Topic: Kawai VPC1 faulty key weighting

I measured the MIDI output of each key with ~140g weight, results are alarming:
PunBB bbcode test

Can any fellow VPC1 owner please sanity-check my results - press G#1 and A2 simultaneously and look at the MIDI output. In my case A2 will have ~20 points higher value (first major jump on the plot).

Looks like there have been units with exact same problem before:
http://forum.pianoworld.com/ubbthreads....040/1.html

Re: Kawai VPC1 faulty key weighting

I had same issues with my VPC1. Ditched it in the end.

Much prefer my Yamaha P515. Much better manufacturing quality control.

Warmest regards,

Chris

Re: Kawai VPC1 faulty key weighting

I had the same issue as well. Ten minutes after connecting the keyboard for the first time I put it back in the box and returned it. It's my impression that quality control for production of the Kawai VPC1 is poor.

I tried a Studiologic sl 88 Grand, which I find to be too noisy and having a too heavy action for my taste. Will now try a Roland FP-90. Quite a quest! :-)

Re: Kawai VPC1 faulty key weighting

Bartes wrote:

I had the same issue as well. Ten minutes after connecting the keyboard for the first time I put it back in the box and returned it. It's my impression that quality control for production of the Kawai VPC1 is poor.

I tried a Studiologic sl 88 Grand, which I find to be too noisy and having a too heavy action for my taste. Will now try a Roland FP-90. Quite a quest! :-)


I also had quality control issues with the Roland FP-90's action. My favourite keys are the ones I own at present.

Warmest regards,

Chris

Re: Kawai VPC1 faulty key weighting

Shame to admit it...but i have been playing on it for ~2 years, i noticed the queerness quite soon..but kind of ignored it because it was my first piano and i could not play anyways...
Now i have grown quite used to this defect, astonishing what humans can adapt to

I am planning to try and calibrate the weights myself..fun project no matter what!

Last edited by gasparka (30-10-2019 16:21)

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

I measured the MIDI output of each key with ~140g weight, results are alarming:

Can any fellow VPC1 owner please sanity-check my results - press G#1 and A2 simultaneously and look at the MIDI output. In my case A2 will have ~20 points higher value (first major jump on the plot).

Looks like there have been units with exact same problem before:
http://forum.pianoworld.com/ubbthreads....040/1.html

Gasparka, can you let us know how you measured the response of the keys? Do you have some kind of device for it? I like your attitude to this problem (fixing it yourself). Good luck!

Re: Kawai VPC1 faulty key weighting

Bartes wrote:
gasparka wrote:

I measured the MIDI output of each key with ~140g weight, results are alarming:

Can any fellow VPC1 owner please sanity-check my results - press G#1 and A2 simultaneously and look at the MIDI output. In my case A2 will have ~20 points higher value (first major jump on the plot).

Looks like there have been units with exact same problem before:
http://forum.pianoworld.com/ubbthreads....040/1.html

Gasparka, can you let us know how you measured the response of the keys? Do you have some kind of device for it? I like your attitude to this problem (fixing it yourself). Good luck!

I stacked a bunch of flat washers to make a ~140 gram weight.
I set the weight on the key while holding the key inplace. Next i let go of the key, so it falls freely from upper position, and record the MIDI value from Pianoteq MIDI viewer.

Re: Kawai VPC1 faulty key weighting

I wonder what will be the range of values if you measure one key several times?

Re: Kawai VPC1 faulty key weighting

scherbakov.al wrote:

I wonder what will be the range of values if you measure one key several times?

Quite consistent, i would say my measurement error is +-2.

Edit: I have removed the "let-off simulation/escapement" ribbons - linear touch is easier to measure. MIDI values were inconsistent before and after the removal.

Last edited by gasparka (31-10-2019 11:41)

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

I measured the MIDI output of each key with ~140g weight, results are alarming:
PunBB bbcode test

Can any fellow VPC1 owner please sanity-check my results - press G#1 and A2 simultaneously and look at the MIDI output. In my case A2 will have ~20 points higher value (first major jump on the plot).

Looks like there have been units with exact same problem before:
http://forum.pianoworld.com/ubbthreads....040/1.html

Have you looked at the VPC 1 editor, the offset might have been changed from all values set to 0? Regarding calibration to the weights I'm not sure how you would do that. The capstan adjustment only serves one purpose, to set the key dip, before adjusting press/hold down several keys (flat hand across keys) and see if the key dip/travel distance is the same, if any key is higher this will have a shorter throw and would result in low velocity triggering. In this case adjust the capstan to achieve the same key dip as the other keys. Don't try any wild adjustments here to try to achieve a faster or lighter playing experience, just make sure the key dip is all uniform.

Re: Kawai VPC1 faulty key weighting

If everything is in order with the mechanism, maybe it’s worth trying to adjust the values using the VPC editor?

