Official Luthiers Forum!

...and voicing to match the top? I've seen mention of such things but cannot wrap my head around how this might be done. Not going to use a sound spectrogometer scopey anything probably ever so can it be done by hand and ear alone? Is there a technique to 'match' it with the top voicing? Simply be feel and touch?

Happy New Year!!

_________________
Thank You and Best To All

There are resonances, 'tap tones', in the back that can be heard and 'tuned', to some extent. These do affect the sound of the guitar. There are limits to all of this, but it's certainly useful to understand. In many cases, for example, the best way to enhance the bass register of a guitar without losing the treble is to shave back braces. Not only is this effective, it's also structurally much less risky than shaving top braces.

This does get technical: at least, it helps to have some understanding of the underlying physics if you're going to work with it. Once you 'get' that you may not need much in the way of technology to implement it. You do need to be aware that what you hear is not always what's there. It can take a while to learn to sort out what you're hearing, and technology can supply shortcuts in that respect.

One of the dichotomies you'll run into is the discussion about 'active' vs 'reflective' backs. This is something of a red herring: the only guitar backs I've seen that are almost totally 'reflective' are on Ovations. If you think of it, the only part of the guitar that is being directly driven by the strings and can produce sound effectively is the top, so on the one hand any energy that gets into the back is being 'stolen' from the top. You would thus expect back vibration to reduce the output of the guitar. This is, for the most part, true. However, as far as I can tell, back vibration can enhance the output in the lowest 'bass reflex' range, roughly below the pitch of the open G string. If the 'main back' tap tone is close in pitch to the 'main top' resonance, this can produce more power in the lowest range. This is not trivial: some researchers maintain that most of the actual power output of the guitar is in this range, so enhancing it can help. There are risks here: getting the 'main back' resonance too low in pitch can cause several sorts of 'wolf' notes. This is where learning to sort out what's actually there, as opposed to what it 'sounds like' is important.

Higher pitched back resonances can also 'color' the sound in useful ways, even when they don't add to the power of the instrument. The object there is not to eliminate them, but to control them. Back vibrations are part of what produces the complexity of the sound of a good guitar.

I don't know of much literature that covers this explicitly. Part of the issue is simply that different people have different ideas about what goes into making 'good' tone, and so work toward different goals. There is also the fact that a lack of real objective understanding leaves things open to interpretation. Many of the folks who use these methods and work 'by feel' do similar things, but describe them so much differently that it's hard to sort it out. Even when you know what's happening it can be confusing. But it can be done.

Larry, get yourself the Gore/Gilet books. That question is more or less answered along with the next 50 you're going to come up with.

Choosing to remain mystified rather than informed is imo the wrong choice.

Thanks Alan, seems a bit mysterious but understandable on some level.

Happy New Year

_________________
Thank You and Best To All

Good advice I'm sure and was hoping my question would help me to be a bit more informed...?

Wishing I only had 50 more questions but I'll keep exploring..

Happy New Year

_________________
Thank You and Best To All

Fair comment, just saying that there's a wealth of info readily available for reference in those books, and if you have a modern cellphone, you likely have a spectrogometer scopey something already.

'Cause I'm sure that while it can be done by hand and ear alone, 'why even', when the important aspects to use that technique can be so easily quantified and manipulated? Everything you need to learn to do it is in the aforementioned books, at least as I understand it.

Ed, Very helpful in every regard and I will not ignore your advice. I don't see myself as a builder who will use the technological stuff to build with but that may change with time and more guitars and the desire to better understand such things.

I just carved some back braces with 'voicing' in mind at least, instead of simply shaping the braces till they looked 'good'. I had never tapped a back before but was fascinated by the change that carving made similar to a top. I was not shooting for a particular note or tone but found a pretty sweet, fairly musical tap and stopped there. By the time I had shaped and sanded I felt I went a bit far but that's only a guess and will build as is.

Curious, do you tap/voice/tune the backs of your guitars? With a "spectrogometer scopey something"? By sound and feel? Both?

_________________
Thank You and Best To All

Yes I do, with a scope thing, with great precision. Not by sound and feel. The sound and feel is the result of getting precisely scopey on it.

