fix kernel spectral densities

[sbfnk]

Description

This PR closes #837

We'll get there in the end.

Initial submission checklist

• My PR is based on a package issue and I have explicitly linked it.
• I have tested my changes locally (using devtools::test() and devtools::check()).
• I have added or updated unit tests where necessary.
• I have updated the documentation if required and rebuilt docs if yes (using devtools::document()).
• I have followed the established coding standards (and checked using lintr::lint_package()).
• I have added a news item linked to this PR not done as it's a fix to the previous one.

After the initial Pull Request

• I have reviewed Checks for this PR and addressed any issues as far as I am able.

[seabbs]

(reviewed the maths in the discourse mainly)

[github-actions]

This is how benchmark results would change (along with a 95% confidence interval in relative change) if 328d6c7 is merged into main:

• ✔️default: 23s -> 23.4s [-8.38%, +11.53%]
• ✔️no_delays: 36.8s -> 26.6s [-83.53%, +28.48%]
• ✔️random_walk: 9.78s -> 9.58s [-9.09%, +5.12%]
• ✔️stationary: 13.7s -> 13.6s [-14.97%, +12.99%]
• 🚀uncertain: 41.3s -> 37.8s [-17.12%, -0.12%]
  Further explanation regarding interpretation and methodology can be found in the documentation.