LuxCoreRender Forums

kintuX wrote: Wed Jun 12, 2019 9:12 am Still appears to be something present...

The prism scene has some very tight light beams coming from corners and some partial sub reflections / refractions, but it looks pretty okay with BiDirVM.

Fox wrote: Wed Jun 12, 2019 6:01 pm

kintuX wrote: Wed Jun 12, 2019 9:12 am Still appears to be something present...

The prism scene has some very tight light beams coming from corners and some partial sub reflections / refractions, but it looks pretty okay with BiDirVM.

With BIDIRVMCPU still looks the same to me. Same glitch.
Scene (for Standalone):

kintuX wrote: Wed Jun 12, 2019 9:12 am Still appears to be something present...
RAW render
Prism2CF_LC2-2b_RAW_VolumeGlitch.jpg

CONVERGENCE mix with SAMPLECOUNT
Prism2LC 00001_CONVERGENCE over SAMPLECOUNT.jpg

Scene
Prism2LC.blend

I took your file and changed Metropolis to 50%,128,3%.
The result did not show any glitch.

FarbigeWelt wrote: Thu Jun 13, 2019 6:17 am

kintuX wrote: Wed Jun 12, 2019 9:12 am Still appears to be something present...
RAW render
Prism2CF_LC2-2b_RAW_VolumeGlitch.jpg

CONVERGENCE mix with SAMPLECOUNT
Prism2LC 00001_CONVERGENCE over SAMPLECOUNT.jpg

Scene
Prism2LC.blend

I took your file and changed Metropolis to 50%,128,3%.
The result did not show any glitch.

Nope. Bit less pronounced, yet still glitchy... please, prove with an image next time - is worth more than a thousand words.
Thanks.

Anyways, until something drastically changes in code (for volumetrics), i give it a rest.

kintuX wrote: Fri Jun 14, 2019 5:10 pm
Nope. Bit less pronounced, yet still glitchy... please, prove with an image next time - is worth more than a thousand words.
Thanks.

Anyways, until something drastically changes in code (for volumetrics), i give it a rest.

Please have a look at the images- Maybe I have changed something else but I don't remember anything relevant. More than thousand words... Sure? One Mpixel image uses much more than thousand words (=64 bit or 8 Bytes) space.
Human imagination can produce wonderful things in mind with just a few words input. Parroting phrases can make nonsense. Too often an image does not mean anything if it misses a short description. But a simple sentence can be explicit. Think about it.

However, your file Prism2LC is not set up properly for volumes. I don't know to what the missing external volume of the prism effects. And the mirrors have not any volume defined at all. I have switched your env volume to external volume and added a heterogeneous volume with scattering of 0.001 to the prism. Due to your answer I have download your file again and it is right, the adjustment of Metropolis did not help against the glitch. I am currently looking for the change that fixed the issue. (I should keep track of changes the next time.)

FarbigeWelt wrote: Fri Jun 14, 2019 9:15 pm Please have a look at the images- Maybe I have changed something else but I don't remember anything relevant.
Prism BiDir Metropolis denoised 640 samples bevelled.pngPrism BiDir Metropolis raw 640 samples bevelled.png
More than thousand words... Sure? One Mpixel image uses much more than thousand words (=64 bit or 8 Bytes) space.
Human imagination can produce wonderful things in mind with just a few words input. Parroting phrases can make nonsense. Too often an image does not mean anything if it misses a short description. But a simple sentence can be explicit. Think about it.

Word - purely abstract. Thought - simply subjective. Image - just objective. All in all, totally relative.

FarbigeWelt wrote: However, your file Prism2LC is not set up properly for volumes. I don't know to what the missing external volume of the prism effects. And the mirrors have not any volume defined at all. I have switched your env volume to external volume and added a heterogeneous volume with scattering of 0.001 to the prism.

Not my scene, put no effort into it. Only touched it to please my wandering eye, but to my surprise, minor problem appeared. Or so it seemed - well, so i thought.

FarbigeWelt wrote: Due to your answer I have download your file again and it is right, the adjustment of Metropolis did not help against the glitch. I am currently looking for the change that fixed the issue. (I should keep track of changes the next time.)

Took time to go into to details. Yes, scene is an abomination for pro-creativity
After couple of hours analyzing & diagnosing all posible variations i found it!!! GOT NOTHING TO DO WITH VOLUMES !!!.
It's the freakin' laser beam! Sometimes it shows as visible, but sometimes not. Depending on angle, wattage and efficacy, such a glitch can occur.



& That's all folks!

Thank you for investigating this issue.

Please open bug reports in https://github.com/LuxCoreRender/LuxCore/issues/new with the following:
- short description of the problem
- scene showing the problem (.blend or, even better, result from LuxCore scene export in text format)
- render output images showing the problem
- a description of the steps to make to replicate the problem

kintuX wrote: Sat Jun 15, 2019 10:11 am Took time to go into to details. Yes, scene is an abomination for pro-creativity
After couple of hours analyzing & diagnosing all posible variations i found it!!! GOT NOTHING TO DO WITH VOLUMES !!!.
It's the freakin' laser beam! Sometimes it shows as visible, but sometimes not. Depending on angle, wattage and efficacy, such a glitch can occur.

Today, I played around again with this as well (hadn't followed the details of this thread very well teh last few days), and after installing the new 2.2beta2.

I see the same glitch, and it also appears to be conencted to the laser beam size, and the camera axis. Based on what I see, I also believe it to be caused by the light-rays rather than the camera rays. For very wide lasers, the direct beam appears to decay quickly. After another object (e.g. a glass lens), the "decayed" suddenly reappears at a mor realistic intensity. The dark band may still appear, depending on how the beam is being focused.

I have made a simple file to show the problem. I can open a GitHub issue as well.

Being late to the party won't stop me reporting my own bug version

It looks like after passing through a refractive body reflection does not work properly:
When I activate the mirror in viewport render only the leg of the beam reflecting off is rendered but the part between prism and mirror suddenly disappears
Edit: Solved by CodeHD, missing volume assignment for mirror, no bug!

epilectrolytics wrote: Sun Jun 16, 2019 12:53 pm looks like after passing through a refractive body reflection does not work properly:
When I activate the mirror in viewport render only the leg of the beam reflecting off is rendered but the part between prism and mirror suddenly disappears

Have you checked if all your interior/exterior volumes are set up correctly? Especially check the camera volume. Usually this is the problem.