You might find this produced a positive result. However, a negative result wouldn't necessarily disprove the theory because violins may appear simple, but they are acoustically very complex. Take for example the Chladni patterns for violin plates
and even the strings can be made from different composite materials for different strings. The strings alone can make a significant difference to the sound of a violin. Has your one microphone picked up all the sounds in all the frequencies and directions and has your mechanism tried all the different bowing techniques for all the notes etc.
An experiment to prove that a particular listener can detect the sleepiness of a violin:
Buy a couple of violins. Have several violinists play them until they're warm, and record multiple notes per violinist per violin. Then store half the violins away for a year, while the others are played daily. After that, record again a couple of notes per violinist per violin.
Have a listener listen to pairs of recordings, where the two recordings in the pair were recorded by the same violinst on the same instrument, but a year apart. Ask the listener for each pair if the instrument sounds like it's fallen asleep between the two recordings.
The null hypothesis is that the listener will get the sleepiness property right on 50% of the pairs.
Wouldn't it be both simpler and more accurate to simply have the control group of played and unplayed violins, and then hand them to a number of highly skilled violin players (who haven't been playing these instruments) and see if they can identify which have been played and which haven't by playing them? (If you're worried there might be tells on the instruments themselves, do the same experiment and see if the violin players can tell the difference without playing them.)
Doesn't work for me. You have assumed no loss of sound in the recording and playback process. There is a great deal of difference between a live instrument and a recording.
What's needed is a mechanistic ability to distinguish the important sounds. This might be a computer analysing a microphone signal and saying 'this is the same sound'.
The funny thing is that I think that 99%+ of the listeners will not even be able to tell apart two violins, let alone whether or not one has not been played for a while.
I was composing this while limmeau replied, so some overlap.
We would initially need to calibrate the mechanism and microphones against expert violin players to ensure they pick up all the same nuances of sound. This is because experts hear more than non-experts.
"Two groups of expert instrumentalists (violinists and flutists) listened to matched musical excerpts played on the two instruments (J.S. Bach Partitas for solo violin and flute) while their cerebral hemodynamic responses were measured using fMRI. ... We found an extensive cerebral network of expertise, which implicates increased sensitivity to musical syntax (BA 44), timbre (auditory association cortex), and sound-motor interactions (precentral gyrus) when listening to music played on the instrument of expertise ..."
http://cat.inist.fr/?aModele=afficheN&cpsidt=20951540
So if the experts could demonstrate any statistically significant ability to hear certain aspects of the sound, such as being able to distinguish between different instruments, this would have to be replicated by the mechanism and microphone technique. We would have to try to test a fair range of aspects of the sound.
We would then need a reasonably large sample of similar violins, of a quality for which we might expect the effect to occur, half of which would be put away and the other half played for our test period.
The mechanism could then be rerun and get a fairly good result.
Q: What is the difference between a viola and a violin ?
A: The viola burns longer...
If Hermann Helmholtz could be wrong about some of this stuff then I have no doubt that people are capable of deluding themselves in to having the ability to hear things that aren't there.
From "The Acoustical Foundations of Music" by John Backus, page 207:
For example, the excellent quality of Stradivarius violins has been attributed to a varnish of almost magical properties used by the old instrument makers, whose secret has been lost. However, resonance curves made on instruments in the varnished and unvarnished condition show that as normally used, the varnish itself has rather little effect on the tone.
What effect it has is more likely than not to make the instrument worse.
"people are capable of deluding themselves in to having the ability to hear things that aren't there". I agree totally. I was assuming a mechanistic way of testing for sound similarity attached to the microphone that would give an objective result. This mechanism needs to be calibrated against human experts.
I know, but even though it has been disproven ages ago this stuff gets repeated ad nauseum.
Being able to hear sounds that can't be verified by instruments is not very scientific imo.
That said, there is a lot of stuff going on outside of the range of 'normal' recording and playback equipment, which is why a live performance of a church organ (for instance) is on a completely different level than the best recording of the same.
Violins however fall nicely in the range of modern recording equipment and if you can't record a sound from a violin it probably isn't there.
http://www.phys.unsw.edu.au/jw/chladni.html
and remember that this is just one aspect of one component. Other people study directional tone colour
http://www.josephcurtinstudios.com/news/strad/apr00/Gabi_str...
and even the strings can be made from different composite materials for different strings. The strings alone can make a significant difference to the sound of a violin. Has your one microphone picked up all the sounds in all the frequencies and directions and has your mechanism tried all the different bowing techniques for all the notes etc.