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 - JulioCayetano

Pages: [1] 2 3 4
1
Here you can check a simple comparison, same Turbidity (=2) for all, same altitute (=0) and just changing the multiplier of Volume effect multiplier.


2
Can you just increase strength of the effect at low turbidity and get same result (minus magenta tint)?

I mean, get denser "fog".
After a few minutes testing the Fog with RC4, I noticed that, at Turbidity=2, increasing the Volume Effect multiplier not only makes the Fof effect denser, but produces unexpected results, like bands within the Zdepth range that are more dense and more heavily tinted and other areas that are less dense. I´m not sure what is happening, but doesn´t look like a natural effect to my eyes. I will post examples as soon as possible.

3
Testing the RC4 to check if the Magenta tint is gone. After a quick test in a simple scene, the same I used to test RC3, the issue seems to have improved a lot, but it isn´t fully gone. Maybe this is too subjective. Let me show an example:
RC3 vs RC4 comparison at Turbnidity 4: the improvement is massive.
RC3 vs RC4 at Turbidity 6 (or anything above 4.5), and the improvement is noticeable but still looking a bit weird to me. You can check the refraction pass to see how those glass surfaces are tinted when they shouldn´t, right?

4
Finally I had the time to render some crops and take some screenshots of the production scene I was talking about when commenting about Displacement Artifacts.

Here I´m attaching screenshots of the Facade geometry in 3 versions: the Original Geometry coming from Rhino, that same object with a Subdivide Modifier (Delaunay mode) applied, and a new object Re-Exported from Rhino with higher tessellation. Then you can check the crop render of that facade for each version, in 2 different Displacment qualities (ScreenSize=1.2 and ScreenSize=0.5) to check if quality affects this issue.
To my surprise, I was wrong, and with subdivision, the artifacts are getting more noticeable, instead of less. And, as you may see, the re-exported version of the facade, has more Displacement artifacts than the original geometry. Another weird thing is that, in the higher Displacement quality versions, artifacts are more noticeable (not sure if this is related to some kind Moire effect?).
Now I´m a bit lost here, if anyone can give some tips in how to improve the geometry, or what things to avoid, to make these kind of artifacts less noticeable, that would be welcome.  How do you deal with these kind of surfaces in your proyects?
Thanks in advance

Hi,

Could you provide this scene?

Instructions on how to send ti us can be found in my signature.

Cheers,

Rowan

As I commented when talking about the Atmospheric perspective effect Magenta Tint, I cannot share the scene, due to NDA restrictions, sorry. On top it is super heavy and complex, not a good one for testing these issues. When I find a gap I will try to reduce it to the minimum and at the same time avoid NDA restrictions by taking only a small part of their design.

5
Finally I had the time to render some crops and take some screenshots of the production scene I was talking about when commenting about Displacement Artifacts.

Here I´m attaching screenshots of the Facade geometry in 3 versions: the Original Geometry coming from Rhino, that same object with a Subdivide Modifier (Delaunay mode) applied, and a new object Re-Exported from Rhino with higher tessellation. Then you can check the crop render of that facade for each version, in 2 different Displacment qualities (ScreenSize=1.2 and ScreenSize=0.5) to check if quality affects this issue.
To my surprise, I was wrong, and with subdivision, the artifacts are getting more noticeable, instead of less. And, as you may see, the re-exported version of the facade, has more Displacement artifacts than the original geometry. Another weird thing is that, in the higher Displacement quality versions, artifacts are more noticeable (not sure if this is related to some kind Moire effect?).
Now I´m a bit lost here, if anyone can give some tips in how to improve the geometry, or what things to avoid, to make these kind of artifacts less noticeable, that would be welcome.  How do you deal with these kind of surfaces in your proyects?
Thanks in advance

Honestly; in this circumstance I would definitely not be using displacement, and just modelling the geometry. With floor generator it won't take you long...

I agree that´s the only way that this facade would look perfectly clean, I really like modelling or re-modelling everything with correct topology and UVWs when possible. The issue in this kind of projects is that big and/or complex facades are coming from the architects several times during the process. The design of the facade is not really closed when we are working on the visualization. Let´s say that we receive the facdes once per week or every two weeks... this is the case when I discard re-modelling the facades from scratch or trying to add this kind of relief with floor generator, Railclone or whatever. In this kind of projects we really need some kind of straight Export/import process that is as automatic and quick as possible.
In any case, thanks a lot for your suggestion.

6
Finally I had the time to render some crops and take some screenshots of the production scene I was talking about when commenting about Displacement Artifacts.

