Brainstorming A Diy Concertina Midi Controller

[Łukasz Martynowicz]

I would look at it from a different direction. If you make the standard construction such that the different button-containing end plates can be constructed independently and simply attached into/onto an unvarying "core", then

• If someone wants a Hayden, attach the Hayden ends and ship it. If someone wants an English, attach the English ends and ship it. Etc.

And if someone wants two or more layouts, all that's necessary is to include each set of ends.

That is exactly what I meant by "in form of detachable electronic keyboards". Furthermore, you don't have to make whole endplates or drill holes in pre-cast ones (casting isn't realy an option, as it requires large runs to be economicaly efficient) - you can just design an endplate with replaceable insert. Apart from englishes, all (?) other systems use handstrap/handrest construction and only Dipper custom uses long three rows layout, so they would all fit in some well defined area.

As to making instruments symetrical in terms of button number: for technical reasons I'm assuming a 64 button "poll" to divide. With 64 button duets you will want (at least I would) to have a slightly more buttons on the right hand side. That defines a split of the poll between sides, wich is then fixed for any other keyboard, as some wires must go through the bellows and this cannot be overcome by interchangeable endplates inserts. Apart from that, any layout could use this poll to the maximum of (example split) 30 buttons left, 34 buttons right.

One thought though: with MIDI, if the instrument is built as single chanel only, there is less point for large overlap as you won't get more sound from playing same note on two buttons… And many DAV programs don't support multiple channel MIDI recording.

[JimLucas]

I would look at it from a different direction. If you make the standard construction such that the different button-containing end plates can be constructed independently and simply attached into/onto an unvarying "core", then

• If someone wants a Hayden, attach the Hayden ends and ship it. If someone wants an English, attach the English ends and ship it. Etc.
• And if someone wants two or more layouts, all that's necessary is to include each set of ends.

That is exactly what I meant by "in form of detachable electronic keyboards". Furthermore, you don't have to make whole endplates or drill holes in pre-cast ones (casting isn't realy an option, as it requires large runs to be economicaly efficient) - you can just design an endplate with replaceable insert. Apart from englishes, all (?) other systems use handstrap/handrest construction and only Dipper custom uses long three rows layout, so they would all fit in some well defined area.

I believe Colin's franglo uses both bar-strap and thumb loop, and the Wheatstone "double" uses thumb loops like the English (but no plates for the little finger), but presumably something like that would be special order and cost extra. As for that "well defined area", though, it would presumably have to be wide enough for a reasonable Hayden keyboard with enharmonics (more than 7 buttons?) yet "tall" enough for a decent Crane (7 buttons?) or even English (8 or 9 buttons?) layout.

 

But you could also have more than one standard size... in particular, a smaller size for layouts with fewer buttons (30 button anglo?), which would simply not use all the code assignments included in the larger array.

 

As to making instruments symetrical in terms of button number: for technical reasons I'm assuming a 64 button "poll" to divide. With 64 button duets you will want (at least I would) to have a slightly more buttons on the right hand side.

I'm curious about your "technical reasons" for this 64-button limit. Why not, e.g., 128? (A single byte, after all, supports 256 different codes. However, I've never tried to program MIDI, so I don't know whether its format limits you.) That would allow for even more than Brian's suggestion of 65 buttons, plus numerous control codes, and even allow some redundancy or unused codes (like our DNA?). Come to think of it, they make 88-key MIDI "piano" keyboards with added controls. And the keyboard I'm typing at right now, while not MIDI, has 113 keys. So I wonder, why impose a 64-button limit?

 

As for wanting more buttons on the right hand side, I wonder what sorts of arrangements you have in mind and what sorts you're excluding. I might note that many piano arrangements have a greater span in the left hand than in the right, and not just through octave duplication in the bass. My impression is that current duet players tend to go for instruments with more buttons not so much to get extra range at the top but to get extra range at the bottom... and to get more overlap.