Re: Kawai VPC1 faulty key weighting

MeDorian wrote:

Have you looked at the VPC 1 editor, the offset might have been changed from all values set to 0? Regarding calibration to the weights I'm not sure how you would do that. The capstan adjustment only serves one purpose, to set the key dip, before adjusting press/hold down several keys (flat hand across keys) and see if the key dip/travel distance is the same, if any key is higher this will have a shorter throw and would result in low velocity triggering. In this case adjust the capstan to achieve the same key dip as the other keys. Don't try any wild adjustments here to try to achieve a faster or lighter playing experience, just make sure the key dip is all uniform.

Yes, all offsets are 0. I also restored the factory settings, which had no effect.
I will take a closer look at the key dip tonight, thanks!

I was planning to calibrate the weights by soldering extra lead onto some weights and taking some metal out from the others. But first i have to make sure that the problem is actually in the weights.

Re: Kawai VPC1 faulty key weighting

scherbakov.al wrote:

If everything is in order with the mechanism, maybe it’s worth trying to adjust the values using the VPC editor?

Have been thinking about this. Is the error coming from mechanism, sensors, software ...need to figure out how to test this.

Re: Kawai VPC1 faulty key weighting

scherbakov.al wrote:

If everything is in order with the mechanism, maybe it’s worth trying to adjust the values using the VPC editor?

Yes! Go easy though, values of around 3 for example can result in a loss of low velocity values meaning your lowest note might not achieve a value of o. Small adjustments like +1 can give surprising results so avoid values like +10 or -10.

Re: Kawai VPC1 faulty key weighting

gasparka wrote:
scherbakov.al wrote:

If everything is in order with the mechanism, maybe it’s worth trying to adjust the values using the VPC editor?

Have been thinking about this. Is the error coming from mechanism, sensors, software ...need to figure out how to test this.

It could be but see how you get along with the editor. Modifying takes a huge amount of time so unless you have another keyboard to use in the mean time, best to compromise and get at least some evenness to the keyboard without modifying. Adding extra weight would add noise which is not good

Re: Kawai VPC1 faulty key weighting

Do not forget that the editor remembers 5 quick presets, which are switched by holding the power key and one of the top black keys. In quick presets, keyboard speed adjustments are remembered. Try different options.

Re: Kawai VPC1 faulty key weighting

Bartes wrote:

I had the same issue as well. Ten minutes after connecting the keyboard for the first time I put it back in the box and returned it. It's my impression that quality control for production of the Kawai VPC1 is poor.

I tried a Studiologic sl 88 Grand, which I find to be too noisy and having a too heavy action for my taste. Will now try a Roland FP-90. Quite a quest! :-)

I also tried the Studiologic SL88 Studio, which someone on Pianoworld forum described as having an action "equivalent" to the lighter action of an upright, and the SL88 was like a grand.  I couldn't find one on display, so I purchased one using my Guitar Center trade-in guarantee on a Casio that broke.  And supposedly the lighter of the two SL88's, it was unbelievably heavy. 

Especially if you played anything in C# Major, such as Bach's Prelude in C# Major WTC I, or even the D Major Prelude from WTC I, where you have to play deep in the "back of the keys" in the rapid passagework, it was utterly unusable!

Around a half-decade ago, there was someone who was making a controller keyboard, according to a preview website, that was going to be a full piano keyboard that was MIDI-ized; sadly that seems to have never happened; Kawai's VPC1 is a total compromise and while it almost sounds good as a compromise, it's an expensive compromise....too expensive, all things considering.

But yes, the Studiologic SL88 - all of the instruments based on the FATAR keybed are all too heavy, especially if you are playing anything in keys that are not in C Major.

I have been slightly tempted by the VPC1, but the more I read, the more I am concerned of their quality, and the compromise of shortening the keys, when they simply could have made the properly long, and have begun to read more confirmations that they too, when one plays deep into the keys, are problematic, and the posts in this thread also are not comforting!

Last edited by scottusa (04-11-2019 09:40)

Re: Kawai VPC1 faulty key weighting

scottusa wrote:
Bartes wrote:

I had the same issue as well. Ten minutes after connecting the keyboard for the first time I put it back in the box and returned it. It's my impression that quality control for production of the Kawai VPC1 is poor.

I tried a Studiologic sl 88 Grand, which I find to be too noisy and having a too heavy action for my taste. Will now try a Roland FP-90. Quite a quest! :-)

I also tried the Studiologic SL88 Studio, which someone on Pianoworld forum described as having an action "equivalent" to the lighter action of an upright, and the SL88 was like a grand.  I couldn't find one on display, so I purchased one using my Guitar Center trade-in guarantee on a Casio that broke.  And supposedly the lighter of the two SL88's, it was unbelievably heavy. 

Especially if you played anything in C# Major, such as Bach's Prelude in C# Major WTC I, or even the D Major Prelude from WTC I, where you have to play deep in the "back of the keys" in the rapid passagework, it was utterly unusable!

