I know also the 5400 have been replaced by a led one
(5400 II) and about Kodachrome
Do note that 5400 II has a *single* while LED while Nikons have 3
individual RGB LEDs. That's a very important distinction for a number
of reasons! The Nikon approach is far superior (better color purity).
To complicate the comparison the superior Nikon's light source is
countered by Minolta having higher resolution (5400 vs 4000) and
higher bit depth (16 vs 14).
However - and as always - there's a catch and you identified it:
Kodachromes! Generally, they are a nightmare to scan with Nikons. The
"Kodachrome" mode (I have an LS-50 a.k.a. Coolscan V) just doesn't go
far enough. When you run the A/B tests do try *dark* Kodachromes
because they challenge Nikons the most!
ICE doesn't work with Kodachromes, in general. However, it does work
on slides which are overexposed i.e. all the silver has been washed
out, or the uniform, low contrast, areas of the image. But any bits of
silver left in the film confuse ICE as does a transition from bright
to dark where dark, silver loaded areas, are mistaken for debris.
A time consuming approach may be to scan twice and then manually
"paint over" the ICE image with data from a regular scan to correct
areas where ICE fails. Sounds messy but beats having to correct the
*whole* image.
Don.