Wednesday, October 25, 2017

Mea Culpa

 (November 2017)
Mea Culpa.

As said earlier, everything presented so far has been in its nature a hypothesis, reaching toward an understanding of classical Italian violin family design and making methods.   I started looking seriously at violin making around 2009, and this research began in earnest two years later.   It’s now been about a year since I started making the results public.  The research has of course continued, and the ideas have developed and changed to some degree.  

 As I’ve alluded to before, the looseness of classical violin making limits the absolute confidence that can be drawn from any single example.  But each thing learned helps draw a cleaner and more definite picture of practices.  When we look across many instruments and makers, through generations of Italian and Cremona work, we see the same patterns consistently confirmed, and our confidence in the emerging picture of classical making methods improves in clarity, resolution, and detail.   But there will always be some degree of uncertainty and some parts of their processes will like remain unknown.  The hope is that we can discover enough to be able to effectively pick these traditional ways back up and use them to build new instruments.  

In short, some of the things I’ve written about look slightly different now, after continuing the research further.  And the current ‘best picture’ of classical methods is still only a working hypothesis, not settled concrete fact.  Some parts of what I've presented will undoubtedly end up looking somewhat different in time, as research continues. This remains an ongoing process. Hence the mea culpa.    

I thought it might be appropriate to give a quick overview of how things look at this time.

One the most general level:

1)      Designing and building interact, much like the relationship between cooking and recipe.  And mostly the kind of recipe that says things like ‘add 2 parts this to 3 parts that’, or ‘add a pinch of’, or ‘add to taste’, not the kind that says ‘add 3.5grams of’.

2)      The classical makers seem to have preferred to guide every bit of the work by simple geometry constructions or rule and compass, combined with simple integer ratios that can be worked out with dividers as you go.  Always working from the size of some feature already established in the build.

3)      They constantly varied their choices in small ways, but not freehand.  They always achieved their variations by working within the traditional ‘tool kit’ of geometry constructions and ratio choices, by varying the combination of choices and exact application in simple ways.

4)      Such experimentation and freedom  seems to be fundamental to the classical building tradition (but always using or just slightly extending the traditional tool kit).  Since everything is guided by definite choices within a limited range of possibilities, you can ‘repeat the recipe’ or vary any parts you wish while leaving the rest the same.   This traditional approach of overall repetition with limited controlled variations of some parts must have been powerfully helpful in both individual and community learning.  Each instrument is a test of what works and what doesn't. And the 'recipe' approach allows repeating or varying any feature with very fine grain control.


5)      A further realization from this last year:

It seems that in portions of the work they had both ‘execution choices’ and ‘guide choices’. 


I want to look in more detail at this idea of 'guide' methods versus 'execution' methods.  This is one aspect of the classical methods that wasn't clear to me last year when I wrote up the posts on my research.

One example of ‘guide’ methods versus ‘execution’ methods has to do with horizontal placement of the upper eyes for soundholes.   

The soundholes generally have been a difficult thing to study. With the upper eyes for example, the placement is often enough significantly erratic in classical work. This erratctic suggests that perhaps the upper eye location might have been set at an early stage of work, or in a way that tended to lead to drift in the final placement. Perhaps their location was set from the back of the plate before arching was complete?   Only a guess.

But this erratic placement also made it more difficult to untangle what they were doing.  And it was made more difficult as initially it appeared that there were contradictory methods at play.  But eventually it became evident that one method was used to execute the work, the another as only a loose guide to placement.  Also, the execution method could be varied by the maker by selecting different ratios.  For example, the ratio 1:2:1 was most common, but a maker can get only slightly different results using a similar ratio like 3:4:3.  The guide method seems to have been used not to carry out the actual work, but rather to guide the particular choices made for the execution method.

Let’s look at these complications.   The Lady Blunt Strad provides an exceptionally clean and ordered example.   Since everything is so clean on this instrument, it’s easy to see the choices behind the design with a high resolution and confidence. Even so, conclusions should never be drawn from a single instrument without seeing if the same things play out across many instruments.  And even the cleanness of the Lady Blunt can be misleading. Since everything is so well reconciled and crisp, it also looks like all the relationships might be ‘execution’ choices.  Only by looking across many instruments can we learn that one of the methods is only a guide.

Ok. Here is what turns out to be the guide relationship:

We can see that the outside edges of the upper eyes just touch onto lines drawn down from thirds of the upper bout width.   Now, it turns out that the horizontal placement of the upper soundhole eyes in Cremona (and general classical Italian work broadly) tend to be related to these upper bout thirds.  

We can also see in Cremona work that the vertical placement of the upper eyes is also related to the upper bout width.  As here in the Lady Blunt:

So what happens when we look at these two relationships across many instruments and many makers?  In Cremona work, the vertical position of the upper eyes turns out to very consistently be executed in relation to a line drawn the upper bout width down from the top of the body.  Often the eyes are placed touching just below or just above.  Sometimes they're exactly centered on the line.   In some instruments (Strad long patterns often) they're placed at a distance of one or two eye radii from the line.  But we can see that Cremona work tends to execute the vertical height of the upper eyes in relation to this line based on the upper.  For ease, I've been calling this the upper bout square.

