I noticed that if I set the repeat rate to 2 I actually don't get an envelope output frequency that's twice as faster than the one I would get setting the repeat rate to 1.
In fact, the envelope's output frequency @ rate 2 is slightly slower than the one I would get by actually doubling the envelope's output frequency I get @ rate 1 (it drifts of more that 1 tenth of a second after 16 retriggers)
Same applies with repeat rates of 4 and 2...and so on I guess.
I did the above measurements by connecting env2 output to the gate (delay attack and release to minimum) and recording Aalto's output @ env2 repeat rate 1 - then rate 2 - and finally rate 4.