Around a half-decade ago, there was someone who was making a controller keyboard, according to a preview website, that was going to be a full piano keyboard that was MIDI-ized; sadly that seems to have never happened; Kawai's VPC1 is a total compromise and while it almost sounds good as a compromise, it's an expensive compromise....too expensive, all things considering.

But yes, the Studiologic SL88 - all of the instruments based on the FATAR keybed are all too heavy, especially if you are playing anything in keys that are not in C Major.

I have been slightly tempted by the VPC1, but the more I read, the more I am concerned of their quality, and the compromise of shortening the keys, when they simply could have made the properly long, and have begun to read more confirmations that they too, when one plays deep into the keys, are problematic, and the posts in this thread also are not comforting!

Although it is possible to lighten the VPC1 action through modification, as you say, the keys are not long enough. This key length is the fundamental flaw in the VPC1! Is there any keyboard (apart from MP11 for example) with full length keys? There was the Vax 77 ( not sure if this is the correct model?) that never really worked out production wise, and the M piano (Alpha keyboard I believe), again still doesn't appear to be available and also very expensive. Time for the VPC2 I think?

Last edited by MeDorian (04-11-2019 13:24)

Re: Kawai VPC1 faulty key weighting

The VPC2 would only be worth considering for me, if it keeps the price of the VPC1, and the action would be fully regulated and MIDI output properly checked, with a worldwide transferable service guarantee.

Alternatively there are, for a price:

+ Ravensworks VPC1 (plus parts and 40h regulation work)

+ Lachnit MK23 from Vienna, Austria (flkeys.at), IICR they have a voucher partnersip with Pianoteq.
The Lachnit is supposed to be perfeclty regulated and has light sensors instead of rubber dome contacts.

HTH

Last edited by lophiomys (04-11-2019 20:56)

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

I measured the MIDI output of each key with ~140g weight, results are alarming:
PunBB bbcode test

Can any fellow VPC1 owner please sanity-check my results - press G#1 and A2 simultaneously and look at the MIDI output. In my case A2 will have ~20 points higher value (first major jump on the plot).

Looks like there have been units with exact same problem before:
http://forum.pianoworld.com/ubbthreads....040/1.html

Op, I purchased a VPC1 in September of this year, and definitely some keys are weird with velocity. After reading your post I decided to try the same thing on mine so as to provide another data point for you.

I also used a 140g weight. The kawai piano was set to preset #1, which is the default velocity curve and I zeroed out any velocity corrections previously done in VPC Editor. Following similar convention, note 1 = A0, note 88 = C8, etc. I  took three measurements of each note and averaged them to build this chart.

For me the most noticeable wrong velocity is right in the middle of the piano from A4-E4, then after that some random notes that are too loud.

PunBB bbcode test

Re: Kawai VPC1 faulty key weighting

Don't think I'm being crazy here, but wouldn't most acoustics suffer the same problem?  It's inconceivable to think, that with all the clatter and stuff, they'd have a uniform response.  Maybe this is part of what gives them their character,  I've always thought most non fully wooden digital action are too clinical, and I suspect this is the reason.  They are just TOO uniform!

I'm playing all the right notes but not necessarily in the right order

Re: Kawai VPC1 faulty key weighting

peterws wrote:

Don't think I'm being crazy here, but wouldn't most acoustics suffer the same problem?  It's inconceivable to think, that with all the clatter and stuff, they'd have a uniform response.  Maybe this is part of what gives them their character,  I've always thought most non fully wooden digital action are too clinical, and I suspect this is the reason.  They are just TOO uniform!

peterws,

That is a very educated opinion.
You should become a piano technician, and report back after one year in action.

I would predict a great career in the piano industry, big savings to be expected in the finishing process of acoustics!

SCNR,
Lo PHi

Last edited by lophiomys (19-11-2019 10:26)

Re: Kawai VPC1 faulty key weighting

Thanks for sharing your measurement, rdwing!

rdwing wrote:

I  took three measurements of each note and averaged them to build this chart.

... repetition is good! What do you think was your measurement error? Also +/-2 as gasparka estimated?

Re: Kawai VPC1 faulty key weighting

@rdwing
Thanks for measuring, your piano looks slightly better compared to mine. Want to trade?

Anyhow, i have done some additional debugging with some results.
I opened up the piano and inspected the keys manually, at first i did not see anything too strange. I also swapped around some keys and weights but problems did not resolve. So i decided to measure the sensor outputs from G2 and A2 - thats where the major bump is on my piece.

Here is a picture of the setup:
https://i.postimg.cc/5t7rWFGb/IMG-20191106-195625736.jpg

https://i.postimg.cc/7hshg9tj/IMG-20191106-195640634.jpg

I found the sensor locations by following the excellent writeup of "VPC1 Keyboard Scanning" by Jack O'Flaherty. His web page is down currently, but here is the copy:  https://drive.google.com/open?id=1xoiKL...2siMz5vS2M

Sensor data for G2:
https://i.postimg.cc/QxnZbngx/G2-MIDI37.png

Basically the controller measures delay between Sensor 1,2 and 3 activations and uses that to calculate key speed. In previous image Sensor 2 was activated 8,886 ms after Sensor 1, and Sensor 3 was activated 8,097ms after Sensor 2. Or the total key press time was ~17ms. And this resulted in MIDI value 37.

