Changes in the way to specify note velocity in the inspector

• May 7, 2019 - 21:26

Currently there are two ways to specify note velocity: 1) User, 2) Offset.
User is an absolute velocity value, independent of any previous dynamic mark.
Offset is a relative value, expressed as a percentage (though without including the % sign).

Changes requested:
1) Replace the current term "Offset" by "Offset (relative)" and add the % sign in the dialog so that is clear what it means, keeping its effect as is currently.
2) Add a new category: Offset (additive) meaning that the indicated number is the number to add to the velocity, whatever value has been previously reached as a combination of last dynamic mark and the cummulative hairpin(s) effect counting from that mark on.
3) Include the same changes in the piano roll editor.
4) Make the User choice box show the real velocity, as is the case of the piano roll, instead of 0, so to be informative about the real velocity after the effect of dynamic marks and successive hairpins.
5) Include the formulas in the corresponding handbook entry to eliminate any posible ambiguity, and possibly give an insight on the use of each one. I offer myself to collaborate in preparing such material.

Rationale:
1) An offset is, technically, a value to be added to something, representing a deviation from a standard value, so in the current state, naming "offset" something that is really a relative increment is misleading.
2) This is confirmed in other areas of the Inspector, such as the x and y offset for element location, which is absolute and not relative to current position (if it were relative, the same number would imply a larger displacement at the bottom of the page than at the top).
3) The current implementation of the piano soundfont (not confirmed for other instruments but likely) is that velocity is roughly proportional to psychoacoustic loudness (a good design decision), so a constant increment of velocity implies a constant increment in loudness sensation, regardless of the current dynamic mark. What many users expect is that the addition of a fixed number represent a fixed increase in perceived loudness. As currently implemented, i.e., as a percentage, this is not the case. For instance, an offset of 10 % at a dynamic mark f, with a velocity of 96, implies a change in velocity of 9.6, rounded to 10, which is quite a change, but the same 10 % at pp with velocity of 33 implies a change of 3.3, rounded to 3, which is barely an audible change. With the new suggested category, an increment of 10 velocity units would cause f to be roughly midway between f and ff, and p to be roughly midway between pp and p.
4) Keeping the current implementation is only for the sake of back compatibility, to honor the effort that many people have done so far trying to have the score sounding as they want using the available tool.