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 - Dionysios.TS

Pages: [1] 2 3 ... 51
[Max] Daily Builds / Re: Clouds playground!
« on: 2023-02-24, 17:12:20 »
This issue has been reported multiple times already. No need for yet another report please.

Ok, thanks!

[Max] Daily Builds / Re: Clouds playground!
« on: 2023-02-24, 16:12:43 »
Guys, I've just realized that you can avoid this annoying "clouds(hdri)+volume haze horizon rectangle bug" just by adding refractive surface, for some reason this rectangle is just not visible through refraction, so I come up with creating a huge non-reflective glass plane (see the attachment). I don't know for what reason do you need this :D but maybe devs could implement this fake solution (checkbox "create huge glass plane on the horizon line) until they come up with something better.

I confirm as well, this is a bug!
Did you reported it?


Dionysios -

Your brain has most sophisticated tone mapping engine of them all :] What human brain "sees" and what camera sensor registers are two very different things. I think it should be possible to save Corona render to exr and with some advanced tone mapping and colour grading get very similar results to your phone's photo and that wouldn't be cheating at all.

Yes I know and you're right on this!
The only issue is my workflow, I use Lightroom for years now and the 32bit format is not editable directly there, still and unfortunately, the 16bit is still ok but I was only wondering if such a lighting situation is feasible without additional LUts and heavy color grading editing.

Anyway, tomorrow I'll try the RAW format to see the differences.

Thanks -

I think the problem is not in that Corona would be physically inaccurate, but in fact that you're comparing its result with the smartphone photo. Phones these days does not take single picture with simple tone mapping curve applied, instead they do what's called computational photography, when multiple exposures are taken and very sophisticated algorithms are used to combine those shots and apply complex tone mapping. It's simply no match to your average Corona render with few tone mapping operators. Try to take RAW photo with your phone and you will likely see completely different picture.

Yes, this is absolutely true and I thought about it.
Tomorrow the weather will be the same and I'll try the RAW format. In any case, all the rendering engines they simulate the photo camera lens and and sensor effect but looking with my eyes the feeling of the light was the same!

Excuse me for insisting on this but I find the Corona daylight result too white on some time positions.

Thanks again!

Thanks for your answer.

In Juraj's post the question was on the very blue sky.
Here the problem is a bit different I guess, in the hour of day I am testing, the Daylight system produces a very subtle effect and the separation of the two lighting elements is not strong as it should. Sun is not warm enough, sky is not light blue (not strong deep blue).

Hi guys,

I am sorry for this post but I am doing some lighting tests lately to find another artistic mood and in my mind I want to be Physically corrected and use the tools we have now available in the engine without cheating too much on the lighting parameters, colors, ecc.

Unfortunately, I've noticed that the Corona Daylight system is not very consistent on the lighting simulation of some specific timings during the day.
For the architecture photography, sometimes is good to take photos near 8:30 - 9-30 in the morning or later in the afternoon (I am not talking about the golden hour).
I took some samples photos here in Paris with my mobile phone, a Samsung S20+ and I'am trying to replicate the beautiful and warm lighting in the engine.

I attached the file "Real_Photos_Paris.jpg" to take a look. The light is warm, guite orange and the skylight has a beautiful and different color contrast from the sun.

Well, doing the same in Corona doesn't produce at all the same lighting effect, the light seems too bright and too white. I've played with all the available parameters: White Balance, Color Tint, the Skylight parameters in general and have no luck! I've also noticed that the White Balance doesn't separate beautifuly the sun color from the skylight color as in reality happens when we take a photograph. The skylight is less blue, the sunlight is less yellow/orange, I am refering on timing at 8:30 - 9:30 in the morning.

The only way to acheive a similar result is to cheat the engine and the technique in general! Use an HDRI bitmap as skylight, make it very blue (Kelvin setting) and bring the Sunlight to 3.000K and get the result!!! Yes, 3000K! But this is cheating right? But the effect is quite good! See attached image: HDRI_Sun_Test.jpg.

So to conclude, now that we have our lovely clouds and our Daylight system, are we sure the combination of the two elements, Sun & Sky is correct? Is it Physically and Chromatically correct?

I don't want to use the HDRI + Warm Sun technique too much, HDRIs are beautiful but I've found sometimes their limitations.

What do you think? I ask too much?

Thanks in advance,

Dionysios -

Hi Rowan and thanks for the update.
I'll let you know if the issue persists.


Dionysios -

Ooooh! Good news then!

Thanks Rowan,

Dionysios -

BTW, I use the Daily Build for Corona 10 - 2 Jan 2023.

Hi guys!

It crashes when I select the preset "Old Photo" from the dropdown list in CIE.
It crashes with 3ds Max in default state without any scene loaded.
It crashes in CIE standalone with any CXR file loaded.

I've attached the 3ds Max minidump file.

Hi guys, non error message in 3ds Max, it just crashes!
But in CIE standalone I get an unhandled exception occured error.

I've attached the image.

I also confirm that the crash occurs at the office and at home as well.


Dionysios -

Hi guys,

I've noticed today that 3ds Max + CIE standalone they both crash when I select the ToneMapping preset "Old Photo" in CIE.

Can you confirm this?


Dionysios -

Fixed in the latest daily!

You can grab it in the usual place :)

Hello Rowan, great news!

Thanks for your support,

Dionysios -


Was this working okay in a previous driver version? Or was it always broken?

If you send over the scene I can forward it over to nvidia.. Please can you let me know your GPU + driver version?



Hi Rowan,

The driver version is the 52.98 and the GPU cards we use are: 1080 GTX, 3080 RTX and 4090 RTX.
Sometimes the Nvidia Denoiser fails to clean properly the images; it cleans them but it adds some other artifacts, it seems that it fails to predict the correct results.
I repeat, it doesn't happen always but lately I've notice it more.

I confirm that in the past this Denoiser worked perfectly except of another case where was producing the same defect and then was fixed by Nvidia.

Now the bad part, I can't share a 3D scene due to some restrictions we have in the office (RPBW Paris). Till next Friday I am very busy, eventually later, I can create a sample scene and try again, if the defect appears we can send that scene to them.

Thanks for your support as always.


It's been fixed internally but it didn't quite make the cut for the last daily build.

Hopefully we'll be able to get it in the next one :)



Thanks Rowan!

Pages: [1] 2 3 ... 51