Now the same weight on A2 results in MIDI 52, and the plot shows part of the problem:
https://i.postimg.cc/jdDFsqNt/A2-MIDI52.png

There is switch bouncing - third sensor triggers too early, and the controller thinks that this is a fast key press.
BTW: Looks like the controller only considers sensor 2 and sensor 3 for calculating the key speed.

Anyways, then i opened the switch compartment to see if anything is jammed up:

https://i.postimg.cc/tg6zf7Lm/IMG-20191106-200438824.jpg

https://i.postimg.cc/FsngHjvB/IMG-20191106-200747144.jpg

I didn't really do anything to the switches, but the problem was gone after i reassembled the rubber parts. Looks like the Kawai switching design is not that robust, and i believe the problem can reappear at any time on any piano.

Sadly this only solved part of the problem. Key calibration is still quite messed up. Now it looks to me that the problem is in key dip (as already suggested here). I have done some experiments but the process is quite complicated honestly (small change results in big MIDI change)..will let you know if i get some good results.

Also, i have access to brand new Kawai CA78 - this has the latest Kawai action. I will be performing the 140g test and will report it here shortly. Quite interesting to see how it looks, i mean the piano costs almost 3000 EUR, i am expecting good curves!

Re: Kawai VPC1 faulty key weighting

gasparka,
thanks for your informative posts.

For whatever it's worth,
A few years ago i had an Kawai MP7 - different action (RH2), similar rubber dome contacts:
When a freshly replaced unit developed unreliable key triggering after a few months,
a warranty repair was scheduled in a qualified repair shop nearby.

The technician there replaced all the rubber dome contacts AND
painstaikingly cleaned the contact surfaces on the electronics board,
mentioning that little grains of (graphite) dust would potentially disturb the sensors,
even though I was not able to see any dust or dirt at all.
Afterwads that MP7 was working, though it was not checked for even key action triggering.

HTH

Re: Kawai VPC1 faulty key weighting

First, thank you gasparka for those insights, very interesting!! I have to read your post more thoroughly later.

I "measured" my old Kawai ES3 this evening just for comparison. My weight was 140.5 g and the placement can be seen on the foto.

https://i.postimg.cc/sgrQVFjB/140g-weight-velocity.jpg
https://i.postimg.cc/256rFTjj/ES3-velocity-140-5g-chart.png

Cheers

PS: I did no averaging, just a single shot.
PPS: Kawai ES3 with factory default velocity curve. PTQ with default velocity map y=x.

Last edited by groovy (21-11-2019 20:24)

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

Also, i have access to brand new Kawai CA78 - this has the latest Kawai action. I will be performing the 140g test and will report it here shortly.

Will wait! Really interesting! I'm try now select what to buy CA78 or Roland LX708.

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

@rdwing
Thanks for measuring, your piano looks slightly better compared to mine. Want to trade?

Anyhow, i have done some additional debugging with some results.
I opened up the piano and inspected the keys manually, at first i did not see anything too strange. I also swapped around some keys and weights but problems did not resolve. So i decided to measure the sensor outputs from G2 and A2 - thats where the major bump is on my piece.

Here is a picture of the setup:
https://i.postimg.cc/5t7rWFGb/IMG-20191106-195625736.jpg

https://i.postimg.cc/7hshg9tj/IMG-20191106-195640634.jpg

I found the sensor locations by following the excellent writeup of "VPC1 Keyboard Scanning" by Jack O'Flaherty. His web page is down currently, but here is the copy:  https://drive.google.com/open?id=1xoiKL...2siMz5vS2M

Sensor data for G2:
https://i.postimg.cc/QxnZbngx/G2-MIDI37.png

Basically the controller measures delay between Sensor 1,2 and 3 activations and uses that to calculate key speed. In previous image Sensor 2 was activated 8,886 ms after Sensor 1, and Sensor 3 was activated 8,097ms after Sensor 2. Or the total key press time was ~17ms. And this resulted in MIDI value 37.

Now the same weight on A2 results in MIDI 52, and the plot shows part of the problem:
https://i.postimg.cc/jdDFsqNt/A2-MIDI52.png

There is switch bouncing - third sensor triggers too early, and the controller thinks that this is a fast key press.
BTW: Looks like the controller only considers sensor 2 and sensor 3 for calculating the key speed.

Anyways, then i opened the switch compartment to see if anything is jammed up:

https://i.postimg.cc/tg6zf7Lm/IMG-20191106-200438824.jpg

https://i.postimg.cc/FsngHjvB/IMG-20191106-200747144.jpg

I didn't really do anything to the switches, but the problem was gone after i reassembled the rubber parts. Looks like the Kawai switching design is not that robust, and i believe the problem can reappear at any time on any piano.

Sadly this only solved part of the problem. Key calibration is still quite messed up. Now it looks to me that the problem is in key dip (as already suggested here). I have done some experiments but the process is quite complicated honestly (small change results in big MIDI change)..will let you know if i get some good results.

Also, i have access to brand new Kawai CA78 - this has the latest Kawai action. I will be performing the 140g test and will report it here shortly. Quite interesting to see how it looks, i mean the piano costs almost 3000 EUR, i am expecting good curves!

If the electronics are ok then I guess it's the calibration you'll need to do. The foam pads wear down over time so this will increase the key dip travel, more in the most played areas, the middle keys in most if not all cases. If you are to do some calibration make sure to write/log all key weight example W1B(or something?)  with the key number.  The weights use an algorithm and are not in absolute order. You will need (unfortunately) to remove all keys to do this. I would suggest leave the top(highest) octave in place and find a measuring object to match to key dip in these keys (top octave). Each key should then be adjusted with the capstan to give the exact same key dip. The sharps have there own key dip which should be uniform with the top(5 sharps) octave( I remember using a ruler for this laying flat). The final key, the lowest A should have the same key dip than the highest C. There should be no need to measure the weighting of each key. When complete small adjustments in a problem key(s) can be made using the MIDI editor. Edit...if the initial problem was worn pads, then, and if you knew where to find the same foam pads (I never found the material in any shop or online) then no capstan adjustments would ever have been needed and only the foam pads replacing?

Last edited by MeDorian (20-11-2019 17:44)

Re: Kawai VPC1 faulty key weighting

Here is the curve for CA78 (normal settings). Don't know what to think about it. Maybe the argument of @peterws makes sense here. Still, i would rather have clinically clean touch response and add randomness in Pianoteq if i want it.
https://i.ibb.co/MgWvHg5/ca78.png


Also here is my VPC1 in same plot style:
https://i.ibb.co/Dkptdfs/vpc1.png

@MeDorian
Thanks for the tips, i will take it on soon enough.

One significant advantage of the CA78 keyboard is that the key press is not as deep as in VPC1 - 0.9 vs 1.1 cm measured from tip of white key.

Re: Kawai VPC1 faulty key weighting

Hi all,
Definitely the key dip is not uniform across my keyboard. There exist some keys that at rest are higher than others, and some where the dip is not quite aligned. Also have a few keys that make a weird sound when depressed. Given that this is a brand new instrument, I will attempt to have Kawai repair/replace first.

peterws, I wanted to say that an acoustic piano may experience the same issue, but I would expect the technician to fix this during regulation and voicing work, post purchase.

gasparka, what kind of usb scope is that? I need one of those!

Also it is worrying to me that we all tested with about 140g weight, and on my instrument 140g responds around 55 MIDI velocity. This is too loud to play any VST with default velocity curve, so I must add in an exponential shaped curve. Everyone else's instruments seem to respond around 40 for the same weight.

In the meantime, I went through VPC Editor and applied correction to each key. This has evened out the key response, but hasn't resolved any physical issues. Plus, since the correction can exceed 12, I lost access to the loudest and softest regions.

Regarding the measurement error, yes I was able to reproduce +/-2 for each attempt. Usually +/-1.

PunBB bbcode test

Last edited by rdwing (21-11-2019 06:58)

Re: Kawai VPC1 faulty key weighting

rdwing wrote:

Hi all,
Definitely the key dip is not uniform across my keyboard. There exist some keys that at rest are higher than others, and some where the dip is not quite aligned. Also have a few keys that make a weird sound when depressed. Given that this is a brand new instrument, I will attempt to have Kawai repair/replace first.

peterws, I wanted to say that an acoustic piano may experience the same issue, but I would expect the technician to fix this during regulation and voicing work, post purchase.

gasparka, what kind of usb scope is that? I need one of those!

Also it is worrying to me that we all tested with about 140g weight, and on my instrument 140g responds around 55 MIDI velocity. This is too loud to play any VST with default velocity curve, so I must add in an exponential shaped curve. Everyone else's instruments seem to respond around 40 for the same weight.

In the meantime, I went through VPC Editor and applied correction to each key. This has evened out the key response, but hasn't resolved any physical issues. Plus, since the correction can exceed 12, I lost access to the loudest and softest regions.

Regarding the measurement error, yes I was able to reproduce +/-2 for each attempt. Usually +/-1.

PunBB bbcode test

Yes, if it's brand new then don't be taking on any regulation yourself. My Kawai is 6 years old and way past warranty.

Re: Kawai VPC1 faulty key weighting

I still don't get the 140g test? With the VPC1 action been graded and with an algorithm I can't see how the test would work? I might have missed something but the MIDI values are never going to be perfect, I would suspect most midi files have been edited, no way I can play that perfect and don't really want to. What I would be looking for is an action that when played, no notes are obviously too loud or quiet (standout), this in general playing rather than checking the MIDI velocity readout. Again as for the key dip, doesn't have to be a fraction of a millimetre in, just uniform. For an instant more even playing the dynamic slider in Pianoteq can be reduced to around 30, also I think the Pianoteq curve in the VPC1 is best, or possibly a flat curve in the VPC1 and a fast curve in Pianoteq.

