It is difficult to make a comparison between digital colors and real paints; this could be tricky because of different settings of monitors and of the impossibility to see a painted chip and an image on the monitor under the same light.

I've made in this way:
1-I've painted chips of Humbrol paints on a card;
2-I've scanned them aside some pages of a Federal Standard catalogue, on the same file; whatever is the setting of the monitor, the resemblances or differences of chips are preserved;
3- I've inserted some scans from the Akan catalogue, after having checked the good match between one chip (AMT-4) and the closer FS known on other sources (FS-24102, 24151 by Orlov; the shade is identical to 34102, 34151 shown here, but satin); unfortunately I don't owe AKAN paints; I've also inserted my chips that, wherever is possible, are made according to AKAN paints and FS by Orlov, that are in good agree between them;
4- I've compared the colors by copying, pasting and moving chips.
No doubt, there could be some small error in this procedure, but I think that this is small when compared to the uncertainties in deducting the original shades by 70-years old samples.
 
 

Here the chips of Soviet colors are compared to the closer Humbrol matches.
 
 

The khaki utilized before 1937 is similar to 155, just a bit darker. The AII green, when new, could have been similar to 80, but it darkened quickly to 151 or even darker shades. The A-19f , when new, could have been similar to 80, but faded quickly to a lighter shade. The 3B was similar to 66 or, if we suppose that the original shade was more green, to 114. The army yellowish green 2Le hasn't equivalent and should be obtained by mixing sand and green. The army ight green 22lg was similar to 120, nut a bit more emerald-ish. The army green 4BO was similar to 150. The supposed Spanish green, perhaps a variant of AII green, is similar to 151. The AMT-4 and A24m ant their glossy equivalents AGT-4 and A-24g were similar to 151 when new, and lightened becoming similar to 80.