[eric nixon]

That pattern shows that the normmal map has banding (contours) in the gradient despite being 16bit, try a massive blur on it, + a lower normal height.

[dariolanza]

By the way, give it the try to the Displacement approach: Subdivide the plane a bit, and apply that normal map as a Displacement map instead (remove it from the normal map slot).

Keep the Subdivision value low (6 is a good starting point).

Show us the results.

Cheers

Dario

[lifeofdave]

That pattern shows that the normmal map has banding (contours) in the gradient despite being 16bit, try a massive blur on it, + a lower normal height.

True, it needs a bit of a blur, next time will look out for this..

By the way, give it the try to the Displacement approach: Subdivide the plane a bit, and apply that normal map as a Displacement map instead (remove it from the normal map slot).

Keep the Subdivision value low (6 is a good starting point).

Show us the results.

Cheers

Dario

Ok, first time using using displacement so still need to read up about it properly, but did some quick tests..



I'm struggling to get the displacement method to have any impact in terms of warping the reflected light, even though it looks like it should work (for the 40% one at least) judging by the distortion in the material preview scene. I've tried a height of both 40% and 100cm, both give a the same result.

[lifeofdave]

This is how the normal map method material appears in the preview window:

[Bubbaloo]

Make sure to use "pretesselated" displacement. If not, you will need to subdivide your geometry first. Start with a low subdivision level first, and increase as needed. I'd test it with a direct view of the displaced object, instead of trying to gauge the displacement result by the caustics pattern it produces.