Last edited by MeDorian (21-11-2019 15:24)

Re: Kawai VPC1 faulty key weighting

@ rdwing
Scope: https://www.saleae.com/
Also, my measurements were done with Pianoteq touch curve (green led), sorry i failed to mention that. I just tested that with Pianoteq-curve middle C reads 35 and normal curve (red light) reads 57, big difference! In that sense i guess your piano is normal..


@ MeDorian
I think the 140g test has been quite informative. For example, we can see that for CA78 the keyboard is not that much graded - if you press note in low register and high register with same force you will get same sound volume (i think this is a good thing!). Also in some cases it clearly shows that the keyboard needs calibration i.e. when neighbor notes have 20% difference.

We need better ways for testing digital keyboards. Basically digital keyboards function like an Analog to Digital converter - these are often ranked by bit count e.g. 16 bit vs 24 bit audio. But actual performance of converter is shown by Effective Number Of Bits (ENOB) measurement, which basically shows how much bits are left after internal noise, defects etc. Often 24 bit converters only yield ~18 bits of usable data. This is similar to MIDI keyboards, they advertise as being 7-bit converters, but in reality i assume they have maybe ~5 bits ENOB i.e. last 2 bits are more or less random noise. Maybe i am wrong, but surely knowing this information would make piano buying process much simpler.

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

@ rdwing
Scope: https://www.saleae.com/
Also, my measurements were done with Pianoteq touch curve (green led), sorry i failed to mention that. I just tested that with Pianoteq-curve middle C reads 35 and normal curve (red light) reads 57, big difference! In that sense i guess your piano is normal..


@ MeDorian
I think the 140g test has been quite informative. For example, we can see that for CA78 the keyboard is not that much graded - if you press note in low register and high register with same force you will get same sound volume (i think this is a good thing!). Also in some cases it clearly shows that the keyboard needs calibration i.e. when neighbor notes have 20% difference.

We need better ways for testing digital keyboards. Basically digital keyboards function like an Analog to Digital converter - these are often ranked by bit count e.g. 16 bit vs 24 bit audio. But actual performance of converter is shown by Effective Number Of Bits (ENOB) measurement, which basically shows how much bits are left after internal noise, defects etc. Often 24 bit converters only yield ~18 bits of usable data. This is similar to MIDI keyboards, they advertise as being 7-bit converters, but in reality i assume they have maybe ~5 bits ENOB i.e. last 2 bits are more or less random noise. Maybe i am wrong, but surely knowing this information would make piano buying process much simpler.

Did you notice a similar jump in velocities (if not as bad) with the CA78 compared to the VPC1 using the same 140g test? On a different thought could it be USB power issue to the VPC 1, I use the external power supply but had no issues with USB, did notice it powered up faster so I continued using it ever since (about 6 years). edit..just another thought, in the VPC 1 editer, midi output settings, do you have just the keyboard highlighted (as I do)? There is something in the manual that recommends the external power supply when outputting conventional MIDI. Again, I'm sure you know this but it could be something as simple as this. Hope you get things up and running soon with the VPC1

Last edited by MeDorian (21-11-2019 21:11)

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

Also, my measurements were done with Pianoteq touch curve (green led), sorry i failed to mention that.

You should do one finite measurement, there are a touch too much inconsistencies now:
* non-default VPC1 curve
* convention A0-C8, whereas Pianoteq is using A-1-C7 for 88-keyboards.
* let-off simulation in use or not.
* unzoom to full MIDI range (0-127)


For example, we can see that for CA78 the keyboard is not that much graded - if you press note in low register and high register with same force you will get same sound volume (i think this is a good thing!).

... we have determined only MIDI values, not "sound volumes". Loudness, dynamic and coloration can vary in different registers (depends on the calibration of the virtual instrument).

Re: Kawai VPC1 faulty key weighting

Hi Folks,

I would like to test my piano (Yamaha cp33).
Where do I find Pianoteq's touch curve (Stage)?
Or any other method to test the consistency of the keyboard? Without having to open it and testing voltages -something- by connecting to wires.

Thanks

Have a nice day & happy playin'

Yamaha CP33 -- Scarlett 2i2 -- Yamaha HS7 / Sennheiser HD650 -- PTQ 7 Std [Linux/OSX] -- Some instruments

Re: Kawai VPC1 faulty key weighting

groovy wrote:
gasparka wrote:

Also, my measurements were done with Pianoteq touch curve (green led), sorry i failed to mention that.

