Jump to content

Some musings on measuring bellows activity in electronic Anglo concertinas


Recommended Posts

this goes back to the 80s where I vaguely remember that electronic pianos used to determine whether a key was played piano or forte by sampling the time between the key in rest position and the key fully pressed.

 

Sorry for being unclear - imagine (just for approximation, of course, the details as usual will be much gorier) the imaginary bellows sliding on a ground plane that ends at the point where the bellows are half way open. On every outsice bellows fold there is a contact point which is pulled up to high voltage but reads low when in contact with the ground plane. So with the bellows fully closed, all contacts read low potential.

 

Now as the bellows are being pulled apart, successively more and more of the contacs will detach from the ground plane, reporting high due to the pull up resistors as soon as they leave the ground plane.

 

If sampled fast enough, the readings could give you a time profile of how long it took to pull them apart, which should roughly (again, the detail work is the pudding) allow you to compute how strong a pressure was exercised on the bellows.

 

Does that make sense?

 

I may be way off with that idea, apologies again if I am.

Link to comment
Share on other sites

 

2 hours ago, RAc said:

Does that make sense?

 

 

5 hours ago, RAc said:

Apologies if this should be a stupid remark (it probably is).

 

Best,

Tom

Link to comment
Share on other sites

RAc: I think that you're proposing to use the bellows movement speed in an electronic concertina as a model of the air flow through a real concertina reed.

 

That would work fine... apart from a few problems(!):

 

1. You'd need to emulate/model other aspects too - for example, you'd need to have valves that would open up (presumably via servos etc) as you press the buttons as otherwise there would never be any bellows movement anyway. That would be tricky - as the system would have to be extremely fast in order to avoid latency.

 

2. OR if you just had permanently leaky bellows, that resulted in movement proportional to the applied bellows force, then the more leaky they are (and thus the easier the distance measurement), then the less like a real instrument it would feel.

 

3. In either case, how do you measure the bellows opening "distance" when they can bend (ends not staying parallel)? And if you stopped them bending, it wouldn't feel right.

 

There is a way to measure the speed at which the bellows open/close - and that is the flow rate through a valve. That would probably be more effective than any mechanical measurement as it would handle the bending... but you'd still have the first two problems above:

 

1. A complex system of valves (one or many) linked to the buttons, with a flow rate sensor. All the servos and the flow rate sensor need to be really fast in order to get a low latency (a few ms) - so as you press a button, the microprocessor needs to control the servo to open the valve. It also starts playing the note, but at this point the flow rate is zero, so it's zero volume. The flow sensor will then start to measure the airflow, and report it back to the processor, which increases the volume. And all that needs to happen within a few ms, or the instrument will feel laggy.

 

2. A leaky-bellows system, where the flow rate is just a proxy for pressure. Now the microprocessor can measure the applied force is happening even before the button is pressed, so the latency problems are reduced... unless you're reversing the bellows direction (relevant for Anglos). But it won't feel like a real concertina...

 

So I think method 1 is theoretically sound (maybe even superior), but impractical. And method 2 is effectively just measuring the pressure using a flow rate sensor as a proxy. I think the result is that you'd be better off just measuring the pressure or force directly!

 

 

 

 

Link to comment
Share on other sites

10 hours ago, ttonon said:

 

 

 

 

Best,

Tom

Did you mean to add anything to the quote, or is that a subtle hint that it does not make sense AND is stupid? If so, I can take it... 😁

Link to comment
Share on other sites

Posted (edited)
9 hours ago, RatFace said:

RAc: I think that you're proposing to use the bellows movement speed in an electronic concertina as a model of the air flow through a real concertina reed.

 

That would work fine... apart from a few problems(!):

 

1. You'd need to emulate/model other aspects too - for example, you'd need to have valves that would open up (presumably via servos etc) as you press the buttons as otherwise there would never be any bellows movement anyway. That would be tricky - as the system would have to be extremely fast in order to avoid latency.

 

2. OR if you just had permanently leaky bellows, that resulted in movement proportional to the applied bellows force, then the more leaky they are (and thus the easier the distance measurement), then the less like a real instrument it would feel.

 

3. In either case, how do you measure the bellows opening "distance" when they can bend (ends not staying parallel)? And if you stopped them bending, it wouldn't feel right.

 

There is a way to measure the speed at which the bellows open/close - and that is the flow rate through a valve. That would probably be more effective than any mechanical measurement as it would handle the bending... but you'd still have the first two problems above:

 

1. A complex system of valves (one or many) linked to the buttons, with a flow rate sensor. All the servos and the flow rate sensor need to be really fast in order to get a low latency (a few ms) - so as you press a button, the microprocessor needs to control the servo to open the valve. It also starts playing the note, but at this point the flow rate is zero, so it's zero volume. The flow sensor will then start to measure the airflow, and report it back to the processor, which increases the volume. And all that needs to happen within a few ms, or the instrument will feel laggy.

 

2. A leaky-bellows system, where the flow rate is just a proxy for pressure. Now the microprocessor can measure the applied force is happening even before the button is pressed, so the latency problems are reduced... unless you're reversing the bellows direction (relevant for Anglos). But it won't feel like a real concertina...

 

So I think method 1 is theoretically sound (maybe even superior), but impractical. And method 2 is effectively just measuring the pressure using a flow rate sensor as a proxy. I think the result is that you'd be better off just measuring the pressure or force directly!

 

 

 

 

Thanks for taking the time, Danny!

 

I am afraid we are not in sync, though... my ideas are not related to reeds or buttons at all, sorry for being unclear. That part, I believe, is taken care of by existing mechanics... my idea was to take a different approach to digital bellows pressure approximation, so (staying in the spirit of this thread) brainstorming about possibilities to digitally approaching bellows dynamics.

