Macaferri soundhole

[Dave M]

I was enjoying a local gypsy jazz group in the pub the other night when someone asked about the soundhole. The rhythm guitarist had the tiny vertical oval soundhole and the lead the seriously big D shape - Big mouth, Grand Bouche...

Anyway I said it was just aesthetic but I bet I'm completely wrong on that. What will be the effect of such wildly different soundholes actually be?

[kiwigeo]

I'm no expert but looking at the equation that determines a guitar's Helmholtz frequency the shape of the soundhole doesn't seem to be a factor. My take on the traditional circular shape of the soundhole...it's the easiest shape to make a hole which will also look asthetically pleasing to most eyes.

[blackalex1952]

The large D shaped soundhole was from the original Selmer Maccaferri design by guitarist and luthier Mario Maccaferri. The soundbox contained a second soundbox within it and a reflector which directed the air movement from the internal soundbox through the soundhole along with the main body air movement. The internal soundboxes used ro rattle and many owners ended up simply removing them. Later M.Maccaferri ended his contract with Selmer. By this time these guitars were becoming popular so Selmer continued to manufacture them without the internal soundbox and bearing the marque Selmer only. The smaller oval hole and the 14 fret neck on their "modele jazz" appeared. Listening to the guitars in an ensemble the D hole guitars sound like they spread the sound and were originally used for rhythm accompanient. The lead player usually uses a small oval soundhole guitar which tends to throw the sound more directionally with more low mid punch. These days with stage microphones that acoustic imagery is often lost. Originally used without amplification. The soundboards are ladder braced and braces curved. The top has a bend similar to the old neapolitan mandolins. This has the following effects. The bridge height is increased resulting in more string load on the soundboard. The top becomes sort of domed with arching in two directions. My current undersatanding is that these factors push the soundboard resonance upwards and increase the mid frequency projection. The doming also stiffens the soundboard against the string load allowing it to be thinner and lighter which makes for a faster attack. A voice in an acoustic jazz ensemble. (The small soundhole on a good petit bouche as they are called, is in the 90hz vicinity). In the last run of guitars from Selmer the 5th lower brace was left off on these models. This gave more support to the low end. Possibly by allowing greater soundboard movement with a lighter less stiff soundboard and some low mid support to the air resonance. (my mspeculation) the four braced models are most sought after. -Ross

[Dave M]

Thanks Ross and Martin. These particular guitars would just have had the appearance of Selmacs. But I am very interested in the construction of the originals - so completely different from what we think of as a steel string today.

I was thinking that although the big mouth looks pretty big, it's area is maybe not so much more than a circular soundhole, but the petit bouche is seriously smaller than the norm so the air resonance would be lower as you point out Ross at 90 Hz given that the frequency is proportional to the root of the area roughly.

You do have be a bit careful when these questions are thrown at you when a few glasses have been consumed!

[kiwigeo]

You do have be a bit careful when these questions are thrown at you when a few glasses have been consumed!

Just write the equation down on a scrap of paper and pull it out of your wallet at the right moment.....should kill the conversation straight away

[blackalex1952]

Listening to the guitars in an ensemble the D hole guitars sound like they spread the sound and were originally used for rhythm accompanient.

Just to qualify my comment...the guitars I have heard do not have the internal soundbox. The large soundhole models (grande bouche) I have tested have an air resonance as high as a B on the 5th string. I have yet to experiment with grande bouche models and haven't examined many. The higher air resonance would perhaps have some influence on the large soundhole models having a better voice for chords. If anyone interested , posts on this forum by Craig Bumgarner are very interesting. He has posted a little data he collected on GJ guitars. Hopefully he will chime in here in this thread...cheers! Ross

[Craig Bumgarner]

I don't have any great insight on this, but here are a few observations.

My guess about the guitars used by the Hot Club of France Quintet is it was mostly just coincidence. As Ross said, the 14 fret guitar with the small oval hole that Django is known for came later in the Selmer model development time line and the HCFQ being endorsers, so they were given guitars. The early models were the large sound hole models with resonators. As the guitars developed, Django, being the star and I'm sure a co-developer, got new ones more often than the rest of the group. The 12 fret D holes guitars were adequate for rhythm and playing chords. The fingerboards were nearly as wide as classical guitars and the resonators dried up the sound a good bit giving a nice dry rhythm chick sound to overall sound. During the ten years when HCFQ as very popular, the commonly seen arrangement was two rhythm players on 12fret Gran Bouche and Django on his 14 fret PB. Because modern followers of Django want to be just like their hero, lead players usually use the 14 fret petite bouche model and for many years, rhythm players used the 12 fret Gran Bouche (without resonators).

