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 - Yuriy Bochkaryov

Pages: [1] 2 3 ... 5
1
Quote
Hi,

At the moment that's currently how it works. It is not possible to stop the caustics being generated from an object, it's only possible to stop objects receiving caustics.

We'll log this feedback and I'll let you know when we have some more information.

Rowan

Hi
Thanks
It's a good news

I wrote my proposal on caustics a long time ago, then I was told that the idea is good and will be implemented later.
A lot of time has passed and I can see on the board in trello the item - "Include / exclude for caustics"
My first thought - is great! you finally did it! but after reading the description of the function, I was disappointed.
I hope you will make this feature soon.
as well as the new tone mapping, color spaces - we have been asking all this for not the first year.
I would like to see great development in the future.
Corona is an excellent engine, but unfortunately its development is extremely slow now ((

2
Quote
Not sure what you're doing in your scenes but we're not having anywhere near the trouble you seem to be with caustics: https://www.recentspaces.com/latham-pools-collection

I'm not going to pretend that we've had no issues with caustics and the workflow because we have, and of course bugs, there are plenty of them, but you can work around them if you're careful, and the results can be spot-on. As you can see above we've happily rendered tonnes of stuff including 4k Cinemagraphs and all sorts.

I'm not disagreeing with you that the workflow with exclude list isn't backwards, because I agree that it is. But... the new system still seems a great step forward by the devs. I think if you can barely achieve caustics on teapots right now then you have bigger challenges to solve first perhaps.

do you calculate the caustics on the final render, with all the objects at once?
or do an additional render, where you hide unnecessary objects and render only objects that generate and receive caustics?



3
Quote
There are enviroment overrides for direct, reflect and refract, there is also ray switch material. All the tools are already there to use different maps for lighting, reflections and background. Why would anyone want such frankenstein color correct map that you're proposing? That makes no sense to me

I know that there are slots for replacing the environment / reflection / refraction - but this is not very convenient
you need to have 2 slots in the editor material, where you will have 2 variants of one and the same HDRI which will have different settings via CoronaColorCorrect
1 CoronaColorCorrect will be adjusted for light
2 CoronaColorCorrect will be set to replace the background
it is not very convenient to have 2 separate CoronaColorCorrect and it wastes more time
it is faster and more convenient to have both settings for both the light and the background in one shader, haven't you thought about that?
It is faster to work, takes up fewer slots in the material editor and it is more convenient to control

and how do you set up the volum effect for the HDRI which currently only works with the Corona sun?

I can tweak the effects through volumetric BUT we always strive for convenience and speed, adjusting the volum effect in Corona sun is easy and fast, and I would like to have this option for HDRI, because it is fast

4
Quote
Hi,
do you know that you can switch it to include list in the exclude dialog, like all similar lists?
Hi Ondra
is there a way to select an object on which the caustics will be calculated?
for example, here we have 4 mirror objects - can we make the caustics generated from only one?
the function that is now, it works on the adoption of caustics, if I understand correctly, and not on the choice of the object that will generate
I am adding a test scene for 3dmax 2019
Thanks

5
there is another problem - many (almost all) HDRIs do not have a sufficient brightness range, this makes the lighting not very good and in order to improve it, you need to use CoronaColorCorrect, increase the brightness and decrease the gamma - this makes the light better, but also does the HDRI not pretty for the background.
To fix it now - you have to have 2 variants of the same HDRI
The first is the HDRI setting for scene lighting
and the second one is HDRI configured to use slots with background replacement

I propose to make a new environment shader based on CoronaColorCorrect, which will have the same functions as CoronaColorCorrect, but there will be 2 versions inside!
The first set of settings for example will only work for lighting, reflection and refraction.
the second set of settings will work in the background,
 
we will put this shader in the environment slot and it will work separately for lighting / reflection / refraction / - the first settings and separately for the background - the second settings.
it will be very cool - we will be able to adjust the lighting without breaking the background, when for good light it is necessary to strongly distort the gamma or increase the brightness, while the HDRI, due to the gamma and brightness settings, looks terrible in the background

There is also a nice Volume effect function in CoronaSky, which only works with the corona sun so far, but most of the CG Artist uses HDRI for lighting, which makes the Volume effect function unavailable for them.
I suggest adding a Volume effect to the new environment shader so that we can use this cool effect in conjunction with HDRI lighting.

6
Quote
Right now it looks more like an artistic tool rather than CPU helper. It allows you to control that you only want caustics to appear on certain objects e.g. a pool surround, ceiling, table etc. But I have to agree that it feels a backwards solution. In practice we're going to need to always exclude the entire scene models, and then selectively allow those few objects to remain. But as you build/change/increase the scene complexity as you work you're going to definitely forget to keep adding all the undesirable objects to the exclude list. It should just be reversed to an Include list.

now I cannot use caustics in my exterior scenes, even my AMD 3970X cannot cope with it
if I turn on caustics, the rendering speed drops several times, there is almost no caustics itself, or these are different spots, it looks like there are not enough photons in the scene for caustics, there are too many objects with reflections and they do not have enough rays to form caustics
I do not want to render caustics on all objects in the scene, I want to be able to select the object that I need and let my CPU calculate caustics only from this object
this will remove unnecessary load from the CPU, increase the speed and quality of caustics
Now I can hide the display of caustics from some objects, but why do I need it - if my PC cannot calculate it
to hide the caustics from the object - first you need to calculate the caustics)))
where is the logic?
now I can only get caustics on tests with teapots or hide everything from the scene and leave only the elements that will generate caustics and those on which caustics will be displayed - that is, to show the caustics in the picture, I need to make 2 final renders and collect it in layers in photoshop
now caustics are useless for me, too much extra time and actions )