Would a tuner be too high tech, one that might double as the one you tune the strings on your guitars? I was using strobosoft in my video use tap to thickness the plates and it gave me the data I was looking for. A scale for weighing your tops (you can even use a balance beam). A deflection test can be as simple as a known weight and a ruler. Sound and feel works for a lot of builders; maybe after 25 to 30 guitars you may have an ear and the feel to start to shape the sound. It would sure help to calibrate your senses and ear with information you can get with low tech tools and books. In the mean time enjoy the process.

_________________
http://www.Harvestmoonguitars.com

VERY interesting and helpful. Perhaps I'm knocking on the door of my next steps in guitar building? I'm on guitar 15, but quit for 5 years and am literally making every mistake over again, and again, and again, as I try and re-sharpen my skills. Maybe after completing a couple and getting my confidence back I'll start exploring the tech side of sound design.

Again thanks very much and have a Happy New Year.

_________________
Thank You and Best To All

I use a 'tech' version of tap tone tuning on both the top and back of my guitars, and that's what I teach my students. Tap tone tuning, as, say, Dana Bourgeois does it, is simply a method of finding the pitches of the resonant modes of the top and back. He uses a good ear and many years of experience to sort out the frequencies and figure out where to shave braces to get things to work the way he wants. The method I use, the 'Chladni pattern' technique, uses a signal generator and speaker to drive the plates, which enables you to both find the exact pitches and visualize how they are vibrating. It still requires experience to interpret what you're seeing and work with it, but it's more objective in many ways; it's easier to record and transmit the information, and to teach. You're also working with a bit more information, which never hurts.

This sort of 'free' plate tuning does NOT tell you what the pitches of the assembled modes of the guitar will be with any specificity. You're not 'tuning' the plates to particular notes but rather 'tuning them up' so that they work well within themselves, particularly at higher frequencies. With experience you can home in on ranges for 'free' plate mode frequencies that will get you into the ball park on the assembled instrument, but there are so many variables that it's hard to generalize. I have found relationships that seem to work pretty well on the guitars I and my students make, but even small changes in how things are done can alter the outcomes significantly.

Dana also uses 'tap tuning' to fine-tune the top thickness after the box is assembled, and you can do the same with the back thickness or (more productively) bracing. Again, he uses the 'feel' of things: particularly, looking for a specific sort of 'kick' when tapping the top in different places. I was able to 'track' him through the process at a GAL convention back in '92, using my signal generator to find the assembled resonant modes before and after he did his adjustments. The 'kick' he seeks is evidence that various resonances of the wood and air are 'coupling' strongly because they have been tuned to happen at nearly the same pitches. We were going to write up the whole thing afterward, but the top of the guitar he used cracked on the way home, and had to be replaced.

Anyway, there are ways to do what you're thinking of. The fact that they are not either simple or obvious is simply a reflection of the complexity of the system we're dealing with, and the range of outcomes that define 'good' for different people. This is not rocket science: it's more complicated.

Here is the post on this forum that really game me the aha moment.
http://www.luthiersforum.com/forum/viewtopic.php?f=10117&t=30241

There is a wealth of info here. The main thing that applies to what you are doing has to do with simply tuning the back to the top. Rule of thumb being to get the back and top to ring together at the same pitch before the brisge is glued on.

I use spindle clamps to temporarily attach the back to the body (after the top is glued on and overhang is trimmed)
Then compare the tap tones of the two. Take off the clamps, carve braces and repeat until they are tuned together.

Hope this helps.

Dave

Thanks Dave, I actually remember that post from those many years ago but forgot the details. Thanks for posting the reminder and the advice.

_________________
Thank You and Best To All

Alan, one question and one comment.

I have seen multiple references to free plate tuning, which I hope to try once I get that far on this first guitar. Would attaching the top to the sides (without back attached) provide a better situation for tuning the modes? The sides alone will not be as stiff as the fully assembled body, but you would get closer. Putting the sides and top in a mold might be too far the other way, but provide another option. Top modes could be tuned this way, and back done free (or perhaps tweaked through the sound hole after assembly). I have wondered if this isn't normally done simply because it makes glue cleanup difficult for the back, which is much more visible than the top. That is probably important for a commercial instrument, but I'm building for myself and I am far more interested in learning how to produce a good sound! Thoughts?

