Amplitude EG simulation is not right

[SloeComputers]

Both the rate and levels do not seem to be correct. Also not replicating what seems like delayed stage transition behaviour when levels are 0 and rate controls how long to stay in the state?

[SloeComputers]

The EG levels are better after this patch 268831a

[SloeComputers]

There is also sometimes a nasty click at the start of a note.

[SloeComputers]

Seems the nasty click is dependent on some patches e.g. "BRASS. 1" and very likely just a manifestation of the rate problem.

[SloeComputers]

Significant improvements to attenuation scaling in this patch