Artificial Embouchure

[jbtsax]

After almost a year of making and testing prototypes, I have completed a saxophone artificial embouchure device. The intended purpose of this apparatus is to conduct acoustic experiments to see what effect introducing different variables to the instrument will have on the harmonic spectrum and the perceived sound. It is hoped that by removing the "human element" that one will be able to better isolate and measure the effects that different changes make.

Some of the questions I want to pursue include:

-Does removing the lacquer from a saxophone make a difference in the harmonic spectrum?
-Does adding weights to the neck above the tenon make a measurable difference in the sound?
-What effect do resonators have on the sound, and what if any difference do the materials make?

If anyone has any other suggestions, I would be happy to add them to my list if possible.

More photos of the device can be seen at this location: http://s162.photobucket.com/albums/t...%20Embouchure/

John

[pete]

I, for one, think this is gonna be really, really kewl.

There have been a couple of sax playing robots out there, but they don't use a sax (reed) mouthpiece, so they don't sound "right". This apparatus would probably make it "right".

jbt, do you think you could post a sound sample or two? I'd be eager to hear the result!

BTB, you mention "lacquer". I'd also recommend plating and construction material: bronze, copper, brass, plastic, sterling silver, King S20 vs. Silversonic, plated/copper/sterling/wood/brass necks, etc.

I'm assuming that the mouthpiece could be changed. That'd also be something to check.

The only one "problem" is embochure. I'd Think that what your apparatus does is create an environment to get the reed to vibrate. One would ask what kind of embochure this emulates.

[jbtsax]

Thanks Pete. I'm not sure the device would be useful if one could not use exactly the same saxophone with only one variable changed. Lacquer or plating would work. I suppose you could change necks as well.

Sound samples will be forthcoming. I can tell already that I need a "beefier" compressor to keep the tank at a useful pressure. The main problem I am confronted with is the noise of the compressor. I may end up making a sound deadening chamber, putting it in a separate room and using a long air hose.

As far as I know the robotic saxophone playing was done with an apparatus like mine, but much more sophisticated to enable articulation and phrasing with computer controlled solenoids and such.

My device definitely has an "Allard Embouchure" rather than a "Teal" since there is no pressure from the corners in. There might be a way to create the rubber band around the mouthpiece embouchure artificially that one can control, but I am not smart enough to do so.

John

[Gandalfe]

Wow John, just wow! And imagine the sound test for different resonators and pads. It just boggles the mind.

[pete]

Oh. And modified for clarinet to see if gold plated keywork is better than silver plated, nickel plated or nickel-plated plastic .

[MartinMods]

John,

If you have ever messed around with synthesizers and waveforms, you know that, regardless of the type of waveform and what it's overtone ratios are, a static, absolutely steady tone, without modulation, is just about the most boring thing imaginable. It's exactly the same with the saxophone, or any other wind instrument. It's how the horn, mouthpiece, reed, resonator, pad, or whatever you are trying to decide if you like or not, responds to the attack and every little embouchure and air support variation, that tells you if it is "IT" for you or not. Even playing long tones, its how the thing responds to the way you shape the note.

With a static embouchure and air source, you may be able to discern spectral differences between different test subjects, but that extremely isolated instance of information will not tell you anything about what that means for the player, using it to make music. It's like looking at one atom of an elephant. YOu can say there are 4 protons and 16 electrons, but it won't tell you anything about what color the elephant is, and what it looks like to the kids at the zoo.

I think, in order to get some really usable information, that directly relates to how we play the instrument and what we like to hear and feel as players, you are going to have to come up with a computer controlled mechanical embouchure and articulation device.

Good work though. It's a start.

[jbtsax]

Thanks for those observations Lance. They are certainly true. The scope of my studies are going to be very limited. I intend to record various notes on a saxophone and create a harmonic "footprint" sound spectrogram. Then changing only one variable on the saxophone the same notes will be recorded at the same pitch and air settings. Then the sound spectrograms will be analyzed to see if there are discernible differences in the pitches and/or strengths of the harmonics.

There will be no attempt to replicate what we hear or feel as players. The hope is that the machine will produce a more consistent tone production than a human player although admittedly the quality will not be the same. In that way the "human variable" can be eliminated when comparing different trials.

John

[pete]

MartinM, the argument is essentially ... well, ridiculous, at some level.

