Automation--could it deceive experts?

[bean_fidhleir]

There are a lot of subtle, computer-guided tools available today. If someone wanted to do the programming, in principle they could absolutely copy any existing instrument they could lay hands on. They could also, again at least in principle, create additional instruments that would fit in with the maker's existing oeuvre.

Could experts detect the fakery without using sophisticated chemical, dendro, or radiological tests?

If so, what would give the fakes away?

As devices for making music, how would they probably compare with hand-mades by the average successful maker, and why?

[Craig Tucker]

It can't be done.

Someone has misinformed you.

[bean_fidhleir]

Why not? (I should perhaps have qualified my 'absolutely' by disclaiming being able to exactly copy the wood, but I hope that's not your reason)

[Craig Tucker]

Hmmm, perhaps I'm wrong about this...

Why don't you go ahead and do it, then report back to us?

[MANFIO]

Oh yes, I'm interested too, I would love having a machine made Andrea Guarneri viola (cheap, of course, after all it would be a machine made instrument...). I would love also a DG Cannon copy, as well as the Cremonese by Strad.

And... ... just for a change, a copy of a Manfio violin, if it works I'll comission about 100 of them in my first order. I presume that a different machine will varnish the instrument.

[violinmark]

I am no expert (for sure) But I think that probably not.

The shape is just the beginning.

The subtleties of surface finish and tool marks, wood used, finish, assembly, ageing, and the skill derived from making 100's of violins that goes into deciding how a particular piece of wood top or back should be thicknessed yielding of course --- the sound---

Kinda like looking at somebody elses relatives. They all look and sound different and yet somehow you can tell they are part of the same family.

Although I do think that we would all love to have a very accurate plastic or plaster strad "casting" that we could copy and use as our model (instead of posters)

Maybe that is why a handmade Lamborghini(sportscar) can sell for $300k vs. a factory/machine made Corvette at $60k

[Regis]

It really depends on the skill of the "experts" involved.

I think that it has already been done by machine...$29.95 on e-bay including case and bow. The qualification for those "experts" though is to have spare $29.95 + shipping/handling.

In the past I have written and AI system that we could train to be the expert that validates the computer generated violin. Assembly????? let me see, perhaps we could bypass assembly and run that cnc arm through one of the F holes manking the violin out of 1 piece of wood. Then we could simply make really authenic looking fake glue joints.

Noooo, I think Jeffry would spot that pretty quick to,,,

Well, as an after thought, if we kept the violin in the computer, we could do it. We could generate images on the screen and tones that would even pass the Jeffery and Michael inspections. And, they would be inexpensive to reproduce.

[MANFIO]

Hi Regis, I think I would spot the Manfio copy on the spot... And that with the visual aspect, the sound would be much, much more difficult to replicate, after all even my own violins differ a litle from each other.

[Andres Sender]

The reason why this can’t be done now and probably won’t be able to be done for a long while is that what makes instruments distinctive is a whole field of traits which result from a wide range of tools and ways of using them. You can program a computer to do some pretty sophisticated things, but to get the same tool marks etc. etc. you’d have to replicate a human arm and have it use hand tools. Even if it’s possible the expense would be vast. I doubt it’s possible right now–humans use their tools with constantly varied force and direction based on feedback both tactile and visual, last I heard robot technology hasn’t matched that.

[MANFIO]

Yes Andres, that's it. We have machines making industrialized food but there is no machine making high cuisine food (unfortunatly...). The same happens with violins.

[bean_fidhleir]

I'm a bit dismayed to see the dismissive responses. Were they perhaps meant to teach me not to pose thought-experiment questions to my betters?

[MANFIO]

Hi bean! Sorry for my skepticism... we are sometimes toooooo objective here. You are wellcome and you do can ask whatever you want. I beg your pardon if I had offended you.