With the horizontal placement, the situation is a little different. Many times the relationship to thirds of the upper bout is very exact, but also it’s often only approximate.   

We can see another method also apparently at work in the Lady Blunt:

Here, we can see the inside edges of the upper eyes appear to fall on a simple ratio across the center bout, along the very important line down form the upper bout.  This kind of relationship is again seen across many many classical Italian instruments.  These kinds of ratios across the center bout are also found in the cases were the relationship to the thirds of the upper bout is only approximate.  So, while it remains possible that workers sometimes executed the horizontal position from upper bout thirds, it appears that much more consistently they the executed the work using a ratio across the center bout, using the thirds of the upper bout only to guide the exact choice of ratio to execute.

We can also see that the distance between the upper eyes emerges as a key to the rest of the work in the soundholes.   The key distance might between the inner edges of the eyes, their centers, or the between the outer edges – as in the Lady Blunt example.   Often, this key distance is the one the matches the upper bout thirds.

We can also see that the upper eye diameter derives from this distance:


Another relationship that turns out to only be a ‘guide’ has to do with the bridge line.   We will find in classical Cremona work that the bridge line is generally ‘near to’ 4/9ths the body length from the bottom of the instrument.

In the Lady Blunt, we can see this relationship is not really followed.  Stradivari as experimented in this instrument with a significantly higher placement.   Though most Cremona instruments, and most of Stradivari's instruments place the bridge line much closer to this guide. The Amati paper template also shown places the bridge line much closer to the guide.  In the Amati paper template it might actually be the method used to find the bridge line.   


However, in actual Cremona instruments the 4/9ths method turns out to only be a 'guide'. What we see as the actual bridge line execution choice in classical Cremona work is directly related to the vertical placement of the upper and lower eyes.  Both the upper and lower soundhole notches, giving the bridge line, are just placed midway between the eyes.   In placing these notches, we see a great example of Cremona makers getting varied results by changing the exact application of a simple rule.

For the Lady Blunt, the inner soundhole nick is placed at mid height between the bottom of the upper eye and the lower eye, and the out nick is mid height from the bottom of the lower eye to the top of the upper eye.  Typically, each side is worked independently, with the nicks following the variances in the eye locations.  In the Lady Blunt, you can see the work is not absolutely accurate, but close:


I’ve used the Lady Blunt to illustrate as it’s such a clean example.   For the horizontal placement of eyes, the relationship to thirds of the upper bout and the spacing in simple ratio across the center bout seem to be redundant.  Looking at one instrument, we can’t tell if either or both of these were actually used in making the instrument, perhaps these are merely accidental relationships?    So to clarify, we have to look at similar relationships across many instruments, and across multiple makers.

When we look at these things across many generations of classical Cremona making, what emerges is not a completely simple picture.  We see a range of different ratio choices across the center bout, sometimes grossly executed, and we usually but not always see a relation to thirds of the upper bout.   My hypothesis at this point is that they generally executed the work using a ratio across the center bout, but that they had an awareness the distances between the upper eyes should not be much different than a third the upper bout width.

Similarly, the notches on the soundholes appear to be executed by the simple method of placing them at a height mid between the upper and lower eyes.  But the lower/inner notch also represents the body stop or bridge line. They seem to have been mindful not to stray greatly from the 4/9th of the body level for this.

Let’s look at the same set of details across several more instruments.

Here is a 1574 small violin by Andrea Amati:


We can see that the bridge line guide of 4/9ths body comes very near to the executed line placed mid height from the center of the upper eyes to the bottom of the lower eyes.   Similarly, the thirds of the upper bout guide is very near to the execution ratio of 1:2:1 across the center bout.   
Also note that most of the important calculations in this instrument are from the mold line (or 3 purf widths in from the edge) instead of from the outer edge.   Most all calculations in Cremona work appear to be either from the outer edge or from this line 3 purf widths in, which approximately represents the mold line.   This holds true even when later generations make the inset of the purfling 3 purf widths instead of the 2 favored in earlier generations.   Thus the ‘mold line’ moves from the inner side of the purf to the outer.

A hundred years later, we see Ruggieri making almost the identical choices in a violin:


He achieves slight differences in the way we always see in classical methods.  He uses completely traditional guide choices, but varies the exact application in small ways.   He bases the vertical position of the eyes on the upper bout width from edge to edge instead of the mold width.  And he uses the more modern smaller sizing of the eyes by basing them on a 1/8th ratio were Andrea Amati used a 1/6th ratio.

And again in the 1700 Ward violin, we see Stradivari using virtually the same choices.


The only really notable difference here is that Stradivari has significantly mislocated the bass side upper eye.  It’s horizontal position is considerably in closer to the center line than expected.  This is not some subtle plan.  Such looseness is common in classical work, and essentially random. 