Here I´m attaching screenshots of the Facade geometry in 3 versions: the Original Geometry coming from Rhino, that same object with a Subdivide Modifier (Delaunay mode) applied, and a new object Re-Exported from Rhino with higher tessellation. Then you can check the crop render of that facade for each version, in 2 different Displacment qualities (ScreenSize=1.2 and ScreenSize=0.5) to check if quality affects this issue.
To my surprise, I was wrong, and with subdivision, the artifacts are getting more noticeable, instead of less. And, as you may see, the re-exported version of the facade, has more Displacement artifacts than the original geometry. Another weird thing is that, in the higher Displacement quality versions, artifacts are more noticeable (not sure if this is related to some kind Moire effect?).
Now I´m a bit lost here, if anyone can give some tips in how to improve the geometry, or what things to avoid, to make these kind of artifacts less noticeable, that would be welcome.  How do you deal with these kind of surfaces in your proyects?
Thanks in advance

7
Doesn't look like the 2.5D Displacement issues have been entirely fixed.  Using RC3 - see attached

You can clearly see the artifacts in the render and the corresponding mesh in the viewport that has a denser mesh where the artifacts are occuring.

This is disappointing as there is currently no alternative and what I was dreading the most.

The displacement is only subtle and it's on a flat plane, so not even curved geometry this time.

This really really needs to be fixed before you release the final version without the old displacement.

I´m suffering these very same artifacts on a couple of facades in the same scene where I rendered my Atmospheric persp Magenta tint issues. I tried to increase screen size a lot for this object, but it didin´t work. I tried to subdivide my poly object with several subdivide methods, but it just made the issue a bit more subtle. The problem I find is that it is not easy to subdivide an object that comes from external sources with very nasty topology (as this client´s Rhino file). So I end up with uneven distributed triangles density and shape, and Corona 2.5 Displacement strugles with it.

We love using Corona, it is our only engine for production, but sadly, we always complain about Displacement quality internally while we work. Maybe this is just subjective but we feel that Displacement has always been a weak point of this nice software when compared to other engines. I´d love to have an alternative method to avoid using this Displacement method, maybe something like vRay´s 2d displ. or FStorm´s parallax bump (not sure if they would help, but I´d really like to give them a try).

Later I will try to find a gap to post a crop render example of the real production shot, so you can see how visible these artifacts are on my facades.
Thanks

Or Coronas original displacement that pretty much had zero issues except RAM consumption ... but it worked.  Come on guys, you said this was fixed, what scenes are you testing it on?

Yeah, original method was better on this topic, but for us RAM usage was really an issue, so we coulnd´t use that method with enough quality, so the results were poor too. We added 128gb of RAM to all PCs and even with that, we are usually too near to 80-90% of usage. Of course, we are really bad optimizing our scenes, so it is not only Coronas fault ;)

8
Doesn't look like the 2.5D Displacement issues have been entirely fixed.  Using RC3 - see attached

You can clearly see the artifacts in the render and the corresponding mesh in the viewport that has a denser mesh where the artifacts are occuring.

This is disappointing as there is currently no alternative and what I was dreading the most.

The displacement is only subtle and it's on a flat plane, so not even curved geometry this time.

This really really needs to be fixed before you release the final version without the old displacement.

I´m suffering these very same artifacts on a couple of facades in the same scene where I rendered my Atmospheric persp Magenta tint issues. I tried to increase screen size a lot for this object, but it didin´t work. I tried to subdivide my poly object with several subdivide methods, but it just made the issue a bit more subtle. The problem I find is that it is not easy to subdivide an object that comes from external sources with very nasty topology (as this client´s Rhino file). So I end up with uneven distributed triangles density and shape, and Corona 2.5 Displacement strugles with it.

We love using Corona, it is our only engine for production, but sadly, we always complain about Displacement quality internally while we work. Maybe this is just subjective but we feel that Displacement has always been a weak point of this nice software when compared to other engines. I´d love to have an alternative method to avoid using this Displacement method, maybe something like vRay´s 2d displ. or FStorm´s parallax bump (not sure if they would help, but I´d really like to give them a try).

Later I will try to find a gap to post a crop render example of the real production shot, so you can see how visible these artifacts are on my facades.
Thanks

9
Hi,

I easily replicated the Magenta Tint issue in a very simple wood scene, with a very simple building with glass openings.
As you can see, the interior of the building seems to be flooded with some kind of strange dense magenta fog. I will share the scene, but you can easily replicate it in 2min.
I used everything by default, except:
-Altitude: 40m
-Volume Effect: 1.5
-Turbidity: 2, 3, 4 & 6. see attachment.

Thanks

10
Hi everyone,

We have been using some of the Corona 7 daily builds in some of our latest production archviz projects. We are very happy in general with all the nice new Corona 7 features, like the PhysicalMtl, etc. All dailies have been pretty stable for us during these last weeks.

