Typical/maximum velocity values

[Dan Morris]

Wondering if I can get some help with an informal experiment...

I recently observed that my two controllers (an Axiom 25 and a Yamaha P120) both top out around velocity 100, even when pounded far beyond the force I'd actually play with. The Yamaha maybe hits 105 with a real pounding. Of course when I open a soft synth piano in particular, and play a key using an on-screen keyboard (typically velocity 127), I hear a completely different tone, which suggests I'm not seeing the entire tonal palette of most soft synth pianos.

This is all well and good if there's an industry convention about where "practical" velocity tops out, and all controllers can reasonably agree on this, so instrument designers design accordingly. But if there's not even a de facto standard, then at least the "out of the box" experience you have with a new virtual piano is completely dependent on your keyboard's velocity behavior.

I'm separating this from the velocity curve, which is somewhat configurable on both of these keyboards, but in practice all available curves top out around the same place; the curves only really affect the bottom end of the velocity range.

This got me thinking about whether there in fact is any consistency across keyboards wrt velocity behavior. So here's where the informal experiment comes in... can I get folks to reply to this thread to say:

1) The model of controller you're using

2) The maximum velocity you see when you "play pretty hard", like as hard as you'd play for a high-energy passage (but still as if you're playing for musical purposes)

3) The maximum velocity you see when you hit the keys as hard as you're comfortable hitting them, possibly beyond a real playing range. Or alternatively, if you see that at some point you can't really get the velocity any higher, the point at which you see it stop increasing with increasing physical force/velocity.

Unless otherwise stated, I'll assume you're using your controller's default velocity curve.

Everyone will presumably have their own way of looking at velocity values, but FWIW I find MIDI OX to be the easiest way to get a simple view of velocity values in Windows (velocity is "data2" for a note-on event).

I'll add my own reply later today.

Thanks!

-Dan

 

[happyrat1]

I'm somewhat too lazy to perform your experiment at the moment, but if I recall correctly I've read of similar experiments on other forums where your results are pretty much nominal.

As you said, velocity curves are usually adjustable on most keyboards and from a practical standpoint most keyboards top off around 100 max by all reports.

From an engineering standpoint I can see where this makes sense. While digital electronics consists entirely of absolute values within the circuitry, the mechanical membrane switches used in most keybeds are something of an analog wild card. It actually does not surprise me to think that the engineers built in a bit of tolerance to accommodate the physical differences between individual switches and left an overhead as a safety margin to avoid overdriving the outputs in a particularly dynamic piece.

If you truly wanted to experiment and get some valid results you'd probably have to jig up some sort of uniform key striking device with an adjustable measurable striking velocity and then testing it on every key on a given unit.

Otherwise, simply putting out a general call for uncontrolled experimental results is not going to tell you much except that a wildly random sampling is going to give you wildly random results.

Gary