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.

Topics - v.p.vlasenko

Pages: [1]
[Max] Bug Reporting / PhoenixFD Ocean texture + Corona decal
« on: 2021-12-16, 23:28:25 »
The height parameter in ocean texture set to vector mode is ignored if corona decals are added to this object.

All simple, check the images.

Top view, corona camera, not orthographic. Render with default corona material as override and volume effect 10 - increased to see result better.
The second image with greatly increased contrast in PS to see where that volume effect produces the strongest result.

Additionally - not sure but seems like in some situations it produces the wrong color, like in a clean scene with sun and default corona material as override - adding volume effect gives some red/purple tint to the scene.

The above is true for at least a few latest DB.

And few questions:
1. aerial perspective still not working with volumetrics, will that be fixed in some nearby future?
2. is it possible to make it true volumetric (checkbox to switch between shader and volumetric)? Or maybe it'll be even if we can plug some map into volumetric material scattering/absorption.

1. Black edge where object with refractive material contacts other geometry - image attached.
2. Setting opacity to 0 makes fog like material.
On the second image water material have opacity 0, but underneath surface have blue color.
The third image is a preview but the sphere is still barely visible.

Both problems appear on physical and legacy materials.

Corona V7.

I have few materials. When applied to object separately speed ranging 7-14M. When mixed in multilayer material with corona distance - speed is 700k +/-

Same problem with opacity. I have a landscape model + a lot of buildings with pools. Bottom of pools below ground level - instead of cutting geometry I used corona distance to cut it during render - speed dropped twice.

Maybe that's correct behavior with opacity, but 10 times slowdown with mixing materials looks strange.

Nothing to explain here - moving vdb or changing settings very slow, looks like reload each time

And the problem with the visible bounding box still here - with global volume or inside/intersecting with another object with volume mtl.

example of PhoenixFD ocean

This could be solved by switching to 3D displacement previously but not anymore

Same weird result I had with prosimplex.

The problem looks similar to already posted here by mahjoub but I created a new topic because of:

1. It happens for a couple of years already and I'm using the latest daily builds, so the problem here for a long time.
2. Happens in completely (almost) different scenes.
3. Not looks like it bound to resolution.

Almost different is because those scenes have something similar - a lot of objects and a lot of materials (and big scale, but I don't think that's the problem). It never happened with productviz scenes and some small archviz like couple of buildings or some interiors.

So the problem: if I run IR and moving or creating objects, moving camera - all fine most of the time, let's say 95% it works fine - I can stop IR and all good. BUT, if I do changes to some materials - it works fine until I stop IR, then max becomes locked for a long time (5-40 min). And again that not 100% cases but it almost for sure, that after doing changes it'll lock max.

Sometimes that happened if I even didn't make any changes, so just run IR and stop it after some time lock max as well, same for regular render - but that happens rarely.

The problem occurs even with isolated objects, so hiding everything except one certain object will behave same way, so I have to move that object to another scene to work with it and then back.

Another interesting thing - sometimes I still can work for a couple of seconds after closing IR and then max became locked.
At some point in those scenes I stopped using IR and did regular test renders instead with cancel at the end, not stopping. Can't say for sure, but seems like cancel work better - I can't remember freeze after cancel, but can't guarantee that.

And one more thing I tried - selected all objects in scene (seems like there were around 12k of them) and applied one material to all. IR seems to be running ok after that. I did 3-4 runs, applied some materials to few objects during IR and seems like all fine. Again - can't say for sure if the problem eliminated, as it does not occur in 100% of cases.

Summary of the bug: scenes with high object and materials count, freeze occurs mostly after doing changes to materials during IR, but sometimes it happens without any changes and with regular render. PROBABLY reason is a lot of materials.

P.S. I'll create few more topics with bugs that annoy me for a long time and with newcomers found recently. That about 2,5D displace quality, openvdb and corona distance.



The problem is simple - adding CShading_Caustics in photoshop produce completely different result than on raw render.