 

That defines a split of the poll between sides, wich is then fixed for any other keyboard, as some wires must go through the bellows and this cannot be overcome by interchangeable endplates inserts.

Why must wires go through the bellows? I'm typing at a wireless keyboard. I believe it uses radio frequencies, but infrared or even visible laser light (which wouldn't be visible outside the bellows) should work just as well. It has been suggested that connection to a synthesizer and/or mixer/amplifier/speakers could be wireless. If that's done, then why not use it also for communication between two ends?

 

Apart from that, any layout could use this poll to the maximum of (example split) 30 buttons left, 34 buttons right.

 

Ah, that unequal split again. Unequal split isn't the best fit for an English (or Linton) layout.

 

One thought though: with MIDI, if the instrument is built as single chanel only, there is less point for large overlap as you won't get more sound from playing same note on two buttons… And many DAV programs don't support multiple channel MIDI recording.

What's DAV? (Wikipedia gave me many possibilities, but none of them seemed to fit this context.)

 

I'm certain that the main reason for the overlap is not to get more volume on particular notes by playing them on both ends simultaneously but to enhance playing through alternate fingerings, either to avoid or to take advantage of parts crossing between the hands, depending on the circumstance. More overlap is desirable because it provides more options for musical control.

Edited March 19, 2014 by JimLucas

[Łukasz Martynowicz]

My mistake - it is DAW - digital audio workstation.

 

Have you missed that this is the DIY thread?

The answer for almost ALL your question is that I try to come up with ideas that I'm able to design and make - you are just wandering around on a conceptual level. Also, I'm designing with primary a Hayden layout and my own needs in mind. So no wireless communications in the bellows as this would require two power sources, two sets of electronics, etc.. and would rise total cost. As for the number 64 - it is because multiplexer used in HARDWARE have 8 bits and the most optimal use of connectors available on a small Arduino unit I'm using is thus 64 (8x8 grid, which uses 9 multiplexers, so you need to controll 6 outputs and one input). For 88 key you have to add 3 more standalone multiplexers and for 128 you have to double the whole deal (the outputs are common, but the additional input is independent). With piano synthesizers 88 key is a musical standard, with concertina 60+ keys we're getting to largest sizes manageable with handstraps, so I never even considered going 128. [This is of course trivial with mass produced, commercial instruments with large market, so you can have full-size accordion controllers with 212 buttons and zilions of other controls].

 

As to exchangable keyboards - of course it has to be tall enough for at least 6,5 buttons to accomodate a minimal Hayden (6 + row shift). With transposition capability, you can play on minimal Hayden keyboard with only single accidentals (in E.T.) by just transposing to the desired key. It is 8,5 if I wanted to benefit from enharmonic capabilities of a Hayden in other tunings. And I realy don't aim at building small, 30 key Anglo box, I'm perfectly fine with a 7-8" box.

 

English and Wheatstone Double are a different family boxes and would require a different endplate altogether, so the whole idea of interchangeable insert obviously does not apply. At least not in form of a single, totally universal design. Everything has some limitations.

[JimLucas]

My mistake - it is DAW - digital audio workstation.

Thanks. Found it. Now I understand... sort of, since it's not something I'm personally familiar with. Unless Audacity counts, but even with that I'm still at a primitive level.

 

Have you missed that this is the DIY thread?

 

The answer for almost ALL your question is that I try to come up with ideas that I'm able to design and make - you are just wandering around on a conceptual level.

Exactly. But different individuals have different knowledge, skills, and tools with which to embark on a project, as well as different desires for the ultimate result, and so both the process and the details of the result can vary considerably. I, unfortunately, have neither workshop nor tools at this point in time, and my skill set and desires are apparently quite different from yours. But I'm very much interested, and the best I can hope for at present is to learn as much as possible by asking questions and presenting ideas for review by you and others... ideas based on my own background. That's what I've been trying to do. If I'm lucky, maybe I'll be able to try what I've learned a couple of years from now.

