Official Luthiers Forum!

Just posted on TLB about the aesthetic aspect to designing a new body shape.

http://www.theluthierblog.com/articles/ ... itar-body/

How do you do it? What's your normal starting point?

Nigel

http://www.theluthierblog.com
http://www.nkforsterguitars.com

My starting point is to guess at some basic dimensions. I have a text file of dimensions of various standard shapes, and another of all my own designs thus far, to give some rough guidelines. Upper bout, waist, lower bout width, body length, depth at neck and tail, scale length, joining fret, soundhole diameter. I work in inches, and the dimensions I choose are often influenced by what are nice looking numbers

Then there are various details like whether I want square shoulders/tail or rounded, whether to include a cutaway or not, asymmetrical shape or not (for fan frets). I generally try to make the curves look as natural as possible, not just a series of circles and flat sections, or bezier curves. Sometimes I do the initial sketching/fiddling with the shape on computer, other times on paper at 1/4 scale. But the final refinement of the curves and vertical position of the upper bout/waist/lower bout are done during the full scale drawing.

I take more of a "big picture" approach in design. Try to consider the 2D shape, 3D shape, the specific woods I have in mind, and inlay theme all together to get the overall "personality" of the guitar right. But it does take a lot of mental processing time. Especially if I end up being unable to use a particular piece of wood, I've been known to get stuck for days trying to make up my mind between whatever options I do have available, and how to re-balance everything else

Also, I bend on a hot pipe and assemble freestanding, largely for the purpose of minimizing the work/cost/storage space needed to create new shapes. I rarely make the same thing twice.

I'm a big fan of the notion that you should never build a guitar that you can't get a case for. Custom cases are expensive, and making your own is a drag. I won't say I always follow this rule, especially cor 'one-off' stuff, but if you're working on something that's intended to be a standard design, it's a good thing to keep in mind. Obviously, since the standard cases are made to fit standard production models, those dimensions are a good place to start. That said, I've never been a fan of the shapes that Martin uses, and some other manufacturers have done things that I'm not keen on copying either. So the thing for me is to come up with a shape that I like that fits the major dimensions of the usual production boxes closely enough to fit the cases. There are several ways to do this.

I've seen layouts, especially for Dreads, that are all straight lines and circular arcs. It works, sort of. I find it boring, but that's just me. In practice you won't end up with that anyway. I've also seen layouts that use a succession of circular arcs. Often the attempt is made to relate them by some simple proportional rule, like the 'Golden Section'. I often find that the closer you adhere to the rule working this way the less I like the shapes, but that's an aesthetic judgement. Given enough circles you can match any curve arbitrarily closely: the trick is to justify all the circles.

Back in the 80s there was an article in the Catgut Society 'Newsletter' about using mathematical curves to draw guitar bodies. Basically, the author was multiplying an egg by a hyperbola to get a waisted shape with a narrower upper bout. Some years later I wrote a BASIC program that allows me to change the parameters of the various equations and look at the resulting shape on the screen. I've added various tweaks over the years when I've needed them to get things like a convincing 'Jumbo' shape, but all of my guitars are designed with this program. I like it because it gives a continuous curve, with no flats or simple circular arcs.

The output of this sort of thing is some sort of cubic spline, more or less, and most vector drawing programs, like 'Illustrator', produce curves this way. With one of these things get pretty simple: you can locate the major points or lines, like the upper and lower bout widths and the waist, on a grid, and connect the dots with curves. The program provides 'handles' you can use to adjust things, and, when you get something you like, you can print it out full size on several sheets of paper.

If you're more traditionally minded you can use a bent stick. An article in the Violin Society journal several years ago showed how you can use this to reproduce the shapes of violins and early guitars (like the Rawlins Strad) quickly. I'm pretty sure you could get a nice looking Dread this way, but not so sanguine about a Torres style Classical. Then again, I haven't tried it. I'll note that some other researchers have been arguing that the major points along the curves can be found using 'Golden Section' layouts, or some variation thereof.

Fred Carlson, who's about the antithesis of 'mathematical', outlined the method he uses in a GAL article some years back. He starts with the strings (which is always a good place to start) and works out from there. He has actually made up a form that has a center line for the strings, marked off at either end where the nut and saddle go, and lines parallel to that indicating the 'usual' waist, upper, and lower bout widths, and the upper and lower limits of the box. These are just useful guidelines, but if you stay more or less within them you'll at least end up with something you can hold on to. Fred comes up with some pretty artistic shapes.