Regarding your rocket science comment, the two are not that far apart! I have done a couple of tap tests on launch vehicle components, and once got to do one on a fully assembled rocket at the pad.

A top with the sides attached is still effectively a free plate (with added stiffness and mass around the periphery), so no it doesn't provide a better solution. Once again, get the Gore/Gilet books!

If you tap before the bridge is glued on, the stiffest transverse brace on the lower bout and 20 to 30 grams of mass are still waiting to be added to the soundboard. It's been observed that the fundamental resonance (0,0 mode) of the soundboard often doesn't change much when the bridge is added, as the added stiffness and the added weight try to move the resonance in opposite directions. But that's more of a coincidence than you may want to rely on and doesn't read at all on the higher resonances. Tapping, or other tests of the soundboard before the guitar is playable, have to target what the luthier's experience shows to be a good waypoint, but, until the guitar is finished, it won't be possible to target the final destination.

There is a big difference between the 'free' plates, where the edge impedance is effectively zero, and plates that are mounted on a rim, where the impedance of the edge is very much higher than that of the plates. Also, a 'free' plate can bend along the edge, while a 'bound' plate can't. In essence, all of the 'free' plate modes that have node lines that are not parallel the the plate edges are suppressed, and even those with node lines parallel to the edge are changed in pitch when they're attached to a rim.

In my view, the 'character' of the sound of the guitar; whether it's a Dread or a Jumbo or an OM or a Classical, is largely fixed by the low end mode behavior. A guitar that has it's low frequency 'main top' and 'main air' modes roughly an octave apart at G ~98/~196 will have a different overall tone than one where those modes are nearer A=110/220. You can look at these modes easily on the assembled guitar, using, say, Chladni patterns, and find the peaks and relative activity in a spectral plot. Small differences in the pitches of the these don't necessarily translate into audible differences in tone unless they happen to fall exactly on scale degrees: the standard 'guitar wolf' is the 'thuddy G' you get when the main air peak is too close to that pitch. It's hard to predict the exact pitches of these modes from the 'free' plate data. Larger differences account for a lot of the differences in character from one guitar to another.

What I think of as the 'quality' of the sound of the guitar, how 'good' it is given it's basic character, seems to be much more a function of how well it works in the high frequency range, particularly the top. There is a paradox here, which I believe was pointed out by Tim White back in the '80s. On the one hand, high end response seems to be the thing that correlates best with 'high quality' sound. However, we don't have any direct control of the high frequency response characteristics of the guitar. You can set up the stiffness to weight ratio of the plates to establish a certain bias in the frequency distribution, but there is simply no way to make a top, say, so that it will have a strong resonance at a particular pitch in the range above 800 Hz or so. In part that's because this is a 'resonance continuum; there are so many resonances of all of the different parts of the guitar that they overlap, and couple so strongly that it's impossible to attribute any particular output peak to some structure with any confidence. You can go around after the fact and make alterations, but it's an ad hoc procedure where you try something and see what happens. The paradox, then is this: if no maker has direct control over what it is that produces a 'high quality' sound, how is it that some makers are consistently able to produce higher quality guitars than others?

I think that the answer is in some form of indirect control. This could take several forms, but they all come down in some way to establishing some sort of 'balance' between elements of the top (in particular) that enables it to work well once the guitar is assembled. 'Free' plate tuning can be one such procedure.

Again, if you compare the modes of a 'free' plate with the modes of the same plate that has been secured around the edge in some way, you'll see that the 'free' plate will have a larger number, and wider variety, of modes that can be visualized using Chladni patterns. Once the plate has been secured it will tend to have modes that are very similar to every other assembled plate: whatever information was contained in the modes that were suppressed when the edges were secured has been lost. Many of those 'lost' modes have antinode areas (which are bending in vibration) that are much smaller than any that can be easily accessed with Chladni patterns. you could do so with holograms, but most of us lack the technology to do so (priced out an optical bench lately?). Those hard-to-visualize modes occur in the high frequency range we'd like to be optimizing, and some of them, at least, can resemble the sorts of patterns we can see on 'free' plate modes at frequencies that are easy to visualize. It makes some sense, then, that optimizing those higher-order 'free' plate modes could, perhaps, enhance the frequency response of the assembled top.

