Back to the Shapespark home page

Mirror is reflecting wrong Reflection Probe

Hello, I have some scenes with reflection probes on each room, but every time I have to make a mirror reflect, it reflects what’s on a different probe, not the one in which box the mirror is located.

I noticed it after the last update, so I don’t know if it’s relative to that.

Thank you

The light probe bounding box is not taken into account for object to light probe assignment. The mirror should use the closes light probe from which the mirror is visible. If mirror is not visible from any light probe position, the closest light probe is used.

Perhaps a light probe that you would like to be used by the mirror is not assigned to it, because there is some wall or other object between them blocking the visibility? You can try to move the light probe a bit and see if it fixes the problem.

Could you send us a link to the scene where you see this problem (here or via email to ?

@jan, will it be possible that we can have cubic light probes in addition to the spheric light probes? Twinmotion uses this system for reflexes in the renderings and I see that it has both types of light probes, spherical and cubic. The cubics allow a non-distorted reflection on windows and mirrors or on flat reflective surfaces, makes a considerable improved in the reflected image in that types of surfaces.

Actually Shapespark has cubic light probes only, they are just shown as spheres in the editor. In Shapespark they light probes allow for non-distorted reflections but only of things that align with the light probe bounding box (walls, windows, etc.), the geometry inside or outside of the bounding box is distorted. Do you see that with Twinmotion such geometry is also non-distorted?

You can see the diferent type of distortion beetwen them in this video
We try here to make a parallel reflection in mirror but is not very realistic

@jorgearq, in this particular scene the inaccuracy of reflections is caused by the limitation of light probes irrespective of their type (cubic or spherical). To produce realistic reflections of objects at different distances (eg. a standing person in the foreground, and a wall in the background) full 3D information is necessary. However, images captured by light probes are like panoramas - they cover all the surrounding but they are only 2D - so they do not produce enough information for fully realistic reflections.

This is true for light probes in general, so a similar scene in Twinmotion will suffer from similar inaccuracies and distortions of light probe based reflections.

Ok, I got it, thanks :+1:t3: