An interesting collaboration on
A new flute
The pitch analysis showed some of the usual irregularities but nothing I haven't dealt with before. Indeed, it positively beckoned. But did I really have time to make a new set of reamers to make this flute? Couldn't I cheat just a little?
I mapped out the bore shapes and compared it with some of the other flutes I had made over the years. There were some peculiarities I couldn't hope to match without the proper reamers, but maybe they were small enough to not be critical. I didn't have to clone this flute for the purposes of initial discovery - just approximate it. If it continued to look promising, I could then make the reamers and do the job properly. Indeed, fudging it promised to give us an idea of just how accurate a copy has to be to have reasonably similar playing characteristics.
A nice piece of gidgee (Acacia Cambagei) of just the right dimensions was found quivering in fear under the saw bench and was rapidly slashed into bits. The rag-tag collection of reamers converted it into a plausible approximation of the original. Although the temptation to "make a few little changes" was high, I opted to keep it similar, stupid. Grey would be back in town in a week; let's see what he thinks needs fixing before we mess with it.
As is my usual practice, I opted to make a keyless version in the first instance. Much less work, just in case it turns out to be a complete dud.
The Firth, Pond & Co 6-key
original in cocuswood and silver (above)
You'll notice in the comparison above I've thinned the head a little on each side of the embouchure, and reduced the overhang beyond the stopper, both measures to reduce the possibility of making the flute "head heavy". This can be an issue with keyless flutes with a short foot, especially when the body is elegantly thin as with the Firth, Pond & Co. The barrel appears to be a little lower, but that was just so I could make use of my usual socket cutters. It has no acoustic implications.
So what did I think of the copied instrument? I thought it was good. Surprisingly good volume, good intonation (a few quirks but nothing offensive and nothing that couldn't be tweaked), snappy articulation, and very good balance of volumes. Let me extrapolate on these points.
It's certainly the normal assumption that large-hole flutes are louder than small-hole flutes - they sure give that impression. We should thus expect this flute to be very quiet, and yet it's not. Hmmm ...
Articulation is all about how quickly or slowly notes will form. Or perhaps more importantly in our musical context, how precisely can ornaments such as rolls, cuts, cranns, etc be articulated? My impression is that small hole flutes are "snappier" than large holed flutes. Perhaps the changeover from "covered" to "uncovered" occurs in a shorter vertical distance, and therefore less time? More hmmm ...
Balance of Volume
It's always seemed to me that, contrary to popular 19th century opinion, a flute with identical sized holes all the way along would not display perfect balance of loudness on each note, but indeed display a gradual reduction in loudness with increasing length. The obvious example is the Boehm flute. There's no way that the lowest few notes are as loud as the top ones.
And I'd expect a flute with small holes to do the reverse - the body notes will be a little subdued, but the bottom end note (which takes advantage of the whole bore diameter for venting) will honk like the proverbial gander in the proverbial pratie hole. Now that's been my supposition, but is it true?
Is this the opportunity to find out?
The tuning of the original was not perfect (is it ever?). And because I was loath to make substantial changes before Grey reviewed it, neither was the copy. But it was better - I did work the undercutting harder than Firth & Pond did. What was interesting was that it seemed to need intervention a good deal less than most of the English flutes I had copied. By any yardstick, it was eminently satisfactory. (I might yet just give it a little tweak later...)
So Grey returned a week later to rejoin the fray. The obvious first thing to do was see what he thought of the copy...
You can hear Grey playing my version of his flute - The Sunny Banks.
You'll remember I stewed this thing up using a dodgy set of reamers. If Grey ( the original flute's owner) felt the copy was a good work-a-like of the original, this must tell us something about the level of precision needed. Just how different are these bores? Check them out on this extremely exaggerated chart.
Hmmm, decidedly sloppy work, you might say! and indeed there are few points where the original (in bold) and the "copy" (thin traces) coincide. I could probably have got the match better, even with "borrowed" reamers, but you'll remember time was of the essence. Still, they follow the same general shape and rarely deviate more than 0.5mm - about 3 to 4% at worst. Now normally we aim for a precision in the 0.1mm area - better than 1%. It probably means we could afford to relax a bit, but, as 0.1mm accuracy is fairly easily achieved, why would you? Aiming for anything better would be wasted effort, especially as wood movement with the seasons will negate any higher precision.
With those matters out of the way, we settled down to study some of the issues pertaining to the original identified earlier:
The Loudness and Articulation studies threw up some very interesting stuff which really requires further investigation. Preliminary results are given though.
We measured the loudness of three flutes:
Ideally we would have done this outdoors to minimise reflections from walls, floor and ceiling, but unfortunately the day was far too windy. Some of the lumps and bumps on the curves below are probably due to standing waves in one or more of the three axes of the room. None the less, as the three flutes were measured under the same conditions, the general relativity should be valid.
Ignoring the bumps then and looking for trends, we note that all three flutes follow the same general pattern - a reasonably flat response above F natural and a diminishing response below it. Even that could be a room effect, indicating reducing support for lower frequencies below the room's "cut-off" frequency. But the important thing to note is that the small holed Firth, Pond & Co (pink trace) is not significantly quieter than its bigger sisters.
Tabulating the results:
We can see that the Pratten's copy ends up an average 2dB louder, while the other two were very similar. Keep in mind that 1dB is the smallest change detectable by a trained listener on a constant tone, and three dB is the smallest change detectable in normal program material (eg music) for the average listener. An apparent doubling of sound intensity is 10dB. In the light of those definitions a 2dB change is not very significant.
