HDR artefacts problem

So, some time ago we ported PBR (from jme repository) in our project, but there we faced with not cool problem:

HDR images show dotted artefacts as you can see here:

This problem occur in some HDRI-s, but some work pretty stable (it is 2 different HDRI shot in same place with same camera in different light conditions):

This one is really fancy:

I am intended to make my own HDRI in blender (particularly), but HDRI from blender in jme looks really unstable. I dont know what occure this problem, other softwares show all of my HDRI-s correctly.

Here is HDR photo, which we have tested and which has problems:

Also, is there any chances to see .exr support in jme? EXR seems to be more precise, and it is also good open alternative, I guess. OpenEXR - Wikipedia

up

Uhm, you “ported” a WIP extension supposedly into your own build of jME, provide no code and expect somebody to know what the issue is…

We use PBRisComing branch. You can reproduce this problem using TestPBRLighting.java, which contains in this branch. You can replace this line on these examples

Spatial sky = SkyFactory.createSky(assetManager, "Textures/Sky/Path.hdr", SkyFactory.EnvMapType.EquirectMap);

for testing our case.

I’ll look into it. I’m not home right now

You are always helping out! Thanks :smile:

up

I cannot reproduce the issue as you described with TestPBRLighting.
Also please don’t bump your threads.

I had something similar to the latest screenshot: a storm of red green and blue pixels on a pbr surface.
Does this sound related to your problem? I can’t really tell from the picture, since, there, it seems that the dots are floating or on the screen, while in my case they were on the texture…

When it happened i didn’t care too much since i was working to something else, also note that it was a really old version.
In few days i will probably work with the new pbr and then i will eventually add more informations.

Could be a GPU specific issue. What video card are you getting these artifacts on?

Nvidia geforce 750M with 361.28 proprietary driver