Brainstorming A Diy Concertina Midi Controller

[Don Taylor]

Matthew Vanitas recently started a thread polling for folks who might be interested in buying a Hayden midi concertina:

http://www.concertina.net/forums/index.php?showtopic=16394&hl=

 

I am afraid that I have hijacked his thread by starting to talk about doing some sort of homebrew, DIY, midi concertina project. Let's move the discussion about a DIY midi concertina from Matthew's thread to this one and leave Matthew's thread to its intended purpose.

 

So, to reply to Łukasz's last post here:

 

There is one problem with iDevices - you have to build a true standalone MIDI controller based on Arduino (or any other microcontroller), which sends MIDI messages over a MIDI cable and this is a lot of computation for Arduino (and my board had too few inputs and outputs to make that happened). The reason for that, is that iDevices require deviceID for anything other than MIDI accesories and this is not easily obtainable. At least that's how it was three years ago, this could have changed. That is why I wrote about Android devices, which have no such limitation. But yes, a smartphone as a "brain" for such concertina is probably the most versatile solution. But it is a path towards commercial production, not a single, DIY controller.

 

In my concertina Arduino serves only as a gatherer for different values, which it then sends in raw form to Processing driver running on host system. The reason for that is that Processing has some neat MIDI libraries, Arduino does not. This requires a host capable of running my code and support for a software MIDI device. Back then that was true for OSX, Linux and Windows.

If you were starting over from scratch then would you use an Arduino/Shield now, or would you use a Raspberry Pi instead? I don't think that RPi was available 3 years ago when you started your project. Maybe keep using an Arduino as a device controller and use a RPi to run the Processing driver and generate midi o/p.

 

A lot more power and capability in an RPi including some synth libraries so it might be possible to use it for an in-concertina synth.

[Łukasz Martynowicz]

The solo use of RPi might be tricky, as it is a computer not a microcontroller, so it has no ecosystem for connecting raw electronics to it.

I have an old Arduino Duemilanove http://arduino.cc/en/Main/arduinoBoardDuemilanove#.UyTK01w4NGJ . If I were to build completely new instrument now, I would probably aim at Arduino Yun or Tre http://arduino.cc/en/Main/ArduinoBoardTre#.UyTJ1lw4NGI , http://arduino.cc/en/Main/ArduinoBoardYun#.UyTJxlw4NGI , depending on what exactly can you run on Yun Linux implementation. This way I would probably try to build a self containing instrument with optional MIDI capabilities.

 

Yet, right now I'm aiming at home studio use and not full-blown stage/gig applications, so connecting pure Arduino via USB and then work on software MIDI is the best option for me. I could just use a laptop as a synthetizer if I ever need to play elsewhere.

[Don Taylor]

Three hinged cylinders will allow same degree of freedom - they will only not bend around the leg in the same fasion, but relative movements of concertina ends can be the same as the real thing.

Do you mean three cylinders in parallel with each other connected by pivots at the inside ends of the concertina. These connection points on the ends forming a triangle?

 

If so, do you put pressure sensors inside each cylinder and (somehow?) combine the three pressure readings to give you one value for setting the volume?

 

Don.

 

(As an aside: Do you suppose that if Charles Wheatstone were alive today, he would be making electronic instruments)?

[MatthewVanitas]

(As an aside: Do you suppose that if Charles Wheatstone were alive today, he would be making electronic instruments)?

 

In all seriousness, undoubtedly. The guy wasn't at all folky, retro, or old-school; the qualities that lead modern folks to play acoustic concertina. He was a real forward-leaning guy on the cutting edge of instrument design. Also, outside our small community he's far better known for all kinds of inventive gear; given how keen he was on electricity in general, electronics would've been right up his alley. Or maybe he'd even discover a yet further step beyond electronic music!

 

Just the lede of his Wikipedia bio shows what kind of guy C-Dub was:

 

 

 

Sir Charles Wheatstone FRS (6 February 1802 – 19 October 1875), was an English scientist and inventor of many scientific breakthroughs of the Victorian era, including the English concertina, the stereoscope (a device for displaying three-dimensional images), and the Playfair cipher (an encryption technique). However, Wheatstone is best known for his contributions in the development of the Wheatstone bridge, originally invented by Samuel Hunter Christie, which is used to measure an unknown electrical resistance, and as a major figure in the development of telegraphy.

[malcolm clapp]

(As an aside: Do you suppose that if Charles Wheatstone were alive today, he would be making electronic instruments)?

 

Doubt it. I would suggest that he would probably be workng for Micro$oft......

