annoying situation, e.g. in the case of many objects, such as books .Link to the scene after converting, original vray materials in the editor  https://drive.google.com/file/d/1QV2Bx3_FdbIcoBxPBFkn4r9vbNAEY-mY/view?usp=sharing

Hello Skalpel3d. Thank you very much for your feedback. To be honest the metal / non-metal detection was probably one of the most difficult things to implement in Corona Converter, since CoronaLegacylMtl has no "metalness" parameter (and for a long time also VRayMtl had not "metalness" parameter), therefore there is no "clean" / "exact" way how to determine whether the original material was intended to represent a metal or not. As a result, Corona Converter can only make some "educated guess" based on available parameters which sometimes unfortunately leads to false negatives or false positives. We are aware of this limitation and we are already working on some improvements, however we decided not to include them in V7 anymore, since it will require huge amount of testing to make sure the new heuristics works better in all / most cases than the current one. Once the testing will be finished, we'll include these improvements to Corona Converter bundled with Corona Renderer 8 daily builds.

4
Hi

RC5

Again the same tree test. Thanks, this time render result is closer to original, but the shader structure is still messy. Sorry, but I can't see the logic in this Mix map chaos, even simple opacity map from original material became two mix maps (original bitmap + CoronaColor?) plugged in to Front/Back map. Actually all simple bitmaps are converted to Mix maps - original bitmap + CoronaColor with no affect of CoronaColor map in those mixes... This structure is very difficult to handle which makes it not usable for further adjustments. It's easier to remake the shader manually from original material, than to handle this mess.

The test scene was uploaded some time ago - 1624024326_converter7-vray-comparison.zip

Hello Bormax, we tried to make some changes in the Corona Converter code to create simpler shading networks when converting VRay2SidedMtl. These changes are included in Corona Converter which is shipping with Corona Renderer 7 RC6. Please let us know whether the shading networks created during conversion now look as expected or whether there is still some room for improvement.

5
some materials are converted with refraction at 1! For now the Rc has disappointed me. I hope the final version of Corona 7 is good.

Hello kanyno, thank you very much for your feedback. If the original materials do not use refraction and the converted materials become refractive, this is indeed a problem, but as romullus already mentioned, we are not able to fix this issue unless we have the scene available. Please, would you be so kind and send us the scene in question? We'd like to have a look at it and to try to fix / improve Corona Converter accordingly.

6
Hey,

I intended to use the Corona Converter tool  mainly to convert Vray materials into Corona materials. But I always get a mess (so many unecessary nodes) and the result is not convincing. Eventually, I end up making the conversion manually which is very time consuming.

Did the corona team plan to make major improvements in this field ?

Thanks !

Hello mos2787, thank you very much for your feedback. Please do you use the latest Release Candidate of Corona 7? Over the last couple of "Release Candidate releases", we tried to address the issues you mention. Also, please could you provide us the scene in question?  We'd like to have a look at it and to try to improve Corona Converter accordingly.

7
Convert by class in corona converter v1.45 does not convert 3Ds Max Physical Material (compile error: Unexpected end-of-script).

I am still on Corona 6 hotfix 2 in 3Ds Max 2022, has this already been fixed, or is this a known bug?

Hello Bjoershol! Thank you very much for your feedback! We weren't aware of this problem and we'll try to fix it as soon as possible.

8
[Max] Daily Builds / Re: 3dsmax Material Id Channel
« on: 2021-06-21, 12:57:37 »
We are losing (3dsmax) Material Id Chanel after conversion. Corona Legacy Mtl with id channel = 1 always converted with mtl ID channel = 0. So I lost my mtl channels in materials for postproduction in my scene.
Corona G-Buffer ID override works fine.

Hello FixelFakes, thank you very much for your feedback. Please, is this something which used to work as expected in previous versions of Corona Converter and only stopped working in the latest versions?

9
Done
1622111539_converter6-7-comparison.zip

Bormax I tried to convert the scene you sent us using Corona Converter shipping with Corona Renderer V7 RC1 and the shading network generated during conversion is simpler. In case of this scene, I noticed that there is still a problem with some duplicate nodes being generated during the conversion, which is caused by known bug in one MAXScript function (on 3ds Max side), we are aware of and already trying to fix, however this should have no negative impact on the visual result.

10
hello my friends!

still the same issue with falloff map ...

corona 6 - 1 falloff map in difuse chanel - OK
corona 7 - 1 falloff map in difuse (converted) - lot of CoronaMix maps.

why? wouldn't it be easier not to convert this type of map? (falloff) or rather, create a Coronafalloff?

Hello lzanlorenzi , this issue with lot of CoronaMix maps being generated should now be fixed in Corona Converter shipping with Corona Renderer V7 RC1. Please let us know whether it works as expected :)

11
A material made with Corona 6 consists of 1) after conversion with the latest converted, the newly created material contains many maps and nodes 1) + 2); That's so frustrating; Is there any workaround?

The Corona Converter shipping with Corona Renderer 7 RC1 now contains changes which should result in much simpler shading networks being generated during conversion. Please let us know if it does or does not work as expected or have some ideas for improvement :)

12
Done
1622111539_converter6-7-comparison.zip

Thank you very much!

13
Hi
Not the 1st time mentioned, but here is comparison of Corona 6's and Corona 7's conversion results. Corona 7 tries to keep front/back colors which is really good, but gives a mess of mix maps which (IMHO) have no sense. The tree model taken from Evermotion Archmodels 207 collection where material structure of all models is about the same.

DB 2021-05-20

Hello Bormax, please could you provide us the scene in question, so we could make sure it works as expected even for this particular case?

14
Another issue is incorrect recognition of metal materials, non-metal in the case of vrayblend conversion turns the materials from the machine into metallic

Hello Skalpel3d, thank you very much for your feedback. Please could you provide us the scene in question? We'd like to have a look at it.

15
hello my friends!

still the same issue with falloff map ...

corona 6 - 1 falloff map in difuse chanel - OK
corona 7 - 1 falloff map in difuse (converted) - lot of CoronaMix maps.

why? wouldn't it be easier not to convert this type of map? (falloff) or rather, create a Coronafalloff?

Hello lzanlorenzi, thank you very much for your feedback, we're already working on changes which should result in simpler shading networks being generated.

Pages: [1] 2 3