Updated: Apr 7, 2018
Years ago Randy Machany, dulcimer player with the group No Strings Attached, created what became the first attempt at creating a midi-hammered dulcimer or at least something that had some of the functionality. If you’d like to read a description of that seminal work, check out the pdf of Randy’s thesis. Randy contacted me many years ago to show what he had created and enlist my help in making it better. At that point I didn’t have a clue what I could contribute. I’m not that much further along right now but do understand some of the key impediments to a more useful version. The material included on this page are documents and links to material which others might find useful. If there is any thing you’d like to contribute, please contact me and I’ll include it.
What got me going again, at least briefly, was Dan Landrum’s interest in developing such a dulcimer. He felt it would be very useful for teaching, transcription and certainly unique performances. Dan has a lot on his plate being a top notch performer and editor of Dulcimer Players News so he hasn’t been able to consistently put the time in. I on the other have spent a lot of time attempting to solve some of the problems. My failings are a lack of electronic and programming experience. I certainly am able to build the infrastructure but I could use more help with sensors, electronics, microprocessors and of course midi itself. As a result the project is stalled. If you feel like contributing, get in touch.
An exchange with harp builder David Kortier
I’d like to start by sharing a conversation I had with David Kortier, a maker of fine harps and the developer of a midi harp. What he has learned and was willing to share certainly contributed to the dialogue and the possible development of a midi-hammered dulcimer.
James – I have a customer who wants me to build him a midi hammered dulcimer. Currently I know about enough to be dangerous. I was aware of your work on the midi harps since I too build the occasional harp so I thought it would make sense to contact you. Pretty amazing what you have accomplished in that arena! I’m looking for suggestions on what would be a suitable controller for a multiple pickup instrument or did you develop your own? What sort of pickups would you recommend for a hammered dulcimer (I was thinking of using short pieces of cable piezo)? If you have any other suggestions as to things I should consider, I’d appreciate it.
David – Good to hear from you! Here’s the story – I have worked all along with an electrical engineer who has designed and produced the controller boards for me. We are now in the fourth or fifth generation of design, and have evolved to something that could possibly be useful for applications other than a conventional harp. Instead of one large universal board that is suitable for any harp of any size, we now are modular. Each board accepts 12 inputs (twelve pickups), and you can daisy chain as many together as needed. Each board can be set to output whatever notes are desired, either by a 16 position selector switch on the board, or by connecting to a PC and running an app that changes the settings via a USB cable. The move to this modular design was partly an attempt to bring the price down, but unfortunately we haven’t sold enough of them yet to benefit from manufacturing in larger numbers. The current price is $750 per board because we are amortizing the design costs as we go, and the boards are hand populated in my engineer’s shop. If we ever get up some momentum this cost would drop to about a third of present. So, if you have 36 notes to output, such as a full sized harp, you would need three boards, or a cost of $2250 just for the controllers. How many notes does the average dulcimer have, anyway?
Now, the other part that you need to know is about the pickups. To successfully trigger MIDI events, you need isolate each string from its neighbors. My first attempt at a MIDI harp worked, surprisingly, but I got clusters of notes instead of individual notes. So, the process for the last 15 years or so has been to develop a transducer that captures the vibrations of the string it is against, but reject or not be sensitive to any other vibrations (or electrical interference). And, they need to be consistent in output, so you don’t have strong notes and weak notes. The controller boards have individual gain and threshold adjustments, but having the pickups be more or less consistent makes the whole job a lot more successful. I predict that your idea of using short lengths of piezo cable would sort of work, but I assume you would have difficulty isolating notes. Maybe not…
The way I make pickups is to place the active element, the piezo bit with foil shielding and coaxial output wire attached, into a mold and pour epoxy resin to encapsulate it all. The trick has been, first of all, to know exactly where it needs to be situated, and then to get it consistently placed there.
I currently mold them in two pieces. The electronics are in one bit, a piece that looks like a short piece of half-round moulding with a wire coming out the end, and the other bit is a rectangle that has a notch on one side at about 45 degrees, to hold the active part in the correct orientation to the string. It occurs to me that the one bit with the element in it might be appropriate for amplifying a dulcimer, without the other part. I will take some better photos and send them to you.
