Today I Learned: Tornavoz

Okay, I'd been thinking about this sort of thing for a while, but didn't know it had such a cool name (that I'll never remember).

Background: I like playing Low G. I started on a Tenor, and even then I found Low G tuning to be "boomy". I've since experimented with a Concert, and found that the fretboard size suits me better... and, of course, the body is even smaller, to the Low G boom is definitely not any better. I'm forever studying ways to improve this problem.

Okay, so if you know anything about Helmholtz resonance, you'll know that the body of any closed-body stringed instrument (such as a ukulele) is effectively a Helmholtz resonator. That simply means it's a cavity that has an opening in it, and it happens to resonate at a particular frequency. Sort of like blowing into the opening of a bottle... hit the right frequency, and it's going to resonate.

For the ukulele, that resonant frequency is normally intended to be just BELOW the frequency of the lowest note that the instrument can play. For a Concert uke, that note is C4 at a frequency of about 261 Hz.

When you string said Concert uke with a Low G, the low note is now a G3 at about 196 Hz.

So, there's our problem. We have an instrument that wants to resonate at something less than 261 Hz (maybe 250?), and that's why all of those notes below C4 tend to be varying degrees of "boomy". They're in that "diva" resonant range of the body and sound hole, and they want to go all Celine Dion on us and take over everything else.

I've studied Helmholtz resonance for other purposes, so I already knew that there's another parameter to the resonator. It's not JUST the volume of the chamber and the area of the opening... it's also the length of the tube created by the opening, which is called the "port".

I've always wondered if this could be leveraged on the ukulele, but never got around to researching it. Today I did, and I found that term: Tornavoz. (and I'll leave reading about the history and origin of that as an exercise for the reader) It's EXACTLY what I'm talking about. Basically putting a "neck" on the "bottle" that is our ukulele body. The neck is inverted, it goes inside the hole, but that's basically what it boils down to. There's even a cheap (but, kinda fancy, looks like a velocity stack on a race car intake) piece available for acoustic guitars called an O-Port that would be a quick and easy way to add one... sadly, the smallest size is 3.5" diameter, way too big for a uke.

Now that I know that this is "a thing", and it seems to do exactly what I'd expect it to do... makes the bass sound better by getting it out of of that resonant frequency range... I'm ready to experiment!

First, I looked up a Helmholtz Resonator Calculator (because, why math if you don't have to?) This one seemed to get the job done. (it's actually the second calculator on the page)

I plugged in some estimations of the size of a Concert ukulele body and fudged it around to get the result of 260 Hz. (H=26cm, W=12cm, D=8cm, Port Dia=6cm, Port Ln=0.2cm) The exact dimensions aren't critical, it's the resultant volume of the chamber that's critical. The port is relatively accurate to the uke (60mm Dia, 2mm top), so it gave me something to work with.

Changing just one parameter, the length of the port, from 2mm to
4.45cm (2.75"), changes the resonant frequency to 192 Hz, which is a tick below our target "Low G" note of 196 Hz. Cool, the math works.

Now, what have I got laying around the house to experiment with? Cardboard tubes? Nope. Too small or too big, nothing in the Goldilocks zone. Plastic cup? Hmmm... close, but just a little too big. Expired vitamin bottle? Exactly the right size! I even had to peel the label off of it to get it to fit into the hole. (spent more time cleaning the adhesive off of the bottle than anything else!)

After cleaning up the bottle, I cut the bottom off of it, then trimmed it to the 1.75" size (I'm 'murrican, I think in inches). Spread the strings out and shoved this newly repurposed "Tornavoz" device into the hole. And...

It seems to work. The low notes seem to be more tame.

I've got some other stuff to do on my uke soon, and I'll try to maybe get some video (I can't play worth crap, but I might be able to make some sounds and get them recorded) with and without the Tornavoz.

Stupid ukulele tricks.

Physics is cool.

I just thought to try to check the resonant frequency of my uke body by muting the strings and tapping the back of it with the sound hole over my phone with the tuner app running.

With the Tornavoz in place, I got a very consistent 161.3 Hz. Well, that's cool. Now I gotta pull the thing out and see if it's actually different without it! So, I wiggled it out, and checked again. A little less consistent, but always right around 211 Hz.

Now, if I go back to the calculator and massage my body size numbers to get 211 Hz. (which gave the body a volume of 3.8 liters... which doesn't seem impossible) And then change the port to 4.45cm... I get 155.13 Hz.

And if we adjust that down to 4.13cm (because my cut isn't perfect, and it's a little less than 1.75"), it's 157.89 Hz. And that's pretty darned close for a hand-cut medicine bottle jammed into a ukulele sound hole!

Really glad I didn't have to do all the calculations by hand, but physics is cool!

And I just tried to record some comparison sound samples. You'll have to take my word for it, there is definitely a difference. To my weak ears, just plucking individual notes doesn't sound a whole lot different. May be slightly less volume with it in? Hard to say. Maybe a little less umph behind the bass notes? Again, hard to say. Maybe even slightly less sustain to the bass notes? But, it might just be that the volume is slightly less.

But... when you strum some chords that have those bass notes in them, THERE you can tell the difference. The bass notes go from being very much in the foreground "muddying the water" to being more at the level of the other notes and blending in. Otherwise, things sound very much the same.

I was recording with my phone, which has an excellent camera. But, apparently, not so excellent microphones. There's a background hiss, sound isn't great, and a ton more excuses. I just don't feel that the sound is of sufficient quality to bother posting.

Oh, geez. I need to quit geeking out on this. More info pulled from that calculator page.

