Are all MIDI keyboards the same?

In my experience, the character of a sound influences the feel of a keyboard more than the converse. There certainly are differences in the way Pianoteq responds to different keyboards, but there are limits. The main difference would be in the Note On velocity value that is sent in response to a given physical velocity of the key as it passes through the sensors. This is hard to measure precisely, but you can get a sense of it by playing a very familiar piece as consistently as possible, recording the MIDI (maybe several times from each keyboard) into a DAW, and examining the captured velocities. DAWs commonly have a view that shows velocity as a sort of histogram with a vertical bar for each note on a scale from 0 to 127 over time so you can get a sense of the average velocity and overall variability in different sections of a piece which is going to translate to dynamic and tonal changes in Pianoteq.

As you probably know, since it's a hot topic around here, Pianoteq has a velocity curve function that allows you to modify its response to velocity from the keyboard and also an overall dynamic range setting for each preset such that you should be able to largely compensate for the differences between keyboards, and get a response that feels and sounds right to you. If the Casio sounds 'thinner' it's likely that it generates a slightly lower velocity for a given physical velocity than the Roland.

The second most important difference in the case of Pianoteq would be whether the keyboard sends variable release (Note Off) velocity such that the rate that the sound is damped varies with how quickly the key is released. This is even harder to measure as no DAW that I know of has a view that will show Note Off velocity. But so long as a MIDI keyboard is not just sending some fixed value and has a release velocity curve that is consistent with the Note On velocity curve, I would think there should not be vast difference in the feel and sound of Pianoteq.

Beyond that, and the significant difference between the response to a continuous pedal vs. a non-continuous one, any sense one might have to Pianotec sounds different with two different keyboards is probably more about the physical feel of the keys, the way they accelerate/decelerate, the depth of the stroke, how firmly they bottom out, the mechanical noises they make, etc.

As for the idea that the sustain would vary from one keyboard to the next, there's really no technical basis for that. So long as the key/pedal is held far enough down that no damping is being invoked, the amplitude and duration of the sustain and evolution of resonances and tone over time for held notes/dampers should be identical. A quick search found comments indicating that the decay time of the CDP-S110's internal voices has been altered, but this would not affect Pianoteq's response to MIDI.

One other difference that might come into play would be how quickly and consistently the keyboard (and its USB drivers if applicable) transmit MIDI messages. MIDI latency affects how "connected" the keyboard feels to the sound source. This is usually dwarfed by the output latency of the audio interface but could vary slightly from one keyboard to another.

EDIT: Love the username BTW. ;^)

Thanks for the reply Brundlefly.  The Casio is a much cheaper keyboard than the Roland. It doesn't sent "note off" for instance. I just wonder what else it doesn't send. The rapid decay of each note makes it more difficult to play a reasonable legato, and this does seem to be the case through PTQ as well. I quickly learned that I have to hold the key a little longer (quite good practice in some ways). I think you are to some extent confirming that not all keyboards are going to sound the same, for various reasons. Do you think it would be possible for reviewers to take this on board?

My user name is nothing to do with music, I'm afraid. One day I turned over a paving slab in my garden and there was an ants' nest under it.  The ants went into panic mode, trying desperately to move their eggs to safety, and the name just came into my head.

Strange. I checked the MIDI implementation chart for the CDP-S110, and see it sends a constant Vel=64 with Note Off. So the decay of the note after key release is going to be fixed at a medium value. But I usually think of legato as playing notes with a slight overlap in time so that the release of one note actually occurs after the attack of the next in which case a short decay time should not be that noticeable because it's hidden in the attack of the new note.

From your description, it sounds like the issue might be more that the Note Off is sent as soon as the key leaves the keybed - or very close to it - such that you have to really deliberately hold the key all the way down to play legato which would indeed be a little awkward.

Interesting. The manual indicates only that there is a setting for it but does not say where to find it. It seems an odd choice to completely disable it by default but explains why I could not hear an effect when I experimented briefly. I think I did stumble on the alternate curve types when I was first poking around the UI, but did not fully register that the Note-Off curve was fixed at 127 and promptly forgot about it. I will have to give it a try. Thanks for the heads up!

If your keyboard does send variable note-off (and I think most Roland keyboards do, FP-30 certainly does), try pasting the whole of the following line into the Velocity Curve section:

Note-Off Velocity = [0, 41, 127; 8, 41, 127]

This is the note-off curve that I use, and it does make a difference. The reason I don't take it all the way to 0 is because I had some problems with the piano sounding as though it had defective dampers!

The Yamaha CDP-725 transmits a fixed Note OFF value of 64 according to the manual:

You can verify it in Pianoteq's Midi event monitor. After Pianoteq has received that value 64 it maps it per default to 127 (see your NOTE-OFF velocity curve in Pianoteq). This means per default a key release "sound" is modelled as if you released each key as fast as you can.