Don’t feel that you are intruding on my domain by producing a MIDI instrument. I would be delighted if other makers started offering MIDI instruments, even, or especially, harps. Part of my problem in getting up that momentum is that very few people have seen such a thing, and simply don’t have the imagination to envision the possibilities. Not only would my sales be easier with others out there, but if it worked out that my controller boards and pickups were wholesaled to other makers, the unit cost would come down significantly, which of course would make the instruments much more affordable. So give this some thought and let me know what further question we would need to address. I will help you any way I can.
James – Thanks so much for your response. You have obviously put a lot of effort into the development of these things and I appreciate your willingness to share some of your expertise. I was a little taken a back at the cost of your controllers but understand that when anything is built custom, it ends up costing a lot more. Was there nothing off the shelf that even came close? http://www.midi-hardware.com/ I also have been looking at what is being done with the Arduino. These things are dirt cheap ($20-$50) and amazingly flexible. They have spawned an explosion of applications as everything about it is open source. It is capable of being adapted to midi. The small ones, the Arduino Uno are only capable of having 6 piezo (analog) inputs but there is another board which can handle 16 and I’m assuming that they could become modular to accommodate more inputs. I have no idea of whether they are capable of handling all the parameters that you’d desire in a midi controller and if they are fast enough to handle the amount of data you’d want it to handle. I’ve just been doing some exploring. Can a player of your midi harp change instruments at the harp/controller? Do you have any links to performers doing exciting things with your midi harps? Ultimately I find that is what drives interest is someone doing good work.
As far as pickups go, I’ve already done some experiments using the cable piezo located in a slot below a conventional 1/8″ acetal saddle (see attached) The string bares down on the saddle and activates the piezo. I understand the need for isolation so it would be a slog dealing with all the connections. Something akin to what you have done by embedding piezo material would also work but be more time consuming. What piezo material are you using and where do you get that? Most of the piezo material I use comes from Measurement Specialities.
The largest problem besides finding a suitable controller is dealing with string spacing. Hammered dulcimers unlike harps have far too many strings crossing bridges that are quite close together. This characteristic makes it almost impossible to create individual piezo sensors to accommodate each course (pair of strings). I’ve enclosed a photo of a 3/16/18/9 (one of my larger hammered dulcimers) to show what I mean. There are three courses (pair of strings) grouped together at the 6 inch mark which illustrated the problem. I’m inclined to stick with dulcimers which have greater space between courses or even create an exclusively midi instrument ( like your midi harp) which would utilize just one string per course and would have reduced tension and adequate string spacing.
I also have been looking at software based solutions to midi interfaces which respond to pitch such as JamOrigin. These solutions which will work with any instrument entirely negate hardware but according to my customer have some latency. As computers get faster though latency becomes smaller and smaller possibly making them more viable.
David – I have been aware of the websites that you mention for a long time. In fact I purchased such things when I was just starting out with all this. What I discovered was that I have no capacity for working with programming, or connecting these things together so that they actually do what I want them to. My only option was to pay someone else to sort this out. I found the engineer that I mentioned on the web, I contacted him and told him what I wanted to do and how it was to work, and he took it from there. All I had to do was keep sending him money. We have worked together for more than ten years now, and every time I have run into (yet another) problem, he has always been able to remedy it, often by simply emailing me new code to patch into the processor.
The catch is that this plan is only viable if there are customers buying the product, willing and able to pay what it costs. There have been a few, enough that I haven’t lost TOO much money. However, since the economic downturn, call it what you will, there have not been many willing and able. I have threatened, in my own mind of course, to simply walk away from the MIDI harp. But then I get a few inquiries and I am back in the middle of it. However, as I get older, I find that I don’t have as much energy as I used to, and I would like to focus on the things that give me the most satisfaction. Working with beautiful woods, carving and shaping it, and hearing the sounds that the instruments make, this is what matters most to me, and if I had any sense, a debatable point, this is what I would stick to.
I am telling you all this to explain why I am so generous with information about MIDI implementation. I feel that it is a very exciting area for music making and I don’t think it has been exploited at all. Making a few ocean sounds at the beginning of a tune doesn’t count as exploiting the medium, IMHO. But I would enjoy it almost as much if someone else makes it happen.
The short answer to your question about whether the Arduino, or Roman Sowa’s system would work is yes (probably). With a little bit of technical savvy and some experimentation I think a system could be put together, at a very modest cost.
The sticking point is interfacing with the vibrating strings. Most of the other types of MIDI instruments out there in regular use (pianos, accordions, etc.) are based on simple on/off switches. When you press a key, it closes a circuit and the processor is stimulated to spew out a particular MIDI message. Or, they are connected to variable resistors (pots), so that when the resistance is changed, the message is changed.
