[sdantonio]

I'm a long time Rhino user and I have an interesting problem I would like to explore with someone much more knoledgable than I am.

The problem is acoustic guitar design, for simplicity imagine a box 20"X15"X3" deep. On one of the 20X15 surfeces (the top) cut a 3" hold 1/3 from one end along the lengthwise axis. On the same surface place a vibration generator to drive the surface and generate sound. Assume all the other sides are infinitely rigid perfect reflectors.

Now assume the second 20X15 (the back), although being prefectly ridgid can be CNC carved to any desired shape. What is the optimum shape to focus as much sound as possible out of the box through the hole? Now change the shape from a box to a guitar body. Im assuming this optimum shape would be some variation on a parabolic dish, but I want to get it as exactly matching the guitar shape as possible.

Is this optimisation algorhythm something Rhino can do with one of it's ray tracing addons? If not, are there any programs that anyone can recommend to solve this problem?

I originally wanted to post this on one of the Rhino3D forums, but whenever i go to the McNeel site and try to access the forumd they all come up as blank screens.

Thank you,
Steven

[awerby]

Rhino is a pretty versatile 3D modeling tool, but it basically assumes you know what you want to design beforehand. Generating a surface that's optimized for a certain kind of acoustical projection might be asking a bit much. (Of course, as soon as I say something like that, someone comes up with a new plug-in, but this sounds like a pretty complicated problem which will take working out theoretically before you can actually model it). Each frequency, as you probably know, reflects a bit differently, so you need to decide which of them you want to select for and design your reflecting dish accordingly. If it was me, I'd probably make a few different backs and try them out to see which sounds better, but I'm not mathematically inclined. If you're better at equations than me, you might be able to nail it on the first try with the help of a book like this (and maybe a program like Mathematica): The Physics of Musical Instruments: Neville H. Fletcher, Thomas D. Rossing: 9781441931207: Amazon.com: Books

Andrew Werby
www.computersculpture.com

[BurrMan]

So for me you answered your own question. Yes, a "dish" is the projection and capture method, but your dilema is the projection is irregular. So with rhino,
I would setup the modeling of the shape in the "box" and keep everything straight and symetrical so you can easily have your centerline of the projection, then
you can use rhinos flow tools and flow the static propjector to the irregular surface (matches the output hole). The hard part will be to generate the centerline of the reciever (output hole)
and create the flowee surface with that centerline in mind.

[sdantonio]

Thank you for the responces,

Andrew, I already have the book. It is a great book too.

Burrman, I agree. It is not a trivial problem. The problem being, the solution may be such an insignificant improvement over the back that is currently in use for this particular style of guitar that it may not even be worth doing the simulation in the first place. I was hoping that it coued be automated in some relatively simple way. What I don't have is the time to loft several dozen surfaces and check each one individually. Mathematica may be a better choise and Andrew pointed out. I also have proposed this as a problem on a physics forum inder the photonics sdesign section hoping that some canned ray tracing program may be available since this problem appreas to be closely related to mirror design.

Steven

[mmoe]

I see what you are thinking, but I'm not convinced that there is a whole lot that can be gained over the hundreds of years of design that have already been performed. Ideal projection of sound does not necessarily equate to the most pleasant projection of sound. If so, the materials used would not have nearly the impact that they do, but because we hear acoustic instruments and expect certain tonalities, the little things like changing a maple neck to a mahogany neck make all the difference (keeping in mind that their shape could still be identical). I worked at Gibson Guitar's acoustic division in Bozeman, MT for quite some time and there were some instruments that just had "it" and some that just sounded soul-less, even though they were essentially the same shape. Some of the factors that could be attributed are the little things. How long did the guy sanding guitar A spend sanding the top or back down vs. the guy sanding guitar B? How tight or wide space is the grain of the top and how well quartered is it? How dense was the wood (you could see quite a variation between different shipments of mahogany for example). Is the finish thick or thin (when buying my personal instrument, I measured the thickness of the clear coat by taking samples off the top where the bridge would be located later).