 

Also, I'm designing with primary a Hayden layout and my own needs in mind. So no wireless communications in the bellows as this would require two power sources, two sets of electronics, etc.. and would rise total cost. As for the number 64 - it is because multiplexer used in HARDWARE have 8 bits and the most optimal use of connectors available on a small Arduino unit I'm using is thus 64 (8x8 grid, which uses 9 multiplexers, so you need to controll 6 outputs and one input). For 88 key you have to add 3 more standalone multiplexers and for 128 you have to double the whole deal (the outputs are common, but the additional input is independent).

Okay, and again thanks. I have a better idea of where you're coming from... and a number of new things to learn more about. One thing that I think confused me, though, was your talk about interchangeable end plates and replaceable button inserts. I understood that as suggesting commercial production, which might justify something less "off the shelf" in other aspects, as well.

 

With piano synthesizers 88 key is a musical standard, with concertina 60+ keys we're getting to largest sizes manageable with handstraps, so I never even considered going 128.

I think some of our current players (too bad Dirge has taken a bye) would disagree with you about what's "manageable" -- especially if one could have as many buttons for the left hand as for the right, -- but that's really a side issue in view of what you've clarified as your own goals.

 

As to exchangable keyboards - of course it has to be tall enough for at least 6,5 buttons to accomodate a minimal Hayden (6 + row shift). With transposition capability, you can play on minimal Hayden keyboard with only single accidentals (in E.T.) by just transposing to the desired key. It is 8,5 if I wanted to benefit from enharmonic capabilities of a Hayden in other tunings.

Once again you're back to Hayden-only design, where I thought that for a while you had been speaking of something more universally interchangeable. No problem here.

 

And I realy don't aim at building small, 30 key Anglo box, I'm perfectly fine with a 7-8" box.

That suggestion from me was once again in conjunction with the thought that you were thinking in terms of multiple-layout commercial production. I think there would definitely be a market, but I guess that would be for someone else to consider.

[Don Taylor]

I think that a little conceptual conjecturing is good even if the ideas are unlikely to be implemented.

 

Don

[Łukasz Martynowicz]

@Jim: when such individuals will embark on the common project, I'll be more than happy to read in full technical detail about such ideas as internal and external wireless data transfer in DIY project. But it sounds like an overkill to me, when you can simply have a couple of wires going through the bellows or in some other connector.

 

You're right about general ideas on commercial product, and yes, I was thinking about broadening the market for potential product, but keeping the design simple enough to be able to produce it on a small workshop scale. Think of it as adaptation of traditional concertina hand-building approach to the world of MIDI instruments. Even "large batch" of "mass produced" MIDI concertina would be more likely in the order of 10-50 instruments than a 1000. I'm just trying to keep it real.

 

The whole idea of interchangeable endplates came to my mind because I have an already modular design:

 

 

As long as I'll follow the same connection pattern, I can make any physical layout by just swapping the button plates, so I could easily accomodate various duet systems already.

 

I didn't thought about 30b Anglo mostly because I'm aware, that for Anglo players size does matter significantly, and with the whole interchangeable endplates idea they probably wouldn't be very happy with an instrument in a size of a large duet "heavy tanker". The innards of course support any type of concertina up to the upper limit of buttons, be it 64 or any other number.

 

And by manageable I meant a physical constraints of finger reach and handrest/handstrap design, not a musical manageability of a large compass. I'm using a combined thumbstrap/handstrap approach now on my Elise and plan something completely different on my acoustic Hayden I'm building - similiar to Thomas Restoin's thumbstrap+wriststrap, but more suited for movements required on large Hayden (mostly increasing a reach of pinkies and allowing for easier playing close to handbar).

 

Hope that this claryfies my approach.

 

[Don Taylor]

Lukasz

 

Interesting picture.

