it couldn't possibly ever work, at all, for several reasons.
stereo width processors that work by mid/side processing have to convert the audio to mid/side from stereo first, before processing it to 'enhance' the stereo or mono signals, then they convert it back to an l/r signal for the final output.
so most of the 'magic' of a stereo enhancer is coming from what is actually simple processing done to the signals after they have been converted to m/s. mostly it's just simple volume adjustments to either the stereo difference and/or mono signal. other things might be in there too, like the portico stuff has some phase controls.. but it's a phase control. is this or that phase control going to be worth sampling? it's a phase control.
so basically the 'enhancement' is coming from a volume control, or maybe also a phase control. so even if you could sample the effect, you would be sampling a volume control. the 'magic' is in that it's done to a mid/side signal instead of an l/r one.
the actual mid/side conversion (and back) itself that these things do, simply cannot be sampled. it just can't. it would be like trying to sample a process like reversing a track in your DAW. it's a process. NAT can sample tone, dynamics, and harmonics, not complex processes like that.
furthermore, the tones used for sampling are sine sweeps, and are identical between left and right channels. so they are mono. if you ran that into one of these enhancers, it wouldn't even have any stereo 'difference' signal to work with itself. but even if it somehow did, and you got out a stereo enhanced set of tones, loading those into nebula would not re-create the process of converting any audio coming into it into m/s format, applying whatever processing the hardware did, then back to l/r.
it'd basically be magic if that happened.
it'd be like if you could 'sample' the act of taking the sine sweep tones used in sampling, converting them to another sample rate, processing them somehow, then converting them back, and you expected nebula to actually convert the sample rate of any audio coming into the program made from that.
a typical chorus effect is another thing altogether, and if there is any 'stereo enhancing' going on there it's because the different channels (l/r) are being delayed differently from one another.
Thanks for the clear explanation man. I understand now. So if chorus can be sampled with NAT, why are there so few? I use chorus on so much stuff. I wonder why it hasn't been more developed on the NEb side.