About 15 years ago, however, players started breaking out of the stereotypes and using different models. 14 fret Gran Bouche models came out and were used by some lead players. Rhythm players started using 14 fret Petite Bouche guitars. In many groups, the lead and rhythm responsibilities are shared between the guitarists if there are more than one. If there is only one, they, of course, have to hold down both jobs. Today, it is quite common to see any arrangement. As Dave M said in his OP, he saw the rhythm player on a PB and the lead on a GB, just the opposite of what it used to be. Today, the majority (90% ?) of both lead and rhythm players both use 14 fret PB guitars as they offer the most flexibility and have the best sound.

Just to show the exception proves there really are no rules about this, Roger Chaput, one of the HCFQ rhythm players played a 14 fret Selmer with a round sound hole:

http://gypsyguitars.com/inventory/makers/selmer/ Top one on the page. I'm currently building a replica of this for a customer on request.

As mentioned, the larger the sound hole, the higher the air resonance. This should make for a brighter instrument that makes for a drier sound without too much bass to muddy up the rhythm sound. It should be noted, however, the sound the player hears with the larger sound hole is MUCH bassier than what the audience hears. This is experienced as a woofy sound when chords are struck. With the very large D hole, the player will often complain about hearing nothing but raging bass. I think this is because with the larger sound hole, a disproportionate amount of the air resonance is being heard by the player. The audience though hears more of the projection off the top and so to them, the sound is brighter. In my opinion, not hearing what the audience hears is not such a great idea. A good player can control the woofy sound with his playing, but it gets to be a struggle sometimes. In a way, it can be argued that the arrangement Dave M saw is actually better, PB for rhythm and GB for lead.

I've played Selmer #101, the first Selmer guitar made, a 12 fret Gran Bouche with a resonator and I didn't hear any of the excess bass I heard in the non-resonator models, probably because the resonator's internal volume is a lot less than the guitar body itself, plus the resonator partially blocks the hole. Resonators are an interesting idea, but as Ross says, they were fragile and high maintenance, so in most cases they were removed.

Django played Selmer guitars. Once the early days of model experimentation had settled down, Selmer built almost nothing but petite bouche. Other builders in Paris, however, built all kinds of variations. Round holes, larger oval holes, medium size D holes, Bean holes, heart shaped holes and so on. My take is the shape is not important to the sound, but size (area) is quite important. I once did an experiment with a piece of scrap top spruce and some small clamps to progressively close off the a GB sound hole. The sound altered noticeably with each reduction. Fingerboard extension into the sound hole effectively reduce the size of the sound hole.

[Dave M]

Craig thanks a lot for those insights.

I think the Macafferi designs are interesting not so much for the sound they could produce which can be quite harsh, as the approach he took to getting more volume. This might be a useful input to designers today looking to increase loudness. As I understand it the tops were pretty thin and he used just transverse bracing...? As you say the shape of the soundhole is probably immaterial but the area certainly is. Yes I noted the wide fretboards.

I made the point to the incredibly talented young lead player on this occasion that the style shares with Flamenco, which he also plays a little, an enormous work rate. There can be few other guitarists who produce so many notes per minute!

[blackalex1952]

Thanks Craig. For those who haven't seen the guitars that Craig makes do a name search on the web. Beautiful work.
There are two other aspects of these guitars that now deserve a mention. 1)The strings, archtop construction and scale length 2) The playing techniques
The guitars are made with a tailpiece, so they give the associated bridge movement in the vertical direction only. Being ladder braced they support long grain soundboard movement and suppress cross grain movement. The scale length is quite long, particularly on the 14 fret models-625mm. Some makers, for example, Busato, used even longer scale length. This results in a higher string tension when compared to the kind of scale lengths used by the American style archtop guitars, with a corresponding effect on the tone the strings produce. The strings are made with a steel core which is wound with silver plated copper.
This combination results in a unique string tone, and a softer feel which is compatible with the longer scale length. The action heights favoured by Gypsy Jazz guitarists are around 2.5mm to over 3mm. This is because of the picking style. The volume and tone required with a low action chokes the strings.
A stiff thick plectrum is used, in some cases 3.5mm thick usually no less than 2mm.
The picking style comes in my opinion from the ancient world and evolved from the need to produce volume from acoustic stringed instruments. The style can be observed in the case of Indian Sarod players, Ouds, Bouzoukis, Mandolas and other instruments. Rather than using alternating up and down strokes with the plectrum, the strings are played predominantly using down strokes, usually following through and coming to rest on the next string. This results in a "digging in" feeling followed by the back pressure of the rest string. Great for going from chords to single strings. Players like Jimi Hendrix, Stevie Ray Vaughan and most old school US jazz players use this picking as well. In the US it is often referred to as economy picking, but is often a less disciplined approach. Arpeggios are often played by "sweeping" a series of rest strokes across the strings and playing more along the fingerboard than across it with the left hand. Up strokes are used, but the golden rule is not to change strings on an up stroke. This means that in the case of ascending notes the sweep picking needs to be done thus: in the case of a down/up/down movement on a string(ie 3 notes on that string) the next string is approached by following through what would be the down rest stroke with a downward stroke on the next string. In the case of 2 notes on a string, ie down/up, the next string would be a down stroke. When descending across strings, to satisfy the rule, a down/up followed by a string change to the next (lower) string requires a down stroke, satisfying the rule. In the case of 3 notes on a string before moving down to a lower string down/up/down won't work as it doesn't prepare the picking direction according to the rule. Usually on the three note string, the movement is down/down/up then the string change gives the required down stroke. The plectrum hand wrist is held up, so that there is no strain on the complex 7 boned wrist joint, and the much simpler ball and socket joints in the two forearm bones are used for the fine movements. The pick is held slightly cocked forward. This technique is difficult to learn at first but becomes pretty natural after a while. The GJ guitars seem, to me, to respond to this style bio feedback wise, giving a sense of certainty. The result is a more relaxed feel, great accuracy and control, speed, ease of changing from chords to notes, evenness of volume and most rhythmically important notes occur on downstrokes with the corresponding punch and clarity. Also greater volume with less hand and wrist fatigue. I can now play using alternating down/up strokes and also the GJ picking style. I find that on light string low action electric guitars like my Strat, I use a combination of both picking styles and a lighter pick as I don't feel the rest stroke back pressure as easily sometimes. I have however, graduated to heavier strings on my electrics these days-better tone, more sweep control and higher output from the pickups.
here is a summary of the rules from the net:
When you change the string, play a downstroke.
- When you play multiple notes on the same string, play alternate picking.
But if you play triplets, play down-up-down_down-up-down_and so on.

- The last note of the lick is a downstroke.

and of course, stay always relaxed.
https://www.youtube.com/watch?v=UahB8mGXI7g
Gypsy Jazz guitarists are renowned for the learning and practicing of arpeggios and their use. I have invented a word for the style when one is learning- KLEPTOARPEGGIOSIS! Should be covered by the disability pension. -Ross

[Craig Bumgarner]

Dave, top thickness will vary depending the builder and how many braces the gypsy guitar has. This can be as few as two ladder braces only (!!) or as many as five ladder braces plus little support braces in the bridge, head block and sound hole areas. Fewer braces, thicker tops. I've seen tops as thin as 2.0mm and as thick as 3.6mm, both were outstanding. Thicker tops and fewer braces gives a more direct fundamental tone. Thinner top and more braces adds to complexity. Deflection testing & specific mobility are a big help in understanding the balance of the top thickness and bracing (Thanks Trevor!).

Alex, thanks for the shout out! I agree with you on what we call "gypsy picking" here in the US. Good execution of gypsy picking is probably more important to the gypsy jazz sound than most of the guitar building nuances builders like me fuss over. Good players in this style can make a cigar box guitar sound good. The scale of most 14 fret gypsy guitars is 670mm. Busato and Favino made some that were 675mm and I've seen one at 680mm, but that is getting pretty extreme. 660mm is sometimes used for players with smaller hands, but much less than that, the tone changes away from the preferred GJ sound.

The 12 fret grand bouche models typically have a shorter 640mm scale and they definitely do not sound the same as the 670mm 14 fret models. We talked about the effects of sound hole size and shape earlier and how this affects lead and rhythm playing, but there is no doubt the longer scale is preferred for soloing because it is louder with better clarity, at least given the way these guitars are typically set up.

