I don't think Somogyi was the first to use a dummy side set to clamp a top onto, but I can't say for sure. That's still not the same a having the top on a guitar, since the back also affects things. Also, in my experience, temporary mountings are never quite a stiff as permanent ones.
I've been using Chladni patterns, the 'tech' version of tap tuning, for decades. It gives the maker a lot more information to work with than tap tones, although it takes longer. It's easy to record what you've done, and easy to communicate to others. The equipment needed has become far easier and cheaper to get than it was forty years ago: now you can use a free app on a tablet or 'phone to generate the tones, and amplifiers and speakers are easy to come by. I've tracked Dana through a workshop that the local luthiers set up, and we end up looking at the same things, although, as I say, the patterns show you more.
The problem with relating 'free' (off the guitar) plate information to the sound of the final instrument is a difficult one. Basically, it's possible to make two guitars using the same wood and get them to sound 'arbitrarily close' using Chladni patterns, but they'll probably never be exactly the same. If you change anything, no matter how minor, they will probably diverge more. The guitar is extremely complex in the way it works, and there are lots of variables. By the time you get to about 700 Hz or so you're in a 'resonance continuum', where there are so many interacting resonances, and they overlap in pitch so much, that there is, apparently even in theory, no way to predict exactly what the guitar will do above that pitch. Fortunately, the low range, up to the 'continuum' seems to establish the 'character' of the sound: you match that by getting the lowest resonances of the top, back, and air to match the desired pitches and strengths, and there can be several ways to achieve similar ends. Sound 'quality' seems to depend more on the high end resonances, and what seems to matter most there is how many of them there are to an octave, and how low the loss factors are: you don't necessarily need to have them at the same pitches to get similar quality. People seem to pick out 'different' guitars on the basis of those high end pitches. That's where post assembly fine tuning comes in.
I really try to avoid that, for two reasons. First, I find I can usually get 'close enough' to the desired sound (the right 'character' and good 'quality') without it, and, second, it reduces the ability to use the 'free' plate data to tune the sound. In the 'pair' experiments I've done I have not been able to make two that sound 'alike', but they're always very similar, and the differences are not such that players or listeners have had a preference for either one: they're just 'different'. I find it so much easier to trim braces before the box is closed that I'd rather develop that approach to get the sound I want. It does seem to be working.
I'll note that the factories are not trying to match any particular sound, and, IMO, probably would not want to if they could. So long as the instruments bear a 'family resemblance' (Martins don't sound like Taylors) and the quality of sound is reasonably good they'd rather offer some variety. They can do that by sticking with proven designs and reasonable quality control, and rely on the natural variation in wood properties to provide a range of sounds. Carving all the braces for a particular model so that they look about the same works to that extent. Everything they make will be somebody's 'Holy Grail', and they only have to find that person. A custom maker has to know how to home in on a particular character of sound, which is roughly like nailing jelly to the wall...
|
) can post a pic of their setup. It's sort of like a mold but lets you trap the top plate and the mold provides the mass and so it's no longer a free plate. The braces can be carved as if the box is closed. I'm not sure who first came up with this, but I think they teach this method at Galloup.