Bug RTR StackLit clearcoat approximation base layer lighting hierarchy issue

The current RTR StackLit clearcoat approximation seems to use the Sky instead of Reflection Probe for the base layer lighting hierarchy, which results in bright materials in the dark.

What I've tried
- Lighting & reflection probes baked with the lights disabled
- Switching the ambient mode from static to dynamic did not make a difference
- Increasing the light cluster range did not make a difference
- Switching RTR fallback hierarchy settings did not make a difference
- Switching between RTR performance/quality did not make a difference
- Increasing RTR bounces from 1 to 2 did not make a difference
- Disabling "Dual Specular Lobe" did not make a difference
- Disabling "Coat" on the material looks correct, as expected
- Increasing the HDRI Sky exposure, makes the problem worse

Further testing
Recently, PR #6107 and #6103 by slunity changed the behavior of StackLit clearcoat + RTR. To make sure this is not a newly introduced regression, I checked out on commit df37bc5415b3deee5c6bacaf6ecf2dc2077c69e0 (November 22) before those PRs were merged, and the issue was already present there.

Extra information
The material used is CarPaint_DarkRed from the Measured Materials Library.

Example

Lights on



Lights off



I can submit an official bug report through fogbugz if needed.

Thanks,
m0nsky

 

I created a new 2022.1 project for the bug report and noticed the issue wasn't present there, so I went back to investigating a little more. (issue is present in an older project)

I recreated the lighting settings, HDRP asset and global settings in that project and wasn't able to reproduce it anymore. I think this can be closed for now as it doesn't seem to be related to the package/shaders.