My First Velocity Curve - HOW I DID IT

Thanks for posting your methodology.  Do you end up with the ceiling and floor on your velocity curve with your keyboard for acoustic pianos as well?

Hello,

interesting idea to use Pianoteq's VU-meter for velocity calibration!
Smart idea to define a medium loudness of a preset with the VU-meter and to find the noteon velocity, that triggers it.

A VU-meter more or less corresponds to the human hearing, so this could work eventually. Have to try it myself. One problem could be, that a noteon velocity not only changes the volume level, but also the timbre / frequency-spectrum, which also influences the human loudness perception.

Thanks for sharing your idea!

Late edit: The Limiter switch should be OFF in this process. Yes! I had overlooked this - thanks for your contribution!

It is more result than a choice, how I understand it: Klicking a key on Pianoteq's virtual keyboard at the top of the key gives a small green bar on the VU-meter. Yes Klicking on the bottom of the virtual key a large green bar / audio level. Yes The half between that two audio levels is assumed to represent the "normal" loud level.This is what I assumed.

The next assumption is that mf-mp in Pianoteq's velocity diagram represents the subjectivity of the Pianoteq crew, how their presets should sound at a middle velocity level (named mf-mp on the y-axis). In another thread I had shown, that the middle between mf and mp stands for the number 64).

At this point we are having two representations:
The "normal" audio level (by VU-meter)
The "normal" velocity (64 aka mf-mp by Modartt) I didn't choose 64 because it was half-way house between 0 and 127. It just happened to correspond to the middle of the range the keyboard produced.

The glue for both parameters is a third parameter: Pianoteq's input velocity ( = output velocity of your MIDI keyboard).

With the VU-meter we can identify which input velocity triggers our defined "normal" audio level. If we found that to be 50 for example, we now can map that to Modartt's medium 64, and bam! we have the middle dot of our new curve [50; 64]. Yes, using the keyboard velocity and the middle of the found range is indeed the middle dot of the 'new' curve.

That is, how I understand your method. If it has been a misunderstanding, it is the draft for my own new method! ;-)

The lowest and highest points in the individual curves are trivial, this is what PTQ's midi calibration assistant outputs already or is found manually by several users (with some subjective variations of course). I wouldn't say they were trivial, but vital as some keyboards might be very good at low velocity and not so good at high. I am simply trying to map the range of the keyboard to the range of available volume offered by Pianoteq. By using a set middle of 64 in the case of a keyboard range of 1-100 say would produce a 'quiet sounding' curve. Would you not agree?

DEZ wrote:

I didn't choose 64 because it was half-way house between 0 and 127. It just happened to correspond to the middle of the range the keyboard produced. In other words 105-23 = 82. This is the limited range of the keyboard. Half of that is 41. Now because the range starts at 23 we add 23 and 41 to get... 64.

... that I don't fully understand probably. You tried to play in the middle range of the VU-meter and then you find/read 64 in Ptq's  midi monitor?
Or did your read 50 then? Looking at the VU-Meter I hit keys on the keyboard and got them to strike the mid-point and then checked the velocity in options = 50. I can now set my node on the curve as 50, 64.

The funny thing is, we come to the same result, because my proposition has been, that the medium audio level in the VU-meter represents mf-mp, which is ... the equivalent to y=64.