Piano World Home Page
Some people may have followed my odyssey in the DIY real grand action controller? thread. It started two years ago with me researching the possibilities of creating my own MIDI controller from a grand piano action and optical sensors measuring velocity at hammers, similar to Yamaha AvantGrand and Kawai Novus. From that thread you will see that I was rather unprepared for that, had to learn electronics almost from scratch, had to learn how to do piano action regulation. I also had to learn how to work with wood and metals, etc. I'm also a father of a 2.5 year old, a software developer with a permanent job, and on top of that I purchased a Yamaha N1X one year ago. So, expectedly I didn't have a lot of time for this project which is why it took me so much time. But ultimately I managed to finish it (well I need to make a cabinet now). Whoever is interested, can follow details in that other thread. No more words, here's the result:

Simply outstanding!
Absurdly impressive. How would you say it compares with, say, a Novus10?
Amazing !
Very good job.
So how much do you think you will sell your CyberHybrid 1 ?
Originally Posted by BachToTheFuture
So how much do you think you will sell your CyberHybrid 1 ?


Cybrid?
I was going to ask you guys for name suggestions but I already see some good ideas laugh
Cybrid & Sons? wink
Congratulations CyberGene. That is a superb "Proof of Concept."
Truly impressive, an awe-inspiring effort, and an awe-inspiring result. Congratulations!

And, also, beautiful playing. Inspiring in its own right.

As for a name ... CyberGenie?
Impressive work cybergene, it seems you nailed it.


A few questions, how does it compare to your N1x? How can I order one? How does it pair up with pianoteq? eek
Awesome accomplishment CyberGene!
Well done CyberGene, amazing work.
I am impressed, nice work!
Genybrid ?
Wow! Very nice. Let me know when you start to take orders.
Wow !

I'm in absolute awe. What's your next project going to be ? I can't see how you'll top that one !!!

If I could make a couple of suggestions, how about fixing cold fusion or if you get a spare afternoon, could you resolve the Grand Unification Theory ?
Originally Posted by AndyOnThePiano
Wow !

I'm in absolute awe. What's your next project going to be ? I can't see how you'll top that one !!!

If I could make a couple of suggestions, how about fixing cold fusion or if you get a spare afternoon, could you resolve the Grand Unification Theory ?


LOL 😁
There'd be a Nobel Prize in it for you.....
As my grandmother used to say “I’m flummoxed”. Amazing Gene.
Wow, very impressive indeed! Having followed the other DIY thread, it has been quite a ride with such a great conclusion!

(I’ve visited these forums for a while, deciding on my next piano, but now I just had to register in order to add my congratulations to this amazing accomplishment... Btw, ended up with the N1X, coming in a few weeks!)
Really impressive work and very cool project- congratulations to the awesome result and nice playing too!
+1 for a comprehensive HowTo Guide pls!
The piano looks and sounds great.

I think there is a market out there for this thing. If you can reliably source grand actions and improve the process so as to reduce the amount of manual labor involved, there is a nice business opportunity here.
Originally Posted by NoMeGa
Wow, very impressive indeed! Having followed the other DIY thread, it has been quite a ride with such a great conclusion!

(I’ve visited these forums for a while, deciding on my next piano, but now I just had to register in order to add my congratulations to this amazing accomplishment... Btw, ended up with the N1X, coming in a few weeks!)

Congrats on the N1X! I wonder where Tyrone is, he hasn’t posted in a while and he keeps statistics on N1X owners.
So the CG1X is finally here! That's an amazing achievement CG! Very impressed!
This is amazing! So how does the action compare to your N1X? smile
Impressive work!
I don't think I've ever read so much acclaim here on the board. But it's well and entirely deserved. Congratulations CG.
WOW
personally, I'd like to know the details and what you went through to create it. I'm sure I'm far from the only software engineer or EE around here who is curious what you went through to build it.
Fantastic job really, congratulations CyberGene !
@CyberGene: Here a very important question for you. Is your new creation a hybrid?
Thanks for the kind words everyone! Glad what I did is well received smile

Let's answer some questions.

- Comparison with the N1X. I was gonna say the N1X had slightly better touch response. But I couldn't live with that statement just like that. I was exhausted and too busy the last days. I got some good sleep today and decided to do something I should have done earlier, measure some MIDI values. So far I've been only "tuning" the piano aurally by turning the trimpots and listening to the sound from the Garritan CFX. Turning the trimpots is like a touch curve but on a per key basis. Anyway, I played the N1X and my controller side by side and immediately noticed that the N1X has wider touch response range. And so I measured the MIDI values. N1X gives MIDI values between 10 and 110 (you can produce quieter and louder velocities, but harder). And I measured mine and immediately realized it was generating velocities within a compressed range of 20-90. And then I slapped my face because I remembered I forgot to use the logarithmic multiplier I had in my code. I experimented a little, this time with a MIDI monitor on my iPhone and easily found the value that would give me a similar range to the N1X and voila! The pianos are indistinguishable in terms of velocity response! This is simply fascinating. I will admit I didn't expect my controller to be that good, honestly! The only difference is of course the actual touch. The N1X is a brand new, well regulated keyboard from a world renown piano manufacturer. Mine is an unknown action, probably 100-120 years old. I replaced many things and I regulated almost everything but there are still some things that may need more regulation or repair. Also, there are no lead weights in my keyboard. All that being said, it feels VERY good, it's certainly not cr*p. It's only different than the N1X, not certainly worse. The feel is also just a bit more springy compared to the N1X but I suspect the felt bushing that the rear part of the keys rest on. I ordered two types of felt from China and the more mushy one arrived first. I now feel a bit sorry for not waiting for the second one but... Anyway, it's not that bad. It's not the bottoming out that's affected but rather the dampening of the keys, when they reach the top position - instead of firm, it's a tiny bit more springy.

- Challenges. Well, basically I encountered all kind of challenges, as you can see in the other thread. I studied Physics at the university but I never liked it and I started working as s software engineer in my third year and barely finished my university courses. I've chosen the least desirable specialization "Optoelectronics and semiconductor manufacturing" because there were free positions and while one may say it's exactly what's needed to make a digital piano with optical sensors, well... I didn't like it and don't remember ANYTHING! I started learning electronics from scratch. I didn't know Ohm's law. Also I didn't know what pull-up resistor is which is why I designed my first version of PCB-s without pull-up resistors laugh And spent days scratching my head what was wrong. Anyway, I redesigned the PCB-s. So, everything was like that - I had to learn and read materials on the Internet for basically everything, starting from Wikipedia smile I improvised with aluminum profiles, cut them with a hacksaw and regular hand drill. I wasted a lot of them until I managed to make something good and stable. I didn't expect that I will have to regulate the action, nor restore it. That frustrated me a lot and was the reason why I almost abandoned the project for half a year. Ultimately I sat on my a** and started reading the piano book by Reblitz, watched YouTube videos, etc. There were also challenges with writing the Teensy code. I am a lazy Java developer and although I work mostly on backend stuff (multithreading, algorithms, data structures) I'm not used to perform such a low level code optimization nor am familiar with CPU-s, controllers, etc. And the guys on the forum, well, some are helpful, but mostly are a bit "elitist" and what they would tell me is: there are examples, look at the examples!!! Whatever wink

- Sharing the code, schematics, etc. Well, I need to think of all this. It turned out good. I spent quite some time to design it, make some choices, gather some know-how. I wouldn't make another one myself again for sure, it's too exhausting. But I may turn this into a sellable kit, or a Kickstarter project or something like that. I'm open to suggestions smile
Good job. How do the optical sensors work? I believe the Yamaha disklavier uses gray scale flags to sense key and hammer velocity.
Originally Posted by LarryK
Good job. How do the optical sensors work? I believe the Yamaha disklavier uses gray scale flags to sense key and hammer velocity.

I use optical proximity sensors at the hammer shanks.
Originally Posted by CyberGene
Originally Posted by LarryK
Good job. How do the optical sensors work? I believe the Yamaha disklavier uses gray scale flags to sense key and hammer velocity.

I use optical proximity sensors at the hammer shanks.


Can you show us what that looks like? This page describes the anatomy of the Disklavier:

"Record Mechanism
Disklaviers use a number of different sensors to determine the timing, speed of movement, and position of keys, hammers, and pedals.

Unlike consumer-quality retrofit systems from other companies, contemporary Disklaviers are distinguished by their ability to record incremental positions of the left and right pedals. It is necessary to capture incremental pedal movement in order to provide realistic playback.

Today, Yamaha’s standard and PRO model Disklaviers use sensors under every key as well as advanced gray-scale sensors on the hammer shanks in order to determine the timing of notes, the velocity with which the hammers hit the strings, and the speed of the release of each key. At the present time, these Disklavier models are the only recording pianos on the market that record with this level of precision.

Contemporary standard and PRO Disklaviers capture the complete range of piano expression and even record brushed notes (silent notes that result from slight key movements that do not produce audible sound)."

from: http://yamahaden.com/anatomy-of-a-disklavier
Yamaha’s solution with sensors at hammers and under the keys is a unique one. I can hardly compete with a 20-30 years of iterative development by the world’s biggest electronic and acoustic piano manufacturer, a huge Japanese corporation that also makes consumer electronics and vehicles smile

My solution is closer in principle to Kawai Novus in that I only measure hammer velocity. I’m not saying my solution is as good as either AG or Novus but a rough comparison between my N1X and my own DIY controller affirms that my solution is comparable in terms of dynamics. Whether that will be consistent and durable in the long-term is another story though.
Originally Posted by CyberGene
I remembered I forgot to use the logarithmic multiplier I had in my code. I experimented a little, this time with a MIDI monitor on my iPhone and easily found the value that would give me a similar range to the N1X and voila! The pianos are indistinguishable in terms of velocity response! This is simply fascinating. I will admit I didn't expect my controller to be that good, honestly! The only difference is of course the actual touch. The N1X is a brand new, well regulated keyboard from a world renown piano manufacturer. Mine is an unknown action, probably 100-120 years old.
Nice catch!
Originally Posted by CyberGene
And the guys on the forum, well, some are helpful, but mostly are a bit "elitist" and what they would tell me is: there are examples, look at the examples!!! Whatever wink
hahaha.The "unexpected" result of stuffing a bunch of piano players in a room.

That said, I sense that virtually all the commenters were trying to help. Some, like me, didn't have the skills or time to fully vet their ideas but wanted to present ideas that might be worth considering. Some of the comments were stellar and on-point. Overall, I thought the community was very positive on your project. You might miss some of that excitement being "in the weeds" and with English as a second language (your English writing is superb by the way).
newer_player, I meant the Teensy forum, not PW. All here was very positive smile

P.S. Didn’t want this to be critical of any forum. I’ve learned a lot from all the forums smile
It is an awesome project - congratulations once again!
Quote
and what they would tell me is: there are examples, look at the examples!!!


that's just their version of "get a teacher!"
Originally Posted by U3piano
Quote
and what they would tell me is: there are examples, look at the examples!!!


that's just their version of "get a teacher!"

