Wavelet sharpening is essentially multiscale sharpening.

A simple way to understand it is:

1. Start with the image.
2. Make a blurred version of it with a low-pass filter.
3. Subtract the blurred version from the less-blurred/original version. That difference is a high-pass layer containing detail at a particular scale.
4. Repeat this with progressively stronger blur levels to create several detail layers, from fine to coarse.
5. Adjust those layers, then recombine them to form the final sharpened image.

So the RegiStax layers correspond to different detail sizes: lower layers typically affect the finest details, higher layers affect broader structures. Increasing a layer boosts contrast in details of that scale; reducing noise controls helps suppress noise that also lives in those frequencies.

The exact slider names are software-specific, but conceptually they are changing how strongly each frequency/detail band contributes to the reconstructed image.

A good practical approach is to:

• start with the finest-detail layers,
• increase them gently,
• watch for halos and amplified noise,
• then adjust broader layers if needed for larger features.

In short: wavelet sharpening is like splitting the image into several high-pass detail bands at different sizes, editing each band separately, and adding them back together.