[Regis]

Please don't take this jousting as dismissive of ideas. I understand the computer side of your question (almost too well). Asking the same question on advanced technology forum would have gotten very different answers. Mostly because they could formulate the computer-machine motions necessary to rough out the task. But, the same folks would not understand the details that a violin expert would look for to differentiate violins.

"Yes Andres, that's it. We have machines making industrialized food but there is no machine making high cuisine food (unfortunatly...). "

That was the best analogy.

Regis

[Brad Dorsey]

As a result of a recent trip to my dentist, I'm not ready to be as dismissive as some of the other replies you've had. He prepared a cavity in one of my teeth and took a digital photograph of it. I then watched as a tiny computer-controlled milling machine cut a piece of porcelain so precisely that it fit inside my tooth! I see no reason why something similar, but on a slightly larger scale, could not exactly reproduce any violin down to the smallest tool mark.

However, I don't think we have to worry about this process producing counterfeit intruments for several reasons: 1. It doesn't do varnish. Furthermore, if you're trying to copy a varnished instrument, this process would reproduce the thickness of the varnish onto the new wooden parts. 2. If the original instrument has the reeded grain texture typical of many spruce tops, the texture would be reproduced exactly onto the copy. But it wouldn't line up with the wood grain of the new top.

[HongDa]

I think Yamaha or Suzuki use CNC machines over in China. I guess this is so their product is more or less consistant in style and sound.

A good CNC machine capable of doing plates would probably run around 30,000 dollars but also has a limited life use.

Once you paid the machine, the operator,necessary hand finishing and materials cost etc...in the west I bet you'd be looking at around 500 dollars to finish one instrument.

Or you could buy the same thing in China for 100 dollars.

An engineer friend of mine once tried to convince some Chinese factories to buy a manual copy router he had designed.

They just kind of scratched their heads and said..."But human workers are cheaper...".

I guess someday when the price of labor goes up in China, then the violin factories will just move to another area of the world.

[HongDa]

Here's a website of a guy who is doing it and selling his instruments for 2,500. You could pay that price or pay 250 for a decent enough Chinese handmade one.

web page

[Craig Tucker]

"I'm a bit dismayed to see the dismissive responses. Were they perhaps meant to teach me not to pose thought-experiment questions to my betters?"

Only if you can present an argument and then only accept one answer to it. Especially if the only answer you will accept is the incorrect answer.

You asked a straight forward question and got a straight forward answer. When you ask a question like this and get the answer you most probably expected, why act dismayed? If you think the answer I gave is incorrect then you are free to attempt to prove it. But you asked a question and got a swift accurate response. What more do you want?

Be forewarned that there are actual active living violin makers here who will answer your questions based on years of experience in the field.

If you pose “thought experiment” questions you might have, based on no or little actual experience, then why act disappointed or dismayed that the answer isn’t what you may have hoped for?

[Dean_Lapinel]

I think the answer would be more clear to you if you have made violins. Not to offend but I have to assume this question comes from a lack of the "experience".

Once you make a violin you realize that there is nothing that can (adequately) automate the process that goes into working individual characteristics of the tonewoods.

If I weren't so lazy right now, I would search the archives for a comment Michael Darnton made about his approach to arching. Look for that comment and you will see there is much more human thought that goes into making than can be done by machine.

[Andres Sender]

Quote:

---

I'm a bit dismayed to see the dismissive responses. Were they perhaps meant to teach me not to pose thought-experiment questions to my betters?

---

No, but it sounds like your 'experiment' wasn't exactly motivated by the spirit of pure exploration, since you're so dissappointed in our honest answers that you feel dismayed and have to label those answers as 'dismissive'.

The term 'dismissive' is a better fit when applied to a post which brushes aside a series of answers by giving them a negative label and not addressing their content.

