You can replicate Kodachrome very well using 3 color shots on B&W and then using a reversal C/M/Y development scheme. This has been often published. At the end, you bind all 3 together to get a finished slide. There are no tricks and no need to compound special formulas, you just make up what works. Leadly and Stegmeyer published this in the '40s.
And Rudi, I am aware of these two types of endeavor but what I am saying that there is a LOT of talk here and across APUG but little action among those able to do it. They talk, but they don't do.
What you refer to is colour photography by means of taking colour seperations in succession. This is a process going back to the beginning of colour photography and was a important process back then. Look for the work of Adolf Miethe.
The Library of Congress' website is down until Monday, it seems, but if you Google "LOC color photos of Russia" you'll find several links to color photographs of Russia made in the late 1800's using a tri-color process. Digital technology (which can be useful, after all) allowed the Library to scan all three single-color images then very accurately combine the three layers. (I read all this long ago, so I hope I'm remembering correctly.) It's wonderful to see a world that we normally only see in b/w images as being as fully of color as our own.
Color filters can influence the quality of the result. The details are very complex and would require an hour or so of typing and pasting graphs and charts to show the details.
Yes, I was just thinking that maybe the filtration layers would create a different look in the end product compared to a color separation using regular B&W film, and that maybe that would preserve a "Kodachrome look".
The "ditch it" option is the best.