Edited March 15, 2014 by malcolm clapp

[Łukasz Martynowicz]

@Don: Yes. But if you want just bellows pressure information you can combine pressures from each cylinders into a single sensor. Of course if you want to improve the concertina controller to be more than just an electronic version of acoustic instrument, you could use this additional information as MIDI controls. I think that this corded version I posted previously uses each cord as a separate MIDI X/Y control in addition to bellows movement information.

 

One drawback with using air cylinders would be relatively small travel, perhaps equal to three fold bellows (assuming the most ergonomic size of entire instrument)...

 

And I agree with Matthew on Charles Wheatstone, I think that he would be first to push boundaries further. After seeing some of his patents on advanced mechanical reeds (e.g. re-tunable reeds) I think that he might even electrify concertina (as in electric guitar, not MIDI controller).

Edited March 16, 2014 by Łukasz Martynowicz

[Terry McGee]

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

[Don Taylor]

If I were to build completely new instrument now, I would probably aim at Arduino Yun or Tre http://arduino.cc/en/Main/ArduinoBoardTre#.UyTJ1lw4NGI , http://arduino.cc/en/Main/ArduinoBoardYun#.UyTJxlw4NGI , depending on what exactly can you run on Yun Linux implementation. This way I would probably try to build a self containing instrument with optional MIDI capabilities.

@Łukasz:

 

They look to be very nice boards, especially the Tre.

 

I still like your single pressure sensor inside the bellows idea and I wonder if better pressure sensors have appeared since you last worked on this project.

 

What range of pressures did you see inside the bellows? The medical industry uses 0-300mm Hg (about 5.8 psi) sensors in blood pressure monitors. For example:

http://cache.freescale.com/files/sensors/doc/data_sheet/MPX2300DT1.pdf

Response time is stated at 1m. You can buy something close, if not identical for $1.49 each: http://www.goldmine-elec-products.com/prodinfo.asp?number=G15473

 

Did you consider these? Or find something better?

 

Anyway, I have just 'won' a busted Chinese concertina off eBay and am thinking about ordering some of these medical pressure sensors to play around with.

 

Don.

[Łukasz Martynowicz]

I now have two of theese http://www.tme.eu/pl/details/mpxv5004dp/czujniki-cisnienia/freescale-motorola/# - they are one way, so you need two of them, and have a range of 4kPa. Looking through this site I have found a sensor I'll probably give a go when revisiting my project after finishing my acoustic Hayden that I'm building now - http://www.tme.eu/pl/details/d6f-ph5050ad3/czujniki-cisnienia/omron/# - it is a two way sensor with probably the ideal range of +/- 0.5 kPa. All those pressure sensors are differential.

One note - the exact range of over/underpressure you'll get inside your bellows depends on how you'll setup the airflow - it will be significantly higher when you live a hole simulating a single speaking reed or two full chords… I don't have hard numbers on this, but judging from my sensor readouts I could go from as low as 200Pa to as high as 1kPa of overpressure. [i'm giving this numbers from my head now, as at this moment my concertina is dissasembled and can be used only as a dumb MIDI array with no volume or velocity control.

 

Here I found an article on building a MIDI accordion: https://github.com/accordion-mega/AccordionMega/wiki/Accordion-Mega-story - it uses a sensor with a range of 110kPa but from the look of it, it is an absolute sensor. When I was designing my concertina I first used a 50kPa sensors, which failed miserably, because I'm using differential setup, not an absolute measurement.

[JimLucas]

I'm following this and the other (midi-Hayden) thread with interest, though I currently have neither the money nor -- more seriously -- the time to pursue such a project personally.

 

Also, I'm not particularly interested in the Hayden layout, or at least not any more so than other layouts. (One layout that particularly interests me is the Linton system, which is considerably less available than the Hayden... i.e., currently not at all.) But if you folks develop techniques for building a reasonable midi concertina, then they should be applicable to any keyboard layout, so that would be great.

 

In fact, one great advantage of a midi system is that one can place switches ("buttons") wherever one wants, without having be concerned with chambers (size, shape, fitting into an end) or levers (positioning, lengths, collision avoidance, etc.). Assignment of individual notes can all be done in software, so (e.g.) one could change between Wheatstone and Jeffries anglo layouts with the flick of a switch

[Don Taylor]

Lukasz and I have decided not to collaborate on a single project, but we will share lessons learned.

 

Lukasz is interested in building a 64 button Hayden and has a well-developed prototype.

 

I am not wedded to a Hayden layout, but my project only consists of a couple of orders for parts that have not yet arrived.