[Claire Curtis]

Changing to bows for a moment -- The big difference between the French and German styles of bowmaking is their approach to the material. The German style (as I understand it) takes a good piece of wood and cuts it to spec very precisely. The French method takes a stick, which may or may not have imperfections, and shapes it, constantly flexing it and adjusting the shaping as needed. This can results in a bow that deviates significantly from 'spec'. But French bows are renowned for their playing characteristics; German ones are not. (n.b. exceptions in both directions abound -- I'm using stereotypes here.)

This is not to say that bows made to a set of measurements can't be good. They can, especially if you change material to something that does not have the individual differences that wood has. Carbon fiber bows are pretty good, and are happily mass-produced. But the question was whether the products of automation could deceive experts, and that question isn't relevant if you are changing the material.

Similarly, you can take a piece of tonewood and using a CNC router cut it precisely to the dimensions of your model, be it a Strad or whatever. You can graduate it to match. But the original was built with the same sort of interplay and continual adjustment to the characteristics of this particular piece of wood. The copy will not be. The shape of the copy will be optimum for a different piece of wood.

Another analogy -- You are a tailor making suits. You have a suit that fits "Adam" to perfection. So you copy it absolutely precisely for "Bob". But it is highly unlikely that the copy fits "Bob", even if the two seem to be pretty much the same size. Each person requires personal tailoring. (If you're having trouble following my analogy, the person here represents the wood, not a player. Sometimes I get a bit obscure.).

--Claire

[fiddlecollector]

```Similarly, you can take a piece of tonewood and using a CNC router cut it precisely to the dimensions of your model, be it a Strad or whatever. You can graduate it to match. But the original was built with the same sort of interplay and continual adjustment to the characteristics of this particular piece of wood. The copy will not be. The shape of the copy will be optimum for a different piece of wood. ```

Which explains why you can get the odd factory violin or bow that is good.Probably can be explained in terms of a percentage.Sort of like reaching a goal but having no idea how you got there,which is a pointless way to work,of course.

[Jimbow]

Sort of like reaching a goal but having no idea how you got there,which is a pointless way to work,of course.

fiddlecollector

heh, heh! I like that, fiddlecollector.

A classic !

Jimbow

[Jeffrey Holmes]

Quote:

---

There are a lot of subtle, computer-guided tools available today. If someone wanted to do the programming, in principle they could absolutely copy any existing instrument they could lay hands on. They could also, again at least in principle, create additional instruments that would fit in with the maker's existing oeuvre.

Could experts detect the fakery without using sophisticated chemical, dendro, or radiological tests?

If so, what would give the fakes away?

As devices for making music, how would they probably compare with hand-mades by the average successful maker, and why?

---

Hi Bean;

Your question brings to mind the old sci-fi idea of an android passing for a human. Interesting, but not possible... at least with the technology that exists... and then there’s that darn “soul” thing.

I know a well know bow maker in Europe who produced “exact copies” of his Gennaro Gagliano violin in the way you described. Impressive setup. Really high-tech.

The instruments were “copies”, in terms of form, including distortions of the arch, dents, idiosyncrasies in the purfling inlay and head, wear to the edges (although there wasn’t much on this particular original). They didn’t look at all like the original in any other aspect, however. Even on some where the wood selection was close, the varnish was off, patina forced, wood looked too new (despite attempts top hide this). The graduations were a match to .05 mm, but I can’t say the “copies” sounded well, either. They lacked a certain soul...

There have been copyists in the pst that have worked without sophisticated machinery and come much, much closer to being convincing.

Rather than “can we fool the experts”, I think a much more fruitful, or at least practical, question might be: With the existing technology, how decent a fiddle can be produced mechanically?

[MANFIO]

Hi Jeffrey! By the way, some makers that don't do exact copies happen to make fantastic violins that are "inspired" by the original violin. These violins sometimes are much more interesting that "exact" copies. Rocca is a name that comes to my mind in this sense.

[Jeffrey Holmes]

Hi Manfio;

I agree!! My response was to Bean was in relation to "copies" made by machine.