Don't worry, not faulty, it is just the cheesy design. MIDI protocol has reserved 7-bit (0 - 127 decimal) for velocity, but some brands don't care to calibrate their keyboards to use it (I have a Kawai and a Korg that are not better in that respect).

I guess it is easier and cheaper for manufacturers to use let's say a partial range of 80 values instead of 127 for velocity. By mechanical tolerance the keyboard output could be midi velocities 10-90 or 20-100 or 30-110 for example. All within the MIDI standard v=1-127.

Pianoteq (and other virtual instruments) don't know, how calibrated/uncalibrated a connected MIDI keyboard is, so the default "velocity curve" is usually just a straight line from 0 to 127 in the velocity mapper.

The work is outsourced to the user (you! :-)) to "calibrate" your own velocity curve.
Can be hard work, be warned, good luck!

The CDP - 130 has a superb action. I love the way it feels so mechanical. I nearly bought one (wish I had). I'd borrowed a CDP - 120 which has the same action and loved how it worked with Pianoteq. I actually preferred it to the 2nd generation of this action that was in the Casio PS5 which I bought instead.

I have a Yamaha P515 now which is very well built. It does send note off velocities.

I agree; it's unimagineable the the CLP-725 is not capable of generating the full range of velocities. Possibly it has a user-configurable curve that is currently set to limit the velocity output, but the issue may also be technique. Note that "Velocity" is literally a measure of how fast the key is moving. This is not necessarily the same as "hard" (i.e. how firmly the key bottoms). The OP may find it easier to get full velocity by playing stacatto with a very quick finger movement.

Thanks for your replies . I have to take issue with you on speed vs hard. Pianists often like to say that what matters is not force but velocity. According to Newton's second law, acceleration is directly proportional to force.  The key is at rest.  To make it move you have to accelerate it. The greater the force the greater the acceleration and the greater the velocity as the hammer hits the string. It's not actually possible to move the key faster without increasing the force. The great Heinrich Neuhaus liked to talk about F, V and H.  I'm afraid they are just different ways of doing the same thing. Nevertheless, in the end it doesn't matter how you think about it: if it works for you, it's right. Of course, if you are manufacturing a digital piano it's the velocity which the sensors are measuring but these velocities are still produced by varying force on the keys. Anyone like to discuss this further?

It may have to do with the association of "tension" or "clenching" when you say "hard". Meaning no disrespect, but many amateur or hobby pianists seem to use almost their entire arm, making everything rigid when trying to play loud. A trained pianist can play loud but stay relaxed, thus reaching a high key velocity without making a hammer out of his finger.

Of course, equal volume means equal force but the trained pianist uses fewer muscles and relaxes them quicker so it doesn't look "hard".

Something similar happens when playing fast. Most people (including trained pianists) tend to clench up when asked to play something just slightly faster than they actually can. Playing fast requires more relaxation, not more tension. The quicker you can relax your muscles, and the more relaxed you can keep the muscles that are not involved in each keystrike, the faster you can play.

I haven't really thought about joint damage when hitting keys (the right way), although I've broken nails fairly often if I didn't keep them short. I think tendonitis is more common and that can happen to anyone. Also, I don't know the word for it in English but my left ring finger has a tendency to fall out of the proper arching and plays more flat if I don't pay attention, that can lead to damage in the long run. Although I've managed to go the first 33 years of playing without serious problems (knock on wood).

Of course. And to fill this advice with content: An individual velocity curve is healthy. It can avoid too much effort/pounding. Not "all MIDI keyboards" are the same

https://www.modartt.com/velocity_curves
This page should be the answer to our questions about velocity curves, but the more I look at it the less I trust it.
How is it possible that for the same keyboard there are such different curves? Where do these curves come from? Are they purely subjective and proposed by the users (Some of them being concert performers and others  beginners) or are they curves proposed by pianoteq team after measuring the velocity of each keyboard to get closer to a real piano? Only a direct comparison between the keyboard and a real piano is relevant as a good base curve. I personally use a curve for my FP 10 but I have no idea of its realism, is the one listed on the page more realistic? It seems too speed for me but I don't know... it would be interesting if modartt would certify the given curves guaranteeing a reliable starting point for users who don't have a grand piano at their disposal...
Another remark concerning this page, the visual of the curves does not work for me anymore....

Translated with www.DeepL.com/Translator (free version)

As far as I'm aware, in almost all cases the velocity sent over midi is affected by the selected velocity curve on the instrument. So that page with velocity curves is basically useless because it doesn't say anywhere which velocity curves have been selected on the instrument. In the case of the Korg Kronos, it doesn't even say whether it's on the synth action keybed or the hammer action keybed, though you might hope for the latter one.

I like the idea of a page like that (even though Pianoteq makes it really easy ro calibrate your own) but as it is it's basically useless.