Kawai VPC1 faulty key weighting (Page 3 of 4)

This is a link to the picture of the paint damage mainly on the edge of the piano top:

https://www.dropbox.com/sh/ki4pq294qjk8...ZLHOa?dl=0

Question: does anyone have experience with repainting. I think is it a special kind of paint, probably an imitation of a lacquer layer, often used on acoustic piano's. But Kawai never answered my question about this. Thank in advance!

Sjoerd

A question for who played many Kawai and real grand pianos:

What's the difference, in feeling and touch sense, of play a real concert grand and a Kaway DP with Grand  Feel Action II or III (long pivolt point as in a concert grand) ?

Nick,
Thanks for this sound advice! I will now at least postpone the cleaning action until I do suspect wrong triggering, noisy keys etc.

The paint to my knowledge is Enamel on the VPC1? My keyboard has been re-sprayed although as groovy quite rightly pointed out, this is OT (I read it as 'over the top' OTT might be wrong?). I had too much time on my hands and wanted something other than the standard VPC1.

Regarding Beto-Music pics, I think the Grand Feel III looks much improved over II, and way better than the VPC1. The key length on the VPC1 and the capstan are its downfall (curved top doesn't score highly either).

Roland A 88 Mk II looks promising (having MIDI 2.0) as a future replacement or VPC2 if it should ever be made/released. Just thinking ahead, sorry off thread somewhat.

Nick

EDIT: OT means off topic...had to look that up.

This is my first post in months because of one particular person on the forum who was not cordial. I used to love visiting the forum but no longer wish to. You never know what sarcastic remarks will he hurled at you.

So, just try to ignore them, if you can. I've not been able to. It's farewell from me. I have a VPC1 as well and appreciate your wanting to find out more about it.

My best to you,
Robert Cooper Rivard

https://isolatedpianist.hearnow.com/

The octave-wise dislocation could also be due to a bad positioning of the rubber contacts. There are quite a few screws that fix the sensor boards and make sure that the rubber contacts are pressed tightly down against the hammers. One could check if those screws are sitting tight and if they are all there at all.

But, no, digital pianos don't require maintenance or regulation. Lol.

I wish Bechstein would release a grand piano keyboard with open-source electronics for DIY people where piano regulators/service technicians could actually start to come in and service those too, ie. establish some sort of advanced standard for digital pianos where you can keep your high quality mechanics forever because you can maintain it and the electronics without vendor support.... can we contact the EU about this? Who would that contact even be? Because the way digital pianos are constructed nowadays reminds me of planned obsolescence. It doesn't even have to be a law. Just a sticker or something that informs the customer about how dependent he will be on the will of the manufacturer for keeping his stuff in good condition.

This has nothing to do with the topic, but its the only vpc1 i found and i didnt want to create a new thread for it. I lost my Power supply, does anyone has a Photo or can tell me which one of kawais (maybe the ps154?) fits for the vpc1. I know i can play it via USB power, but my USB Ports are limited

I know this is an old thread, but could someone please tell me the current state of the current run of VPC-1 and what the verdict is on it?

Have the quality control and other issues been resolved?

I'm looking at either getting the VPC-1, ES520 or ES920 from Kawai, or Yamaha P515 or Roland FP-60X or FP-90X.  MP11SE was highly regarded as best all-arounder and arguably most closely resembling an acoustic piano action, plus longer key length and great build quality, but at $2800, it's venturing close to entry pricing for a standup acoustic piano.  So I'm thinking I should look at $1800 and below for a digital piano and save for one day having an acoustic piano.  From what i understand, there are some nuances that can be achieved on a real acoustic piano that may not be possible on modern day digital pianos, even though they've come a long way.

I hear great things about Yamaha P515, and they have a solid reputation for consistency and quality control.  I hear mixed reports about the same for Kawai and Roland.

VPC-1 and P515 are kind of old now.. as is MP11SE... Roland FP-90X is enticing because it seems to be the most recently released and therefore has the latest support and features.

I'm looking for something that's the closest to the action/feel of a real acoustic piano, with consistent quality control and reliability for years to come.