Actually I’m not right. It turned out I used code exactly as in the examples. Not fair to say anything bad about Teensy forum at all. I was in a hurry and wanted quick answers but ultimately I had to learn the examples. Apologies to any Teensy forum member reading here 🤝
Originally Posted by CyberGene
Yamaha’s solution with sensors at hammers and under the keys is a unique one. I can hardly compete with a 20-30 years of iterative development by the world’s biggest electronic and acoustic piano manufacturer, a huge Japanese corporation that also makes consumer electronics and vehicles smile

My solution is closer in principle to Kawai Novus in that I only measure hammer velocity. I’m not saying my solution is as good as either AG or Novus but a rough comparison between my N1X and my own DIY controller affirms that my solution is comparable in terms of dynamics. Whether that will be consistent and durable in the long-term is another story though.


Haha, of course! I was just curious how you were measuring the velocity in comparison to how Yamaha is measuring the velocity.

I'll stick with my order with Yamaha unless you can scale up very quickly! LOL.

I still have this nagging question of whether 1024 levels of velocity would make an audible difference. Tyrone says it would not be audible. Since Yamaha does not put the Disklavier Pro version in an upright, there is no way to compare the standard resolution system to the high resolution system. I suppose it might be possible to make this comparison with a grand, but I'm unlikely to ever buy a grand.
Pianoteq can interpret high resolution velocity extension. This makes it easy to compare a 127 levels with a let’s say 1000 levels.
VPCG-1, a name idea, since the vpc2 will never be released or compete with yours.
You know what I'd like to call it? Mine!
If you were starting over today, what are some of the things you would change or rethink about your design? E.g., would you still use voltage sensing, knowing that the values can swing as much as they do with different power supplies?
Originally Posted by Frédéric L
Pianoteq can interpret high resolution velocity extension. This makes it easy to compare a 127 levels with a let’s say 1000 levels.


I want to hear a comparison of regular and high resolution velocity values played back on an acoustic piano. What I really want is a Disklavier Pro grand, lol.
Originally Posted by Gombessa
If you were starting over today, what are some of the things you would change or rethink about your design? E.g., would you still use voltage sensing, knowing that the values can swing as much as they do with different power supplies?

Yes, I’ll redesign power supply. Although it’s currently very stable and haven’t moved at all for two days. I’ll split circles and/or make sure each note module (the 17 big vertical PCB-s) has its own voltage regulator, for instance. Currently the modules are power chained and besides the 0.3V voltage drop (not a big deal but yet) there’s also an increasing ripple starting at 10mV up to 50mV p-p at the last one. But probably because of the comparators built in hysteresis (and I also have small caps at each comparator) I don’t notice any influence at all. All else seems great for now.
Well, I’d love to get rid of the voltage sensing and use metal notch profiles or something similar to the Yamaha and Kawai solution, that would be great. But that would only be possible in big commercial manufacturing. I couldn’t afford this for a one-off DIY project. But for now the solution with voltage and trim pots is surprisingly good and stable.
Originally Posted by LarryK
Originally Posted by Frédéric L
Pianoteq can interpret high resolution velocity extension. This makes it easy to compare a 127 levels with a let’s say 1000 levels.


I want to hear a comparison of regular and high resolution velocity values played back on an acoustic piano. What I really want is a Disklavier Pro grand, lol.


A Disklavier is an acoustic piano... just take one, play through an extended MIDI file, the same with HiResVelocity stripped... and compare.

Perhaps http://www.kuhmann.com/Yamaha.htm has some HiRes file (it has in two format : proprietary Disklavier file, MIDI file)

(I suppose that the grand you want is the CFX wink )

-

It is possible that hearing a prerecorded file is not the good comparison. Hearing what we play live is important. When I play the EWQL Bechstein where the layers are badly tuned, I don’t here what I am playing. I think such an issue is easier to spot while playing than when earing a demo.
BTW, had I finished this controller a year earlier, I wouldn’t have purchased the N1X. I still love it, yet I’m not sure the high price is really justified in my particular case where I use it only with headphones. Yes, it has slight playability edge with the VRM, binaural, ultra smooth action and mature implementation. But I have a lot of fun with an old keyboard and a bunch of PCB-s I soldered in the bedroom. Add to that the wonderful Garritan CFX. What is not to like smile Yep, this is a biased opinion of course 😛
Originally Posted by Frédéric L
Originally Posted by LarryK
Originally Posted by Frédéric L
Pianoteq can interpret high resolution velocity extension. This makes it easy to compare a 127 levels with a let’s say 1000 levels.


I want to hear a comparison of regular and high resolution velocity values played back on an acoustic piano. What I really want is a Disklavier Pro grand, lol.


A Disklavier is an acoustic piano... just take one, play through an extended MIDI file, the same with HiResVelocity stripped... and compare.

Perhaps http://www.kuhmann.com/Yamaha.htm has some HiRes file (it has in two format : proprietary Disklavier file, MIDI file)

(I suppose that the grand you want is the CFX wink )

-

It is possible that hearing a prerecorded file is not the good comparison. Hearing what we play live is important. When I play the EWQL Bechstein where the layers are badly tuned, I don’t here what I am playing. I think such an issue is easier to spot while playing than when earing a demo.

Originally Posted by Frédéric L
Originally Posted by LarryK
Originally Posted by Frédéric L
Pianoteq can interpret high resolution velocity extension. This makes it easy to compare a 127 levels with a let’s say 1000 levels.


I want to hear a comparison of regular and high resolution velocity values played back on an acoustic piano. What I really want is a Disklavier Pro grand, lol.


A Disklavier is an acoustic piano... just take one, play through an extended MIDI file, the same with HiResVelocity stripped... and compare.

Perhaps http://www.kuhmann.com/Yamaha.htm has some HiRes file (it has in two format : proprietary Disklavier file, MIDI file)

(I suppose that the grand you want is the CFX wink )

-

It is possible that hearing a prerecorded file is not the good comparison. Hearing what we play live is important. When I play the EWQL Bechstein where the layers are badly tuned, I don’t here what I am playing. I think such an issue is easier to spot while playing than when earing a demo.


Yes, I know, the Disklavier system is installed on an acoustic piano. I owned one 30 years ago, and I have another one on order now. ;-)

I wanted to perform such a test on an upright, but, like I said, Yamaha does not install the Pro system on any uprights. I've been told that the Pro system would be compromised on an upright but I would like to see whether it made any difference.

It would be great to hear someone attempt this test on the one of the Disklavier Pro grands, if it is possible, which I am not sure it is. How exactly does one strip out high velocity resolution values? Are the values mapped to 0-127? Does Yamaha do this to play back the same performance on Standard and Pro models, or do they record takes on Standard and Pro models?

The grand I want is either the DS5X ENPRO, or the DS3X ENPRO: ;-)

https://usa.yamaha.com/products/musical_instruments/pianos/disklavier/enspire_pro/index.html

I'll probably never own one but it is nice to dream.
One way is the to play a record with a DAW.

If the DAW is Reaper, we can use a script which strip the HiRes Velocity event. With Reaper, you can bypass a script and compare full and reduced resolution.

If the DAW is not Reaper, perhaps** we can install the free ReaPlugs which contains the ReaJS scripting engine of Reaper.

** I have written “perhaps” since nearly all DAW support VSTi, but I am lot sure about VST which transform MIDI to MIDI.
Nice project, and congrats on completing it!
Originally Posted by CyberGene
BTW, had I finished this controller a year earlier, I wouldn’t have purchased the N1X.


Here here (or is it hear hear?)! As polished as the production hybrids are, it really sounds like your homebrew solution is really everything you could hope for (well, maybe 85/88ths of everything) smile

The best part IMO? You don't have to sit on the sidelines pontificating with the rest of us about whether x, y, or z feature is beneficial, you could just do it and test firsthand!

So, how about triggering a note-on velocity 1 so long as the key sensor is activated? That's something neither the AG nor Novus currently do wink
Originally Posted by Frédéric L
One way is the to play a record with a DAW.

If the DAW is Reaper, we can use a script which strip the HiRes Velocity event. With Reaper, you can bypass a script and compare full and reduced resolution.

If the DAW is not Reaper, perhaps** we can install the free ReaPlugs which contains the ReaJS scripting engine of Reaper.

** I have written “perhaps” since nearly all DAW support VSTi, but I am lot sure about VST which transform MIDI to MIDI.


I don't know enough about the extended Midi spec to agree or disagree but I guess you do. smile It would be cool if someone took a Disklavier Pro grand and played back standard and high resolution midi files and performed a blind test to see if most people could hear the difference. Yamaha makes and sells high resolution record and playback Disklavier Pro systems so they must think the difference is audible.
Originally Posted by Gombessa
So, how about triggering a note-on velocity 1 so long as the key sensor is activated? That's something neither the AG nor Novus currently do wink

I tried this for a moment: by extending second distance point. But I lost the capability to play off the jack, so I reverted it. And can easily play off the jack now. (and N1X and NV10 can’t do it or it’s not very easy, so I’m already ahead). Another possibility is to detect whether hammer passed first sensor (damper on) and if sufficient time passed without actual note being produced, send MIDI 1 which is what is on my “backlog” of things to test further although it may slightly slow down my scanning and I wouldn’t risk this for such a low priority stuff.
Hmm, actually I can send 1 immediately and unconditionally on first sensor activation. This is cheap operation. Only wondering what would happen on fast notes: 1 followed immediately by actual velocity. If the VSTs will behave with that.
Originally Posted by CyberGene
Hmm, actually I can send 1 immediately and unconditionally on first sensor activation. This is cheap operation. Only wondering what would happen on fast notes: 1 followed immediately by actual velocity. If the VSTs will behave with that.


That's exactly what I what I mean--note-on 1 with key sensor (though I agree it's not worth a significant hit to sensing speed). EVERY DP, hybrids, included, can technically do this today. And doing so is unquestionably acoustic-like behavior. But NOBODY does this. Why?

I think most VSTs ignore velocity 1 and will play a subsequent note-on correctly (as it would be treated like a triple-sensor strike). A couple are exceptions (I've heard VI Labs Truekeys sounds on velocity 1).
WoW, such a project looks like a lot of fun. laugh
Turned out great, congrats for pushing through all the difficulties.

*chuckles* you studied the same thing I did. I wonder how much of my uni stuff I could remember if I needed to.

Oh and really nice playing. Makes me even more envious than your new toy.
Originally Posted by LarryK


I don't know enough about the extended Midi spec to agree or disagree but I guess you do. smile It would be cool if someone took a Disklavier Pro grand and played back standard and high resolution midi files and performed a blind test to see if most people could hear the difference. Yamaha makes and sells high resolution record and playback Disklavier Pro systems so they must think the difference is audible.


MIDI defines several events : Note On, Note Off, Controller Change, etc.

There are several Controller Change (on a piano, each pedal, but synthesisers use other control : modulation, expression, volume...), each identified by a number. A specific Controller Change, the #88, add some precision. The idea is to send CC88=32 then Note on velocity 45... the actual velocity will be 45+32/128=45,25. Then you add 127 velocity levels between each “normal” MIDI velocity level.