Simply optimizing a box to be capable of the most projection will not insure that it sounds good. I'd take an instrument with fantastic tonal response over an instrument with a larger range of volume every time. I'd recommend focusing most efforts on bracing, selection of wood, care in sanding and care in finishing. Perhaps you already know these things and obviously it's more complex than it sounds, but I would say that the best builders I've ever known are the best because they are meticulous craftsmen about those fundamentals.

Also, FWIW, I'm not sure how well a carved parabolic back or top would perform. It would be an interesting experiment, but I suspect that it will be a bit dead and probably not as strong. When gluing up bracing to flat back to generate an arch, you introduce some inherent stress to the system that contributes strength. This allows for the system to also be lighter in weight and therefor probably more responsive to vibration. Gluing curved bracing to a flat plate could be done based a parabolic form or a spherical form, but I doubt that it makes a whole lot of difference in practice since the final shape will be somewhat reliant on how the stresses of the system balance out. A carved parabola would maintain it's shape, but I really think it will have to be much thicker than standard in order to maintain it's shape once the force of the strings are applied.

[BurrMan]

I seem to think I remember thats what Ovations whole goal was..................

I also agree with mmoe about the resonation of the materials being the key.

[sdantonio]

Hi MMoe and Burrman,

Maybe this time I can get the post to go through without hitting the wrong button.

OK what I’m looking to do is to optimize that back for a Smallman style lattice braced guitar similar to the one used by John Williams. Basically the guitar has a paper thin top (About 1mm which is about i/3 the thickness of a traditional cedar top guitar. The bracing is a carbon fiber/balsa composite. The sides and back are very thick laminate (the back is about .25 thick) so there is virtually no coupling to the sides. They appear to work completely as a bass reflex system. On the top the only part the vibrated is the lower bout, everything else is held rigidly by a plywood frame. With all the plywood and the laminate materials you would think this would sound like crap, but surprisingly they sound great.

The back optimization that I’m looking to do, as you pointed out is similar to what ovation did, their problem as I see it is that they used high damping fiberglass composite for their back which defeats the entire purpose of what they trying to do. Additionally their construction methods are more like furniture construction than guitars (I’ve has Ovations come through my shop that have had almost 3mm on polyester finish on the top as a way to stiffen up a poor piece of wood).

I’m attached a link to one of Dave Schramm’s guitar builds so you can see what I’m trying to do. Dave has been incredibly forthcoming and helpful. The backs he (and Smallman) use are essentially modeled on archtop guitar backs with a catenary lengthwise curve and curtate cycloid widthwise curves (I have quite a few posts here of archtop design and curtate cycloids). I’m just thriving to see if the back shape can be optimized even further (which may be one of those things where any improvements would be so insignificant as to not make a hearable difference). Wu Yin-Jang

One of the things Dave told me when I first started looking into the Smallman design is to forget everything that I know from 25+ years of building because it doesn’t apply to the mechanics of this design.

So, that is where I’m coming from with this design project.

Steven

[mmoe]

It's an interesting build to be sure. Are you looking to apply this to a steel string? I guess that while the overall design is different, I'm seeing most of the traits of traditional concepts present in this design as well. There have always really only been two goals in my mind. First, the box needs to be rigid enough to withstand the tension of the strings. This is typically done by various bracing designs that transfer the forces applied to the weakest part of the guitar (center of the lower bout), to the strongest part of the guitar (waist area). His lower/suspsended bracing design accomplishes this in a new way, which is innovative, but the fundamental principle of what needs to be done has not been altered. In that way it follow tradition quite closely.

The other thing is that you want the top to be free to vibrate. The dichotomy which is usually present is still in play here. The top needs to be free to vibrate, but the force of the strings must be effectively transferred to the strength of the box which requires some degree of rigidity. I'm a bit unclear as to how the bridge interacts with the lattice, but it does appear that it's free floating above the lower bracing system, which provides an interesting reduction to the restrictions normally applied to the top where you are both bracing the box for stiffness/rigidity and transferring tension from the bridge. The results in his videos somewhat confirm what I think would have been the logical outcome, which is that a larger surface with less restriction would enhance the bass response. This is an improvement in the overall balance to a classical guitar, which I typically find to be unbalanced toward the highs, and the video seems to be evidence that the overall balance is much better.

