LuxCoreRender Forums

Hi,

I've just recently installed the new LuxCoreRender 2.0. I was previously using 1.6 a bit, and there is one thing I can't seem to get working in 2.0. I also didn't find any info on this matter that helped me, so maybe someone here knows whats wrong.

I'm trying to render the following setup:
- One beam light (spot or laser, doesn't matter)
- one mirror
- a homogeneous world volume, to make the beam visible

In 1.6, this worked nicely, in 2.0 not really. The beam reflected off the mirror is not visible. I attached two pictures that should show the difference.

I can't really figure out what I'm doing wrong. All settings are pretty much standard. No difference if I set only a world volume, or also an external (+internal) material to the mirror. No difference between mirror and glossy shader either.

Thanks for any help,
Johannes

Seems like the attachments go missing if there are two dots in the file name...
Should work now.

v2.0 contains a bug in the camera's volume auto detection, if I remember correctly.
You could use v2.1beta3 instead.
Or you have to specify the correct volume in the camera settings manually.

Unfortunately, that didn't work. Still the same image with v2.1beta3.
If i maually specify the camera volume also. If i select anything other than the world_volume the scattering disappears completely, so it seems to be auto-detecting it. Must be something else.

Can you upload the scene?

ok, I just made some progress actually!

Attached is a test file I set up to have something simple and from scratch. Only a light, mirror and camera (and a file size that can be uploaded...)

After all, I just got it to work by choosing BiDir instead of path tracing.
I remember trying that with v2.0 yesterday, where it made no difference.

I have access to my machines at this time, so I cannot check your attachment.
But I observed some things you may have to consider:
- use clear volume with ior of air for camera and world
- set a box with solid frames aoround your optics. you may use solidify and apply it to basic cube.
- define material of the box as follows:
— internal volume homogenous,
— external volume clear like the world volume. easiest way is a pointer to the world volume.
- now use view port (shift z) to adjust scattering of homogenous volume,
— first with bidir then if you like the result
— switch to path.

in geneneral if you use volume define solid objects, what meanss they have walls with thickness or are thick theirself, say they have volume. Avoid planes or hollow objects with plane walls.

I did try your suggestion, but again that makes no difference.

- I set up a clear and homegeneous volume
- set world volume clear
- Create a light, mirror and camera
- Create a cube. Place mirror inside cube, light and camera
a) outisde and
b) inside cube.
- assign clear volume to cube external, homemgeneous to cube internal
-- Result: same as before

When using Bidir, it again workls as expected. Scattering starts and ends at the cube boundaries. The mirror works.

When using Path, scattering is seen between the light and the mirror, starting from the light source. I.e. the cube is ignored completely.

CodeHD wrote: ↑Tue Dec 11, 2018 8:05 pm When using Path, scattering is seen between the light and the mirror, starting from the light source. I.e. the cube is ignored completely.

It is just normal and expetect, path tracing is not able to render caustics of not intersectable light sources (like point, spot, lasers, etc.). If you replace the light source with an area light, something will start to appear but it will still very hard to render for a path tracer. For this kind of renderings, you must use BiDir.

Ok yes that again helps getting to the bottom of things.

A quick check reveal that 1.6 (using the experimental luxcore as I did before) defaults to bidir. 2.0 defaults to path.
I hadn't put much thought now into the different raytracing approaches (I know caustics are handled better in some than others, but I'm not an expert on that).
Something else that made me not realise this before is the fact that the 1.6 bidir render looked very similar to the 2.0 path - out of the box at least. But that might be an issue of how I set the lights up for 2.0 after switching...