I have just tried the new Atmospheric perspective effect in one of our current complex exterior scenes (urban situation). I find the effect to be really nice, fast to render and effective, thanks a lot for the hard work! However, I´ve experienced some weird behaviours, something like a strong Magenta tint, while tweaking the turbidity values. The most curious thing is that it is super visible through reftactions on glass surfaces. I´m a bit confused, not sure what is happening here. Please find attached some crop redners of one of the shots. I´m using a default CoronaSun and the new "improved" Sky model, with everything by default, except from:
-Altutude: 40m
-Volume effect: 1.5
-Turbidity: variable from 2 to 6 (read each file name to see value)
Hi,

Can you share this scene with us? Instructions on how to do so can be found in my signature.

Thanks,

Rowan

It is not possible, I´m very sorry, for NDA reasons. That´s why I posted a croped version of the original render. Nevertheless, I will try to use another scene, and replicate the issue, specially the refraction tint. and share the scene.
Thanks for your effort.

11
Hi everyone,

We have been using some of the Corona 7 daily builds in some of our latest production archviz projects. We are very happy in general with all the nice new Corona 7 features, like the PhysicalMtl, etc. All dailies have been pretty stable for us during these last weeks.

I have just tried the new Atmospheric perspective effect in one of our current complex exterior scenes (urban situation). I find the effect to be really nice, fast to render and effective, thanks a lot for the hard work! However, I´ve experienced some weird behaviours, something like a strong Magenta tint, while tweaking the turbidity values. The most curious thing is that it is super visible through reftactions on glass surfaces. I´m a bit confused, not sure what is happening here. Please find attached some crop redners of one of the shots. I´m using a default CoronaSun and the new "improved" Sky model, with everything by default, except from:
-Altutude: 40m
-Volume effect: 1.5
-Turbidity: variable from 2 to 6 (read each file name to see value)

I'm seeing the same thing, anything above 3.5 turbidity seems to add a really strong magenta tint.  Also the altitude setting, when the tooltip refers to 'observer altitude' - not quite sure what this is referring to or how to implement it correctly for the correct effect

Just to add to this, I think it would be worth while for a dedicated helpdesk article on using the aerial perspective and what effect the settings can have, I'm seeing some quite heavy colour shifts just from changing the altitude value, currently feels like it's going to take quite a lot of trial and error fiddling with turbidity and altitidue to get the right effect.  Green and Magenta tints are really obvious in different parts of the image depending on what you change.  For example, 1.5m altitude has relatively neutral haze, but magenta in higher up shadow areas, change the altitude to 10m, the magenta is more subtle but the haze is now shifted heavily towards green/yellow, change to 25m and the colours swap over.  And it gets weirder if you throw turbidity changes into the mix too.  I don't currently have time to render examples etc, but will if I get a chance

I just wanted to point out again, that the most weird thing that I found while playing a lot with all parameters is the heavy refraction tint seen on cars and facades window glass surfaces. If you check my Refraction Pass iamge, you will notice that most of the heavy Magenta Tint the beauty result has is coming from the refraction of those elements. This really doesn´t make sense in my head, atmospheric perspective effect should not be stronger inside cars and rooms than outside, right?

12
Hi everyone,

We have been using some of the Corona 7 daily builds in some of our latest production archviz projects. We are very happy in general with all the nice new Corona 7 features, like the PhysicalMtl, etc. All dailies have been pretty stable for us during these last weeks.

I have just tried the new Atmospheric perspective effect in one of our current complex exterior scenes (urban situation). I find the effect to be really nice, fast to render and effective, thanks a lot for the hard work! However, I´ve experienced some weird behaviours, something like a strong Magenta tint, while tweaking the turbidity values. The most curious thing is that it is super visible through reftactions on glass surfaces. I´m a bit confused, not sure what is happening here. Please find attached some crop redners of one of the shots. I´m using a default CoronaSun and the new "improved" Sky model, with everything by default, except from:
-Altutude: 40m
-Volume effect: 1.5
-Turbidity: variable from 2 to 6 (read each file name to see value)

13
[Max] I need help! / Re: Realistic thatch roof
« on: 2020-09-28, 16:25:57 »
thanks GeorgeK,

AO... sounds like a nice tip, i´ll give it a try.

14
[Max] I need help! / Re: Realistic thatch roof
« on: 2020-09-28, 12:44:07 »
Hey, we have done some testing going the Scattering Way. It is just a simple Vertical Garden preset from iToo FP, just a few tweaks like scaling it down and making it denser. It has a very simple CoronaMtl that has some translucense and very soft anisotropic reflections.
It is starting to look much closer to real life refs. What do you think?
Thanks

15
[Max] I need help! / Re: Corona 6 W O W
« on: 2020-09-26, 13:42:21 »
+1
This issue of the colorpicker maybe isn´t a big deal of course, but for me personally is always really annoying not to be able to pick colors from the other screen. That would we very nice if it could be fixed for the next version of it, the "faster color picker" that I saw in the Road Map.

Pages: [1] 2 3 4