 

How do you, or anyone else who has played it, find the feel of the buttons compared to real physical buttons?

 

They are clearly going to feel different, but is that a real problem once you have got accustomed to their feel?

 

I am trying to decide what to do about buttons and I do like the flexibility of arrangement that can be had from using electrical switch buttons rather than using optos on the existing physical buttons.

 

Maybe I should ask this question of the expert players: Other than the air button do you use the buttons in anything other than an on/off state? Do you manage to press the buttons partially in so as to control volume on a button by button basis?

[Wolf Molkentin]

Don, half opening resp. closing the buttons (and thus pads) will result in bending the tone, be it deliberately (which is working quite fine with the lower reeds) or by accident.

 

Best wishes - Wolf

[Don Taylor]

Don, half opening resp. closing the buttons (and thus pads) will result in bending the tone, be it deliberately (which is working quite fine with the lower reeds) or by accident.

 

Best wishes - Wolf

Ok, is this a feature that is actually used and would be missed?

 

Well, actually it makes no difference as even if I re-use the original button mechanism and use optointerruptors to detect the button state then I would still only know if the button was off/on. I can't tell how much a button was depressed.

 

If this feature is required then it would have to be emulated with something like a pitch bender on a midi keyboard.

 

The air button is different in that it would remain as a mechanical means to relieve bellows pressure.

Edited March 19, 2014 by Don Taylor

[Łukasz Martynowicz]

Don - have you ever played a harpsichord? Unlike a piano, there is a distinctive moment, when the plectrum plucks the string. Same goes with switches, but to a lesser extent - you can feel when they switch. And they have only 2mm travel. As for playing - such keyboard is very unforgiving in terms of a rhytm and you'll need some time to get used to that instant response. This can be corrected in the software, to some extent, via seting a longer note attack time.

 

And bending notes via half-pressed buttons on a MIDI concertina would require analog, travel sensitive buttons which I don't know even exist. Roland accordions use a single common bar for note bending.

[MatthewVanitas]

Regarding button speed, Dirge had a really interesting observation in the 2009 thread about S-Wave's MIDI EC:

 

 

What I didn't expect is that I was almost immediately playing things crisper and faster than I do normally, it was really noticeable. I wasn't aware of the mechanics of my concertinas holding me back (they are both Aeolas) but they clearly are. Dean tells me others have found it tightened up their play, because it is so precise you can't blame the concertina for broken chords or poor timing so you sort it out.

 

If I read that right, what makes the MIDI a mental shift is that it's better than the acoustic.

 

Anyway, the 2009 thread is really worth reading.

 

 

 

 

If this feature is required then it would have to be emulated with something like a pitch bender on a midi keyboard.

 

I'm not sure that I personally would have great user for a pitch bender, but when I was envisioning extra knobs/switches on a MIDI, for a handstrap model I thought a little tensioned pitch-wheel by the thumb would be convenient, so long as it wasn't somewhere it'd be accidentally brushed.

 

That is one thing that's going to annoy me when I "upgrade" from the Beaumont to a trad-reed: I really like the air button on the handrail (as shown in my avatar). Though probably not patented, I take it that using that same placement for other makers' concertina would be in poor taste, "biting their style"?

[Don Taylor]

These guys have an approach to making a velocity or pitch-sensitive keyboard:

 

http://www.youtube.com/watch?v=KfjEYu79J-4

 

It involves having two layers of keys and timing how long it takes between the press of the upper button and the subsequent impact of the upper button on its corresponding lower button. The video explains this better. Not sure what happened to this project and it would be beyond what I could do, but still interesting.

 

I am looking at my Korg microKEY midi keyboard and wondering about putting a Dremel tool to it... It has 37 velocity sensitive keys, a USB powered midi controller and costs less than $80.

[Wolf Molkentin]

As to bending, I'd easily guess that this would be a welcomed feature anyways. With the acoustic concertina, it isn't that widely used, or even known. However, I've met one EC player who does it nearly all the time, without really taking notice of that himself except when questioned...