Lots of ways to do it, and they all work for somebody.

Alan let me say that was a very interesting read. Not saying I completely understood all you said. But liked the concept.

Hey Alan, could you point to any more information on using the Golden Section for guitar design? I would also like to hear from anyone who uses something like Inkscape for coming up with outlines. All of mine have been done with compasses, trammels and French curves and there has to be a better way.

Why bother with all the French curves, bent sticks, computer splines, etc.? Just draw the sucker. Your brain can compute more complex, natural feeling curves than anything Look at it, and if it's not quite right, erase a line and redraw where it should be. So much easier than all those indirect methods.

Unless you do something real extreme most people will not think it is new or unique, most mods still either look like a D, OM or jumbo or a 12 fret, with the parts moved a bit or reshaped slightly. I am with Al's thinking, make sure it will fit in an available case. Most people are more impressed with sound than looks and the sound enhancing mods are not seen.

Fred

_________________
Fred Tellier
http://www.fetellierguitars.com
Facebook page http://www.facebook.com/pages/FE-Tellier-Guitars/163451547003866

Something that may interest some is a little program called G-Thang by Liutaio Mottola.

It is a free Windows program and it seems like a nice little tool. I have played around with it a bit but have not used it for planning a guitar but it seems like it does everything including fret scales and fretboard layout.

http://www.liutaiomottola.com/formulae/gthang.htm

From the web site:


Cheers,
Bob

For those wanting simple computer based methods (rather than CAD), there's G-Thang as per Bob's post above, which is based on linked circles and there is Jon Sevy's guitar design Java app, which is based on cubic splines.

Once you get your eye in, you see that most traditional designs are based on linked circles. You can get very close to the same shape with either app if you put the effort in.

_________________
Trevor Gore, Luthier. Australian hand made acoustic guitars, classical guitars; custom guitar design and build; guitar design instruction.

http://www.goreguitars.com.au

+1 for G-Thang

_________________
Cambrian Guitars

"There goes Mister Tic-Tac out the back with some bric-brac from the knick-knack rack"

Some fascinating stuff there.

I think most of us "layout" pretty much the same. Most of us start out with the scale length and work out. It's how we develop the curves that interested me in the recent post:

http://www.theluthierblog.com/articles/ ... itar-body/

When I worked for Sobell, we either used existing molds as heavy "French curves" or I held the 3' rule in a curve whilst Stefan drew around it. That would be cut out and taped to the wall (we usually used brown wrapping paper or thin card.) then the "adding a bit here" and "trimming a bit there" would begin.

So how do you decide what looks right and what doesn't? Do you rely on mathematical formulae or do you use your instinct?

How do you decide what looks good, and what is ugly? What's your "signature" look?

Nigel
http://www.nkforsterguitars.com
http://www.theluthierblog.com

Cool discussion!

Being into classic designs myself, I find a model I like, find a good head on photo and blow it up to the proper size and print it. Then I will make a full size tracing and modify that as I like. The real important part for me is hanging the shape up on a wall and staring at it for weeks, months while I work and making changes when I see something I don't like about it. Then it hangs there till I find something else. Eventually, I like the shape and transfer it to a mold.
It's something you have to live with, so it's worth all the staring...

I like to draw it out on CAD ,makes it easy to change shapes and design features to see how the final guitar will look , not to mention being able to make templates.

_________________
MK5acoustics.com

I pretty much do this. I'll start with the string length and position the bridge in relation to the lower bout. Then I'll decide the greatest width of the lower bout, waist and upper bout. Dependent on the overall affect I'm looking for I'll move the position of these in relation to each other.
The rest is drawing, free hand, a 'fair curve'. Then it's a matter of refining the curves. It all gets transferred and cut out onto a thin modelmakers ply template. Even at this stage I can be making small refinements to the outline. I've literally done dozens of outlines so I'm pretty quick at it these days. The most difficult one that I've done is probably one for a Vihuela, largely because historically there is so little to go on. Then again some might say it gives greater freedom.