Willing to go with a controller instead if the price is right, but also hear some people say the hassle of a controller + computer + VST setup can interfere with being able to play out ideas quickly, which can impact creativity.  I don't know how big a deal or a hassle it really is, and whether that can be mitigated by having a computer and VST setup ready to go and always on?

Thanks in advance for your inputs!

I have the Yamaha P515 (White). I love the build quality, probably one of the best, if not the best builds out there. However, it was not designed to be a controller so the response from Pianoteq is not optimised. I continue to work on the velocity curve.

Warmest regards,

Chris

I too was surprised by a lack of calibration of the NoteON MIDI velocity. In particular there are two octaves that sound much louder and I had to use a velocity offset of up to -20 with the VPC1 Editor (= obviously by doing this I lose the dynamics upwards and this is not a definitive solution if you want to play also strong).

I have carefully read the various types of problems or user customizations. Currently the 1.05 firmware allows you to do many interesting things (for example the two sensor mode is very useful if you want to have a better sense of keyboard connection until you reach the bottom of the key) but the velocity compensation should be done better than a simple offset that decreases the dynamics. In practice you need to be able to adjust the time interval (ms) between the sensors and the debounce threshold, maybe they will do it with the next firmware. By doing so, the dynamics (0-127) would remain intact.

After reading the excellent "VPC1 Keyboard Scanning Prepared by Jack O'Flaherty January 31, 2014 "I have the possibility to build a MIDI encoder with the following electronic board
https://www.olimex.com/Products/OLinuXi...e-hardware
With all those contacts I was already using it for the diode arrays of three organ keyboards (two sensors) and a pedal board. It wouldn't be a problem to reuse it for a single piano keyboard with three contacts.

In relation to my experience with VPC1 I already have some ideas to implement:

1) set the time interval between two adjacent sensors for the minimum and maximum velocity possible (key by key and NoteON/NoteOFF)

2) set anti bounce threshold (NoteON/NoteOFF)

3) set the scanning frequency of diode arrays (currently set at 11 kHz as mentioned by the document)

4) 7bit / 14bit (CC # 88 hi-res midi) mode

...

x) what you want

When I find a used or out of warranty keyboard then I will start the work, we hope it will help to a greater customization for the end user...The keyboard is very elegant and with a good feeling, in my opinion it deserves more than what Kawai offers today.

Hi cauldron, -20 is a huge amount to offset, the keyboard must be in a bad way. This would be like moving the condition slider in Pianoteq to full worn, similar anyway. Good luck there, I'm sure you'll sort it.

Nick

Edit; Not quite as bad as I thought, on my keyboard, if I play note-on value of 80, a -20 offset will be around 50 note-on value (approx). This is just as I say a rough test.

RUBBER CONTACT STRIPS ARE THE PROBLEM
Hi everyone.
I have Kawai CA-78 piano which should have the same contact boards and contact rubbers as VPC-1.
There is a problem in the rubber contact height, I think. It seems Kawai is not able to produce them exactly the same.
Why do I think so? Because swapping rubber contacts helped to alleviate my problem. :-)
In picture 1 you can see my midi output when I bought the piano.
In picture 2 you can see my midi output after I swapped the rubber contacts according to their MIDI value for the same weight of 380 g.
Originaly, notes A1-G#2 were a bit louder and notes A3-E4 were significantly louder (picture 1).

The rubber contacts come in one-octave strip (12 notes) and can be easily divided into 4-note segments. However (!!!) Left(L), Middle(M) and Right(R) segments can not be interchanged!!! The spikes on the rubber and holes in the board are slightly different for L M and R segments, so "L-rubber" must go to "L-holes", M to M and R to R. Otherwise you will get some notes extremely loud while others dampened (tested).

I marked each rubber segment with a number from 1 to 22 and measured its MIDI value in one spot of the keyboard (A1-G#2) with the same 380 g weight.
Then I aligned them according to their increasing MIDI value.

Original sequence of rubber segments was 1-2-3-4-5-6-7-8-9-10-11-12-13-14-15-16-17-18-19-20-21-22.
New sequence is 19-20-21-13-14-15-1-2-3-7-8-9-22-17-12-16-5-18-4-11-6-10.