This would be very difficult to 'prove' in any rigorous way. However, we do know that with practice experienced luthiers can achieve some measure of control over the high end sound quality of their guitars, so some such mechanism has to exist. I'll note that there are a lot of ways to approach this, not all of which need to be conscious or rational. At a very basic level one can simply learn what 'feels right' as the top is flexed and trimmed, and correlate this with desirable tone through experience. 'Tap tones' give some of the same information that Chladni patterns do, and can, with practice, apparently become fairly reliable guides. Even just scraping or sanding on a top gives some audible feedback as it drives high frequency modes, and if you can learn to stop removing material when it 'sounds right', even if you don't define that it terms of pitch, it can be a useful quality control measure.

I've been 'free' plate tuning guitars for a long time. The feedback I get from my customers suggests that the plates that have a lot of 'well formed' modes tend to produce 'better' guitars. I do have a few somewhat more objective results, often from 'matched pair' experiments, that give targets to aim for, and (even more useful) a few to avoid! Data gathering is slow when it takes as long as it does to make a guitar: one would like to do some experiments on a simpler system to speed things up, but even then a lot of time is required. Logically this all makes some sense, and my experience seems to align reasonably well with that, but I also have to keep 'Feynman's Dictum' in mind: "You are the easiest person for you to fool". FWIW

Alan, Your response is full of great information. Thanks for continuing to share your wisdom on this forum. The free plate tuning that you coached me through has led me to achieve consistently wonderful instruments that have great tone and projection.

LarryH, do not let yourself be dismayed by the technical talk. Your original post spoke to your reluctance to embrace the technology and science involved in making your guitars. I (and I am sure many others on the forum) can recall a time when I would see a response or post by Al Carruth and my eyes would actually roll back in my head with confusion. I think I went cross-eyed for a week after one of his explanations of chladni patterns.

But after fifty or so guitars, I think I am getting the hang of it. You will too.

Good luck and keep seeking knowledge.

Dave

Sorry I couldn't read Alan's response - I was cross-eyed

Yeah great advice and can see myself getting there eventually but am as you suggest a bit overwhelmed by the info.

_________________
Thank You and Best To All

Thanks for the explanation, Alan. Let's see if I understood well enough to summarize.

It sounds like my idea of trying to tune the top with the edges fixed might help me to visualize the finished modes better at the low end (both in terms of shape and frequency), but those are mostly governed by the overall size and shape of the instrument anyhow, so probably not much to be gained in that region. At the high frequency, these modes can't be visualized or measured reliably in either condition, free plate or completed instrument. However, there may be some correlation between certain low frequency mode shapes and the behavior at high frequency, which has been explored in the form of free plates rather than attached. So although that could be re-explored from the perspective of plate+rim, why reinvent the wheel?

Along the lines of high frequency behavior relating to high sound quality, do you have any more specifics on what measurables tend to "sound good"? Significant high frequency output relative to low end? Even high frequency response, or perhaps uneven?



Yes, in addition to finding it difficult to identify and alter any particular high frequency mode, at high enough frequency there will be so many modes contributing to the response that even if you could completely eliminate one, you may not hear a difference! In the room acoustics world they have what is known as the "Schroeder Frequency". Once you see three or more modes contributing within the half-power bandwidth, the response is considered diffuse from that frequency and up. Design in that range shifts from attacking individual modes to more broadband efforts.

As in 'Schroeder Diffuser'? I've fooled around a little with those.