If you filter out the CC88, the velocity resolution will be adjusted to 128 levels.

There are multiple Reaper scripts already done. Without any programmation, select MIDI CC mapper, and reassign 88 to 89. The event will then be ignored by the Disklavier which don’t know what to do with such a Control Change.

Note: 1/ Reaper can be fully evaluated for 60 days. 2/ I have no shares on Cockos which edit Reaper.
Originally Posted by Frédéric L
Originally Posted by LarryK


I don't know enough about the extended Midi spec to agree or disagree but I guess you do. smile It would be cool if someone took a Disklavier Pro grand and played back standard and high resolution midi files and performed a blind test to see if most people could hear the difference. Yamaha makes and sells high resolution record and playback Disklavier Pro systems so they must think the difference is audible.


MIDI defines several events : Note On, Note Off, Controller Change, etc.

There are several Controller Change (on a piano, each pedal, but synthesisers use other control : modulation, expression, volume...), each identified by a number. A specific Controller Change, the #88, add some precision. The idea is to send CC88=32 then Note on velocity 45... the actual velocity will be 45+32/128=45,25. Then you add 127 velocity levels between each “normal” MIDI velocity level.

If you filter out the CC88, the velocity resolution will be adjusted to 128 levels.

There are multiple Reaper scripts already done. Without any programmation, select MIDI CC mapper, and reassign 88 to 89. The event will then be ignored by the Disklavier which don’t know what to do with such a Control Change.

Note: 1/ Reaper can be fully evaluated for 60 days. 2/ I have no shares on Cockos which edit Reaper.


So, you need a keyboard device which resolves to 1024 midi velocities, don't you? How many keyboards resolve keyboard velocities to one of 1024 values? Then, of course, in the Disklavier's case, you need solenoids which can accurately play back 1024 keyboard velocities. I'd love to own a Disklavier Pro but I doubt it will ever happen.

Who else uses extended midi velocities?
Unfortunately, very few keyboard have an high res velocity extension. VAX77, and some Casio. Perhaps the Cybergene DIY keyboard wink of course, the Disklavier Pro...

You can have some highres velocity files (and normal corresponding files) in http://www.piano-e-competition.com (not all years, search the MIDI Yamaha XP logo).

The MIDI XP is a specific Yamaha extension (but Pianoteq can be set to understand it)
I think the Casio high resolution midi scheme was a farce. Someone tested that and concluded the extra velocities were noise "randomly" added. I think the post was at PianoTeq forums last year; he said Casio no longer advertises that "feature".
How far do you drive to work each day?
Me: 4.9 miles
Hi-res fanatic: 4.9132344821376091198271 miles
The practical difference: zero
This is the Wikipedia entry for the Yamaha Disklavier Pro:

"Pro[edit]
In 1999, near the end of the Mark IIXG model series, Yamaha introduced the Disklavier PRO. A key selling feature of this model was the claim of greater recording and playback accuracy than had been possible with previously available models.[9] These instruments recorded not only hammer velocity (as MIDI note-on velocity) but key down velocity and key up velocity (MIDI note-off velocity) as well. The instrument was also capable of recording and reproducing key movements that resulted in no audible sound.

Before the PRO, Disklaviers were limited by design, like all MIDI keyboard instruments, to working within a 0–127 range of values for note-on velocity, note-off velocity, and incremental pedal movement. To break this accuracy limit, Yamaha's Disklavier engineers pioneered a unique use of normally undefined MIDI controllers for the purpose of substantially extending the range of values for note-on/note-off to 0–1023 and for pedal movement to 0–255. In Disklavier lingo, this "extended precision" data was referred to as "XP" data.[10]

The recording and reproduction quality of the PRO have been validated by the International Piano-e-Competition, formerly known as the Minnesota International Piano-e-Competition.[11] In 2002, the Piano-e-Competition used the Disklavier PRO on two continents to enable Yefim Bronfman to participate as a member of the competition jury from Hamamatsu, Japan, 6,000 miles from where the competition was taking place in St. Paul, MN. Following each solo performance, synchronized MIDI and video files were transmitted over the Internet, and Bronfman was able to watch performances on a large screen while the local piano reproduced the playing.

Since that time, the Disklavier PRO has been used by the competition to enable pianists to participate in a screening-round of the competition ("virtual auditions") by submitting a video-synchronized performance recorded on a Disklavier PRO. All rounds of the competition are recorded on the PRO and made available as downloadable files from the competition's website.

The original PRO was the first model Disklavier grand to include the silent system. Ever since the instrument's introduction during the Mark IIXG model era, newer versions of the PRO have been available in subsequent model series and have been known as Mark III PRO, Mark IV PRO, and E3 PRO."

from: https://en.wikipedia.org/wiki/Disklavier#Pro

I'm willing to try out at Pro grand if someone wants to deliver it to my apartment. smile
Congrats CyberGene - great work!

What does your sister think of it? wink

Cheers,
James
x
Originally Posted by Kawai James
Congrats CyberGene - great work!

What does your sister think of it? wink

Cheers,
James
x

I do have a sister. And a brother. But yeah, I know what you’re asking. The lollipop sis likes naked pianos wink
By a “popular” demand, here’s a closeup video of the sensors and their working:


(And a shameless bump of course 🤛🏻)
Only just caught wind of this. That's a LOT of work, and super.

Do you intend to create/sell a kit in future?

I'm quite a fan of DIM (do it myself). smile
Originally Posted by MacMacMac
How far do you drive to work each day?
Me: 4.9 miles
Hi-res fanatic: 4.9132344821376091198271 miles
The practical difference: zero


This. Especially since you dont have 1024 recorded layers, it is just interpolation. And since people already cannot reliably detect the difference between velocity level 67 and 68, for example, having eight more interpolated levels in between is just pointless. Hi-res MIDI is useful for controlling filters on softsynths and such, not velocity on piano.
Adele H, the brand which sells an high grade portable (and foldable) piano (the Piano Phoenix) has involved concertists and they seem to have asked for a 1300 levels piano.

However, I have to admit that I am unable to trigger two successive notes with the same velocity, and even velocity in a tiny -2/+2 range. Then the velocity quantization is not an issue for me.
You're on thin ground there, sleutelbos!

Yes, I agree with you.
But there are "golden ears" who tenaciously cling to the notion that more is better. Always.
There are even some who think that more is better even when admitting that it makes no discernable difference.
I can internally have around 10000 velocity levels. I won't use them. 127 is enough smile Period.
About the video, thanks a lot !!

How do the sensors measure a velocity ? Two sensors per hammer (and an other to switch the note off), or a single sensor which measure multiple values (and we start a timer at an analog value and stop it at an other).

On the top of the sensors, I see 5 weldings. A common (ground), two led emitters and two light sensors ? I am right ? I suppose the timers have to be adjusted to the geometry (it is hard two have two notes with the sensors at the same distance precisely... and even more near the fulcrum of the hammer where the precision matters).
Originally Posted by MacMacMac
But there are "golden ears" who tenaciously cling to the notion that more is better. Always.

Then these are badly trained golden ears. I've worked with "golden ears" in the past and they knew what to listen for... And yes, it had to do with audio. Compressed audio.
Even dog ears can detect compressed audio. Anything recorded in the last half-century is compressed. If it's recorded it's compressed ... save for the all-too-few special edition stuff.

If people say they can't hear the compression ... then I conclude they've never heard uncompressed recordings. And that's quite possible.

It think it might be that more people have seen genuine Apollo moon rocks in the museums than have heard real uncompressed recordings.
Originally Posted by CyberGene
I can internally have around 10000 velocity levels. I won't use them. 127 is enough smile Period.
You never know until you try. I sense this is not just an audio issue.
More than 127 levels internally can be useful. The timer to MIDI level function is not a linear function, then we can have some precision lost depending of the slope.

The higher the MIDI level, the higher the timer resolution will have to be if we want a given relative precision. (If MIDI levels are the logarithm of the time or something related, the relative precision can be the important parameter).
I'm not sure I have the aptitude to do something like this. I definitely don't have the desire. Bravo for seeing it through. Your hard work and dedication to this project is inspiring.

God Bless,
David
I wasn’t clear. I use the highest possible internal precision. But I translate that to regular MIDI and don’t see any reason whatsoever to use high-res MIDI.

Originally Posted by Frédéric L
About the video, thanks a lot !!

How do the sensors measure a velocity ? Two sensors per hammer (and an other to switch the note off), or a single sensor which measure multiple values (and we start a timer at an analog value and stop it at an other).

On the top of the sensors, I see 5 weldings. A common (ground), two led emitters and two light sensors ? I am right ? I suppose the timers have to be adjusted to the geometry (it is hard two have two notes with the sensors at the same distance precisely... and even more near the fulcrum of the hammer where the precision matters).

I use CNY70, it’s a 4-connector device (a LED and a phototransistor). The fifth thing is a SMD resistor.

So, I measure multiple values with the same sensor. I set them through trim pots. This may seem like too difficult but it isn’t actually. One point is hammers at the stop rail. I have a software service routine that would turn on the LED when the distance is reached. So I hold the hammer at rail and turn the trimpot until the LED is turned on. I use spacers for the other distances. After that it’s pretty good and will need just a touch up. I’ve already worked out a routine for setting the two points very precisely. It’s a manual labor and needs aural “tuning” (like comparing how bright the sound becomes of two adjacent keys through trills or fast triple repetitions, etc) but heck, that’s how it works and why it’s so cheap smile It certainly needs some skills and getting used to but is IMO much easier than, say, a real aural tuning of an acoustic piano.
Originally Posted by David B
I'm not sure I have the aptitude to do something like this. I definitely don't have the desire. Bravo for seeing it through. Your hard work and dedication to this project is inspiring.

Indeed, I'm impressed the project was completed at all, especially after acquiring the N1X... that ought to have been the death knell.

Amazing dedication!

Btw, maybe it's just me, but CG's pinky seems to bend the wrong way in that video. At least as a beginner, my teacher always says to keep the fingers curved inwards and fairly stiff, and not be allowed to flex outwards.

Great playing, as usual.
Originally Posted by navindra
Btw, maybe it's just me, but CG's pinky seems to bend the wrong way in that video. At least as a beginner, my teacher always says to keep the fingers curved inwards and fairly stiff, and not be allowed to flex outwards.

Great playing, as usual.

Yes, this is an old bad habit. The risks of self-education 😟 My mother in law (a piano teacher) thinks it’s too late to try to fix it now and that attempting it can even affect my playing in a bad way.
Originally Posted by CyberGene
Yes, this is an old bad habit. The risks of self-education 😟 My mother in law (a piano teacher) thinks it’s too late to try to fix it now and that attempting it can even affect my playing in a bad way.

Your playing sounds great, so it doesn't matter.
Originally Posted by johnstaf
Originally Posted by CyberGene
Yes, this is an old bad habit. The risks of self-education 😟 My mother in law (a piano teacher) thinks it’s too late to try to fix it now and that attempting it can even affect my playing in a bad way.

Your playing sounds great, so it doesn't matter.

