TECHNOLOGY ON TRIAL

ALL PHOTOS AND IMAGES GUILLAUME SCHÖNAU

The science behind art is fascinating because it allows direct access to the basis of a design. To help us in our understanding, we can use different techniques for observation and analysis, from measuring and drawing, to copying and moulding, to scanning and reverse engineering - measuring an object and then reconstructing it as a 3D model. What technique we use depends on our goals, the type of information we need, and the time we have available.

Violin makers today have access to many analytical tools and technologies that were not available to makers of the past. These methods allow us to study and analyse classical instrument making by accessing and storing a large amount of detailed information. The way this information is collected offers clear advantages over the methods previously used: it is non-invasive and allows for extremely accurate and relatively quick determinations, sometimes without even having to remove a single string from an instrument.

In this article I want to share my experience as a user of such technologies. I emphasise the word ‘user’ because I don’t possess particular skills in computer science. And as with many new encounters with technology, I found the experience had both advantages and disadvantages, as well as raised several questions.

I was fortunate to have a Francesco Rugeri cello of 1690 in my work area (figure 1). Its owner, a medical specialist, has an extremely busy schedule and is particularly anxious about any kind of intervention on the instrument, including routine bonding maintenance. I like the instrument model a lot, despite its small size. It has a very beautiful sound and its overall look is very attractive. I decided to use it as inspiration to create my next cello. My first thought was to produce a copy of the instrument as precisely as I could. But the details of the process eventually discouraged me from this idea.

How do you gather a maximum amount of information, in a very short time, with minimal instrument manipulation and without making the owner’s heart beat too fast? The logical answer would have probably been to do a CT (computed tomography) scan. The usefulness of such technology in stringed instrument making hasbeen demonstrated several times, but in my case, accessing a CT scanner was likely going to be extremely difficult.

Instead I approached a company called Creaform which specialises in reverse engineering. They were able to scan the cello and create a 3D data file. The device they used was a portable 3D scanner called MetraSCAN 3D. It works by laser triangulation, simultaneously tracking several points and recording distances. A computer then compiles and analyses the data in practically real time and produces a virtual recreation of the scanned object. Distances are calculated from fixed points, identified by reflective circular markers applied to the object. The more numerous the fixed points, the better the spatial scan results. Since I could not apply the markers directly on the instrument, I designed and built an adjustable cello stand (figure 2). After scanning, it was necessary to erase or ‘clean off the cello stand image to obtain only the image of the instrument; the surrounding space, which also tends to appear in the scanned image, also had to be erased. Scanning the cello took approximately 50 minutes. It wouldn’t be far-fetched to say that in the future there may be such scanning applications for laptops or mobile phones.

I used software called Meshlab (free) and Rhinoceros (for which I got a free 90-day trial version) for reading the scan files (OBJ, IGES, STL and 3DXML). Most of the available software is complicated to use for an amateur like myself and the purchase price is often very high. Depending on the program used, it is possible to obtain any point-to-point measurement to an accuracy of five microns.

The resulting image can be seen on a computer (figure 3). The drawing, however, consists of thousands of points connected together like a very elaborate and precise grid. For a CNC (computer numerical control) machine, this is not a surface, and an extra ‘filling’ step is required to inform the computer about the surface zones between the grid points.

I found the scanning process interesting, but apart from obtaining the resulting measurements, I didn’t think that observing the cello on my computer added anything meaningful to my work. Up to this point, photos, measurements and drawings would have given me equally valid information.

IT WOULDN’T BE FAR-FETCHED TO SAY THAT IN THE FUTURE THERE MAY BE SUCH 3D SCANNING APPLICATIONS FOR MOBILE PHONES

FIGURE 1The subject of the CT scanning: a 1690 cello by Francesco Rugeri

FIGURE 2The cello rests on an adjustable stand for the scanning process

FIGURE 3How the cello appears on a computer screen post-scanning

FIGURE 4Replica of the cello top plate made from RenShape

I decided to produce replicas of the cello’s table and back on a five-axis CNC machine. As I didn’t have such equipment at the workshop, I used the services of a company called Usimm in Montreal. They used my file to make a replica of the top and back using a polyurethane material called RenShape (figure 4). I also took the opportunity to ask them to make the ‘traditional’ arching templates from the file. The precision of these, made from 3mm plastic, was remarkable.

The RenShape replica allowed me to have a reference model in my work area while I worked on the wood version. It’s more practical than having the original instrument, and the result is very similar to what a moulded version of the instrument would be. For me, this is where the technology provides a big advantage: you get a mould, an exact copy of the original instrument, without any contact.

I admit that there is a disconcerting aspect to this process. The CNC machine can create an exact replica of the top structure in much less time than I could ever do by hand. Usimm also mentioned that if I supplied them with wood, they could produce a perfect CNC-machined replica of the top - possibly a profitable project as long as you were making a sufficient quantity.

I took the liberty of lengthening the model by 12mm without affecting the vibrating length. So my instrument is ultimately now a cello inspired by Francesco Rugeri and no longer a copy in the true sense of the word. I also made another change, opting for a quarter-cut maple piece for the back, instead of Rugeri’s chosen willow. I reasoned that with the harsh climate of northern Quebec this was a more prudent choice. Perhaps Rugeri would have also used a maple back if the option had presented itself, but that’s just speculation.

THE COSTS AND ACCESSIBILITY OF THIS TECHNOLOGY LEAVE IT PRIMARILY FOR INDUSTRIAL USE

I had another dilemma: to what degree would I replicate the top-structure deformation? I had exceptional spruce for the soundboard, with a density of 359 kg/m3for an elastic modulus of more than 10GPa (gigapascals). Once again, I took measurements directly from the original instrument as well as from the 3D mould, and opted for what I would call a modern interpretation of the top structure. I decided to keep the same arching height, but not to reproduce the deformation.

To measure the plate thickness, I used a Magic Probe thickness gauge. Unlike with CT scanning, the portable 3D scanner does not have the ability to see inside the instrument.

Looking back at the experience, I figure that we are only at the beginning when it comes to an effective scanner. The technology certainly exists, but the costs and accessibility leave it primarily for industrial use. For this cello project the scanning and image-filling cost CAN$2,876, and making the top structure with RenShape cost CAN$1,575. Such technology may evidently be profitable, but only for large production runs. And although useful for collecting information, this technology could not be called a necessity, as great copies can be produced by simple observation and summary analysis.

My project with the Rugeri cello provided answers but also raised questions. To what extent will mass-production tools replace craftsmanship? Will the CNC machine become just another tool in workshops, considering that it does not replace the actual human knowledge needed to produce quality instruments? Will the computer modelling process undertaken before producing CNC-machined plates be one of the new methods used by stringed instrument makers? Would it be wiser to use such methods for instrument restoration, in order to reduce our interventions and be less invasive in our approaches?

Any finished instrument is the result of all the decisions made by the luthier at each stage of production. The use of computers, scanners and CNC machines has taught me that technology offers wonderful opportunities to analyse what has been done in the past. But for making my own instruments, I still don’t think it’s necessary.