[Wolf Molkentin]

Don, I believe what you are describing to be the very common feature of a velocity-sensitive keyboard...

[Don Taylor]

I just stumbled across this article by Jim Plamondon about why his Thummer (Jammer) keyboard project failed. Interesting reading.

 

http://blog.igetitmusic.com/2009/09/what-killed-thumtronics.html

 

Setting aside the business reasons for its failure, I picked up on a couple of technical things to consider:

 

1) Use of motion sensors.

See: http://www.youtube.com/watch?v=lWK1d9fzlVQ

I think that Alistair Anderson could make use of motion sensing ...

 

Very hard when Thummer was being developed, dead easy now because they are off the shelf items in every cell phone.

 

2) Substituting midi 'Channel Pressure' (in our case bellows pressure) for key velocity and after touch as found on a piano.

[Łukasz Martynowicz]

Looks like I have reinvented the Channel Pressure when I have tried to make a quick-but-decent velocity-sensitive "Hayden piano" conversion out of my concertina innards a few months ago This works, and I have been experimenting with using a pressure sensor and specially crafted pressure pad placed below the keys.

 

Don: this is a standard solution for velocity sensing. A came across another solution, involving pressure sensitive resistance based strips, but this best works for linear keyboards.

About stripping down your Korg, I think it might be hard to incorporate any form of bellows dynamics into it, but it might be worth to take it apart. Just in more reversible fashion than using a Dremel tool

 

But the whole motion sensing idea could be a great way to build a futuristic, bellows-free MIDI concertina, I must dig a bit more into it. Ideally, with two motion sensors it could be possible to have two independent dynamics for left and right hand notes. No more unsteady accompaniment

[JimLucas]

Don, half opening resp. closing the buttons (and thus pads) will result in bending the tone, be it deliberately (which is working quite fine with the lower reeds) or by accident.

Ok, is this a feature that is actually used and would be missed?

I'm only aware of one person -- the same one mentioned by Wolf -- who demonstrates an ability to do this consistently in a controlled fashion, so I don't think it would be missed by the general population of concertina players.

 

Also, on a "real" concertina, the effect is obtained through a combination of incomplete button depression and bellows pressure, and I don't think it would be possible to implement it with anything like the same feel as the original. Then again, for those who can't already do it on a "real" concertina, that might not matter.

 

Bending on a "real" concertina does seem to be a difficult technique to master, so in thinking about your question, I had a thought that if it were possible to implement this in a MIDI concertina, it might become popular. But I'm not sure it's worth the effort, at least not in the DIY MIDI forms being discussed here. Maybe Don, Łukasz, and others can judge that.

 

Like velocity and aftertouch, bending only makes sense if implemented separately for each button. And like aftertouch, it's a parameter that should be variable throughout the duration of any given note. I think there's the rub. Is it possible implement more than one such parameter independently on the same button? If so, can it be done in a way that's comfortable for the player and doesn't interfere with other aspects of playing? And if not, which parameter do you think would be most important? Bending? Volume? Vibrato? Something else? Well, I suppose the choice might be different for different different styles of music or individual musicians.

[Don Taylor]

If there was any push to come up with a DIY-MIDI instrument I'd be interested in being involved. I can offer general electronics skills, but unfortunately, my programming skills lie mostly in now-dead languages...

 

Incipe

Fac aliquid fieri expectat

Ait: "Oportet quod plus est quam vita"

finem faciam

Desine

 

Terry

It has occurred to me that this project will come to naught if we don't have some really good sound samples.

 

I don't yet know enough about what is involved in creating and managing sound fonts, but I think that we ought to be trying capture the best quality samples from the best instruments that we can lay our hands on.

 

I expect to post more about this later as I think that this would be an important project in general, but is this an area where your expertise could play?

 

Don.