 

I suspect that my first attempt will be a 20B Anglo and I don't expect to get anything working until next winter. If I get anywhere then I do plan to 'open source' the design and the code so that others can build their own or develop it further.

 

As far as keyboard layouts are concerned, I am hoping that the bulk of the design will be independent of keyboard layout. I am also hoping to make it possible to experiment with ideas such as from Chris Ghent - pitch adjusters on small keyboards.

 

I want to make a midi concertina that can be made from a regular concertina, in my case a cheap Chinese box, and I want to keep the parts cost low.

 

Don.

[chris]

Er, is this thread in English?

Chris (active member of the Luddite Party of Great Britain)

[Łukasz Martynowicz]

If using a matrix keyboard and a flexible note mapping in the software, then a single hardware can support any keyboard layout, as long as they have same physical buttons arrangement. With a bit more work it is possible to build a completely flexible and system-independend platform, on which non-physicaly equivalent systems could be exchanged in a matter of a simple end plates variations, either in form of detachable electronic keyboards or button+lever platforms. That includes virtually any system, be it uni- or bisonoric, duet or english-style split sided layout.

 

At this point my prototype serves that purpose well, as each button has an independent note mapping, though I plan to explore possibilities of saving some buttons by combining enharmonic equivalents together, to further enlarge the compass of simultanously available notes.

[MatthewVanitas]

If using a matrix keyboard and a flexible note mapping in the software, then a single hardware can support any keyboard layout, as long as they have same physical buttons arrangement...

 

That was one thing I was pondering: which keyboards are cross-compatible in terms of physical layout.

 

The Linton system (interesting thread on it) looks like it could use the same physical keyboard as the Maccann, just different encoding of the notes on the computer. I suppose a 40b Anglo could also re-program as a Jeffries Duet?

 

For Hayden, the main one that jumps to mind is the Harmonic Table layout that C-Thru music (a maker of MIDI keyboards) is promoting. But I mainly know of that system due to folks doing the reverse: buying a keyboard intended for Harmonic Table, and re-programming it to Hayden layout. There are probably some other hex-based scales out there, and I'd imagine both hex-based systems, and Hayden itself, would be the best "crossover design", something that would lure non-concertinists to the instrument; pitching it as a "portable jammer keyboard with dynamic control", plus visually more interesting onstage than a flat keyboard, could be a method of bringing MIDI players over.

 

 

 

 

. With a bit more work it is possible to build a completely flexible and system-independend platform, on which non-physicaly equivalent systems could be exchanged in a matter of a simple end plates variations, either in form of detachable electronic keyboards or button+lever platforms. That includes virtually any system, be it uni- or bisonoric, duet or english-style split sided layout.

 

While such would be awfully clever, do you reckon there are enough folks that play multiple layouts to justify the additional technical challenges of detachable endplates?

[Łukasz Martynowicz]

 

. With a bit more work it is possible to build a completely flexible and system-independend platform, on which non-physicaly equivalent systems could be exchanged in a matter of a simple end plates variations, either in form of detachable electronic keyboards or button+lever platforms. That includes virtually any system, be it uni- or bisonoric, duet or english-style split sided layout.

 

While such would be awfully clever, do you reckon there are enough folks that play multiple layouts to justify the additional technical challenges of detachable endplates?

 

 

 

I thought about this mostly by a mean to broaden the market for such MIDI instrument so it could reach a criticall mass to start ANY production. With this approach only a small part of production process would require system-dependand design and everything else would be exactly the same. But to make such approach easier, this should be a plain switch buttons design and not a true lever mechanism. And this could prevent some folks from buying it, because switches of suitable size have only about 2mm travel and don't feel like a real thing.

 

And as I share your enthusiasm on Hayden (I realy liked your term 'Haydenites' ), I'm a bit skeptic about it's glorious future. There were some attempts made on comercialising isomorphic layouts in form of Axis jammer you've mentioned, and at least one of Hayden jammers never went into mass production stage (or had a really small run) and they both hadn't get a noticeable attention. From my personal experience, based on numerous conversations with different musicians, our small concertina comunity may be the largest group that is at least aware of this layout, let alone interested in it. Whether we like it or not, we live in a piano world and even as well established layouts as various CBA systems don't have their fair share in the MIDI keyboard market (a quick google search gave me three hits on flat-keyboard synthesizers with accordion keyboard, one of them analogue and discontinued more than 20 years ago; there are also fully-blown Roland reedles accordions and that's virtually all there is).

[JimLucas]

If using a matrix keyboard and a flexible note mapping in the software, then a single hardware can support any keyboard layout, as long as they have same physical buttons arrangement...

