groovy: the velocity curve I use for my ES7 is
Velocity = [0, 5, 26, 51, 83, 112, 127; 0, 0, 32, 64, 96, 124, 127]

Ok, thanks, but what is the effective range, that your Kawai ES7 is able to transmit? That was my question and as AKM mentioned, maybe a curve/your curve was chosen with the intention to MIDI-compress/expand the response.
(I'm thinking of eventually getting an ES8 in the near future (if I should like the action)).

Velocity = [0, 5, 26, 51, 83, 112, 127; 0, 0, 32, 64, 96, 124, 127]

Velocity = [0, 8, 28, 51, 80, 118, 127; 0, 1, 32, 61, 96, 126, 127]

It implies that I can _reliably produce_  MIDI values from 8 to 118 -- so, a range of 110 out of 127 possible values (for the first byte).

Hey, that's fine and good news (for me)! The velocity-range seems to be well designed by Casio.

Thnx

cpcohen wrote:

AKM wrote:

Velocity = [20, 105; 0, 127]

Interesting -- that's reasonably close to the curve I came up with, for my PX-350.

This implies, that the PX150/350 series has a velocity-range of ~ 105-20=85. The MIDI-Standard suggests 127 velocity-values, so the possible range seems to be reduced to 67%.

This is the incorrect conclusion. That velocity curve mentioned was just an example for the discussion about how to hard compress the incoming velocity values using the Pianoteq editor. This is definitely not the velocity curve I'm using actually. Casio keyboards have absolutely fine velocity response with no reduction whatsoever.

cpcohen wrote:

AKM wrote:

Velocity = [20, 105; 0, 127]

Interesting -- that's reasonably close to the curve I came up with, for my PX-350.

This implies, that the PX150/350 series has a velocity-range of ~ 105-20=85. The MIDI-Standard suggests 127 velocity-values, so the possible range seems to be reduced to 67%.

Is anything known about the situation of the Kawai ES7/8 in this regard?

Thanks

The curve I'm using, based on watching what MIDI values I could actually play reliably:

Velocity = [0, 8, 28, 51, 80, 118, 127; 0, 1, 32, 61, 96, 126, 127]

It implies that I can _reliably produce_  MIDI values from 8 to 118 -- so, a range of 110 out of 127 possible values (for the first byte).

The keyboard will produce MIDI values from 2 ( I think) to 7, but they require a lighter touch than I have.  And the 118 is as high as I can go with pretty heavy "FFF" playing.   Someone stronger might get higher values -- or I could change the "Touch" value in the keyboard's menu.

Remember -- the PX-x50 keyboard produces _2 bytes_ of velocity data, and Pianoteq accepts it.  So you're working with 127 x 128 possible velocity values, not 127.

. . . There's enough resolution there for _anybody_.

.      Charles

AKM wrote:

Velocity = [20, 105; 0, 127]

Interesting -- that's reasonably close to the curve I came up with, for my PX-350.

This implies, that the PX150/350 series has a velocity-range of ~ 105-20=85. The MIDI-Standard suggests 127 velocity-values, so the possible range seems to be reduced to 67%.

Is anything known about the situation of the Kawai ES7/8 in this regard?

Thanks

Just paste this (the line only, without my borders):

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

Velocity = [20, 105; 0, 127]

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

onto your Pianoteq velocity window using your mouse right button.

Actually I'd suggest you to experiment with Velocity curve adjustments, it's not a rocket science and one can easily guess how it works, just add and move some dots.

Interesting -- that's reasonably close to the curve I came up with, for my PX-350.  I use it with "Touch = 3" on the DP.

.    Charles

You could do that if you had a PX-350 with "Line In" jacks, as I do.  And you'd find that the built-in loudspeakers are very, very inadequate for showing what Pianoteq can do. 

.             Charles

PS -- I suspect that the PX-150, X-stand, pedal, EV ZXA1 powered speaker (which _does_ work nicely with Pianoteq), and laptop, _all together_, would weigh as much as a "stage piano" from several makers.

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

Velocity = [20, 105; 0, 127]

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

onto your Pianoteq velocity window using your mouse right button.

Actually I'd suggest you to experiment with Velocity curve adjustments, it's not a rocket science and one can easily guess how it works, just add and move some dots.