Doesn't matter which mode I select - the result is different. I tried screen and linear dodge (add), but both at 100% produce too bright (more like atomic explosion) result.

Renders in png 16 format, and working in 16 bit.

Any suggestions?

[Max] Bug Reporting / Caustic solver not working properly
« on: 2020-05-22, 16:59:01 »

I've encountered a problem in one of my projects and created a simple scene to figure it out.

What we have: plane 1000x1000 with default corona material, a simple tube with chamfer (highly reflective material) and corona light.

For first renders I've reduced max photons/iteration to 1k (5000k default) to see what happening GI/AA - 4. All renders are caustics render element.

1. tube radius 190 - everything fine.
2. tube scaled down, all good.
3. scaled down, all good(seems like, but not sure) had to increase photons to the default value.
and here the problem:.
4. scaled down to radius 2,2. Camera much closer and photons increased to 50000k - with default value I can't see caustics. Saved after 2 passes.
It looks like in this scenario no new photons generated after 2nd or 3rd pass - I can't see changes in photons distribution. And with each pass caustics become dimmer.
The second render is the same but at 50 passes. After some time caustics will disappear at all.
Now the even weirder thing. @ pass per render.
5. Same camera position as previous, tube scaled down. Caustics disappear much faster, it seems like they will go after 5 passes (don't remember now).
Next image - everything the same as in previous except camera - moved it closer. Now caustics not visible and disappear very fast. For this render I increased EV to 6 instead of 3.

So what I've found - at some point (size of objects, size of lights, the distance between them) caustics disappear over time instead of refining (looks like no new photons generated after few passes). And even weirder - it depends on how close the camera to those objects.

The scene I've written about at the start - a kind of studio render with some reflective objects. When the camera sees all scene - I have caustics that will not converge over time but disappear instead. But when I move the camera closer to see details - I don't have caustics at all or it disappears very fast.

I'm using last daily build - but I checked couple previous builds and v5 - the same problem. All of this true for default setting as well.

And the last one render - that seems like 20-30 min of render on the simplest scene, low res and 18cores... caustics still noisy. Is that normal?

Hi, I encountered such problem:
CoronaLayeredMtl, simple base material with diplacement (3ds max noise or prosimplex). As far as base material applied directly to object or CoronaLayeredMtl without any other layers - everything fine, I have correct size and look of displaced noise 0,005 (pic 1). But if I add any material to layer 1 (in this example same material as base but without displace) - noise looks different - bigger and less detailed.

I didn't checked other procedural materials, just noise and prosimplex, both use Object XYZ mapping.

And I just remembered - seems like I had same problem with PhoenixFD Ocean texture, did a fast check - it lost a lot of details when used in CoronaLayeredMtl.

pic 3 - only base material inside CoronaLayeredMtl, CoronaDisplacementMod used with phoenixfd ocean.
pic 4 - same material connected to base and layer 1, CoronaDisplacementMod used with phoenixfd ocean.

So CoronaDisplacementMod not helping as well.

[Max] Bug Reporting / Glossy reflection+refraction bug
« on: 2017-09-21, 12:21:51 »
Hi, not sure if that happens only in latest build, but here's what we have - wrong reflection with reflection glossiness below 1 and refraction enabled.

[Max] Resolved Bugs / Non 1 refraction glossiness change IOR
« on: 2016-07-16, 12:22:09 »
Probably I wasn't clear enough in first description or that happened due to 2 problems reporting at once... anyway - denoising crash fixed while glossiness bug aren't.

Seems like non 1 refraction glossiness without caustics change IOR. With caustics everything fine.

That visible in material editor as well.

1. Glossiness bug.

3 files attached.
-Glossiness 1 (correct)
-Glossiness 0.99 (wrong)
-Glossiness 0.99+caustics (correct)

Seems like that happens in all 1.5 daily builds.

2. Denoise + render elements active (AO, Mask, Beauty) cause crash at least in latest daily build

Both happened in more complex scene but can be recreated in such simple (plane, box, sphere, phys camera and white background).

Pages: [1]