 

Of course you are right that a fixed rod wouldn't be practical, so my next idea (again, rough sketch with a lot of refinement to be done) would be two rubber tubes fitting into each other as frictionless as possible (of course, there must also be some mechanical way to prevent the tubes from being pulled apart at the end). One would be attached to the left hand pane, the other one to the right. In the inner tube there would be a magnet at a fixed position, and into the outer tube there would be reed* switches or hall sensors worked in in a spiraling pattern such that the distances between the sensors would be as small as possible.

 

As the bellows are being worked, the magnet would trigger the switches in a time pattern from which (I believe) the force applied to the virtual bellows can be deducted, which the MCU could consult with to determine which MIDI volume level to apply to a button being pressed at that time.

 

If there were several of those tubes in effect, even tilting bellows could be detected.

 

All of that is purely mechanical, there is no air flow, leaks or anything like that involved. In fact, that construction would also work in a vacuum, but admittedly, playing concertina in a vacuum does not sound like a whole lot of fun in the first place... 

 

I hope this makes it clearer. But again, I am doing ok as a software designer, but a lot of my mechanical engineering projects look much better on paper than in practice, so apologies for possibly generating useless noise. As I pointed out earlied, I myself am rather hesitant to spend a lot of time on musing about the best way to approximate bellows movement digitally to translate that to an "authentic" concertina sound (I prefer to spend the time practicing my real concertina). But brainstorming ideas to make that happen is right up my alley.

 

 

*for all those reading who are not familar with electronics: Unfortunately, "reed switches" have the same name as reeds, but they are something completly different, so I try to avoid the technical term to minimize confusion. A reed switch is an electronic component that opens or closes when a magnet comes into proximity.

Edited by RAc
Link to comment
Share on other sites

Posted (edited)

I too have a long history in software design.

 

Please note that the following has no relevance to real concertinas only electronic ones.

 

I apologise for the lack of diagrams, pure laziness of my part.

 

My basic idea is for a force based model. So I see the concertina as a pneumatic system. Force is applied to the bellows by the player pushing or pulling. I am guessing that I can ignore pneumatic impedance in the system. When no buttons are pressed that force is balanced by raising the temperature of the air inside the bellows. So the law of conservation of energy is obeyed. As soon as a button is pressed the force is balanced by the production of a sound wave. The magnitude of the sound wave is related to the applied force. If another button is pressed then half the force goes to button one and half to button two. If there was no reed behind button two then the listener would experience a reduction volume of the note from button one, and so on.

 

Back in the real world we are measuring the applied force with some kind of sensor and deriving a number by analogue to digital conversion. This number is a representation of the instantaneous force and unit less. We need to convert this into a value that we can give as a volume parameter to the hardware producing the sound. This parameter is still unit less.

 

So if we call the gadgetry producing the sound of a particular reed a voice.

Let V = the volume of the voice.

Let B = the number of currently pressed buttons.

Let F = the current applied force.

 

V = F * (some sort of calculation) / B

 

Does this make sense?

Is my physics correct?

Any other comments?

Edited by wearyhacker
Typos
Link to comment
Share on other sites

I believe we must be careful not to mix finger pressure and bellows pressure. On the concertina, it is possible to hit a single button with a rather strong, crisp motion while at the same time exercising little pressure on the bellows (silent staccato effect). I am tempted to believe that apt players can even hit two buttons at the same time with different strength.

 

From what Jody wrote earlier, I was under the impression that the idea behind your measurement is (alternatively to an air pressure sensor) to determine how much pressure would be in a physical bellows if played with the same force so to best approximate a "natural" concertina physics (or in programming terms, how to determine which MIDI volume level to assign a pressed button ). The idea that came to my mind was to propose an even different approach to that issue.

 

However, if the idea is rather to sample how hard a particular button is hit, then this is a completly different discussion, methinks.

 

In fact, thinking about it, the "perfect" digital concertina would need to constantly sample many more parameters than we care to think about; for example, the attack on every single button (not only on/off) as well as the status of the virtual (or physical depending on how the instrument was built) bellows in more than one place.   

Link to comment
Share on other sites

On 5/20/2024 at 11:08 AM, RAc said:

I believe we must be careful not to mix finger pressure and bellows pressure.

 

Yes - though if one wanted to measure attack (velocity), the Melodicade MX project has a nice setup using pairs of switches and timing (but that is again different to after-touch pressure)

 

 

On 5/20/2024 at 11:08 AM, RAc said:

In fact, thinking about it, the "perfect" digital concertina would need to constantly sample many more parameters than we care to think about; 

 

I think it was here on c.net, in another thread, that somebody wisely commented that every thing that can be controlled must be controlled. The point being that, beyond a certain point, the problems of having too much to control outweigh the advantages.

 

In most instruments there is a trade-off between complexity and expression. For example, the way that the piano plays a note according to the strike velocity, and then the note decays away exponentially with the only remaining control being a simple "on/off", is far more "liberating" than it is limiting.

 

 

 

 

 

Link to comment
Share on other sites

7 minutes ago, RatFace said:

every thing that can be controlled must be controlled.

Indeed, it seems that's where our society is headed.


“The best way to control the opposition is to lead it ourselves.” – Vladimir Lenin
 

I wonder if the societal result stems from this technical ability/conviction. 

 

Regards,

Tom 

Link to comment
Share on other sites

3 minutes ago, ttonon said:

Indeed, it seems that's where our society is headed.

 

Though the natural conclusion is that one should try to avoid situations where control is necessary!

 

It occurred to me some time ago (about the time my daughter turned from a child to teenager!), that with children we have a huge amount of influence and almost no control (not in the long term, anyway). Perhaps the same should be true for musical instruments!

  • Like 2
Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...