Good morning, everyone.
Larry, I can tell you definitively that the Disklavier is not designed to produce routinely note-on velocities across the full spectrum of 1-127.
I have no knowledge as to whether one Disklavier model behaves differently than another. I tend to think not. In other words, it would have been counter productive for Yamaha to design some models with a velocity profile that does not match the others. If they did, how could they sell recordings that would play properly on every model?
My first advice is to ensure that your Disklavier is (1) properly regulated from a mechanical perspective and then (2) calibrated using its internal, closed-loop calibration functions.
From having studied thousands of Disklavier recordings—including all of those recorded in the Piano-e-Competition over many years—I can confidently tell you the following:
—Casual playing will yield velocities in the range of 35/45 to 80/85.
—Dynamically expressive playing may widen the range to 30 to 90/95.
—Extremely expressive, concert-quality performances may range from 25 to 105/110, the highest ranges generally supported by thick chords.
Note-on velocities beyond 110 are extremely rare and sound very unmusical unless supported by a thick, thunderous texture of notes.
Note-on velocities below 25 are usually the result of key presses that were not well controlled, key bounces, and brushed notes. In other words, they happen but are usually not intentional.
I have never heard an explanation from Yamaha about the philosophy behind its mapping of actual hammer (or key) velocities (measured in meters per second) to MIDI values. I suspect that the philosophy was to construct a mapping of velocities that can actually be generated naturally or artificially (say, by slamming the keys with your fist or slowly pulling down the key) and thus provide some margin beyond the range of normal playing.
Out of the thousands of recordings that I looked at from the Piano-e-Competition, I only saw a handful that exceeded 115. I don’t think that I ever saw one above 122.
Interestingly, after more than 40 years of MIDI experience, the industry seems to be converging on some standards in this regard:
This convergence involves the participation of Kawai, Steinway, and Yamaha and addresses note velocities as well as incremental pedaling.
This velocity issue goes off the rails when one factors in digital pianos and keyboards with semi-weighted and lightly weighted keys. When a pianists play those instruments, they are in an artificial environment in which they can boost or cut back the volume by twisting a knob or moving a lever. Thus the measured key velocities have no real-world analog. I think that explains the reason for why—outside of the world of acoustic MIDI pianos made by Yamaha and companies with a similar design velocity—you find MIDI recordings with extraordinarily high velocities.
Georg (aka PianoBench)