There are so many great looking designs out there that, for me, it was always driven by one or more design elements I wanted to incorporate.....great cutaway/horn, lovely waist transition, shoulder slope, whatever. I think at that point, it's art ... not math.
I find a good frontal image as Haans says, (actually I'll make it my desktop to stare at it for a while), then crop it exactly to the outside body dimension & paste it into Turbocad (in a box drawn to those same dimensions) and trace a line drawing. From there, use cad tools (bezier curve etc) to adjust the lines here & there, smooth things out & make it your own. (You can also play with soundhole shapes, fingerests etc to see what works & then generate a full scale pattern to make it)

_________________
Dave
Milton, ON

My first few guitars I did low tech by using a scaled grid on a picture of a model I liked to draw the full sized shape and then modified that shape on the grid. Lately I have used G-thang and have found it to be fairly easy to use, for those of us without cad skills. The nice thing about Gthang is the ability to save multiple shapes with subtle differences so I can easily compare how a change affects the overall asthetics. And it is free.

As far as general shapes i have found I prefer more smooth curves in the body of the guitar with very few straight lines.

Kent

I've not done it in quite some time now as I have two designs that I like and have pretty much stuck with for many years now but when I did it I had in mind classical and steel string designs in mind and I traced a lot of guitars to look at on poster board. Then I just drew it. I like curves so there are no straight sections on these guitars. I've tried some other whacky designs but not only is a case hard to come by but you are spending a LOT of time producing something that is not going to be 'normal' or traditional and hence not marketable. After doing several experiments I've come to the conclusion that there is a reason why most guitars look like and feel like most other guitars, it just simply works.

I just got an interesting Guild in for a repair, a 1960's era guitar, that is interesting in design. It's all curvy with a small body but with very deep sides. I'm going to trace this one out as I've not seen one quite like it before and it sounds pretty good actually. Add that to my tracing collection and maybe some day build it as is or tweak it slightly. That's how I do it anyway.

Here's my take on electrics:
Since nothing really affects the tone other than the pickups, I concentrate on appearance and improving my woodworking skills.
My designs all look like Teles and LPs. I am a product designer by trade, but I lack artistic aptitude.
I choose a scale then make a cardboard model of the neck and body to check the fit for playability.
When my pieces are rough cut, I'll check the balance and dish some material out of the body if needed.
By using a certain bridge and headstock, and can somewhat control the string compliance.
I think low string compliance aids sustain, attack, and clarity.
Pickup placement has a big affect on sound.
I admire you fellows that can understand, predict, and implement strut and top carving to achieve a certain tone. I don't think it is something I'd be able to do in my lifetime, but I really enjoy studying your methods.
Not much to add to this thread other than thanks for sharing your ideas.
Dan

_________________
wah
Wah-wah-wah-wah
Wah

I forgot to mention my friend who found a picture of a particularly nice looking Asian-American swimsuit model, and copied the upper and lower bout curves from that. He certainly ended up with a nice guitar shape.

+1

_________________
Todd Rose
Ithaca, NY

https://www.dreamingrosesecobnb.com/todds-art-music

https://www.facebook.com/ToddRoseGuitars/

I've used the "paper doll" method - cutting the plantilla out of paper and refining the shape with a pair of scissors. The fold line is the center line, so the shapes are at least symmetrical. My body shapes are generally pretty conservative. Most of the guitars I build are slight variations on established shapes.

This makes some great reading!

I am wondering if you cut your shape from brown paper, hang it on the wlla to eyeball for a while, changing the shape every now and then, till you like it, is that shape one that you've grown custom to or is this a shape that is instantaneously appealing?
So if someone fresh to this shape walks in does he/she immediately like it?
So maybe better to not hang at the wall for a long time but take it out every now and then, stare at it, change a little put it away and repeat after a couple of weeks?

aram

That is exactly how I do it. Draw it out, along with several versions. Maybe six in total. Look at them every day. Trim, add and discard. After a couple of weeks you have a winner.

One chap I knew (a bass maker) used to do it too, he'd start off with something new...trim....add....trim, over a few weeks. Then he'd say "Perfect!" And take a step back to take a look at his new masterpiece. "Arggh!!!" It was usually a Fender Precision!

Nigel
http://www.theluthierblog.com/luthier-books-offer/

Well, if it doesn't fit in an existing case, the folks at Ameritage will make you one. And for not a whole lot more. They are going to build me some custom harp guitar cases (that don't look like suitcases)