Using WinNessie to adjust QRS PNOscan - Helping Pianoteq's simulation

I use Pianoteq via a QRS PNOscan MIDI sensor rail that I installed into my old Steinway.  I have previously written notes here in my attempts to adjust Pianoteq's settings to best sound like my piano.  Still, even though I could adjust the velocity curve in Pianoteq, it seemed odd to have to do it there instead of at the source, the MIDI strip itself.  Furthermore, to best mimic my piano's keys and how they make sound with strike and release, etc., I needed to do this through the QRS software, which is not clearly documented.

Here is a note that I have send to the QRS folks describing what I have learned, and what confuses me still.  Perhaps if any of you have a QRS PNOscan on your piano, this may help you also:

================

Wim and Dan -

Thanks for your communications a week or so ago. I have finally gotten some time to spend working on the QRS PNOscan settings, and I will tell you what I have done, what I have learned, and what I still do not understand:

WHAT I HAVE DONE:

Appreciate, first, that I am playing on an old Steinway upright that has a relatively new action that was installed about 25 or 30 years ago. The result is that my piano has limited dynamics, and once I get to a mezzo-forte key-strike on the acoustic piano, I have to hit it really pretty hard to get it to play any louder. So, if I am going to try to match this with PNOscan, I am looking for a curve that is convex upwards and plateaus as I go along. Additionally, I like the idea of matching the ability of my acoustic piano to push the keys down very slowly, so that they do not make a tone but they still lift the dampers. Various of my simulated and sampled software pianos can take advantage of this, playing resonances only on the strings where the dampers are lifted, even if they themselves were not played.

My current settings are:

Controller number – 0
controller value – 0
MIDI channel – 1
MIDI program – none
Linear velocity gain – 85
Expo velocity gain – 127
Expo velocity shape – 6
key trigger position – 22
key release position – 17
re-trigger window – 8
minimum velocity – 0
velocity offset – 20
after-touch gain – 60

WHAT I HAVE LEARNED:

Dan and Wim, while you told me to start with linear velocity, that actually was almost the opposite of what I wanted as it kept me from reaching my curve when I was concentrating on that setting. That brought me to Expo velocity shape: For one thing, I did not understand at the beginning that the Expo velocity shape has no effect unless Expo velocity has a value greater than zero. Once I learned this, then it became clear that the Expo velocity settings were what I needed to play with to get my convex upward curve that plateaus as I move towards forte. And, since I wanted to get the maximum effect of a nonlinear curve, I kept increasing Expo velocity until I was at 127. I then started lowering the curve number from eight, which is linear, until I got to 6. This gave me a nice curve. Unfortunately, no matter how hard I struck the keys, I would still not hit a maximal MIDI value of 127. I realized that in order to get to this, I just had to increase the linear velocity incrementally until I got what I wanted. At this point, with reasonable performance, I wondered why the keys were still "too hot" despite what I thought was a reasonable key trigger position. It turned out that the bottom of the curve was too high on the y-axis. This is when I learned that the velocity offset works in a negative direction. I kept increasing the velocity offset until I could very slowly push the keys down and make no sound but hold the damper's open. This came at approximately 20. Unfortunately, that also reduced my maximal values, so I had to increase the linear velocity a bit more. As for the key trigger, key release, and re-trigger positions, these were set while leaving the stop-rail out so I could hear both the acoustic piano and the electronic piano at the same time – I just adjusted them for simultaneous note on, note off, and repeat notes with rapid keystrokes. Once these were coordinated, I had the ability to play both acoustic and electronic pianos simultaneously, essentially simulating what can be done with the Yamaha trans-acoustic series of pianos – quite rich! And, when I switch between acoustic pianos and electronic pianos, my muscle memory for hitting the keys produces very similar results.

WHAT I STILL DO NOT UNDERSTAND:

Although I have not figured out how to use after-touch in the Native Instruments software, such as for their synthesizers, I did find that it works fairly well in Pianoteq, especially when applied to the electronic pianos, etc. I set after-touch on PNOscan to the value 60, fairly randomly, and I am not sure what the benefits were detriments are to putting it higher or lower – can either of you teach me something about using after-touch?

Additionally, what are the settings of controller number, controller value, MIDI channel, and MIDI program? I am not a MIDI expert, but understand that there were 16 MIDI channels, so that would answer that value. What about the three other values?

Another setting that I completely do not understand is hysteresis. I have tried adjusting it up and down. Sometimes, if I move it too far down, it resets itself. If I move it up, the number stays changed, but I cannot appreciate any difference in the sound.

    And about my 'organ-keyboard' settings that I told you about...

I had mentioned that I used one of the banks of settings in PNOscan to set up for having key-strikes at the top of the action, so I can make playing on my heavy Steinway action to simulate using a very lightweight action on electronic organ, with a fixed velocity output. I did not do a greatjob with that yet, and probably will keep working on the piano-style settings, with your help on responses to this note, before I go back to simulating the organ. Still, it's a pretty cool thing to do to be able to use the QRS PNOscan to make even a heavy manual keyboard as featherlight as a Hammond organ.

I look forward to your responses and hope that I might have helped you guys or some other people via what I have learned.

- David