Thanks, John! 👍🏻
Good job. Really good job. Whats the price and where do I order? crazy
Originally Posted by spanishbuddha
Good job. Really good job. Whats the price and where do I order? crazy


I'd suspect that being a software developer probably pays better than hand building digital pianos to order laugh
Originally Posted by Chrispy
Originally Posted by spanishbuddha
Good job. Really good job. Whats the price and where do I order? crazy


I'd suspect that being a software developer probably pays better than hand building digital pianos to order laugh

Quite true 😌 Seems like I may just share the designs and the code for free.
Originally Posted by CyberGene

Quite true 😌 Seems like I may just share the designs and the code for free.


or you can find a partnership with a brand, for example kawai mass produces the grand actions, and you sell the kits to us with all the instructions on how assemble them.
Kawai will not want to do this, they have AnyTime technology, if they were interested in selling a controller with real action, they would not need anyone's help. By the way, Petrof used to sell a grand piano action MIDI controller.
So now, is there anything missing that you want to build in? Soft/sostenuto support? BT Midi for built in page turner? Case and music desk?

Are you going to break out a polyester spraying station or epoxy resin studio?
I have the triple pedal unit from my ES7 (with more realistic hard springs) that I will integrate when I also make a cabinet.

For some reason I’m so exhausted after proving to myself I could do the hardest job, I kind of lost interest in finalizing it (cabinet, etc) and so it might be another two years but at least it’s a perfectly playable piano right now, albeit with only 85 keys and a very soft (but continuous) sustain pedal 😁

It has its quirks though. If I switch off the main supply first (and keep the Teensy, i.e. the main controller that sends MIDI, switched on to the computer through the USB cable and since it’s powered through the USB) it would make a sudden screech of death: it will play all the 85 keys at once with a value of 127 to scare the sh*t out of you 😱
Can we agree on the fact that building the ‘Cybrid’ was not an easy/cheap task?
If so, should we re-examine the long-standing belief that Yamaha and Kawai hybrids are ‘overpriced’?
Originally Posted by Pete14
Can we agree on the fact that building the ‘Cybrid’ was not an easy/cheap task?
If so, should we re-examine the long-standing belief that Yamaha and Kawai hybrids are ‘overpriced’?


Can't speak to easy, but didn't CyberGene say the action cost $1000 and all of the other parts to make it work cost $50?

Of course, it's not a polished production product and BOM alone doesn't account for the cost/time he put into it, but for someone who really wants an acoustic action, and can source the action, this might be a real cost-effective option if they're willing to get their hands dirty.
So the belief stands: Yamaha and Kawai hybrids, overpriced! wink
Originally Posted by CyberGene
For some reason I’m so exhausted after proving to myself I could do the hardest job, I kind of lost interest in finalizing it (cabinet, etc)


Looks like the 80/20 rule is kicking in. Take a little break then make that final push.
Have you heard a piece called Naima by John Coltrane? Well, OK. But I'd bet you don't know CoroNaima 😛



Testing the dynamics of the Cybrid on a rainy day after Bulgarian gov declared a state of emergency with people having to stay at their homes. (Not that I used not to stay at home all the time anyway 🤣)
Originally Posted by CyberGene
Have you heard a piece called Naima by John Coltrane? Well, OK. But I'd bet you don't know CoroNaima 😛



Testing the dynamics of the Cybrid on a rainy day after Bulgarian gov declared a state of emergency with people having to stay at their homes. (Not that I used not to stay at home all the time anyway 🤣)


the sound of those bass notes was really THICK, pro quality.
Wonderful playing CyberGene. I also like the artistic video work.

This is a good case both for the Garritan CFX and the Cybrid piano. It sounds inspiring to play.
Originally Posted by pold
Originally Posted by CyberGene
Have you heard a piece called Naima by John Coltrane? Well, OK. But I'd bet you don't know CoroNaima 😛



Testing the dynamics of the Cybrid on a rainy day after Bulgarian gov declared a state of emergency with people having to stay at their homes. (Not that I used not to stay at home all the time anyway 🤣)


the sound of those bass notes was really THICK, pro quality.

Garritan CFX is THICK and SICK! 💪🏻 I spent quite some time calibrating the velocity touch response on a per key basis (through the trimpots) exactly to the CFX (it’s still with the default linear curve) and it’s really even better than what I remember. I think Garritan CFX is an exquisite piano once the touch response is good enough. To people who hasn’t liked it: it’s really worth it to try to improve the touch curves either in the software and/or in the controller.
Congrats, I never thought you'd finish. Nice playing too, I've been meaning to learn that Scriabin Valse.

I actually think your proximity sensor is just fine. The optical gate thingy Yamaha uses is very regulation sensitive as well, I don't really see the advantage.
Originally Posted by trigalg693
Congrats, I never thought you'd finish. Nice playing too, I've been meaning to learn that Scriabin Valse.

I actually think your proximity sensor is just fine. The optical gate thingy Yamaha uses is very regulation sensitive as well, I don't really see the advantage.

Thanks! TBH, I had no big expectations I’d be able to finish it either 🤣 For some reason I had a boost of energy and desire to finish it in the last month or so. I love that Scriabin Valse, can already play it entirely from memory but still struggle with the middle part and can’t go through it without multiple errors 😕 Besides, it has Bb7 and C8 in the score and I miss these keys...

I used my logic analyzer yesterday to do some measurements. I scan the entire keyboard in 36 microseconds. This is 2 microseconds per group which is made of 1 microsecond delay that I have in order for the pull-up resistors to bring the voltage up and 1 microsecond for the program execution. It’s pretty good IMO. To put things in perspective, the hammer travels for 300 microseconds between my two measuring points when having MIDI velocity of 127; 1300 microseconds for MIDI 64.
Well, I know it's the least interesting thing in the world nowadays and even I started losing interest in my own creation 🤣

Anyway, the last time I wrote I measured the scanning speed through a logic analyzer and it was 36 microseconds. I couldn't live with that because it meant I was losing accuracy at high velocities. More precisely, it meant that if I played a note with a MIDI velocity of 110, it will be measured with anything between 110 and 113 which isn't very cool. Things were a bit better around MIDI velocity of 64 where the error was around 1.

I realized I used the smallest possible programmatic delay of 1 microseconds to wait for the pull-up resistor (because without a delay all was messed up and the controller would act weird with ghost notes being played all over the keyboard). But I actually calculated the RC-delay of the resistors and it was well below 0.1us, so waiting for 1us at every group was a huge waste of time! And so, I just replaced this wait with a single redundant digitalReadFast (that takes around 0.1us) and voila! I almost halved the entire keyboard scanning cycle to 19.5us! This means that the error around MIDI velocity of 110 is now slightly less than 1 which is acceptable and the error around MIDI velocity of 64 is 0.3. Much better! Not sure if I can really feel the difference but my imagination says the Cybrid is now even more responsive! smile
That is nearly a 2x boost in scanning performance!

If you play big chords now, is there much delay among the sound output of the notes by Garritan CFX? Is the USB scheme the biggest bottleneck? I thought there was a lot of forum chatter on USB timing from the competitive gamer community like 5 years ago.
I haven't heard any audible chord note differences. BTW, the Teensy MIDI library supports MIDI stacking, which is to group multiple note ON messages in a single USB MIDI package and I use it. Whether it's understood by Garritan CFX I don't know though. Maybe it splits them in single events and executes them one by one, which is the more probable case anyway since it replays samples underneath. But in any case I don't believe I can hear any problems. Not sure how to test it too, maybe artificially with a code that would send a chord in the grouping manner, record the WAV and then analyze it somehow to see if I can notice separate notes being played in succession rather than simultaneously.
Originally Posted by CyberGene
I haven't heard any audible chord note differences.
Playing experience is all that counts. The new Yamaha N1x nearby is a helpful benchmark.
Originally Posted by CyberGene
BTW, the Teensy MIDI library supports MIDI stacking, which is to group multiple note ON messages in a single USB MIDI package and I use it. Whether it's understood by Garritan CFX I don't know though. Maybe it splits them in single events and executes them one by one, which is the more probable case anyway since it replays samples underneath.
I don't see much multi-core activity on Garritan CFX with any of my computers, and second core activity was limited and principally from OS and drivers, with the other cores in deep c-states (sleeping). This is supported by the general advice for pro audio engineers at GS forums (single VIs played live benefit from fast CPU speeds but not number of cores. Big multi-track productions can benefit from more cores).
Originally Posted by CyberGene
Not sure how to test it too, maybe artificially with a code that would send a chord in the grouping manner, record the WAV and then analyze it somehow to see if I can notice separate notes being played in succession rather than simultaneously.
That is a good idea. You could use a big extruded aluminum bar or one of CyberDaughter's toys to press a lot of keys simultaneously.

Alternatively, I suppose you could inject parallel signals near the sensors and see the results. There are a lot of delicate electronics waiting to be destroyed so take care.
CG, your Cybrid is being ‘featured’ at the Pianoteq forum (by a member).
I wonder if they know how much you love Pianoteq! grin
Oh! The humanity!
Originally Posted by Pete14
CG, your Cybrid is being ‘featured’ at the Pianoteq forum (by a member).
I wonder if they know how much you love Pianoteq! grin
The irony laugh Well, regardless of my attitude towards Pianoteq, the controller is a good one and is certainly one of the better ones for people who use Pianoteq (or any VST) smile
Congratulations on an amazing accomplishment. Also I enjoyed hearing the Scriabin poem. I played the opus 32 in a recital in my teens and never played them since and I'm now 65.

Thank you for sharing such an amazing creation.
The results speak for themselves. Wonderful playing CG
Originally Posted by CyberGene

I don't why, but whenever I hear this piece, I keep waiting for it to turn into Radiohead...

^ Ando, hmm, there's certianly something in common indeed smile I love "Everything in its Right Place" in all possible arrangements, and of course the original which gives me instant comfort, same with the Serenade for the Renegade. Wondering what is the common stuff in these modern pop/jazz tunes that are often interchangeable between artists as varied as Radiohead, Brad Mehldau, E.S.T., etc.
Fantastic! Given some of the fundamental skills you had to develop along the way, it is impressive even to have it work at all, much less to have such a good result. Well done.
Cybergene,


Great project! Sounds great too with the Garritan CFX Samples.
What configuration are you using: Full/Compact > Classic/Contemporary/Player > ?

To what extend does your approach differ from that of Lachnit (MK23) who also works with optical sensors?
This is the full version of Garritan CFX. Full -> Classic -> default. I reverse the stereo image to performer, enable partial-pedaling and re-pedaling and turn down pedal noises to be as low as possible without being completely turned off. I also use a “fix” on the re-pedaling timing that’s been described here.

I don’t know how Lachnit solution works but it uses a Fatar action, so the feel should be different.
Thanks for your response.

Is your DIY project now ready to be build by others who would like to get their hands dirty or do you have to finalize and prepare it first?
Well, to be honest, it wasn’t created exactly with the idea to be reproducible although I created it in a way that allows it to be mounted on basically any grand piano action. However it was a rather one-off project and has its quirks because of the design that requires a lot of calibration (and re-calibration) with trimpots. I’ve already come up with better solutions based on the same principle, probably for a second version. But I wouldn’t commercialize it in the way it is now because it’s just to awkward and people will hate it smile Well, as a free and open-source project it might actually be much better because some clever heads can contribute great ideas to improve it, especially in the electronics part which wasn’t a strength of mine to begin with...