That was one thing I was pondering: which keyboards are cross-compatible in terms of physical layout.

I was thinking about that myself, though my thoughts went in somewhat different directions from yours. (See below.)

 

The Linton system (interesting thread on it) looks like it could use the same physical keyboard as the Maccann, just different encoding of the notes on the computer. I suppose a 40b Anglo could also re-program as a Jeffries Duet?

Cross-compatibility has its limitations, but approximate cross-compatibility might still be acceptable in some circumstances. E.g.,

• In order to be fully chromatic, the Maccann has a few "outlier" buttons both above and below the 6-wide arcing rows. In the Linton scheme they wouldn't fit, but they could simply be ignored. On the other hand, the left hand of the Maccann has fewer rows than the right, while on the Linton they're equal. So a 46-button Maccann would fit only 3 octaves of a Linton, and even a 55-button Maccann would be missing one button (in the left hand), for the highest A of a 4-octave layout.
• Jeffries duets and anglos (of 38 or more buttons) use identical button layouts. There cross-compatibility would be complete.
• A Crane layout (here's where I started to think about cross-compatibility) could be converted to a multi-row 5-wide anglo. A 55-button Crane array could have up to 5 different single-key rows spanning both hands and an extra row in the right hand. Or it could be programmed as a 20- or 30-button and all the other buttons ignored. Or it could be programmed as a 20- or 30-button with an arbitrary (or "well thought out") scattering of additional notes on the "extra" buttons.
• Another potential concern regarding cross-compatibility, though, is the fact that none of the standard layouts is a truly rectangular array. Anglos and most duets have differing curvatures to their rows, while anglos and Haydens have "rake" or "slant" to their columns. So "compatible" may not be quite the same as "standard".
• But the Hayden/Wicki has both "slant" and "parallel" variants, and the rows of the "original" Crane/Triumph are arcs while in the Crabb/Jeffries version of that layout the rows are chevrons. So those sorts of variations shouldn't be game stoppers; they can be gotten used to. (Though I personally find the chevron variant of the Crane layout more comfortable than the arc variation.)

With a bit more work it is possible to build a completely flexible and system-independend platform, on which non-physicaly equivalent systems could be exchanged in a matter of a simple end plates variations, either in form of detachable electronic keyboards or button+lever platforms. That includes virtually any system, be it uni- or bisonoric, duet or english-style split sided layout.

While such would be awfully clever, do you reckon there are enough folks that play multiple layouts to justify the additional technical challenges of detachable endplates?

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.

[MatthewVanitas]

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.

 

Good point; I guess what was rubbing me is that the ends appear to be one of the more investment-heavy parts in the instrument. If you cast them out of resin or whatnot, you need to have molds made, and having to make/buy a different mol for each settup wouldn't be cheap. But then again, I suppose if you cast a "generic" end with any switches/fretwork/knobs, etc you could drill the action holes into said end in whatever format you pleased, install whatever find of rail/rest at whatever angle, etc. I guess it just depends on how much your R&D locks you in to a corner.

[JimLucas]

• In order to be fully chromatic, the Maccann has a few "outlier" buttons both above and below the 6-wide arcing rows. In the Linton scheme they wouldn't fit, but they could simply be ignored. On the other hand, the left hand of the Maccann has fewer rows than the right, while on the Linton they're equal. So a 46-button Maccann would fit only 3 octaves of a Linton, and even a 55-button Maccann would be missing one button (in the left hand), for the highest A of a 4-octave layout.
• A Crane layout (...) could be converted to a multi-row 5-wide anglo. A 55-button Crane array could have up to 5 different single-key rows spanning both hands and an extra row in the right hand. Or it could be programmed as a 20- or 30-button and all the other buttons ignored. Or it could be programmed as a 20- or 30-button with an arbitrary (or "well thought out") scattering of additional notes on the "extra" buttons.

I've just realized that the above descriptions assume that a midi duet would have the same exact layout as a standard duet with reeds. In particular, they assume that the left hand would have fewer buttons than the right hand. But the reason that's the case on "real" concertinas is that the reeds and chambers for the lower notes require more space. Well, since a midi instrument doesn't have reeds or chambers, there's no reason why it can't have as many buttons for the left hand as for the right.

 

So what would that mean? For a midi duet one could/should provide as many buttons for the left hand as for the right. So a 46-button Maccann should be transformed into a 50-button, a 55-button duet (Maccann or Crane) would become a 60-button, etc. This could have significant ramifications not only for cross-compatibility, but also for composing arrangements with a "standard" layout.