So far the best paper I've seen that relates to the high frequency response was one that was put out by Mathews and Kohut. I don't think I have a copy: it was some time in the '70s iirc. They were looking at violin sound, but I think a lot carries over. For their experiment they made a number of op-amp parametric filters, on the order of two dozen. They used these to shape the tone of a solid body violin. They found that what seemed to matter most in the high frequency range was the number of resonant peaks per octave, and the overall Q values. Too many peaks tended to be too 'even' sounding, as there was not enough difference in the overtone structure of adjacent notes. Too few made it too uneven. They also found that an overall peak-to-dip ratio of about 10 dB was most liked. The actual spacing of peaks within an octave; linear or logarithmic, didn't seem to matter. I'll note that, in the 'matched' guitars I've made, where the peak numbers and peak to dip ratios were the same, there was no perceived difference in 'quality', but people could easily distinguish the instruments in blind listening tests.

Another interesting study relating objective measures and subjective quality was that done (again, on violins) by Dunnewald ("Deduction of objective quality parameters on old and new violins", 'Journal' of the Catgut Acoustical Society, 1.7:1-5, 5/91). He used objective measures of response of a large number of instruments to relate objective measures to preferences of a limited number of qualities ('harsh' and 'nasal' sound, based on relative amplitudes in different frequency ranges). Instruments that scored low in 'harshness/nasality' and had high output in their lowest frequency range tended to be judged as 'good'. 92.5% of old Italian instruments were judged 'good', as against 26.2% of 'hobby' makers, 30.7% of Master instruments before 1800, 19.1% of Master instruments after 1800, and 8.4% of factory made instruments; so the derived parameters match reasonably well with the judgement of the market.

Considering how useful this sort of information is, it's surprising that so few such studies exist. They do, of course, take some effort.

I have a different view of what I'm trying to do when I tune a guitar top. None of this contradicts what Al does but looks at the problem from a different angle. We've had many discussions together on this. Consider what a perfect guitar top would be. It would vibrate with little energy being added to it and it would vibrate evenly at most frequencies without huge gaps or spikes.

Now consider what it is we have to work with. We have a piece of wood that has different stiffness along the grain and across the grain. It has different dampening along and across the grain. The stiffness and dampening may not be even as you move along or across the grain. These parameters as well as the weight can vary between different pieces. The guitar top is shaped like a snowman and has a large hole just above the narrowest point of the snowman. It needs to withstand the forces of the strings pulling up on it in the middle of the lower bout. In order to make the guitar stand up to these forces, it has several braces to add strength without adding an unreasonable amount of mass. These braces can be placed in many, many different places and each has their own stiffness and weight at different points along them.

What I want to do is to thin my top and braces so that it vibrates easily without large peaks and valleys in the frequency spectrum. To help me determine this, imagine a chladni pattern where there is a line an inch or so in from the perimeter of the lower bout and cutting across the top just below the sound hole to make a ring. There is a line running across the upper bound just above the sound hole. This shape is the ring and a half mode. The frequency that gets me this shape should be in the 240-260Hz range. This is my goal.

Think about a top that has blobs of putty dropped on it randomly across the top. Some of the blobs are big and heavy some are smaller and lighter. The vibrating behavior of such a top would be very irregular and irregular tops don't sound good. Untuned tops have these blobs but you can't see them. The chladni patterns of such a top would still be there but the shapes of the patterns would be irregular and diffuse.
The objective in tuning the top is to thin the top, place the braces and adjust stiffness of the braces in such a way as to make the irregular chladni patterns more regular. Here are a couple of chladni patterns from a recent build:


The image on the left is very diffuse while the image on the right is quite regular and well defined. The pattern on the right is where I want to be. Notice how the ring is closed with no gaps and the pattern is well formed. The pattern on the left has open lines and very poorly defined ones at that. These two patterns are from the same top at different frequencies but the pattern on the right is the only one I care about. Once I get my top ring and a half pattern looking something like that, I'm done tuning the top.

I have "Fixed" all the the "lumps" and I have a top that is going to behave well once it is on the guitar.

That 'ring+' is the main indicator, and I use it too. What I have noticed (or believe I have noticed!) is that the more other patterns you get that are nicely shaped, narrowly defined in pitch (high Q) and active, the better the guitar seems to be. Or, at least, that's what my customers tell me....

The ring + indicates the main monopole?