I’m really more than happy with the result because it’s most certainly the most complex thing I’ve ever created (including in my job as a software engineer because this project required so many and so different skills that I never had before: learning how grand piano actions work, learning how to restore an action, replace parts, how to regulate it fully; learning electronics, learning how to design PCB-s, how to solder SMD that I did all by hand, more than 5500 joints; learning how to write software for Teensy and how to optimize it for very fast execution; learning how to work with wood and metal, how to build a custom frame from aluminum profiles...). I’m so happy that I really created something with my hands that is at the same time working pretty well, giving me pleasure and also very easy to explain to other people because it’s just a piano for them smile

With all that in mind I really reached a personal goal and that was what I started with. And I’m so exhausted now to even continue with that. Especially in this corona situation. I will think of it after world resumes.
Hey CyberGene,

I feel like this may have come up before, but it's been a while: I was browsing YT last night and came across this video: https://www.youtube.com/watch?v=TIo8yoieX14

[Linked Image]
[Linked Image]


1. that's a grand action piano controller!

2. Not only that, but that looks to be a CONCERT grand action!

3. It doesn't use a hammer shank rail, it uses a stepped hammerhead rail

4. The sensing mechanism isn't visible, but it doesn't appear to have any special sensors visible on the hammers, so I assume it's a fairly pedestrian "PianoDisc" type setup with dual/triple rubber domed key sensors under the keybed?

Just wondering if anyone has seen this, or has any additional info on this setup.
I think somebody posted this or a similar video in the first thread I created. It’s certainly a PnoScan or a similar system. BTW these systems use optical proximity sensors under the keys, they are contactless and I believe very similar in principle to how my solution works, only it’s under the keys and not at the hammers. Very nice action! I bet this action alone might be more expensive than a N1X 😁

BTW, I started publishing and describing my project in GitHub but didn’t finish it only for being lazy. Wondering if there are really people interested in it, so that I know whether I should bother?

I always hate it when I have to create a documentation in my job as a programmer and seems I’m hitting the same problem here 🤨 I can simply publish my PCB designs and the Teensy code but there’s still a lot to explain, many considerations, etc and without that only experienced DIY-ers would manage to recreate it.
This proves a point I made some time ago ...
Originally Posted by CyberGene
... only experienced DIY-ers would manage to recreate it.
If we can't build it, your company, CyberDyne Systems, will have to do so.
The power of Cyber compels you! smile
Please let us all know when your product comes on the market.
I'll be buying serial #2 (assuming that your prototype is serial #1).
Originally Posted by CyberGene
Wondering if there are really people interested in it, so that I know whether I should bother?
I think the project is fantastic and there are a lot of other fans here.

Hollywood music producer Mike Verta has a similar hobby project. He is making his dream studio controller to run orchestral VIs. He has spent a lot of time on the controller and hired some engineers and designers. He said he might sell a few dozen to his friends, which will cover part but not all of the hobby cost. At $5,000+ the market is tiny. I think the controller is almost done pending some IP. FYI the controller is not designed for classical piano work.
Hi CyberGene... Long time lurker here. Following your threads on the DIY project for a long time and finally coming out of the shadows to tell you that I would love a github of the project. There might be 20 guys interested in a joint order of a chinese action. Who knows?
I'm definitely interested and have been following this for awhile now. Been dabbling with sampling my grand, but would just love to use the action to trigger other piano libraries.
Wow and I thought your CFX sustain fix was impressive. Way to go! One question if you don't mind with the hammers not hitting anything can't you feel that bounce in your fingers? I know they're there for the weight but I would think the hammer should hit the sensors no? Thanks Cybergene MooganDavid
Originally Posted by MooganDavid
One question if you don't mind with the hammers not hitting anything can't you feel that bounce in your fingers? I know they're there for the weight but I would think the hammer should hit the sensors no?

You don't feel the hammer hitting anything in an acoustic.
Not sure if anyone else also knows how to read Chinese here, but I noticed a few similar projects in the Chinese video streaming website Bilibili (like YouTube). I'm going to link a few here.

This video demo just looks like what CG does here: https://www.bilibili.com/video/BV1bs411q7L4

This one is also quite interesting: https://www.bilibili.com/video/BV1pJ411X7af. Looks like he is using a Fatar action and connecting that to some chips (for sound module maybe?) and he builded his own cabinet with speakers and amplifiers!
I think in the linked video a system with keyboard sensors is being used, probably PnoScan because I don’t see hammer sensors anywhere. PnoScan is a commercially available system that can be purchased and requires a piano technician to install it. However the hammer stop rail is maybe made by the guy himself and looks like mine, and like the one in the AvantGrand and Novus.

In my solution I have hammer sensors and I believe that’s the most reliable way to determine the actual velocity of the hammers and whether there’s a strike.
One suggestion I have (improvement) is using voltage references for the voltage sensitive parts of the circuit. From the top of my head, CG had calibration issues due to voltage variation of the wallwarts he was using. I would add 4.0V LDOs (TI has som they are not expensive).
Originally Posted by EVC2017
One suggestion I have (improvement) is using voltage references for the voltage sensitive parts of the circuit. From the top of my head, CG had calibration issues due to voltage variation of the wallwarts he was using. I would add 4.0V LDOs (TI has som they are not expensive).
Great, I had you in mind when thinking of power supply improvements. This is what’s good about open source smile I’m absolutely convinced most of my decisions are suboptimal and it was only due to stubbornness that allowed me to finish it in a working way. Hopefully this project can be improved.
Hi, CG, if you choose making the project public the way it is, I suggest you add on the README.md a "Known issues and improvements" section as a heads up for anyone that wishes to replicate the design.

Quote
Great, I had you in mind when thinking of power supply improvements. This is what’s good about open source smile I’m absolutely convinced most of my decisions are suboptimal and it was only due to stubbornness that allowed me to finish it in a working way. Hopefully this project can be improved.

About the "suboptimal" I disagree (I cannot speak about the stubborness part though as I do not know you in person - ROFL). It is a huge achievement (more so if you do not have an electronics background AFAIR). You have all my respect for that. And the outcame was excellent based on the recordings you posted. As with any DIY project we always find better ways to make things after we have learned their shortcomings. I hazard to say the same happens with professional projects as well, no matter how thoroughly specified they are.
Hi CyberGene,

Your project and accomplishment is simply fantastic! And I say that as someone who is somewhat similar to you: I studied Physics at the university, currently work as software engineer and (obviously on this forum) play the piano. Ah, and I have young kids and I did/do many DIY projects with wood, metal, electricity and electronics, but only when "forced" by the wanted result, I don't like any of them "per se" like some people do. Heck, I even started studying the same book you mentioned for acoustic piano action regulation! Are we doppelgangers? No, wait, we're not: I did not build any DIY hybrid controller, even though I thought on and off about it since I thought of buying a digital piano....

Huge congratulations on such an accomplishment. Regarding your question on publishing the project, I think some people would like that. It really depends on how much work it is, and how hard it is from both a labor and electronic part procurement point of view. I may do it.

In the USA, it's easy to get (old and out of tune) grand pianos for free. They usually have loose tuning pins and perhaps cracked soundboard. But who cares? Often enough they have workable actions, so that puts a lot of price relief to people with spare time. Putting a junk grand piano to recycle/dump costs about $200-300 (transportation included), so for the not-so-rare cases in which the action is good, that's an excellent price to start!

If you want to think of commercializing your instrument (and I think you MUST :-) ), I believe leveraging WNG is a great option. They make fantastic action parts and they have a network of experienced technicians. These people specialize in two things:

1) servicing pianos sold by Mason & Hamlin (the only manufacturer who uses WNG action at the factory, I think they make a few hundreds per year), and
2) custom building actions from any piano you want.

Are you getting where I am going already? A couple of years ago I asked for an estimate of such a rebuilt action and I got a $6500-$8000 range depending on what I wanted exactly (parts and labor included). I think a full action with no constraints of an existing instrument could be lower than the lower end of that range. Much lower if you ask for 10 of them, not to mention 100. The fact that M&H already makes full instruments could help a lot for cabinet, distribution network, etc.

Yes, as someone else wrote in this thread, software engineering pays more than hand-building artisan hybrid pianos on request. But perhaps designing hybrid pianos for a company could pay at least the same (if not more) as being a software engineer -- and be more fun, so why don't you think about it seriously? I'll be happy to be your humble apprentice, if you need one :-)

So congratulations again, think about it... and now let me go back to practicing on my NU1...

Cheers,
D.
Originally Posted by Del Vento
Regarding your question on publishing the project, I think some people would like that.

Some people?? I would say lots of people. Maybe a few people initially, then it will start to get more attention with time, suggestions, improvements, building diaries, vlogs about it etc.
Originally Posted by Del Vento
Hi CyberGene,

Your project and accomplishment is simply fantastic! And I say that as someone who is somewhat similar to you...
Thanks, so kind words from someone who’s indeed like a doppelgänger to me 🤝 Appreciate it!

OK, guys, point taken, I will keep on describing it and publishing it. A particular pain in the arse is the .MD format on GitHub which is very limited. For instance you can’t inline YouTube videos. Heck, even this forum allows it 🤣 Maybe I should create a wiki or a blog somewhere. (But mostly you should know I’m one of the laziest people on earth and trying to find excuses to not write a doc...).
Just adding another vote of interest. I honestly wouldn't do anything with it unless and until I get ahold of a grand action myself (and that would require a lot of home commitments ;)) but I would certainly love to download and review your project (and save it offline for the coming apocalypse).
Nice work on that keyboard CG! All that hard work paid off. Good luck with your project.
Originally Posted by Harpuia
Not sure if anyone else also knows how to read Chinese here, but I noticed a few similar projects in the Chinese video streaming website Bilibili (like YouTube). I'm going to link a few here.

This video demo just looks like what CG does here: https://www.bilibili.com/video/BV1bs411q7L4

If you read the comments (I translated them), he's starting a trial production with a piano company as of April 2020. So this actually may become a product.

However, if CG publishes his project, I'd be more interested in using the open source data to make my own....
CG, why not using WordPress? I never heard about github.
Originally Posted by pold
CG, why not using WordPress? I never heard about github.

GutHub is basically a frontend to a Git-based project system and git is version control system, it's used for open-source projects to allow for multiple people to commit source code changes, etc. I decided to use it because it contains the source code I wrote for the Teensy controller that is the main board for my project, it's the one that has a MIDI library and a USB and it's also the one that scans sequentially through the PCB-s that I designed and those PCB-s deal with translating the analog proximity information from the sensors into digital binary signals.

Since people may improve both the software code and the PCB designs, the best way is to have version-control system, so that any change/improvement/fix by any person will be stored and traced into the git log. This is the way open-source projects work. In a static blog such as Wordpress I can put a lot of information but it will be a static one, what I put there and other people won't be able to collaboratively contribute to the project.