Even stepping over to Venice, we are still going to see many of the same methods as in Cremona.  Here in a 1721 Gofriller violin, we find again nearly the identical set of choices for these details of the instrument:


The only less common thing we see here is that the executed bridge line is considerably higher than the 4/9ths of body guide.   But we saw the same departure in Strad’s Lady Blunt.  Not incidentally, both are from the early 1720s.  And this is also typical of the tradition.  We tend to see the same sort of experimentation among different makers at roughly the same time.  

Even when we get to the wild Guarneri Del Gesu, we find his work is actually much more the same and traditional than it is different.  Mostly, his work and his variations are all within the traditional ‘tool kit’.   Like his fellow makers, he doesn’t want to freehand his experimentation.  Only in his last few years does his looseness and his push to take some features past the normal bounds final lead him into corners were he breaks the normal process, though it seems he still tries to find solutions using the common tool kit. 

Here we see his 1741 Vieuxtemps, and the details we’ve been looking at are still entirely familiar and consistent with Cremona tradition.


What emerges most strongly as we look at many features across many instruments and makers is how consistently and prosaically these classical makers approached the many specific features of their work.  Every detail is resolved with simple compass arc and straight line geometry, combined with very simple proportions.   And the solutions are all very traditional.   Even when a maker takes some small new turn in the design, the way it’s accomplished is always either entirely traditional, or a very simple and consistent extension of the tradition.

And we see constant experimentation as part of the tradition, along with trends, fashions, and ultimately changes in the tradition.   For example, we can consider the purfling insets.  In the earlier generations, most of the violins have the purfling inset 2 purfling widths from the outer edge.  This makes the important 3purfling distance which stands in for the mold line occur on the inner edge of the purfling.   And like with any feature, we see some experimentation.   There are some violins by Nicolo Amati that experiment with insetting the purfling 3 widths from the outer edge.  And there are instruments of larger size from the Amati and Guarneri family that also show this.   Some large instruments even show a 4purfling width inset.  But it’s Stradivari that eventually starts making most of his violins with the 3purf width inset.   And Del Gesu mostly follows this lead, though both of them at times make with the older 2 purf inset style.

This business of trends and change might have something to do with the situations were both a ‘guide’ method and a different ‘execution’ method appear to be used.  I somewhat suspect that ‘guide’ methods might in some cases be the lingering remnant of earlier ‘execution’ methods that were displaced as the traditional practice evolved. 

In any event, the ‘guide’ methods tend to be very simple.  The ‘execution’ methods tend to offer more potential for slight variation.   For this reason, it often requires very clean and high resolution sources material to correctly analyze ‘execution’ methods.   When the source material is lower quality, or the workmanship is very loose it will not always be possible to confidently discover the ‘execution’ method.  However, the ‘guide’ methods will still appear to more or less ‘fit’.    So this calls for another ‘Mea Culpa’.   Sometimes, I’ve mistaken a ‘guide’ method to be the ‘execution’ method.  And, sometimes I haven’t properly separated ‘execution’ methods that are very close to each other in result.

The lower vesici are an example of that.  The ‘guide’ idea for the lower vesici is consistently a 1:1:1 vesici piscis.   Only as the research continued and the resolution of my analysis improved in more and more cases did I realize how frequently the lower vesici are actually executed on the very similar vesici ratio of 5:4:5, and that other execution ratios like 4:3:4 also occur occasionally.

There’s one more point I want to note.   We can all see that one of the surprising things about classical Italian work is how asymmetrical can be.   

Often, we can see very significant asymmetries in the corner levels and to some extent also their horizontal position.  Hargrave has hypothesized that this arises largely from the use of an inside form, combined with a twisting of the rib around the plate pins when mating the ribs to the board for the back, and twisting to get a straight alignment for the neck.  I very much accept this hypothesis.

But classical makers also run their ‘guide’ and ‘execution’ methods from the existing features as the build proceeds.  So, the asymmetries beginning in the corners tend to have further consequences.   In the course of my research, I found that in some cases it appears the maker didn't use the exact same ‘execution’ method on the treble and bass sides.   

Now this almost exclusively occurs in looser and more asymmetric instruments, so I wasn’t sure what to think of these cases.   Was it just very very sloppy work, or was the maker actually forgetting that on the treble side they’d calculated from the 3purfling inset line, and then on the bass side they calculated from the outer edge by mistake?   Mostly I found these cases in Nicolo Amati, Andrea Guarneri, filius Guarneri, and Del Gesu, but less frequently there are cases in other Amati and Guarneri family members, and in Stradivari.  Accidental, negligent, or with some purpose?

I also noticed that these side to side differences in execution didn’t usually introduce deeper asymmetry; instead they tended to balance the asymmetry to some degree.  So, this raises the possibility that at times makers changed the exact application of an execution from treble to bass side when it helped with visual balance in an asymmetric instrument, or kept the execution of features closer to desired locations or guides.  So this again brings up the notion that exact methods of execution might be planned out before building, but don’t need to be, and apparently were at the least open to adjustment as work proceeded.


That’s it for now.  The research continues, but no longer as my primary activity.  I’m back in the workshop again, looking to put these methods into practice.  

No comments:

Post a Comment