Indeed, it isn't enough to explain the perceived difference in loudness of the Pratten's and the Firth, Pond & Co. We have to assume for the moment that the difference is mostly timbrel in nature, although a quick comparison of the spectra of notes played on each instrument did not reveal anything startling. Clearly a lot more work needs to be done here. It is important that we can understand what influences our perception of loudness and fullness in a flute, and what characteristics of the flute contribute to that perception.
The range of loudness (difference between the loudest and quietest notes) seems to bear out the perception that this small-holed flute has a better balance of volume across its range than either of its bigger sisters. More work is needed to confirm this with a series of measurements taken under better acoustic conditions.
Articulation - the flute's ability to respond quickly and surely to changes in fingering or blowing - also proves a difficult matter to quantify. Our perception was that the FP&Co responded significantly more crisply than larger flutes, and that the larger the flute the less quickly it responded. We recorded fast transitions between notes (in the form of ornaments) and viewed the waveforms using the digital audio editor Wavelab. In the general case, the difference between the speed of recovery after the change did not seem significant or repeatable. More work needed here too. Again, outdoors on a quiet day or an artificially non-reverberant chamber might be essential for meaningful transient analysis.
There was one fascinating discovery however. One of Grey's armoury of ornaments is the long crann on the middle D note. Cranns, originally a piping ornament, are being used increasingly by Irish flute players. The fingering pattern for this crann is:
all carried out in fractions of a second on one breath.
Grey was convinced that there was a significant difference in the performance of large and small hole flutes on this crann. Examining the waveform of the final transition as played on a large holed flute revealed a significant period of cyclical instability before the note finally stabilised. But repeating the test with the FP&Co also showed a small amount of the same instability. Sure enough, subsequent testing of a medium sized flute showed a medium amount of the same artefact. So it's a flute issue, not a particular flute issue.
This screen grab from the digital audio editor shows the instability clearly - those three (almost five) groups of about five cycles each shouldn't be there! That was the worst case - a Pratten's copy. To make sure it was not just an artefact of my making, we checked an original Hawkes and found the same. Definitely a flute thing.
While we are not in a position yet to describe what's actually going on there, it does illustrate that we can use a screen-based digital audio editor as an investigative tool for transition analysis. As the crispness of transitions is perceivable by the player and the listener, and can help make the difference between mushy and articulate performance, it seems very desirable that we find ways to explore this tricky facet of flute behaviour.
By now, you may well be thinking that here's two smug, self satisfied, mutually supportive lunatics thoroughly encased in their own little world admiring their joint handiwork and preparing to apply for "national artistic treasure" status in their respective countries. But a chilling thought - what would others think of this flute? Gulp, the National Folk Festival master classes and with it the inaugural Australian Wooden Flute Symposium starts the following day. Are we ready to go public? What the heck ...
It's the end of the day, and we're tired but happy. Grey trolls off with his flute, the copy and an fistful of other flutes that might be useful in a summer school situation. Next morning dawns and I struggle in to the Festival site, to find classes in operation and flutes in circulation. Immediate reaction to the prototype is very positive. Within a few days the first order is placed for a four key version. Another follows a few days later, less than two weeks after the project started!
What's in a name?
So clearly the flute has to take its place in my list of models available. But what to call it? Several ideas surfaced and were rejected on one ground or another. Finally we settled on something we all found acceptable - "Grey Larsen Preferred". I particularly liked it because it reflects honour on its proponent, in the way that "Nicholson's Improved" and "Pratten's Perfected" reflect honour on those great performers and enthusiasts for their flutes of days long gone.
Whoa, now that we're talking making keyed copies, it would be handy to know what the keys look like. And to have more input from Grey on what he'd change if he'd his druthers. And to wonder at what we'd found and what yet needs to be done. Graciously, and far beyond his obligations under the grant, Grey spent his last evening in Australia with us. We enjoyed a meal and a chat, got the measurements done and the notes taken, and bade farewell to a new friend we hope we'll see here again soon.
Now just by chance, I happened to mention the project to Seattle flute player Rebecca Deryckx. Rebecca plays a Meacham and Pond flute and was enjoying my assessment of a William Hall flute, another American maker. The need for further research into the history and activities of the US makers came up in conversation, which lead to Rebecca talking to Paul Wells, Director of the Center for Popular Music at Middle Tennessee State University. Paul in turn contacted Robert Eliason, a leading scholar in researching the history of American brass and woodwind instrument makers, and the former curator of the Ford collection. Robert had over the years collected together a bundle of materials on Firth, Hall, Pond, etc with a view to an article, but other obligations always intervened. Robert will make the stuff available to Paul who will complete the project. So one collaboration leads to another.
How do you wrap up a project like this? Frantically short, due to Grey's limited time (this time!) in Australia. Superheated, because of the imminence of the National Folk Festival. But very successful, very satisfying. What did we achieve?
Grey seemed to enjoy the project:
It's a poor study that answers all the questions but raises no others. There are always more questions. This must then be a very good study - we've raised enough questions for several PhD's! They include:
On measurement methodology:
On flute principles:
Interaction between flute and player
On specific flute design:
On understanding the Firth, Pond & Co
I'll endeavour to answer some of these questions as my work continues. Some however will have to await another opportunity to collaborate with a top class and inquisitive player like Grey.
Obviously to Grey, for his unstinting, intelligent and enthusiastic application to the matters in hand, and his kind permission to invoke his name in the titling of the new model,
To Jo Cresswell, Grey's Australian tour manager, for facilitating Grey's involvement in the project,
To the National Folk Festival, for bringing Grey to Australia in the first place,
To Rebecca Deryckx, Paul Wells and Robert Eliason for the promise of more to come,