Besides GitHub can automatically create GPLv3 open-source licenses for the project.
Originally Posted by CyberGene
Originally Posted by pold
CG, why not using WordPress? I never heard about github.

GutHub is basically a frontend to a Git-based project system and git is version control system, it's used for open-source projects to allow for multiple people to commit source code changes, etc. I decided to use it because it contains the source code I wrote for the Teensy controller that is the main board for my project, it's the one that has a MIDI library and a USB and it's also the one that scans sequentially through the PCB-s that I designed and those PCB-s deal with translating the analog proximity information from the sensors into digital binary signals.

Since people may improve both the software code and the PCB designs, the best way is to have version-control system, so that any change/improvement/fix by any person will be stored and traced into the git log. This is the way open-source projects work. In a static blog such as Wordpress I can put a lot of information but it will be a static one, what I put there and other people won't be able to collaboratively contribute to the project.

Besides GitHub can automatically create GPLv3 open-source licenses for the project.

nice to know that, thanx
I think what you've achieved is sort of the "Holy Grail" of "digital " pianos.

Bravo!

Perhaps you should patent it and sell it to Kawai?

I'm a saxophonist first. They are "Swiss Cheese Cones," so their "Holy Grail" is to shut them up while practicing. Yamaha makes it's "Silent Brass" mutes for trumpet and other closed system brass instruments.

I'm toying with several methods to achieve this. Any snappy ideas?

- Jeff
Originally Posted by Skyscrapersax
Perhaps you should patent it and sell it to Kawai?
They have a patented piano like that, the NV10 wink And I own the Yamaha counterpart, the N1X, so the big sharks cannot be challenged by me. But at least it’s a good hobby for handy people smile
Originally Posted by CyberGene
Originally Posted by Skyscrapersax
Perhaps you should patent it and sell it to Kawai?
They have a patented piano like that, the NV10 wink And I own the Yamaha counterpart, the N1X, so the big sharks cannot be challenged by me. But at least it’s a good hobby for handy people smile

Yamaha and Kawai definitely not.

M&H or Steingraeber perhaps.

PianoDisc or QRS (Story & Clark), especially the latter, would be very interested too.

I think this project has potential beyond just hobby for handy people (without dismissing that, as many have said, also in response to my earlier comment).
Well, there are many ways to skin a cat, I'm sure anyone looking to get into the field has a number of different techniques to pursue.

Also, prosecuting a patent is one thing, enforcing it and building out a business based on it is an entirely different level of commitment/effort.

For a hobbyist, it makes a ton of sense to publish it and open source the idea and actually advance and enrich the field (and establish prior art in the process).
Publishing a patent means formulating a set of assertions that none other have applied. I suspect that the DIY here share too much similarities with of AventGrand and Novus that it could be difficult to write such sentences.

A patent is typically written like 1/ An apparatus which has (such little characteristic), 2/ an apparatus like describe by 1/ which has (An other little characteristic)... etc.

The more specific sentences (like 2/) may not protect you enough if the other product is too different (doesn’t share all characteristics). On the other way, the less specific sentences will have more chance to have been already made by a previous system and can’t be protected. This explains the gradation of sentences from the less specific one (1/ a device with a grand piano action which act as a MIDI controller), to more specific ones (2/ a device describe by 2 where the key movement us capture by an optical sensor, 3/ a device describe by 2 where ...)

A patent is interesting if we can describe by word something 1/ which hasn’t been done, 2/ where we have some incentive (a monopoly) too avoid this to be done by others.
Patent is out of the question.

I don’t want to underestimate my own creation and I believe it’s the most complex and interesting thing I’ve ever created (and I doubt I will ever do something as fun and good again) but frankly I haven’t done anything so special or finished. It’s just a niche DIY project which makes it a bit unique but there’s nothing special or innovative in it. I’ve seen hobby projects that are depressingly more complicated in both mechanical and electrical aspect smile With that in mind an open source project is the best thing.
Interesting... Any ideas on the sax mute?

What does your creation weigh?

I didn't know that Kawai / Yamaha had similar creations. I own the MP11.
@CG: glad for your honest and realistic view about your excellent project. Patents and so forth are unrealistic for this case. I'm a biomedical researcher in academia (sensory neurophysiology from intracranial arrays of electrodes), and although sometimes me and my team develop pieces of software and hardware that surpass many commercial solutions, making into a business requires a level of commitmment that is simply not practical. In my field, the current trend is open-source...
A sax mute is a totally different topic.

And it's off-topic for this forum.

Have a look at "electronic wind instruments" with saxophone fingering.

Maybe https://www.akaipro.com/ewi-usb
One more vote for the GitHub repo, mostly because I'm interested in seeing your code smile
Originally Posted by CyberGene
They have a patented piano like that, the NV10 wink And I own the Yamaha counterpart, the N1X, so the big sharks cannot be challenged by me. But at least it’s a good hobby for handy people smile

I am convinced they can be challenged by your project, CyberGene, at least technically. Open source can be very powerful, and both in software and hardware there are great projects that surpassed in many ways commercial alternatives. Git/Github is a good choice. I am looking forward to your repository. I could also help with some aspects if needed.
I would also like to add a vote for seeing this on github! I've been checking in every now and again to see if any code is being released. I have some projects in the works and seeing more details on how you've implemented things would be really helpful.
👍🏻
And I have a question. I guess in part I can answer that from the pictures, but I'd like the full one. Would this project be non-invasive enough to be performed on an actual grand without jeopardizing the needs of the acoustic? I'll bet is much better than the QRS or PianoDisc MIDI strip which they put under the keys..... and I'm thinking of doing it that way, rather than building a separate instrument....
The only “invasion” to the action are the two long threaded screws (not sure about the English name) that hold the hammer stop rail. They required drilling holes in the wooden rail that holds the whippens:
[Linked Image]

Just two holes on each side, not a big deal but you can avoid even that if you create a cabinet to slide the action inside and the cabinet itself has stop rails for either the hammers or the shanks.

The rectangular aluminum frame that holds the sensors and the PCB-s is held by the existing screws of the hammer base rail, I only removed the washers, so the frame acts as washers:

[Linked Image]
Thanks Cybergene for the quick response and the high resolution pictures! Just FYI, I think the name of the "long screw" is "threaded rod" or less likely "long bolt" (I am not native speaker either, even though I live in the USA).

Everything sounds great, but I have two more questions:

1) if the action is inside the grand and it is used to make sound by hitting the strings, the stop rail is unnecessary, right? In one use case, I just would like to capture the MIDI of an acoustic performance....

2) it looks like there is a lot of electronics in the way of the action, which may be a problem for a technician doing regulation. Since you yourself now are a technician doing regulations :-) can you comment on that? If the amount of electronics is indeed a problem, could be the L-shaped aluminum rail holding all the cards in front of the action be replaced with hinges? Those hinges could be normally kept in position by some short flat aluminum rails at the ends (making the L into triangles), but then they could be "unlocked" and moved when needed? Or is removing the front of the frame with all the cards, when needed, not a big deal anyway?

Thanks,
Davide
You don’t need the stop rail if you will be using the keyboard inside a working piano with strings.

I don’t have the original piano, only got the action but I believe it would slide inside the piano with no modifications, depending on how high the front panel is for the action to slide beneath.

The entire frame with everything I created can be detached by unscrewing 5 screws total. Two on each side + 1 in the middle. Those are screws that hold the front wooden rail that holds the hammer bases. I haven’t used all the screws though (there are two more on that rail to still hold it when I remove my frame; I don’t want for the wooden rail to become loose). Here’s a picture of the frame with everything:
[Linked Image]
Wow, this is really good. I think it will give both QRS and PianoDisc a run for their money. Actually I think you should contact either of them and see if they are interested in collaborating to make the project available commercially for the ones who do not want to build it themselves. Well, provided YOU are interested in collaborating with them, obviously.

Like everybody on this thread, I am looking forward to additional tech documentation. And since we both are in software so we know: don't wait for the documentation to be good, because we will grow too old meanwhile ;-)

Just "dump" what you have and let people scramble with it.

Thanks again and have a great weekend,
D.
Originally Posted by Del Vento
I think it will give both QRS and PianoDisc a run for their money. Actually I think you should contact either of them and see if they are interested in collaborating to make the project available commercially for the ones who do not want to build it themselves.

While I agree that this is an awesome project, I do not think this is something QRS or PianoDisc would be interested in or be intimidated by. The main reason is the visibility of the electronics. While it doesn't bother me, I believe that most grand piano owners would not want to have the electronics visible. One of the selling points of the QRS and PianoDisc systems is that the electronics are hidden under the keys.

By the way, this is a project that I am also interested in working on (given the time and the ability to purchase a grand piano action), so I also look forward to additional documentation about the project!
The only advantage my solution has over these is the ability to easily remove it. And measurement at the hammers which is more precise. However their solutions are very advanced in terms of electronics and I believe they are self-calibrating compared to the tedious trimpot turning in my solution. Besides, mounting under they keybed can be standardized because the key widths are standardized, they have ready made PCB strips with sensors at proper distances. In my case the frame is customized to fit a specific action with specific hammer placement, gaps and sections.

So, no, no competition here smile They can sleep well.
Hi Jon,
Not sure what you mean by "the visibility of the electronics", since for the QRS and PianoDisc use case the action is inside the grand cabinet. And if it really bothers people (technicians?) when removing the fallboard, they can be easily made "hidden" by plastic molding box(es).

Regarding your project, as I said earlier there is no shortage of free (or very cheap) grand pianos in the USA. Just take one off craigslist (have to select one with good action and crappy rest), remove the action and pay the removal of the cabinet, straight from the owner's house so you don't have to pay for the move to your place. It should cost you around $200 (may be more or less depending on the local market).

For the time, I'm working on building a machine to stop it, so we all (me first) can have all that we need, ROTFL.

Cheers,
D.
Originally Posted by CyberGene
The only advantage my solution has over these is the ability to easily remove it. And measurement at the hammers which is more precise.

Do not undervalue either. I've heard lots of complains from owner and technicians about both issues. Not sure how yours compare, but on paper (screen) it seems a big leap forward. Especially after hearing how much better the Kawai NV5 is compared to the Yamaha NU1 (hammers vs key sensor).

Originally Posted by CyberGene
However their solutions are very advanced in terms of electronics and I believe they are self-calibrating compared to the tedious trimpot turning in my solution.

They are not self calibrating but require a tedious labor too. And your electronic could be made more advanced if there is enough volume of customers.

Originally Posted by CyberGene
Besides, mounting under they keybed can be standardized because the key widths are standardized, they have ready made PCB strips with sensors at proper distances. In my case the frame is customized to fit a specific action with specific hammer placement, gaps and sections.

That is the only real drawback of your solution, but in my opinion it is a small price to pay and it could be engineerized for the most common string distances and the most common section sizes.

Originally Posted by CyberGene
So, no, no competition here smile They can sleep well.