You should do one finite measurement, there are a touch too much inconsistencies now:
* non-default VPC1 curve
* convention A0-C8, whereas Pianoteq is using A-1-C7 for 88-keyboards.
* let-off simulation in use or not.
* unzoom to full MIDI range (0-127)

This should be more clear now:
https://i.postimg.cc/PqdTDdD3/vpc1-new.png

I agree that it is more correct to measure with linear curve, i will do it tomorrow. Today i already tested the critical spots, and looks like the curve is almost the same, just higher values. But we will see.

groovy wrote:
gasparka wrote:

For example, we can see that for CA78 the keyboard is not that much graded - if you press note in low register and high register with same force you will get same sound volume (i think this is a good thing!).

... we have determined only MIDI values, not "sound volumes". Loudness, dynamic and coloration can vary in different registers (depends on the calibration of the virtual instrument).

Fair point!

Re: Kawai VPC1 faulty key weighting

Thank you very much, gasparka, a lot of work!

gasparka wrote:

I agree that it is more correct to measure with linear curve, i will do it tomorrow. Today i already tested the critical spots, and looks like the curve is almost the same, just higher values. But we will see.

Yes, I would expect a static offset and depending on the shape of the customized VPC1-curve (green led) more or less fluctuation.

Re: Kawai VPC1 faulty key weighting

Here is VPC1 default curve. I only measured white keys, because it is simpler and it is friday
https://i.postimg.cc/MZdsWPPg/vpc1-default.png

I am going to calibrate the key dip and then make final measurements.

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

Here is VPC1 default curve. I only measured white keys, because it is simpler and it is friday
https://i.postimg.cc/MZdsWPPg/vpc1-default.png

I am going to calibrate the key dip and then make final measurements.

There seems a big difference in keys 43 and 45, the 45 key at a guess would have a lower key dip than 43. There is also a possibility of the foam pad having a more worn or damaged in some way, I would suspect key 43 in this case

Re: Kawai VPC1 faulty key weighting

That's the expected positive offset, resulting from the difference between VPC1's Normal and Pianoteq preset-curve around velocity 60:

http://www.kawaivpc.com/en/images/features/ph0801_b.png

Strange, that your measured black and white key curves are parallel. How many rubber/silicone domes has one replaceable stripe? Eight?

@rdwing: Your eight biggest corrections are at eight adjacent keys. Hm.

Last edited by groovy (24-11-2019 13:00)

Re: Kawai VPC1 faulty key weighting

In the upper fotos it seems the contacts are organized in groups of 12 in a VPC1. For example groups of 12 diodes. One octave has 12 notes. Probably the rubber strips have 12 domes also (can somebody confirm this?).

@gasparka
In my opinion your spikes are in octave/12 tone ranges, see my markings in your graph:

https://i.postimg.cc/KYNZyq7H/vpc1-gasp-octaves.png

Would be interesting, whether these octaves match with a single rubber strip or one of those 12-pin balance rails I have seen once in a Kawai CA95.

Re: Kawai VPC1 faulty key weighting

groovy wrote:

In the upper fotos it seems the contacts are organized in groups of 12 in a VPC1. For example groups of 12 diodes. One octave has 12 notes. Probably the rubber strips have 12 domes also (can somebody confirm this?).

@gasparka
In my opinion your spikes are in octave/12 tone ranges, see my markings in your graph:

Would be interesting, whether these octaves match with a single rubber strip or one of those 12-pin balance rails I have seen once in a Kawai CA95.

What exactly do you mean by the "rubber strip", i am assuming the rubber around the switch logic?
Today i measured the key dip difference and there is a correlation. BUT, actually situation gets worse if i try to correct them. So it looks that the key dip is "deliberately" wrong in order to compensate for some other error. I could make the gap even greater and probably get rid of the MIDI value issues but then i have large key dip issues...
Needs further investigation.

Re: Kawai VPC1 faulty key weighting

gasparka wrote:

What exactly do you mean by the "rubber strip", i am assuming the rubber around the switch logic?

Hello again,

probably  rubber contact strip is more precisely (though it seems to be silicone with graphite to me). One example from an unknown keyboard:

https://www.syntaur.com/images/4156-Lg.jpg

The VPC1 has three contacts instead of two per "bubble" and I'm not sure, if it are twelve in a row. I think so, because the Kawai ES8 has twelve per strip and the arrangement is from A to Ab in each octave. -

What do you mean with "key dip difference"? Do some keys go down to 11 mm and some to 12 mm, just for example?

Are those balance rails under the wooden keys also segmented to units of 12?

Last edited by groovy (26-11-2019 16:24)

Re: Kawai VPC1 faulty key weighting

gasparka wrote:
groovy wrote:

In the upper fotos it seems the contacts are organized in groups of 12 in a VPC1. For example groups of 12 diodes. One octave has 12 notes. Probably the rubber strips have 12 domes also (can somebody confirm this?).

@gasparka
In my opinion your spikes are in octave/12 tone ranges, see my markings in your graph:

Would be interesting, whether these octaves match with a single rubber strip or one of those 12-pin balance rails I have seen once in a Kawai CA95.

