Kawai's "Let-off" question - answer received

Just did receive the reply (copied below) from Tom Love, of Kawai, regarding the "let-off" (or, escapement) behavior of not being able to "play off/from the jack" on all Kawai digitals:

--------------------------------------------------------------------------------

Thanks for your email. I was in Japan last week and returned yesterday. Needless to say, I'm a bit jet-lagged today.

I had the chance last week to discuss our current GF action and the let-off issue that you mentioned with the keyboard action engineering team. Here is my understanding of the situation, given my poor Japanese and their decent English.

The design of the GF action does not allow for the let-off playing as you describe. This was done so that the let-off point would be in the correct position. Some competitive actions do allow the playing style that you describe, but to do so, their let-off point is placed in an incorrect point that is lower than that in an acoustic piano. The keyboard team's decision was to place the let-off at the correct position.

Future action designs will take this point into consideration and perhaps they will have both the proper let-off point and the let-off playing style that you describe.

Best regards,

Tom Love
Sr. Director of Online Marketing and Electronics
Kawai America & Kawai Canada

--------------------------------------------------------------------------------

Therefore, all current owners of new models including the ES7, CA65, and CA95, will not be able to get a software update or fix to correct the let-off behavior as this is something that will have to be incorporated into "future action designs" as mentioned above.

The current let-off was designed to be felt at the "correct position" similar to an acoustic, although it still does not account for the other aspect of having the correct behavior in regards to sounding a note from the let-off point, itself.

Therefore, do not expect any changes with the current models.

Es7 has no GF , but the completely differet RH II. I think this item is mainly is about GF based instruments ( and placement of the pivoting point) ?

Yes, this point was discussed with the digital piano keyboard development team last week. in order to recreate the feel of a grand piano action, we agreed that the correct position of the let-off point was very important.

As others will vouch, Kawai's digital piano action team are committed to constantly improving the quality and realism of their keyboard designs, so I expect we may see this 'playing off the jack' behaviour implemented in future instruments.

Kind regards,
James
x

@James,

This is to let you know that I am satisfied with the current CA95 that I own (as my new recording should attest to that)* as there are certain details with the metallic sounds that I have learned to tolerate better.

A lot of this can be easily adjusted within the "Basic Settings" and "Virtual Technician" features. Therefore, it is a very enjoyable piano to play.

Hope this will help to clarify the facts regarding the let-off simulation.

*Recording is in this thread:

http://www.pianoworld.com/forum/ubbthrea...tml#Post1985200

This seems somewhat counter-intuitive for me (if this was a technical explanation at all and not a foreground argument for the marketing decision - sorry for beeing always overly skeptical but I am working for the industry too and know the nature of product developing and marketing).

If the let-off point is higher, as it seems now to happen with GF, it should mean, that more possibilities are left for the placement of both lower sensors beneath, not less. To be able to repeat you have to pass the two lower sensors on the way back. I would think the only thing is needed is to tweak on the algorithms - but I can agree, this is not necessarily a simple task without any risks for the rest of the instrument.(Extensive testing cycle needed, redesign of some other aspects of the instrument could give a more optimal solution, as such a fix....) So if such considerations were also behind such a decision, they could be justified too.

But You never will hear arguments of economy (which should be the primary consideration at a company while managing their product lines), but we should not forget about it if we want to build a realistic picture for ourselves.

Temperament: I think it's the opposite - the let-off point is BELOW the middle sensor. I.e, if we press down very slowly to the let-off point, and then wait for a long time, the middle sensor has already been tripped, so that when we finally complete the stroke, the velocity sensor (which presumably measures the time taken between the middle and bottom sensors) decides that the key press velocity was below the minimum threshold for a hammer strike.

I'm curious about the V-Piano. Does it allow repeats for key releases that are LESS than the let-off point? If so, that means that it too is not completely authentic. ;^) I.e, whilst it allows playing from the let-off point, it ALSO allows another behaviour which is NOT authentic.

I'm also curious to know what the let-off depth for the V-Piano is, and whether it is authentic or not. PV88: you could measure your V-Piano, and compare with some real grands.....

Greg.

Greg, Your explanation is the plausible one for me too, but the above statement suggests just the opposite of that: Your imperfection dilemma with V-Piano is very real: even 3 sensors allow only an approximative modelling. Another approach could be: it would be conceivable to apply a single acceleration sensor only (instead of 3+ position sensors) which could then capture the whole movement curve of the event of playing a key. The algorithm could then take all other things in calculation, as let-off resistance, jack encounter, repetition phase. It seems complex enough, to be sure, but seems possible at least in theory.

How pricey such a sensor solution would be, I cannot tell either.

Attila

Hm..I thought this is how silent pianos work ?

Attila: Good point. I don't understand. Maybe something has been lost in the translation.....

Re the "single acceleration sensor", that's what the Infinite Response VAX77 has, but it's not a hammer action, AND, when I asked them, I was told that it does not actually simulate double-escapement - not strictly, anyway.

And then there's the Alpha piano, which has a real grand action along with pressure sensors in place of the strings. I'm sure it would emulate all this very well, but PERHAPS with a bit of latency due to the fact that the sound generator would be triggered at exactly the same time that the hammer strikes the "strings". (in a real piano, the strings of course will start vibrating immediately)

Greg.

Aftermarket silent pianos typically have a strip that goes under the keys and light intensity reflected off the bottom of the keys informs the strip of the key position. In principle, this method could yield a continuous curve and a solution could be computed for the terminal velocity of the hammer. In practice I suspect there are intensity thresholds that get passed that trigger on/off logic similar to the sensors in typical digital pianos.

Then there are silent pianos that have the hammers break optical paths, triggering on/off switches. This is the way the AvantGrand line works.

There's a guy in the PianoTeq forum that has built his own switches similar in design to midi strips but which apparently uses a more sophisticated algorithm than the 2 or 3 sensor tech we are accustomed to. I don't know how well it works.

I suspect the same. The VAX77 definitely does sample the key position very rapidly, though, and does model momentum.

Greg.

That's cool. And that hi-rez MIDI stuff is cool too.

Does anyone in the forum actually have a VAX77? I once asked here whether anyone has tried that VAX and been able to tell whether the hi-rez stuff makes a difference (perhaps using it to control PianoTeq on using high resolution midi and again with 128 steps). No one replied, though.