Absolutely true: it doesn't matter what the horn's made out of, what it's plated with, etc., if it ain't got that swing, pass. However, if one could take a great horn and say, "If we gold plate it, it's gonna sound darker" or "If we slap a copper neck on it, it'll sound warmer" that would be wonderful.

You'd have to define "darker" and "warmer", though. I mentioned this on my essay-length thought experiment on how to test the theory that these materials make a difference (which I could link to, but I'm lazy).

The "artificial embochure", as presented here, takes care of two variables: the player and the player's experience on the mouthpiece. That's about it. There's the major variable of getting horns of the same make and model in sequential serial numbers (and you've verified that there were no model "improvements" over that range) and that the horns are set up identically. Unless, of course, you decide to take a single bare brass horn and test it through electroplating it silver and then gold, etc.

FWIW, I have heard that the robotic sax playing thing that's been featured on YouTube does not use a reed mouthpiece.

If you tell me that I have two molecules of hydrogen and one molecule of oxygen, I CAN tell you what color it is: there isn't any . If you start adding other stuff, that might change.

[MartinMods]

Originally Posted by jbtsax

Thanks for those observations Lance. They are certainly true. The scope of my studies are going to be very limited. I intend to record various notes on a saxophone and create a harmonic "footprint" sound spectrogram. Then changing only one variable on the saxophone the same notes will be recorded at the same pitch and air settings. Then the sound spectrograms will be analyzed to see if there are discernible differences in the pitches and/or strengths of the harmonics.

There will be no attempt to replicate what we hear or feel as players. The hope is that the machine will produce a more consistent tone production than a human player although admittedly the quality will not be the same. In that way the "human variable" can be eliminated when comparing different trials.

John

John,

I understand. I was thinking of building something similar about a year ago, and these were some of the issues I asked myself. Say you are going to test the difference between flat (F) and domed (D) resonators. You do your test at x air pressure for each and note and compare the harmonic resonance ratios. That test will tell you nothing about what the results of the same test at 2x air pressure will be, and then 3x, and 4x, and so on. You won't be able to say anything usable about the characteristics of the resonators for a player, until you do tests over the entire dynamic range of the instrument. And all those tests will tell you nothing about how the resonators sound when the note is articulated, and articulated over the entire dyanmic range of the instrument, with every possible articulation, and so on. I do not doubt that you will see consistent differences between your test subjects in your controlled limited scope tests. My question is, what can you use the information for practically, if, in actuality, it is the way the test subject responds to the inconsistent human element that we find makes it musically interesting?

Further, and pardon me if I missed it. How did you decide on the size and shape of the upstream resonating chamber? As one notices dramatic changes in tone color and amplitude for any one note, when changing tongue placement and opening the throat, how did you know that your cylindrical resonating chamber is the most characteristic and applicable one, for every note you plan to test?

Lance
MM

[jbtsax]

The artificial lip, container, and airstream is not intended to replicate or imitate the human player. The complexity of that task would be immeasurable. What is hoped that it can do is replicate the exactly the same pitch and tonal characteristics in hundreds of trials---something a human player is incapable of doing. In this regard the volume and shape of the air pressure chamber is irrelevant. There are several research projects that have used a similar device with a clarinet mouthpiece to measure various aspects of the sound in which the container has been all different shapes and sizes---none of which tried to copy the human oral cavity and throat.

Rather than address complex questions I am going after simple ones---such as:

-Does adding weights to the base of a neck produce measurable changes in the sound spectrum?

-What effect on the pitch and sound does closing the key that vents a note 1 mm, 2 mm, 3 mm etc.

-Are there any measurable difference in the tones of a lacquered saxophone, and the same saxophone with the lacquer removed?

A good analogy might be when the acoustic scientists measure the impedance of an instrument. There is no attempt to relate those measurements to what the player does or experiences. It simply reflects what the harmonic acoustic tendencies of the instrument are at various frequencies.

John

[MartinMods]

John,

OK. Just wanted to know. I'll be interested in reading about your results. All the best with it.

Lance
MM

[LampLight]

Wow John. That is really quite impressive. One suggestion: Why don't you pump the air out of it and see if you can discover the electron. - No wait - that's already been discovered.

Kidding aside: Despite any imperfections that arise from having a less than perfect model, I would expect the results to be useful and interesting. Science is never perfect anyway. We do the best experiments we can manage and often have different opinions about how much weight to place on the results.

Kudos to you John.