What exactly do you mean by the "rubber strip", i am assuming the rubber around the switch logic?
Today i measured the key dip difference and there is a correlation. BUT, actually situation gets worse if i try to correct them. So it looks that the key dip is "deliberately" wrong in order to compensate for some other error. I could make the gap even greater and probably get rid of the MIDI value issues but then i have large key dip issues...
Needs further investigation.

I tried playing years back with what I thought were good velocity levels and a key dip that was very uneven to achieve this, it was very distracting so I made my priority the key dip and at the time made some adjustments using the VPC 1 editor. As groovy pointed out, there were two areas, each an octave that give higher readings. I would correct the key dip and lower these areas with the editor (after the key dip is done). The problem can start about here when you've made adjustments throughout the whole keyboard and have to figure out a reference point, again you have to make a choice to which area should be the reference in both key dip and velocity readouts. A key or key area that can hit 127 as well as 0 is the obvious choice. Lowering velocities in the editor is (in my opinion) preferable to raising them, achieving 0 or 1 is much more important than hitting 127 and loosing the ability to play softly

Last edited by MeDorian (26-11-2019 18:38)

Re: Kawai VPC1 faulty key weighting

MeDorian wrote:

The capstan adjustment only serves one purpose, to set the key dip, [...]

Just saw, that the new Kawai Grand Feel Compact (GFC) action does not have adjustable metal capstans. Simply a bump at the hammer and a flat cap at the key end (nylon rubbing on nylon?). Hopefully this construction is more sturdy and does not need calibration for an even key dip, once the awaited VPC2 is released ...

https://www.youtube.com/watch?v=UcZBDSJ9wDc

Re: Kawai VPC1 faulty key weighting

groovy wrote:
MeDorian wrote:

The capstan adjustment only serves one purpose, to set the key dip, [...]

Just saw, that the new Kawai Grand Feel Compact (GFC) action does not have adjustable metal capstans. Simply a bump at the hammer and a flat cap at the key end (nylon rubbing on nylon?). Hopefully this construction is more sturdy and does not need calibration for an even key dip, once the awaited VPC2 is released ...

https://www.youtube.com/watch?v=UcZBDSJ9wDc

Yes the capstan is a poor design, I think I was ahead of Kawai replacing the capstan three years ago but glad they spotted this ( sorry if I sound condescending?). If one is to modify ( not recommended unless you've fully thought out the process) then some form of calibration will need to be done on the actual end key (below the weight) e.g different size risers (plastic) covered with felt or whatever material you think is best. The VPC 1 though will always have a problem with the key length so although things can be improved (with a great deal of time and work) it will never be perfect

Re: Kawai VPC1 faulty key weighting

groovy wrote:
MeDorian wrote:

The capstan adjustment only serves one purpose, to set the key dip, [...]

Just saw, that the new Kawai Grand Feel Compact (GFC) action does not have adjustable metal capstans. Simply a bump at the hammer and a flat cap at the key end (nylon rubbing on nylon?). Hopefully this construction is more sturdy and does not need calibration for an even key dip, once the awaited VPC2 is released ...

https://www.youtube.com/watch?v=UcZBDSJ9wDc

I do hear click noise on the return bounce on the video clip, I tried nylon on nylon and although super fast was too noisy, felt on the end key (self adhesive was good in my case) was a very good compromise as it maintained much of the speed and solved the bounce issue

Last edited by MeDorian (27-11-2019 13:25)

Re: Kawai VPC1 faulty key weighting

MeDorian wrote:

I do hear click noise on the return bounce on the video clip, [...]

I hope these noises in the video are just amplified by the plexiglass, stand, desk and a cheap smartphone :-/

Kawai is famous for quiet DP actions. Hard to believe they didn't have an eye/ear for that on the latest development GFC.

BTW what about the Roland PHA 50 action, isn't it plastic-on-plastic too? Louder/quieter than the Grand Feel Compact action?

Last edited by groovy (27-11-2019 18:14)

Re: Kawai VPC1 faulty key weighting

groovy wrote:
gasparka wrote:

What exactly do you mean by the "rubber strip", i am assuming the rubber around the switch logic?

Hello again,

probably  rubber contact strip is more precisely (though it seems to be silicone with graphite to me). One example from an unknown keyboard:

https://www.syntaur.com/images/4156-Lg.jpg

The VPC1 has three contacts instead of two per "bubble" and I'm not sure, if it are twelve in a row. I think so, because the Kawai ES8 has twelve per strip and the arrangement is from A to Ab in each octave. -

What do you mean with "key dip difference"? Do some keys go down to 11 mm and some to 12 mm, just for example?

Are those balance rails under the wooden keys also segmented to units of 12?

I will take a look at the rubber strips and balance rails this weekend. Key dip - basically yes, actually i measured the hammer resting distance from the "strings", same thing.

@MeDorian
I agree that losing soft notes is worst and uneven key dip also quite bad, so probably high notes must go