As Alex says, these gypsy guitars can be very loud and of course, that was the object. Django is arguably the first guitarist to master and perform jazz solos. It is not a coincidence that Selmer guitars came available to him at the same time he burst on the scene in the mid 30s. The Model Jazz 14 fret 670mm scale Selmer guitar w/ a petite bouche provided an instrument that was loud enough to be used for ensemble performance without the aid of amplification. The harsh tone Alex mentions is a product of the design that permits the guitar to cut through the background and become a successful solo instrument for acoustic jazz. Where a lot of steel string guitars have peak monopole resonances, T(1,1)2, in the low 200hz or below range, most gypsy guitars are up around 220-235hz, some as high as 245. Very bright, but that is what gives them their "cut". Combine this with a very light top that gets a good bit of its stiffness from strong arching in the braces and a heat bent arch across the grain at the bridge (the "pliage"), the specific mobility can be very high yielding quick attack, quick note decay and responsiveness needed for the style.

[Dave M]

Well thanks to both. I now know far more about the subject. Very interesting about the picking technique - I could see the action just as you described it Ross when watching this small group the other week.

And those guitars Craig look superb. I didn't realise this was your stock in trade.

[blackalex1952]

most gypsy guitars are up around 220-235hz, some as high as 245. Very bright, but that is what gives them their "cut"

I recently measured a high end Dupont T(1,1)2 at 242Hz, a Gitane DG 370 at 248Hz and my latest box unbound with a taped on bridge at 237Hz. Bracing not quite tuned as yet, with a back glued on with HHG and paper for removal and adjusting. Definitely in the ball park that I'm currently working towards. An experiment that one...Air resonance is around 90Hz. Poppy seeds look good at 237Hz also.
The Gitane is not built with a pliage after the bridge, rather, it has the fold in the top where the brace below the D shaped soundhole is. This has resulted in the area between the bridge and the tail of the guitar to be flat along the grain, in fact it is slightly sunken under string load. With a rather thick UV cure polyurethane finish, although it is their most expensive model, it is a little tinny...I have had it since they first came out with this model, got it cheap as a knock about and is the reason I started building Selmeroids in the first place! The t(1,1)2 is too high for a start in my opinion. The air is around a C on the 6th string, but it is a D hole. I think it is high but haven't measured any good D hole ones. Maybe you could chime in here Craig? Due to our location on top of the world, not downunder as the American say, hand made Selmeroid guitars are thin on the ground, but the number of players is growing and a well made Aussie GJ guitar is a viable build on the level that I wish to produce guitars.(If we were "downunder", when the kangaroos jumped up in the air they would keep going.... )
-Ross
(Alex is my middle name, my handle was generated by an automatic online username suggestion type thingy years ago and I got used to typing it...too late to change now!)

[Craig Bumgarner]

Ross,

Most good gypsy guitars, at least the ones I like have a T(1,1)2 down around 220-235hz. Higher than that is an indication they are too tight. Deflection is probably only .10-.11mm, so the SM is down around 16 which combined with the high T(1,1)2 will make for an under powered tinny sound.

I go for a deflection of .14mm with a thick top, 3 brace guitar like Castelluccia or Jospeh DiMauro and up to .17mm for a 4-5 braced guitar with a thinner top. This punches the SM up over 20 giving good response and volume. A by product of this is a lower T(1,1)2 unless you have some really extraordinary top and brace wood. Which is okay, cause anything much over 235 can be painfully bright especially if you have special wood that allows that AND high SM. T(1,1)2 in the 205 to 220 range can make for a nicer sound but can also lead to a bit of a tubby sound, for these kinds of guitars at least. I know of only one good guitar in the sub 210 range, a Rino van Hooijdonk petite bouche. At 207 it is a magical guitar.

The way I tune the top is glue it to the side without the back and then tap and deflection test, working the braces down. In doing this, the limitation is usually the T(1,1)2 hitting the lower limit of what I want. At that point, I've usually also hit my deflection target, but if not, I have to make a judgement as to which target is more desirable. Luckily, they usually then go hand in hand.

