In URP, the prepass layer mask can’t render different layers with different camera settings. Given that limitation, your current setup behaves as expected:
- The depth prepass correctly includes the Character layer.
- The ScriptableRendererFeature Render Objects feature correctly renders the Character layer after Opaque using user-defined camera settings.

However, because SSAO relies on the depth buffer produced by the depth prepass, it will produce visual artifacts on the Character objects rendered with a different FOV.

Recommendations:
- Disable SSAO if it isn’t required for the scene or target platform.
- Add an additional Render Objects pass for the Character layer that writes to the depth buffer before the URP depth prepass, using camera settings aligned with the Character rendering (e.g., different FOV). In this case:
* Remove the Character layer from the prepass layer mask.
* Be aware this approach increases overdraw and requires enabling Depth Priming.
* Note that Render Objects does not write to the normal buffer, so SSAO will not correctly affect Character objects.
- The preferred solution is to implement and inject a custom depth–normal prepass prior to the standard URP prepass to render Character objects into both the depth and normal buffers with the correct user-specified camera settings. This ensures SSAO considers the adjusted FOV and applies correctly to the Character layer.

Thanks for reporting this, we will take it into account for our roadmap and see if we can facilitate this use-case.