7
hi
it seems to me that this function is completely useless / erroneous
the idea was that we could select one or several objects in the scene that will generate caustics - this will allow us to save CPU resources, not to count caustics from everything that has reflection-refraction, and now in the PBR era, all materials have reflection, count caustic from only one or more objects that really need caustic, such as water in a pool, glass in a house, vase on a table
I do not understand where you can apply the function - disable receiving caustics. This does not save CPU resources and does not allow the caustics to be calculated faster.
We need a function to disable / enable caustics for unnecessary / needed objects / materials
thanks

PS. thanks for CoronaDecal - the first tests look great!

8
Gallery / «ONE 140»
« on: 2021-10-14, 11:21:46 »
«ONE 140»
Location: RUSSIA / MOSCOW
Year: 2021
Behance: https://www.behance.net/Bochkaryov
Facebook: https://www.facebook.com/CGYuriyB/
Insta: https://www.instagram.com/yuriy_bochkaryov/



9
Quote
Despite being a non-developer, I am breaking the silence! :)

Please check this video, I hope it clearly explains the behavior we are observing here:
Thanks for the info.
Now it looks very strange, but you can get used to it.
The main thing is that the developers do not change anything in the future, which will lead to the breakdown of our materials made now on the new physical shader.

10
The development team is silent and we cannot understand whether this is normal or is it a bug.

11
we still have an open question - why increasing texture brightness starts to work as an invert for IOR
texture through color correction with RGB 500 - works correctly
texture through color correction with RGB 700 - starts to work as inverted. why?
Why does increasing the brightness of the bitmap change the reflection value of the IOR to inverted?

12
Quote
PhysicalMaterial: IOR 1.52 translates as Specular 0.5 (this is industry standard formular, Corona didn't make it up, look Unreal or Substance documentation)
if this is really a new industry standard, then you have to get used to it, this is not a problem. It just looks strange to me and I am worried that it may turn out to be a mistake, I will do a lot of work on my library, and then the development team will recognize it as a mistake and change the algorithm and it will break my whole library, it will be a very unpleasant surprise )

13
Quote
Do you know of some commonly used software that can generate IOR textures
we do not need a program to generate IOR textures, it should work with any texture, with any image, based on the gradation of dark and light, where dark adds IOR and light removes IOR - or vice versa, it doesn't matter, the algorithm subtracts from light to dark, this allows us to define areas with and without reflection on one object through the texture.
For example - a leaf of a tree, the leaf has veins that have no reflection, and another zone of the leaf has a reflection, this makes our picture realistic.
It is impossible for the entire sheet to have the same IOR value, this is not correct.
I’m not saying that your theory is not correct, I’m saying that we probably have an error somewhere in the work of IOR with textures
Please note that when switching IOR mode to Specular mode, the texture starts to work. If everything is correct, then there should be no changes in the reflection when switching different modes to IOR.
I tried a texture with 16 bit depth and the problem persisted.
I also observed problems - when I tried to adjust the IOR through the texture in the color-correction and it did not work, RGB level 0 and level 1000 always gave reflection, but the texture at RGB 1000 was absolutely white, and at RGB 0 it was absolutely black. The reflection was of different strengths, but it was a reflection. Although, within the maximum range of 0 and 1000, I should have had maximum reflection at one value, and complete lack of reflection at the other.
The working range for the color corestore was approximately 120-180 in the RGB, all other values ​​of the RGB multiplier simply did not work, there was no texture on the reflection, it was just a mirror with different reflection strength
Also, if you check the Invert checkbox on the texture, then the card starts working - this also indicates that the work is incorrect, because the invert should swap the reflection and its absence, and not turn the texture on and off.
You need to make tests with a simple texture, with 3d max procedural maps everything works fine, the problem is in raster textures
Another oddity - texture in IOR through color correction with RGB 500 - has the correct reflection pattern, if you make RGB 700, then the texture is perceived as inverted for reflection, but the texture itself remains the same, but the texture itself has become brighter, but not inverted, why is it in IOR perceived as inverted?
I think we have obvious problems with the perception of the range of brightness of the texture, somewhere there is an error in converting the brightness of pixels to IOR values

 

14
New physical material does not work correctly with the map in IOR
I think you need to check and fix it.
Adding a texture to the IOR slot does not draw on reflections.
For the texture to start working, you need to increase its RGB multiplier Level - this is not correct, because the texture has a normal contrast and should already work
Also, if you switch from IOR mode to Specular mode, then the texture starts working.
Please check.

P / S "black - increase, white - decrease" - a bit unusual and probably not logical, but perhaps it was done for convenience using Roughness


15
this should be the correct result.
Right, I see now :)... since it has its own bump it should really be flat over when used & set in such way.
Sure then, a value (slider) to influence it and/or mix with bump under, could be another good feature to have.

Maybe just a padlock on the Bump parameter of the Clearcoat layer would be enough to choose whether the two should have the same behavior as in the first image (Reply #104) or should be independent as in the second (Reply #108).
I think it is a good idea

Pages: [1] 2 3 ... 5