Alright, it seems like your and my intuition agree to some extent. Here is what I'm considering:

Any "time" based controls will remapped from [0..1] to an actual duration in seconds, for instance [0sec..5sec]. It also seems like the sustain knob is backwards in LMMS.  Next, is dealing with the Amount (AMT) knob.  I agree with LMMS that the AMT knob for volume envelope should control a mix between the envelope and 100% gate.  The user can use the channel's volume control if they wish to change the maximum volume. That leaves us with this fomula

volume = envelope_level * amt + (1-amt), if amt >= 0
volume = envelope_level * amt + 1, otherwise

... or something like that. The case for cutoff and resonance filters will be slightly different.  I haven't decided exactly how, yet :)

-- Paul


On Sat, May 11, 2013 at 2:26 PM, John Serafino <lzrblade@gmail.com> wrote:
I try to avoid LMMS's native envelopes for this reason. They do not seem at all intuitive, nor do they mirror any other envelopes I've used thus far. I would expect amt knob to do anything consistent that makes sense. I would expect attack to effect how long the envelope to to reach the instrument volume, which it would stay on for the length of hold. Then I would expect it to take decay amount of time to move to sustain, finally taking release amount of time to drop to zero.

Again, this is merely what I would expect. I can work with envelopes that work different ways, especially if they have an accurate visualization. I just couldn't figure LMMS's out before giving up.