I have a soundBlaster card so all the audio comes from the PC to that card via USB and then to the headphones / monitors.

I never use the NV10's internal speakers. They produce a convincing "piano" sound only with its internal voices but those all have virtually zero dynamic range. I regard them as good for nothing more than 1-fingered tunes, basically - hence, I bought Pianoteq. It would be nice if Kawai would just sell a high quality keyboard without the built-in piano sounds - hopefully it would be half the price.Actualy, it would be nice if the big names would sell just a fully authentic MIDI keyboard although I imagine they would charge huge sums for them.

When I first bought the NV10, I did try routing the SoundBlaster output to the NV10's internal speakers to see if I could save myself the price of Studio Monitors but that sounded awful. The audio circuits in the NV10 are obviously heavily customised to their own requirement.

pdavidow wrote:

yeq30:

Is your setting for volume control twist-knob on the NV10 arbitrary?
Actually, it seems that you are plugging your headphones directly into the computer, so the question is moot.

Don't forget the pedals/dampers.

I never use the NV10's internal speakers. They produce a convincing "piano" sound only with its internal voices but those all have virtually zero dynamic range. I regard them as good for nothing more than 1-fingered tunes, basically - hence, I bought Pianoteq. It would be nice if Kawai would just sell a high quality keyboard without the built-in piano sounds - hopefully it would be half the price.Actualy, it would be nice if the big names would sell just a fully authentic MIDI keyboard although I imagine they would charge huge sums for them.

When I first bought the NV10, I did try routing the SoundBlaster output to the NV10's internal speakers to see if I could save myself the price of Studio Monitors but that sounded awful. The audio circuits in the NV10 are obviously heavily customised to their own requirement.

Is your setting for volume control twist-knob on the NV10 arbitrary?
Actually, it seems that you are plugging your headphones directly into the computer, so the question is moot.

I can only assume that the default diagonal line is the simplest linearisation they can perform from velocity to MIDI value.

Yes, that's what I meant. But why don't you use that assumed "simplest linearisation" for conversion in Pianoteq's velocity mapper? NV10's unresponsive range less than 10-14 and more than 112-113 can be ommitted there directly.

True, doing two conversions is going to involve more errors and uncertainty than one - but I don't know what transform they are doing so it's easiest to wirk on the assumption that either it doesn't matter or I can compensate. Assuming anything else would lead to endless research and/or paralysis by analysis.

They don't offer a "guaranteed raw" option straight from the sensors and I doubt that their "normal" is, in fact, raw. I can only assume that the default diagonal line is the simplest linearisation they can perform from velocity to MIDI value. I have occasionally experimented with using some of their built-in curves and making a Pianoteq velocity curve to work with them but I was never satisfied with the results.The only one that seems to work mostly is the "normal" which gives very similar results to the USER diagonal line but does not generate the full range of values; it stops at around 100'ish instead of going all the way up to 127. The way I usually know that the NV10 has reset to the "normal" curve is that I am not getting the full dynamic range when I play.

I think the relationship between dynamics and physical volume is a bit of a minefield of psycho-acoustic effects etc. In addition, I have found that the Windows volume setting is the only way to adjust for loudness without affecting tonal range. If you use the Pianoteq Volume slider then it seems to move the dynamic range wrt hammer hardness as well at the same time. Also, the Effects page output volume slider on the far right hand side seems to do the same trick.

For example:
I'm interested in the new Studiologic SL73 Mk2 (for having 73 keys and a quick and light key response). But I've seen video clips, where it seems to have the old and common midi controller problem of being too sensitive, f-ff to easily, no good resolution in mp-p (just my impression, maybe completely wrong!!). So probably it needs heavy manipulation of the velocity curve again like the Korg B2. I can't buy that Studiologic, if I don't know if it has a Pianoteq friendly curve or the warranty of a fullrange 1:1-Mapping with good resolution. - Chance missed of the seller / manufacturer.

pdavidow wrote:

dikrek wrote:

1. Lots of pianos out there - logistically what you propose is very hard
2. There are variances between pianos even from the same model

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

Concerning point #1, the keyboard mfgrs themselves can be incentivized to do the the calibration for their particular models. (After all, some mfgrs have already cooperated with Modartt, e.g., Kawai.)

Nobody’s stopping any manufacturer from downloading a Pianoteq demo and offering this for everyone.

That’s a good idea.

Much easier than Modartt trying to do this.

I set the Kawai key touch curve to the default diagonal line first.
ppp    pp    p    mp    mf    f    ff    fff
14    28    42    56    71    85    99    113

A few thoughts about those USER settings of your NV10. Isn't that a remapping already? I think so and have visualized your upper two rows (blue and orange). The diagonal line (green) is the famous 1:1-mapping, which maps an incoming midi-velocity to the same velocity at the output (leaves it unchanged):

But if you put the output of the blue or orange values of your Kawai into Pianoteq's curve editor and remap it again with some logarithmic curve, you have done two transformations. In other words: Two midi-curves in series. That is overcomplex, because a remapping of 127 to 127 values can always be done in one single curve (rounding error is another story).

You mentioned a "normal" curve of the NV10? Isn't that a factory-default midi-response, which is the constant root source of all USER curves and preset curves? Then I would take that raw source and start all over.

dikrek wrote:

pdavidow wrote:

PianoTeq really should have a setting whereby user selected his particular keyboard from a list (e.g, Kawai NV10).
Modartt would then be the one who did all the calibration homework, not the user.

1. Lots of pianos out there - logistically what you propose is very hard
2. There are variances between pianos even from the same model

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

Concerning point #1, the keyboard mfgrs themselves can be incentivized to do the the calibration for their particular models. (After all, some mfgrs have already cooperated with Modartt, e.g., Kawai.)

Give it another 10 .. 20 years and there will be global industry standards in force.

pdavidow wrote:

PianoTeq really should have a setting whereby user selected his particular keyboard from a list (e.g, Kawai NV10).
Modartt would then be the one who did all the calibration homework, not the user.

1. Lots of pianos out there - logistically what you propose is very hard
2. There are variances between pianos even from the same model

PianoTeq really should have a setting whereby user selected his particular keyboard from a list (e.g, Kawai NV10).
Modartt would then be the one who did all the calibration homework, not the user.

1. Lots of pianos out there - logistically what you propose is very hard
2. There are variances between pianos even from the same model

I hope that one day, just the same as totally accurate colour-control is now built into all PC's (it didn't used to be), every preset will just have a button or check-box for "real" so we won't have to mess around setting up presets to sound / feel like a real piano.

Yes, less time consuming and error prone calibration of midi velocity would be on top of my list for the future. Everything else has very high standards already in my opinion. Maybe KI will be a useful tool in this domain. Or the approach mentioned by brundlefly to mirror the response of another well calibrated Digital Piano.

Side note:
My weight response curve cannot be narrowed with a logarithmic trendline in libreoffice (magenta):