BEFORE, the biggest MIDI difference between adjacent notes was 17, it is 7 AFTER. As it is in the higher note range, I am not able to percieve the difference anymore.

It seems that notes in one strip tend to be around the same MIDI value and the difference is only between whole 12-note strips.
But because there are two contact boards, each 44 notes (44 is not divisible by 12), it gets a bit messy.
On the left board, there is a 8-note strip (L,M) at the end (in the middle of keyboard) - segments nr. 10 and 11.
On the right board, there is segment nr. 11 (R) at the begining (in the middle of keyboard) and segment nr. 22 (L) at the end of the right board.
In my case, segments 10, 11, 12 probably do not come from the same octave strip (10 and 11 are much louder).

Picture 3 shows final result after I obtained spare rubber contacts from Kawai technician (after 1,5 years!), measured them and assorted by "sensitivity" (MIDI value with 380 g weight).

Just be aware that there are no "faulty" rubber strips. Rather, there is variability among them. I got 8 new strips from the technician and they also ranged from 63 to 74 MIDI output in the same position with the same 380 g weight (63, 64, 66, 69, 70, 74, 74, 74 to be more specific). So I jsut measured everything I had, assorted it and now I have two complete sets of rubber strips plus two unused 4-note segments. First set ranges from 59 to 66 MIDI value ("quieter", in use, picture 3), second 68-75 (I just keep it as a spare one in my drawer). And I use the louder strips for higher notes, as the hammers are lighter and supposed to be more responsive anyway.
As I say, it seems Kawai is not able to manufacture the rubber strips exactly the same and this is the best I could get.

PS: If there is high variability within one rubber strip, it can be due to glue on the sliptape at the end of the key that got into touch with the bolt on the hammer. Cleaning with rubbing alcohol helps, I would not recommend any lubricant. I am going to replace this  Achilles heel with furniture anti-scratch felt soon. :-)

Sorting the rubber strips according to their increasing "sensitivity" and then rearanging them seems faster and easier IMHO, but I also understand your temptation to do a bit of your own electronic circuitry. :-) Good luck!

Yes, I used 380 g for all my graphs. It was just the easiest solution for me. I grabbed a small PET bottle, filled it with water and I used rubber eraser in between the bottle and the very end of each key. Kitchen scale showed 380 g for bottle+water+eraser, I am pretty sure. :-)
I measured 3 times each key and then took median value.
My piano, Kawai CA-78, has quite a bit longer (heavier) wooden keys than VPC-1, so that might play a role.

One problem I see with SW calibration is that when one specific key moves more slowly (not lubricated or may be glue on the sliptape), you would feel heavier touch, yet "normal" loudness that does not correspond to the "feel" of the touch.
So you would need constant speed, not constant weight for calibration.

Yes, I agree, your solution would work. I have noticed it. I am just not sure how feasible it is...
Using some kind of weight is much easier at home.
If you succeed, share your results, please. :-)

Someone from Kawai should read this thread and do something about it, because apart from the unevenness of response, I am happy with realism of their keyboard. And it can be a dealbreaker for many.

As I wrote before... the rubber strip contacts are not manufactured exactly the same. From your measurement for "ON" and "OFF" note it seems that the 3rd contact (the one that sends "ON" signal) differs the most between individual rubber strips. Does it make sense?

Nice graphs, cauldron, alas not scaled all the same on the y-axis. As I mentioned earlier, 0 - 127 makes all those nice contributions in the thread more comparable.

MIDI ON and OFF is asymmetric on a 3-sensor system. Movement from 1 to 3 triggers ON and return from 3 to 2 triggers OFF.

Another asymmetry might be, that the key is deforming the rubber bubble constantly from 1 to 3 , but if the key lifts faster on return, than the bubble expands freely from 3 to 2. But never saw a slowmotion of the return, so not sure.

-> KAWAI 3-SENSOR SYSTEM

Try swapping rubber contacts (whole octaves, 12 notes together) this way. (From 1-2-3-4-5-6-7 to 4-1-2-7-6-5-3, leave the 4 rightmost notes unchanged).
You should get rid of the "loud" steps and get increasingly sensitive keys going from left to right. :-)
That is what I did (picture nr. 2 in my fist post) and it worked pretty fine.

