I kept banging my head trying to better understand the filter envelope on the XY, until I finally realized just how extremely non-linear the attack duration is. I measured the maximum attack time for both the amplitude and filter envelopes at about 6 minutes (360 seconds).
By eyeballing the attack knob at different points along its range, I got the following approximate times:
- 25%: too short to measure
- 50%: ~2 seconds
- 75%: ~26 seconds
- 100%: ~360 seconds (6 minutes)
These measurements were taken by pressing a key on two different engines, Prism and ePiano, with decay, sustain, and release all set to minimum. Tempo and bar scale had no impact on the result.
These points fit reasonably well on this curve (see screenshot, x is 0 to 100%, y in seconds):
T(x) ≈ 0.0111 × e^(10.386x)
If you were to align the vertical line marking the drop from the attack phase to the sustain phase with the M1 to M4 keys, know that 50% is slightly before the gap between the M1 and M2 keys:
The maximum attack time occurs roughly 40% into the width of the M2 key:
Sorry for the trash pictures.