To review, there are two options with vibrating strings. 1) You can measure the frequency of vibration that is detected, and this is matched up with the correct MIDI note, this is what I call the pitch recognition approach. This is how guitars, violins, and the Camac MIDI harp work. 2) My system does not involve any pitch recognition. Since harp strings never produce more than one pitch (unless a lever or pedal is moved), the presence or absence of electrical activity is enough. The processor looks up on a table what message to send out in response to this event. This is why I am so focused on isolating the vibrations of each string, so the processor doesn’t get false stimulation.
I suggest that you might end up using a combination of the two approaches. Do some experimenting with piezo elements, to get the best separation of notes that you can. Then, perhaps using a pitch recognition software would be very effective. But the key, as I see it, is to get the best input possible, as discrete or separated as possible. You could then wire them into a common buss for simplifying the wiring, but having clear unambiguous notes would be good.
Re: piezo sources- I get piezo film from another company, not Measurement Specialties but similar, and cut it into small bits. The film has a copper plating on both sides, and I solder tiny wires onto each side. This is a tricky time consuming thing to do, but I have not found a ready made item off the shelf. To add to the complexity, the entire assembly has to be shielded with a separate metallic envelope that is connected to ground. Folding the item over so that one side acts as the shielding ground doesn’t work. Bear in mind that I am trying to produce transducers that will provide high quality analog information for amplification. Probably, for MIDI triggering purposes the shielding is not as critical, but better is always better.
One other point you need to understand is about matrixes. This is a way of connecting multiple sources to a limited number of inputs. So, controller boards with only four inputs will in fact support a large number of pickups, by using a matrix to connect them. (Don’t ask me how, I only know that it works. Insert the disclaimer about not understanding electronics here.)
I hope this is helpful. I went on longer than intended because my office is the only place that is air conditioned. Minnesotans don’t handle hot sticky weather very well.
James – I agree with you about the priorities. I get a lot more pleasure doing the acoustic instruments as well. I think the reason I pursue midi and things tangential is to keep my mind agile and to challenge myself. From what you describe about the interest in midi harps, I have a feeling that it may be even worse for hammered dulcimers. Most people playing are very tradition bound. There are very few adventurous souls out there so doing this. I’m doing this more out of curiosity then economics. I’m not willing to invest a whole lot of money in this project since it probably won’t be returned. Thanks for all your suggestions so far. They have informed my searches. I did discover a controller which is still a DIY type but with a lot of very good instructions which seems to have a lot of flexibility. I may just buy the small version (Brain Jr) and do some experimentation. You can see them here under the DIY section. The instructions for everything they offer is on a wiki. Here is the link for Brain V2 which sells for $189. This one has the most capability.
Is this the matrix you are referring to http://www.ntonyx.com/mm10.htm or is it a hardware matrix? I’ll let you know how things progress.
David – The Brain Jr. sounds like it would do pretty much what my controller boards do. It almost makes me want to buy one and see. Actually, they are marketing exactly where we projected our board would go eventually, and the cost point would be about the same.
The MIDI matrix product is a specific example of a matrix in use. In this case, they are connecting 4 MIDI streams into one output via a matrix arrangement. As I said before, I don’t know how these things work, I only know what does and does not work.
Let me know if you get one of those Brain things, and if it pans out. Dave K.
Range of instrument
String Spacing (single string)
Vibrating Length (VL)
Controls on instrument
Volume (could be a pedal or pedals with all controls located there including pitchbend and modulation wheels, volume, octave up, octave down, the ability to change instruments while you play)
Livid Brain V2
I actually purchased the Livid Brain V2, got a variety of sensors and begin trying to get them to work with that microprocessor. I did not have much success. My goals were to create a midi-hammered dulcimer which had the configuration, layout and feel of a hammered dulcimer. As Randy demonstrated, you can certainly create a dulcimer like instrument with piezo or fsr pads but it will not have the same feel as a hammer rebounding from a taut length of wire. What I so far have been unable to find is a sensor that will translate a vibrating string into midi information. David Kortier gave me some of his sensors but I did not succeed in getting the Brain V2 to recognize them. I did spend some time designing what I thought might work in terms of physical layout. I envisioned a compact portable instrument that would be in two halves folded together for carrying.
Pitch Recognition Midi Solutions