And btw if you still want to be able to reach 0 - 127 velocity diapason constantly and easy you can use the Pianoteq velocity curve with hard thresholds to expand it by converting, say, incoming 120 to receiving 127 and incoming 20 to receiving 0 and still have several thousands gradations of velocity to operate - this is where the hi-res shows it potential at max. You'll get virtually zero penalty for doing that. With common 127 value non-hi-res keyboard you will have some noticeable loss of the velocity resolution for such a case: 127-(20+7)=100 steps for the current example.

I wish I knew the procedure for doing this.  Can someone write a detailed procedure possibly with screen shots to show the method of accomplishment?  One thing I did learn is actually the PX-150 can transmit a velocity of 1, and it's accomplished by pushing the key very slowly which is unlike an actual piano were the hammer is flipped up at the string and then the jack is pulled out.  I would like to get an actual 127 which I never could on my KX-88 and now on my Casio.  My general thought or suspicion is that if there is an anomaly in the velocity that is sent, it's more likely to be coming from the Casio than in the reception from Pianoteq which I believe to be superbly designed.  The Casio is very good for what it is at it's price point and I'm a strong advocate for Casio, having owned many different models over the years; nevertheless it clearly is not the "top of the line" in terms of sophistication.  In contrast the Pianoteq is just about as good as it could possibly be.  If I were to suggest improvement to the Pianoteq it would be in the area of screen displays which could and probably should be made larger when working on almost all of the various sound setting adjustments.

I guess it would depend on the actual microcontroller/CPU used and its instruction set. But in general, there's not really anything being "overtaxed" by just attaching another MIDI message to each note generated.

Nah, it wouldn't overtax anything. It's just a 14-byte message instead of 7-byte message, CPU doesn't even feel the difference here because it calculates things either in 32.

Just gonna add that this super wrong. I'm assuming you meant bit where you said byte. First off, no one uses non-exponents of 2 for bit storage/manipulation. A 32 bit CPU can only handle 4 byte operations. Operations on sub 32bit integers are definitely both possible and faster (16 and 8 bit integers are things).

As far as Casio having 16,000 Velocity graduations, that seems highly theoritical as Pianoteq cannot even calibrate my Casio. About the lowest playable velocity out is somewhere around 22 and the highest number seems at best 117.  Just create an instrument that does an honest 1 - 127 (or 0  - 128)  Who could hear 15,001 from 15,002 ?

The ideal geeky solution will be the possibility to manually set the MIDI keyboard sensors lowest and highest measured time thresholds which corresponds to the lowest and highest velocity data to be sent out so one could adapt it to his playstyle. As for now and from my side the PX-150 have rather ideal factory velocity tuning, so even though most of the time the useful velocity diapason is around 20 to 110 (with 127 hi res gradations in between each number for Privias) it is essentially necessary to have some headroom on both ends in case you need and able to utilize it sometimes. It is like removing some keys on both sides of the 88 note keyboard because that keys are so rarely used.

About the built in calibration procedure in Pianoteq at it's current state - I find it useless. But the velocity curve manual control is absolutely essential.

And btw if you still want to be able to reach 0 - 127 velocity diapason constantly and easy you can use the Pianoteq velocity curve with hard thresholds to expand it by converting, say, incoming 120 to receiving 127 and incoming 20 to receiving 0 and still have several thousands gradations of velocity to operate - this is where the hi-res shows it potential at max. You'll get virtually zero penalty for doing that. With common 127 value non-hi-res keyboard you will have some noticeable loss of the velocity resolution for such a case: 127-(20+7)=100 steps for the current example.

=============
Just recently added: LOL, you may actually want to check this thread, all that I just wrote here previously need a huge correction: http://www.forum-pianoteq.com/viewtopic.php?id=3976

AKM wrote:

There is nothing possible to improve, all modern Privia keyboards including PX-150 are the only digital piano keyboards on the market for right now that are sending hi-res 14 bit midi velocity data to Pianoteq. I.e. 16,383 velocity gradations instead of the standard 127.

Infinite Response VAX-77 does that too. Casio is not the only one with that feature.

I just can repeat what I said and it was absolutely correct, just read carefully. a) VAX-77 is not "on the market" - it is actually quite long time out of production (the new model is under development right now). b) VAX-77 is not a "digital piano" or piano type controller - it has 76 keys and some spring mechanics which is obviously not a hammer emulation. So for right now Casio is the only one with Hi-Res as far as I know, though maybe there are some other keyboards that I don't know but not VAX.