I was suggesting you to collaborate with them, not to compete against them :-)

Building a company from scratch is too much work for me to suggest you to do it. You need to have strong entrepreneurship qualities, time, money, etc. It needs to come from inside you, not from someone's suggesting.

On the other hand, talking to them and see if they are interested is easy, not time consuming, and I'd certainly do it if I were you -- so I can suggest that. But obviously is your life, your time and your project, so it's your decision :-)
If PNO or QRS were interested in making this, then I don't think there's anything to stop them.
Unless CyberGene possesses intellectual property rights and is willing to defend them, anyone could use this idea at will.

I think the adjustment thing is an issue. The pots (264 of them!) are a manufacturing burden.
Could they be eliminated by doing the adjustments in software?

I'm not sure whether the visibility of the "guts" is an issue. If the lid is open we can already see piano guts, right?
Anyway, a clever designer could "dress" the Cyber guts to give them a "traditional" look. (I don't possess that kind of clever, but others do.)

I think the main thing is what CG said: the non-standard spacing. Key widths are nearly standard from the player's perspective.
But once you go behind the fallboard, the keysticks splay out left and right.
How much? It varies for each group of keys, and it depends on the piano's scale design.

None of that is standard. Is that the point you're making, CG?
Originally Posted by Del Vento
Hi Jon,
as I said earlier there is no shortage of free (or very cheap) grand pianos in the USA. Just take one off craigslist (have to select one with good action and crappy rest), remove the action and pay the removal of the cabinet, straight from the owner's house so you don't have to pay for the move to your place. It should cost you around $200 (may be more or less depending on the local market).

or you can buy a cheaper upright piano action and turn it into a grand piano action. Or if you find a cheap baby grand you can disassemble everything and pack it inside the car grin
All the commercial solutions use very mature implementations of detecting key velocity through proximity sensors. But under the keys. They only need to use them on the hammers to turn it into my solution, so I can’t give them anything they couldn’t do themselves. I haven’t invented anything special.

And I don’t think I can do a software calibration. The main principle I use is that of turning analog proximity signals into binary digital signals exactly the same as a triple sensor digital piano: I have three consecutive switch lines that are enabled one by one. The controller only scans through them in a loop and I think this is exactly how digital piano logic works. I’m sure if I somehow have a Fatar keyboard or any other triple or double sensor keyboard with sensor switches wired, I can multiplex and feed the signals straight to the main board with the Teensy controller, eliminating the trimpot PCB-s and generate MIDI with my existing code without any change besides coefficients of the velocity calculation.
Originally Posted by CyberGene
I’m sure if I somehow have a Fatar keyboard or any other triple or double sensor keyboard with sensor switches wired, I can multiplex and feed the signals straight to the main board with the Teensy controller, eliminating the trimpot PCB-s and generate MIDI with my existing code without any change besides coefficients of the velocity calculation.

Now, I've tinkered with "maker" bits like this before and been around electronics, IT, and coding for years and years. I understand the words you have used here.

But when you string them together like that, it's either as beautiful as Debussey and as frightening as the Opus Clavicembalisticum! Or BOTH! smile
Guys, instead of trying to convince CG to go the commercial route with his Cybrid, maybe we should focus on showing appreciation for what he accomplished. He was not a piano maker, or an specialist on electronics, and over the course of a few years he managed to craft a functional hybrid piano.

Odds are the commercial version are far more mature developments than his, but in any case his accomplishment is impressive.

CG, as someone who has loves to go into the unexplored with projects, I can only imagine the dedication and effort this took. Congratulations!!!
This is the part that amazes me ...
Originally Posted by greypiano
He was not a piano maker, or an specialist on electronics, and over the course of a few years he managed to craft a functional hybrid piano.
CG said that himself. And that's astonishing. Not only does his Cyber Cembalo work, but the assembly looks as good as a professional prototype. No sky-wired crap. Clean as a whistle.

So how can it be that he claims to be no specialist in electronics? Either he're really an expert in disguise, or he's a prodigy.
Originally Posted by MacMacMac
Either he're really an expert in disguise, or he's a prodigy.


Originally Posted by kimby
Or BOTH! smile
Originally Posted by CyberGene
BTW, I started publishing and describing my project in GitHub but didn’t finish it only for being lazy. Wondering if there are really people interested in it, so that I know whether I should bother?

Since you mentioned Github, I've been waiting for the repo link.

I think it's fine to public unfinished work on Github, so people can start contributing to the project in various ways.

Originally Posted by CyberGene
A particular pain in the arse is the .MD format on GitHub which is very limited. For instance you can’t inline YouTube videos.

As for document or website, you can use a static site generator, and deploy to https://pages.github.com/

Or use https://www.netlifycms.org/ (opensource), and deploy to https://www.netlify.com/ (free).

I can help a bit with those.
A new video from the Cybrid:

Very nice! How will we all do soon, when Cybrid will be moved far from you?

Garritan sounds really nice. It doesn't have a demo version for testing how it would sound on my setup, does it?

Cheers,
D.
Sounds awesome. And the performance has improved! It's practically Sofronitzky now. Though, for effect you should play it using The Maverick VST for proper antique sound...
Lovely expert playing, as usual, but could you explain the attraction of the piece? At the beginning it sounds meandering, almost to the point of being atonal without actually being atonal. I think it might make sense in the context of a movie or comedy (Charlie Chaplin), or maybe a ballet/dance/theater performance, or yes, a poem with words, but by itself it feels hard to situate.
Originally Posted by navindra
could you explain the attraction of the piece? At the beginning it sounds meandering, almost to the point of being atonal without actually being atonal. I think it might make sense in the context of a movie or comedy (Charlie Chaplin), or maybe a ballet/dance/theater performance, or yes, a poem with words, but by itself it feels hard to situate.

Put it in its proper historical context (both globally, and among Scriabin's own background) and it makes perfect sense. The last gasp of romanticism and self-indulgence, with a bit of sweetness and intimacy. The opening has two melodies at the same time at the start of the phrases, and then the ascending scale passage to finish each short phrase. It provides a peculiar buildup on the first and third pages that are resolved on the 2nd and 4th, ultimately dissolving into an embrace of F#, the composer's favorite key. It's even more fun when you pair Op.32 No.1 and Op.32 No.2 together, as the second one is mostly bombastic.
Originally Posted by terminaldegree
Put it in its proper historical context (both globally, and among Scriabin's own background) and it makes perfect sense. The last gasp of romanticism and self-indulgence, with a bit of sweetness and intimacy. The opening has two melodies at the same time at the start of the phrases, and then the ascending scale passage to finish each short phrase. It provides a peculiar buildup on the first and third pages that are resolved on the 2nd and 4th, ultimately dissolving into an embrace of F#, the composer's favorite key. It's even more fun when you pair Op.32 No.1 and Op.32 No.2 together, as the second one is mostly bombastic.

That's a very informative way of explaining the appeal of this piece, thanks!
No problem, navindra. It's primarily what I do for a living, when I'm not posting on the forum, adjudicating festivals and competitions, or writing product reviews.

Scriabin is an acquired taste beyond the early opuses, that really didn't register on my radar until I was in my second year studying at a conservatory. It's very late romantic/early 20th century piano music... for piano nerds. When I was in my 20s I was really taken with his music. I taught a Russian Piano Literature class last fall semester, and the reaction from my undergraduate students to Scriabin was completely mixed-- not all of them liked it, but they generally knew it when they heard it on the listening exams. Your general unease with the opening motive can often be a useful tool in figuring out whether you're going to like a Scriabin work, because he tends to stick with those opening motives throughout so many of his piano compositions from start to finish.

The earliest of his music sounds more like Chopin, while the later music becomes increasingly avant-garde and consumed with crazy imagery, philosophy, mysticism. It's with that stuff that you have to really get "outside yourself" and assume a larger-than-life interpretive persona, to get the interpretations to work in support of the inherent oddity of the music.
Originally Posted by terminaldegree
No problem, navindra. It's primarily what I do for a living, when I'm not posting on the forum, adjudicating festivals and competitions, or writing product reviews.

Scriabin is an acquired taste beyond the early opuses, that really didn't register on my radar until I was in my second year studying at a conservatory. It's very late romantic/early 20th century piano music... for piano nerds. When I was in my 20s I was really taken with his music. I taught a Russian Piano Literature class last fall semester, and the reaction from my undergraduate students to Scriabin was completely mixed-- not all of them liked it, but they generally knew it when they heard it on the listening exams. Your general unease with the opening motive can often be a useful tool in figuring out whether you're going to like a Scriabin work, because he tends to stick with those opening motives throughout so many of his piano compositions from start to finish.

The earliest of his music sounds more like Chopin, while the later music becomes increasingly avant-garde and consumed with crazy imagery, philosophy, mysticism. It's with that stuff that you have to really get "outside yourself" and assume a larger-than-life interpretive persona, to get the interpretations to work in support of the inherent oddity of the music.
This just makes this piece so much more interesting to listen to now... the piece gets better and better with each listen. Everything you've explained here somehow just makes the music better for me, by contextualizing it, and pointing out the reference points. Music is amazing in this way. 👍
Originally Posted by terminaldegree
Sounds awesome. And the performance has improved! It's practically Sofronitzky now.
Wow, thanks so much! You saying that means a lot to me 🤝 I listen to Sofronitsky a lot since he's to me the authority on Scriabin. For this poème in particular I've also listened to a recording by Heinrich Neuhaus who renders it lovely.

Navindra, terminaldegree already did a fantastic explanation about Scriabin and it pretty much sums up how I approached Scriabin myself too. Up until very recently I've found his music rather abstract, a bit atonal, hard to grasp, maybe boring too. The only work that I liked and even recorded on my CA63 some years ago was the Etude Op.2, No.1, a work he composed when 15 years old, so deep and touching, kind of similar to Chopin. But it was only maybe a year or less ago that I started opening my ears more towards his other works and it is definitely an acquired taste. I even remember when I heard this Poème (that I can't get enough of now, I absolutely adore it!) some months ago for the very first time it sounded a bit abstract, and I felt exactly as you: it's almost atonal but then tonal at the same time... I remember I liked the second and fourth pages of it, the arpeggio passages that are impressionistic in feel and more approachable. But then when I downloaded the score and started playing it (and listening to more performances) it suddenly made a lot of sense and to me it's 100% tonal and is so logical! Very weird, this thing happens to me with a lot of Scriabin works from his early and middle period. They start a bit odd, abstract, atonal and with every new listening I love them more and more. Scriabin is now among my most favorite composers, along with Chopin. I can describe him as a more modernized Chopin. And his music is indeed a bit nerdy in that it requires some exploration, I mean to play from the score yourself before you appreciate it.
Originally Posted by CyberGene
Originally Posted by terminaldegree
Sounds awesome. And the performance has improved! It's practically Sofronitzky now.
Wow, thanks so much! You saying that means a lot to me 🤝 I listen to Sofronitsky a lot since he's to me the authority on Scriabin. For this poème in particular I've also listened to a recording by Heinrich Neuhaus who renders it lovely.

