After seeing more posts regarding this issue I realized my initial response was in regards to how I typically use a delay in my signal chain rather than the OP’s thoughts as to how the ‘Mix’ parameter should work. I often run my delays in serial, somewhere after the amp/cab.
IMHO the following represents the best, and by best, I mean most predictable and most intuitive manner for the ‘Mix’ control to work: The levels of the source and the delayed signals should be equal at 50%.
As you move past 50% the delayed signal should become louder and the source signal softer. At 100% you should be hearing nothing but the delayed signal (100% wet). This is so that you can run the delay at 100% wet in a parallel path if you wish, using a mixer block to determine how much wet delay you want in your tone.
One additional point and this is one that several modelers could improve on IMO. Let’s say the equal volume point for the source and delayed signal is 50%. I would like a nice taper running up to 50% providing an audio taper that does not jump too quickly from no audible delay, to too much. Often however I find with my relatively straightforward signal path, that the delay can go from barely audible to extremely audible very early in the taper of the ‘Mix’ control with a miniscule turn of the dial.
For me it is a similar experience to using a linear rather than logarithmic curve on an expression pedal. The jump from inaudible or barely audible delay to too much is more abrupt than it could be and happens too early in the low percentages on the ‘Mix’ control.
In summary I would say that from the perspective of a systematic approach to how the ‘Mix’ parameter functions, I would like to see 50% be equal volume source/delay, and 100% be delay only. I also want the taper to be such that it seems both logical in the numbering of the source/delay ratio (e.g. 12% should not be overly wet), and it is easy to dial in the amount of reverb without small adjustments of the percentage resulting in large jumps in the perceived reverb level .