If you're planning to use this concept for steel strings, the only thing I'd be wondering is if you'll end up with the balance shifted too far to the bass, especially if you try to use the back to focus more vibration to the top. Most good dreadnaughts are a bit boomy already IMHO, and could use a bit more treble response (sort of the opposite problem that classical guitars have. The other thing is that steel strings apply a very asymmetrical force to the top, so you'll have to be mindful of this when designing the lattice. Classical guitars don't have nearly the range of tension and seem to typically have somewhat symmetrical bracing, a trait to which my eyes the lattice you've shown follows. If you think of how flattops are braced, you'll probably have to focus the strength of the lattice in a different manner or the top will pull unevenly (probably resulting in a bit of a twist to the top). It's a pretty cool concept and I'd love to see someone continue applying it in new ways, but have the opposite feeling about using your previous experience. I think that you'll need to take what you know about traditional design, rethink why these designs evolved the way they did, and then determine what you're goals need to be for the configuration of this newer design.

I've always considered a guitar to be both a band pass system and a block instrument rolled into one, and when building one I think of the block concept just as much as the membrane/band pass concept. Think to the sounds you get with wood blocks when you make them vibrate by tapping them with a hammer. They are most potent when the tones are higher. This is probably a critical concept when considering balance. The top vibration in large motions contributes to bass response. Here is where you fine tune your bass tone. The types of woods used for the back/sides/neck and fingerboard probably contribute greatly to the treble response, so here is where you fine tune your treble tone. This is likely why a Brazilian Rosewood body has such wonderful tonal balance in a dreadnaught, at least it does IMHO. Brazilian Rosewood sounds fantastic when cut into blocks and tapped like a xylophone, so perhaps it's use for backs/sides/fingerboards helps to enhance the upper registers in the areas that are solid and less capable of vibrating in ways that produce bass.

Just a random note, Gibson Gospels have veneered sides as well, though in layers of mahogany. Didn't seem to contribute a whole lot in strength, and it didn't seem to have any adverse affect on the tone either. Sometimes a Gospel would be the best sounding guitar of the day, sometimes it would be the worst. The one thing I can say about veneering sides is that you must be very, very careful when assembling and sanding them. It makes binding the guitar a PITA and you have to be constantly aware that you could go through the veneer in a second during sanding/finish prep. IMHO, I see no advantage to it, but lots of disadvantage, and I've worked on hundreds if not thousands of them. Gibson has a high success rate with them, but they also have a small percentage that get painted black, and it's never because they meant to.

Edit:
Actually, now that I think about it more, it was the backs that were veener/laminated. It's been a long time since I've worked there, so some of those details are getting fuzzy. The sides were still steam bent the same as the rest, but the backs were kinda like arch top backs made in the shape of a dreadnaught. Still don't think a laminated side will make much improvement on steam bent for strength though. If you want to consider an arch back shape, perhaps having a look at the Gibson Gospel is not a bad idea. I haven't had one in hand for a while, but I'd estimate that the arch is about 3/4 of an inch (though it's not really an arch exactly, more of a large hump).

[sdantonio]

Hi Mmoe

This is my most recent steel string builds. The lattice in this build seemed to work quite well. The top on that one is only 2.5mm think, so the lattice does allow you to reduce the top thickness t make it more responsive. Could probably have gone thinner, but decided not to push the first try at it. I don't think I would try going with a full blown carbon fiber balse laminate on a steel srting. It you remember the Kasha bracing, it works well on nylon stringed instruments bout sounds like crap on steel string ones. I would be concerned that the CF lattice may have the same ussews with overbalancing the higher overtones. I may give it a go someday just out of curiosity though.

I'll have to read your email carefully and get back to you on the rest of what you wrote. I'm at work now and should get onto doing something productive.

sdantonio93's's Library | Photobucket



Steven

[mmoe]

I like that shape a lot and it should play very well to your concept I suspect. I've never been much of a fan of dreadnaughts, and have always preferred the OMs from Martin as well as the old L-0 from Gibson (the round bodied one, not the smaller later design). Those are the two designs I've based my own builds off of and I think they are more suitable for creating a balanced instrument. Unlike the dreadnaught, which as I said I find boomy, these instruments are perhaps lacking a bit in the bass spectrum and really seem like ideal candidates for the bracing you're working on. It will be interesting to hear more.