Navindra, terminaldegree already did a fantastic explanation about Scriabin and it pretty much sums up how I approached Scriabin myself too. Up until very recently I've found his music rather abstract, a bit atonal, hard to grasp, maybe boring too. The only work that I liked and even recorded on my CA63 some years ago was the Etude Op.2, No.1, a work he composed when 15 years old, so deep and touching, kind of similar to Chopin. But it was only maybe a year or less ago that I started opening my ears more towards his other works and it is definitely an acquired taste. I even remember when I heard this Poème (that I can't get enough of now, I absolutely adore it!) some months ago for the very first time it sounded a bit abstract, and I felt exactly as you: it's almost atonal but then tonal at the same time... I remember I liked the second and fourth pages of it, the arpeggio passages that are impressionistic in feel and more approachable. But then when I downloaded the score and started playing it (and listening to more performances) it suddenly made a lot of sense and to me it's 100% tonal and is so logical! Very weird, this thing happens to me with a lot of Scriabin works from his early and middle period. They start a bit odd, abstract, atonal and with every new listening I love them more and more. Scriabin is now among my most favorite composers, along with Chopin. I can describe him as a more modernized Chopin. And his music is indeed a bit nerdy in that it requires some exploration, I mean to play from the score yourself before you appreciate it.
I think after you've conditioned yourself to accept the modernist aspects of such composers, you can find that a lot of traditional harmony becomes a little boring too. A bit like eating your food with no spices. (or at least you set the bar much higher in terms of thematic content, voice-leading, chromaticism, rhythmic devices, key-changes, modal mixture, etc) I never got completely into atonal music like Schönberg, Webern, Berg. But I did learn a lot from studying it. It made composers like Scriabin very comfortable to listen to. So I think previous exposure has a lot to do with how we hear things.
This is great playing by any standards. Any pianist would be proud of it.
Originally Posted by johnstaf
This is great playing by any standards. Any pianist would be proud of it.
Wow, thanks, John, I respect your opinion so much 🍻
Originally Posted by CyberGene
Some people may have followed my odyssey in the DIY real grand action controller? thread. It started two years ago ... Whoever is interested, can follow details in that other thread. No more words, here's the result:


I've never seen such fluent trill in any commercial dp keybed! (at 6:15) So effortlessly captured and translated! Amazing! Have you made an instructable for this already? So beautiful. cool smile
Josh, thank you. I guess you followed the link to this thread from my signature. I should have linked to a newer thread where I open-sourced the project. Here’s the GitHub:

https://github.com/ekumanov/cybrid

But it’s a one-off project that I believe would be too difficult to recreate by other people. However that project influenced other people to make their own projects and I’ve linked to one of them in the description. I think you should check it out, it’s much more advanced than mine: https://github.com/jkominek/piano-conversion
Yes I did, I have been looking for projects like this ever since... gauging which one's are feasible based on local supplies.
Yeah, it would be difficult , but I don't mind soldering and wiring. However but fabricating custom PCBs would be a challenge though.

Can you confirm though that it can do note repetition and legato like this one here or even better:

Triple Sensor

(scroll down a bit and select triple sensor, for the demo sounds)


Thanks a bunch... smile
Hello,

@CyberGene: I absolutely adore your project. I have been too new to Piano World to be aware of it, but might still read this thread and be enticed to a level where I'd have to strongly manage myself to steer clear of such a project. I'm still trying to learn that 'I possibly could do this (fill in any project)' doesn't necessarily mean 'I should do it'.

Originally Posted by josh_sounds
Can you confirm though that it can do note repetition and legato like this one here or even better

If a VPC-1 like behavior is what you're after, then the wiser, *much* cheaper and *much* quicker thing to do is... drumroll... buy a VPC-1.

Cheers and happy dreams (nightmares included) of sweet projects,

HZ
Josh, repetition and trills depend entirely on the action. I regulated mine very carefully and thoroughly which is why the trills are easy. The sensors simply detect hammer velocity.
There are a few interesting variants of a DIY digital grand piano I see posted occasionally.

CyberGene got it all started!

And I am continuing to chip away at my own.

Its just a hobby so work on it as time permits. Documenting progress on YouTube so once its done (maybe? someday?) I can remember all the pain and suffering smile

Most recent update:

Wow, you guys are really on to something!

I hope we could see a collaboration that leads to a product ‘we’ could buy at some point.

Perhaps something where the customer provides an acoustic action for these ‘systems and sensors’ to be installed and then shipped back (the updated action) to the customer.
stemPianist, great progress! I remember the time I spent working on this project. Although most of the time I was just doing stuff in my head, it used to keep me busy. What a sweet time! I managed to finish it just before COVID and that was a lot of luck on my side since I depended on deliveries and stuff.

I miss those times 😒 For quite some time I’ve been mainly just learning new pieces on the N1X. And I replaced the DIY stuff with my other hobby, the synths but it’s not the same.

So, enjoy it while you can 👍🏻
Thanks CyberGene! And, the videos are a bit behind, so the project is moving along ok.

Pete14 - I would buy one, then don't have to make my own smile For collaboration, I keep thinking to propose a DIY Forum on piano world. Probably lots of creative ideas out there. And could be motivating.
Originally Posted by stemPianist
Thanks CyberGene! And, the videos are a bit behind, so the project is moving along ok.

Pete14 - I would buy one, then don't have to make my own smile For collaboration, I keep thinking to propose a DIY Forum on piano world. Probably lots of creative ideas out there. And could be motivating.

Hey Greg,

Nice to see other attempts at this. Do you have a github repo, besides the YT videos?
Not everybody wants to make things from scratch, but I think very many would buy a pre-designed kit which needs only some assembling and/or some coding, hence sharing the PCB and BOM could help them.
In any case, how is your design different from Jay's piano conversion? From the video I could not tell.

Cheers
BTW, how's Jay's piano conversion going on? I think he doesn't post on the PW or GitHub often, so I'm not sure where he is now, the last I've seen is he already wired one key. I guess by now he should already be finished?
Hello,

Originally Posted by stemPianist
I keep thinking to propose a DIY Forum on piano world. Probably lots of creative ideas out there. And could be motivating.

I'd be interested and would read such a forum, and possibly post some posts on (modest!) mods I'm doing to get my piano setup just those little tads more to my liking.

Cheers and happy projects,

HZ
Originally Posted by HZPiano
Hello,

Originally Posted by stemPianist
I keep thinking to propose a DIY Forum on piano world. Probably lots of creative ideas out there. And could be motivating.

I'd be interested and would read such a forum, and possibly post some posts on (modest!) mods I'm doing to get my piano setup just those little tads more to my liking.

Cheers and happy projects,

HZ

Care to give us a preview of the 'modest' mods you're working on, HZ? :-)
Hello,

Originally Posted by mmathew
Care to give us a preview of the 'modest' mods you're working on, HZ? :-)

😀

Well, maybe, after gaining a lot of additional confidence in my messing and fiddling... I might... Don't get any hopes up too much though...

For instance, there's this multi speaker thing we talked about before, which has now gone from four to six speakers. But I'm experiencing technical hiccups and feel more safe ironing those out before posting any posts on it. If the hiccups prevail instead of me, the whole attempt may be binned. Time will tell. At some moments, like yesterday, I'm tempted to just throw everything out the window and/or pick up and smash the whole shebang across the room.

After which I need a day or two to stop the smoke coming from my ears etcetera, you get the picture.

🤬 > 🙄 > 😬 > 🤔 > 😔

Oh well...

Cheers and happy modding,

HZ
Originally Posted by Del Vento
In any case, how is your design different from Jay's piano conversion? From the video I could not tell.
Cheers

Good question, I have not looked at his in any detail. Pretty sure different parts (except CNY70?). And I don't think there is any way the software or algorithms could be similar. Note that I described the algorithm and my testing in one of the videos and maybe someday, if I ever get it working.... I could dump the entire mess onto github! Or maybe will sell the design for $1,000,000,000 to Yamaha.. haha no. Probably $1 is what it's worth.
Originally Posted by CyberGene
Originally Posted by josh_sounds
Yes I did, I have been looking for projects like this ever since... gauging which one's are feasible based on local supplies. Yeah, it would be difficult , but I don't mind soldering and wiring. However but fabricating custom PCBs would be a challenge though.

Can you confirm though that it can do note repetition and legato like this one here or even better:

Triple Sensor
(scroll down a bit and select triple sensor, for the demo sounds)

Thanks a bunch... smile
Josh, repetition and trills depend entirely on the action. I regulated mine very carefully and thoroughly which is why the trills are easy. The sensors simply detect hammer velocity.

Ahh, I see. So you did meticulous hardware and software adjustments. Thanks for the feedback!
Hoping to get my hands on my own keybed...
Originally Posted by CyberGene
BTW, how's Jay's piano conversion going on? I think he doesn't post on the PW or GitHub often, so I'm not sure where he is now, the last I've seen is he already wired one key. I guess by now he should already be finished?

I don't know where the last two months went (well, actually I do: two online conferences, one of which I led, one big tutorial, bought a very nice grand piano and need to arrange the move which is complicated by my wife wanting me to sell the one I have first, a plumbing emergency with a blown up pipe which I fixed myself, and few other minor or not-so-minor things). So, I haven't touched base with Jay (or made progress on mine, which electronically is a clone of his, but mechanically slightly different) since March.

He does post on github when he has a good update. Last time I heard, he was producing sound with good dynamics for one note, and needed "only" to assemble the hardware for 88 of them (the design was already completed for all of them, in my understanding). In fact we were supposed to meet to share some old-grand-piano parts, but that never happened. I told him that if he was ordering more electronics (presoldered PCB) I was interested and that would have reduced unit price and shipping costs, but I did not hear back on that. As soon as I get a moment from being so busy (which I hope will be in a week or two) I will ask him again.
Originally Posted by Del Vento
bought a very nice grand piano and need to arrange the move which is complicated by my wife wanting me to sell the one I have first

Whoa there! I nearly missed this part. First, congratulations on the new grand!!

Second, what did you get, if you don't mind sharing?
Yeah, congrats on your new piano, what is it?
Originally Posted by Gombessa
Originally Posted by Del Vento
bought a very nice grand piano and need to arrange the move which is complicated by my wife wanting me to sell the one I have first

Whoa there! I nearly missed this part. First, congratulations on the new grand!!

Second, what did you get, if you don't mind sharing?

Originally Posted by CyberGene
Yeah, congrats on your new piano, what is it?

Well, you guys missed it because I wanted to have it in my home before "announcing" (and the logistics is proving to be more complicated than I hoped), but since I have piqued your curiosity I now can't delay. Just make sure you are solidly seated to your chair, because I still fall from mine that I can't believe what I've got!!!

It's a 1930 Mason & Hamlin AA, with some updates (among them new strings and new hammers). I'll post pictures and hopefully a recording, when I'll get it, which as I said is stopped from the sale of what I have now (otherwise the M&H must go in storage for now cry )
Hi CyberGene. I’m not a prolific poster but when I saw the results of your project I just had to leave you a message to say, wow! That is some achievement. I am in awe. Congratulations.
Thank you, Andy 🍻
© Piano World Piano & Digital Piano Forums