What’s With All the Optimization?



Is there a Holy Grail of Flutes?

Shop Notes Blog

As I’m writing this, I’ve just finished a prototype for the first of my “optimized” bansuri flutes. Before I get into what this means, I should mention that these are not the first (so-called) optimized flutes that I’ve made. I’ve done an optimized dizi as well, and my Essential Flutes are also something that I refer to as an “optimized folk flute”. I’ve also been toying with optimizing some xiao (this is an epic in itself). So you might be asking, “Will I eventually optimize everything?” or (even more to the point) “What does that actually mean?

And while it’s a good word for describing what I’m trying to do, it’s also a bit misleading. And it really demands a conversation about whether or not it is either possible or desirable to optimize (i.e. “improve”) any existing traditional flute design, and I intend to explore that idea within this blog as well.

So first of all, what is the point of optimization? Does it rest on the assumption that some existing design is not optimal to begin with, and that it’s inadequacies need to be fixed somehow? It does seem to be implied.

I love making what are commonly called “world flutes”—variations on folk flutes from different traditions around the world. Most of these flutes have long histories during which countless makers and players have experimented with design variations within the limits of whatever tools and materials were available to them. Let us take, for example, the Chinese xiao. I mostly make what are called “bei xiao” or Northern xiao. They are often made from purple bamboo (but also other materials like wood), and they feature a long thin bore. I have an excellent example of one of these flutes from a very capable Chinese maker. When I first acquired this flute, in the first flush of my exuberant ignorance and early exploration (i.e. before I learned how to play the thing) I was already figuring out ways that I might make it “better”. Ah, hubris!

But as the years went by and I became a better player (and a more experienced maker) I realized that many of the design features of the xiao that I initially thought might be enhanced, were already pretty near to optimal. The weakness was mostly in myself as a player. Once I arrived at a moderate level of skill I was able to see that the flute did everything it was expected to do in the context of the music it was designed to play.

Let me digress a moment to say that there is no such thing as a perfect flute. If we assume that a perfect flute plays exactly the right pitches for each note in all available octaves, that it’s easy to blow, and has a totally balanced harmonic profile, with strong resonance on every note, then that might be considered perfect for a woodwind. Needless to say, no such flute exists, even though many makers have tried to create such a flute. Accepting the fact that it’s an impossible goal is a piece of wisdom that all makers eventually have to absorb. I’m still working on really accepting this truth.

So we need to redefine perfection in the context of flutes. For myself, I think a flute is perfect if it can have the fewest compromises in it’s intended function and still be recognizable as what it is supposed to be. So a xiao is recognizable as a xiao, a quena as a quena, a shakuhachi as a shakuhachi, etc.. They embody the characteristics that define them, and at the same they have consistent pitch, a good voice and only present certain idiosyncrasies that are already endemic to the type of flute being made.

Because if we define perfection as I did earlier (perfect pitch, balanced harmonics, strong resonance on every note, etc.) you create a goal that has the potential to completely erase the character of the flute you are trying to make. Let me give you an example:

A few years ago, one of my collaborators (the brilliant Michael Prairie of Norwich University, who has made a particular study of xiao design and acoustics) turned me onto the work of a masters student at the University of British Columbia named Yang Lan. Yang Lan had made the topic of his thesis the xiao and was focusing on optimization of the instrument. I got to read his thesis, of which I understood about 10% or less, since a great deal of it was highly mathematical and focused on the physics of acoustics, so I had to lean on Mike quite a bit to get to the essence of it. But the upshot was that Yang Lan had identified various so-called weaknesses in the traditional xiao design and set about to create a new design that would correct them (minor tuning variations and inconsistent harmonics were among the perceived weaknesses). He did a remarkable job and his results were impressive. And yet….

To achieve his goal of optimization, he remodeled the entire bore of the flute to such an extent that it could not be created using conventional methods. He had to make it from two separate halves of acrylic that were machined using a CNC machine and then glued together. The reason being that the bore was shaped like a sine wave, with wide spots and narrow spots. Finger hole placement and size was also adjusted to be a bit different than a conventional xiao. His goal was to erase those previously mentioned weaknesses. I decided to follow in his footsteps, and he very generously passed on all of his data to Mike Prairie, who shared it with me. How I went about replicating his work is the subject of another blog because it would be a massive digression right now, even though it is very much on the subject of optimization (and it was super geeky and awesome).

I succeeded in copying his model, and what was the result? Well, a really cool flute, actually! It had a strong, reedy tone and really excellent intonation (though the intonation on the traditional xiao is already so good that the improvement here was minimal). All of the notes seemed equally strong and the harmonics really reinforced whatever note was being played. But the trouble was that it didn’t come across as sounding as much like a classic xiao! Avid xiao player Gary Stroutsos tried one of my prototypes that I had sent to Mike Prairie and he said that it sounded “really bad-ass”, but he didn’t really like it very much as a xiao. He said it played more like an Irish flute. And he was right.

So that degree of optimization ended up being something a bit more like homogenization, and this is the risk when one takes any flute design and tries to correct perceived idiosyncrasies, because those very traits are what define it. If all flutes were optimized, you’d probably end up with a bunch of very similar sounding instruments.

The thing is, not all flutes are good flutes. Let me stay with the example of the xiao. The xiao might have a certain recognizable parameters, but within them you can have good and bad instruments. There can be xiao that have poor tone, uneven tuning, weak notes, strong notes, etc., and you can have xiao like the one I own from China that is incredibly good. Near perfect tuning, great tone, sweet voice, nice response, etc.. They are both xiao and recognizable as such. One of them needs some improvement while the other doesn’t. This is where my own notion of optimization comes in.

The trick then becomes, how do we address some of the inherent “weak” spots in traditional flute designs without erasing their character? My own method is to use a very light hand, because in most cases little change is needed. An example of one of my favorite modifications is introducing a slight taper in the bore of a flute that is traditionally cylindrical. Cylindrical bore flutes tend to have a more difficult time getting good intonation through both octaves, which is just physics. The introduction of a mild taper in the head portion of the flute (which is what Theobold Boehm chose to do when he invented the modern orchestral flute) increases the accuracy of the second and third octaves. It also imparts a bit of punch to the response. It does this without altering the character of the flute to such a degree that it no longer sounds like what it is meant to be (dizi, bansuri, etc.). So this modification optimizes the flute in the sense that now it might be easier for a player to use it in an ensemble setting where more accurate tuning is desirable, or it might have somewhat better projection that allows it to be heard in a mix of instruments.

That does not mean that all players will prefer the optimized version. For some, optimization is just an experience of “different”, and not everyone is going to want a favorite folk flute to be different from what they are used to. Different is not synonymous with “better”. And optimized is not necessarily a synonym for improved, because improvement is a subjective thing. But I had to choose a word that would encompass some of these changes, and optimized sounded pretty cool :-)

So the final word on the subject is that less is more, different is not always good, and optimized is not necessarily improved, but I still feel compelled to do it! All in the hope that the quest for the Holy Grail of flute-making might be fulfilled.



Return to Shop Notes

Published Tuesday, November 17, 2020