[blackalex1952]

THANK YOU SO MUCH CRAIG. Excellent info re your methods and findings. I must say that the photos that you post of the instruments that you build have some GJ players drooling over here. My latest build in the meantime, changed as I had removed the back, beefed up the bracing and shaved the bracing in the top, a four braced Selmer size. I ended up with a deflection of exactly .17mm. Measuring the top resonance, however is not so clear without the back in place. I'm hoping you will comment on that. I have spent some time experimenting with two issues, the brace curvature in order to get the unwedged fingerboard neck angle where I have wanted it, and getting the angles right when shooting the two soundboard halves so that, on a thin top with a pliage, there is no ripple effect at the tailpiece end of the soundboard. Particularly with a four braced build and a thin top in combination. I would call thin 2.6mm or less. Also, have you ever edge thinned the top as in the Gore Gilet books? I have done that with my OM builds.
(I really like OM's, and avoid building dreadies, plenty of people build them already.) It occurs to me that ladder braced gutars might respond differently to edge thinning...I seem to recall a post of yours somewhere, of a guitar that you built with a torrefied soundboard. Would love a report on that one also!
I am fascinated by the bombe that Busato mastered. Have you considered a Busato style build and have you any theories re how he executed same. I know that he used multiple pieces in his tops. One of the problems here in Australia is that Busatos and real Mc Coy Selmers are thin on the ground. I have only ever had one genuine Selmer in my hands, and I didn't know then what I now know so I have not been able to assess one subjectively. I have never handled a heart shaped Selmeroid either. Kind regards Ross.

[Craig Bumgarner]

My experience with deflection of tops with back is it will be pretty much the same with the back off or back on. Sometimes the top stiffens up a little after the back goes on and the bindings are installed, so if it is a petite bouche where it is hard to do anything about it after the box is together, better to go .01-.02mm to the more flexible side. In my opinion better to be too lively than not so. Just my opinion though.

Top resonance is trickier with the back off, as you say, but it becomes easier with practice. I attach a pair of magnets, one inside, one out at the center of the bridge area to simulate the mass of the bridge. Most of the time, I seem to get a ~ 3-5hz bump up in the top resonance with the back on and coupled. This will probably vary depending on how you build, but for me it is ~ 3-5hz. So with the magnets and the 3hz adjustment in mind, the T(1,1)2 I get with the top attached, back off, is as close as I can get it with the back off. I've been taping the back on and adjusting it prior to gluing it and this is working okay. The amplitude of the back is weak when taped on, but the resonant frequency is pretty close.

A wedge under the fingerboard end are pretty standard in gypsy guitars and saves a good bit of distortion in the top and side at the upper end. A by product is more space between the top and strings for magnetic pickups. I use the wedge.

Torrefied sitka from Stewmac has been working well for me. I like how hard the wood is, the stiffness to weight ratio is excellent and I like the darker brown color. It is much more brittle, so I use it only on the five brace models which give plenty of support to the top. If you order from Stewmac, include a note that you want a true AAA top, no sap inclusions. So far, I've gotten seven tops from them, 3-4 had sap pockets I was able to work around. I sent one back because I could not. And one pocket started to show up when I was sanding out the top, caught it soon enough, but from now on, I ask for no sap pockets.

I've edge thinned, but while I'm not fully sure of this, it seems to take the top resonance down faster than brace reduction, relative to deflection. As I'm generally struggling to keep top resonance up while boosting SM, the edge thinning takes me in the wrong direction, for these ladder braced guitars at least.

I've not built a Busato copy though I would like to some day.

Yep, there is no substitute to getting your hands on good examples of the style you want to build. I've been very fortunate to been given access to several large collections of vintage guitars over the last 6-7 years and this has been immensely helpful. If you can develop the repair side of lutherie, people will send you their vintage guitars and you not only get to study them as you fix them up, you make money too! Win-win! Easily half the vintage guitars I have studied came to me this way. You also get to see what works and what does not.

I understand about the ripple in the top at the tail end. Do you put a dart so to speak in the center line? I've heard of doing that, but not really tried it. I usually work it out with the caul in this area.

[blackalex1952]

I put the dart in when I shoot the top halves, with this shooting board. It can be adjusted in a few planes with the screws, ie for pliage angle and the lengthwise angle of the soundboard join. For some reason, probably to see if I could get the angles a bit better, I changed it then ended up getting quite a crease at the tail when I made a thin top, four braced, which was only apparent after the tension was applied around the soundboard edges after gluing the completed top to the body. It was also the first one I tried without a fingerboard wedge, adding some stress at the other end of the top as well. So have to make a few tops and glue them to sides to get it just right. If you read my fb pm, you'll understand that I may not have a lot of time to get where I'm hoping to end up with my builds...