Using smart:comp 2 / 3 as a Trackspacer-style spectral sidechain?

[romeosoroka]

Hey guys,

Could someone explain the correct way to use smart:comp 2 / smart:comp 3 as a Trackspacer-style dynamic spectral ducking tool?

I understand that I need to enable external sidechain and use spectral compression, but I’m a bit confused about how threshold and ratio should be approached in this scenario.
Should they be treated more like traditional compression, or more like depth/intensity controls for spectral ducking?

Also, is using spectral compression with an external sidechain functionally the same as learn:unmask, or are these fundamentally different processes under the hood?

Additionally, do these modes rely on the same underlying spectral processing algorithm, or are they based on different spectral models altogether?

Thanks in advance.