Thanks to ribol for the patience and the advice with the images that can be an example for other people (unfortunately I cannot open the keyboard still under warranty, next week I will be able to do it in the authorized service center). When you get the chance I would like to ask if it is possible to see internally the circuit to understand how to connect to the diodes on the side of the three 74HC138 chips (U1-U3) that come out with 22 lines (T0-T21). I hope they can be in DIP packaging so that I can take them out with pliers to apply my contacts. (When I am ready I would like to distribute a public version fully usable for experimentation (the Olimex board indicated is very performing and the USB OTG interface is able to pass up to 40 MIDI events every millisecond on the gadget_midi linux interface!)).

You are right, I have studied the behavior of the VPC1 key without sound, only with a MIDI event monitor.
Practically both in 2 or 3 sensor modes and with the button completely raised, the NoteON occurs only at the end of the button stroke, so to speak when you hear the double escapement and from there you continue to the end (we are therefore talking about sensors 2 and 3, sensor 1 appears not to be used).
It is also interesting to note that it is possible to press the button again while remaining at the bottom of the double escapement (sensors 2 and 3). Many successive NoteON are created (without an immediate NoteOFF). Finally, when the key is raised beyond the double escapement (up to sensor 1) then all the NoteOFF appear with the velocity of the last measurement. Here is an example of MIDI events in time sequence with various repetitions on the double escapement:

note number, velocity, MIDI event
60 23 Notes ON
60 34 Notes ON
60 42 Notes ON
60 53 Notes ON
60 12 Notes ON
60 88 Notes OFF
60 88 Notes OFF
60 88 Notes OFF
60 88 Notes OFF
60 88 Notes OFF

[This also explains some observations I had noticed when playing an A0-A6 cluster and seeing some notes repeated with very low velocity (= this is a small bounce).]

What I can understand from these tests is that there are 2 sensors (2 and 3) after the double escapement and one (sensor 1) before the double escapement. Below are the images that indicate a stroke of about 5mm from sensor 1 to sensor 3. Basically I understand that NoteON means sensors 2->3 while NoteOFF sensors 2->1.

Sensor 1

Sensor 2

Sensor 3

Consider https://github.com/jkominek/piano-conversion instead and see the large discussions about it on PC, e.g. at https://pianoclack.com/forum/d/276-buil...lectronics or https://pianoclack.com/forum/t/technician

Best of luck!

You can process MIDI events after the MIDI keyboard with a professional product available both as software and as hardware:
https://www.bome.com/products/miditranslator
https://www.bome.com/products/bomebox

You can process MIDI events after the MIDI keyboard within a DAW (Digital Audio Workstation): Reaper, Ableton, Pro Tools, Reason, ...
Obviously you have to use the VST / AU / AAX interfaces inside the DAW so Pianoteq will be initialized as a VST instrument component inside the DAW.
Each DAW can have its own pseudo-language for MIDI/Audio processing, the important thing is zero latency real time processing!

Personally I can tell you that I have successfully used the Ardour open source DAW writing some real time Lua scripts but in your case it is much easier to use the Bome products for simulations/production.

Try experimenting with your formulas, if you see that they work I can implement them in my encoder if you are interested in trying it.

Bome Midi Translator uses basic math. You should calculate the square root with Heron's method and use Windows or MacOS, linux is not supported. With linux if you feel comfortable with Perl I would use it in the simulation phase, to see that everything works according to your expectations (your formula is complicated for me).

For the production phase if you have to focus on linux I think you won't have better choices than Ardour DAW. Through Lua DSP Scripts [Audio/MIDI processor - plugins with access to the Ardor session (per track/bus, real-time)] you can instantiate a processor that works synchronously and without delay with the audio/midi stream, ALSA or Jack Audio.

For MIDI event processing there are many Lua DSP scripts in Ardour DAW (inside ./share/scripts/) that can help you understand how the Lua language works within Ardour.
https://www.lua.org/about.html
https://manual.ardour.org/lua-scripting/

These examples can serve as a prototype for your MIDI processing: midimon.lua , midi_remap.lua , _midifilter.lua , ...
This is a very powerful and multi-platform solution that depends only on the Ardour DAW.