So here is my experiment today. Mine was a little noisy so I swapped all the resistors to metal film 1%, that helped a lot. I also socketed all the transistors and have tried a silicon trannie in Q5, works just fine, it acts as a variable resistor to control the gain of Q2 so the affect on tone is minimal. I have tried using lower gain trannies as well KSA1015 for the PNP and KSC1815 for the NPN with gains in the 140 Hfe range. A must is a control over the swell, I did this by putting a 47k resistor in series with a 1M log pot in R8.
I think having a certain amount of gain on tap at anytime (limiting R1 and R9) would be a worthwhile experiment. The input and ouput control more than just the threshhold or output gain they have an affect on the entire circuit. It's interesting that P1 of R9 goes to +9v instead of ground. I haven't taken measurements yet on how the output gain changes the voltage to the collectors of Q2-4.
Regarding the tone, messing around with C5 will change the release but I dont think it would affect the tone much, try messing with C8 or put a passive tone circuit in at the R8 Q3 C8 junction.
However I do see some advantage to putting a variable resistor in series with the collector of Q5 and C5 (perhaps 25KB) to affect the release.
here is the revised Schem