The resonator is, if an electronic equivalent makes sense to you, sort of like a band-pass filter. It gets happy and resonates not just at "a frequency", but also works efficiently within a range of frequencies. There's a "bandwidth" to that filter.

Guess what the bandwidth of my "estimated" Concert ukulele body is? 246 to 668 Hz. That's B3 to E5. Genius! In this case, making the "port" taller (thicker top) even just to 3mm greatly reduces that bandwidth, and making it shorter (say... shaving the area immediately around the sound hole to 1mm) greatly increases that bandwidth AND raises it to over 1 kHz. The sweet spot seems to be right about 2mm. Going to even 1.9mm gives a bit more high end range, while cutting the bottom end off at 283 (C#). Pretty amazing just how "right" the typical ukulele is designed. Who knew?

So, with the Tornavoz in there... yes, the resonant frequency gets down there 50 Hz below our Low G note, which is great. But, our bandwidth? It becomes laser focused! 154-161 Hz! So, the body won't even get into that "sweet spot", and that's why it sounds quieter. The body isn't getting fully excited.

I'm going to leave mine in for a while. Most of the playing I do is quiet practice, anyway. But, man, this is neat stuff!

Very interesting!

The boominess often keeps me from playing my low-g strung uke.

Hey LaurenFL, I'm enjoying this. Please keep geeking out. My Romero concert has a low g and sounds great. I'm guessing it's because of the greater volume (deeper bout than most concerts) makes it resonate at a lower freq. Perhaps I'll plug in the numbers for a few of my concerts - as soon as I read the article and figure out what the two ports measurements (port ln.) are measuring. Also a double bout must cause a problem or two in all this.

Interesting info.
If you really love experimenting, try just taking some non-marring tape (painters tape, masking tape) and cover over the sound hole in your uke. You can also maybe use of a thin piece of cardboard and tape that over. If you really want to go nuts, cover only a portion of the hole.

Here is an experiment Yamaha did:
https://www.yamaha.com/en/musical_instrument_guide/acoustic_guitar/mechanism/mechanism005.html

Then there are side sound ports that are added to some instruments.

I'm glad some of you are getting as much amusement/enlightenment from this as I am!

I've done a bit more reading here and there, and there's definitely more involved in the sound of an acoustic stringed instrument than just the Helmholtz resonance. The body itself is not fully rigid (of course), it has its own resonance that must be considered in addition to the resonant chamber. And, as I think someone mentioned, the shape of the body is also probably a factor. There's a lot going on.

But, I still like tinkering.

I really need to do more with/without testing to be sure I'm hearing what I think I am. But, plunking around on it tonight, I feel like the lows are a little "flatter" (less boomy) and the highs and mid-highs (A and E strings, but also some on the C) seem to have a "sweeter" sound. I made a change to my nut and improved the intonation a few days ago, though... so, I could still be hearing some of that change.

And yet... I just ordered something to make another Tornavoz out of. They make silicone couplers for automotive intakes. I found one that adapts 51mm to 76mm ID. The wall thickness is 4.5-5mm. So, the OD should be really close to 60mm. If it's fat, it'll be snug (it's silicone, so it's flexy). If it's skinny, I can wrap some tape around it to make it snug. But, with a little trimming, this gives me a Tornavoz that's more of a bell shape with a 3" mouth feeding the 2.25" sound hole (which is being reduced to a 2" sound hole). As a plus, it won't add it's own resonance like my plastic pill bottle does. I think I can hear that plasticky sound vibrating on certain noise. (I actually stuck my finger down there to feel it, and it definitely vibrates on certain notes) Silicone won't do that at all, it will redirect air without adding it's own resonance to it. (and will probably damp the movement of the top a little bit)

We'll see how that goes. My wife thinks I'm nuts. It all sounds the same to her!

This is cool. I love trying nerdy stuff like this.

If you find a workable design then you might want to look at 3D printing. Different density and elasticity polymers available and it could be custom made to fit your instrument.

Super interesting, carry on! I’m glad I don’t, just now, feel the need to follow your lead in this experiment, as there are Christmas presents to make, but it’s fun to hear about. So far I’m happy to just not hit the G string quite as hard as the rest when strumming.

Yeah this is fascinating. I have sort of become obsessed with the fact that almost every uke I've played seems to have a range of notes (say spanning 2-3 frets) where the notes sound "off"; Very little sustain or just an unpleasant sound versus a clean pure tone. Last year I returned a K-brand tenor ($2,000 instrument) that I otherwise loved except the E and Eb on the C string sounded terrible to my ears. Other octaves of these notes had the same issue. I've had email discussions with various players and Andrew at HMS about resonant frequency of uke bodies and how they in turn create wolf notes or dead notes. I was somewhat heartened to learn the issue wasn't a 'flaw' but it left me wondering why builders of great ukuleles don't seem to have solved it, or maybe I just haven't found the right builder.. So yeah I'll look forward to hearing about your further experiments

Yeah, I get those dead notes, too. I should probably try to document where they are and see if this mod changes them at all.

LorenFL said:

Yeah, I get those dead notes, too. I should probably try to document where they are and see if this mod changes them at all.

Click to expand...

That would be quite interesting! Thanks

If I understand your setup for this, the cylinder is held in the hole by friction? They do not touch the back? So the vitamin bottle and the silicone tubes add mass to the top and change the way it vibrates.

Plus it adds its own vibration to the back and top vibration. With lesser side vibrations.

The question seems to be, how much of the change is due to resonance the tornavoz and how much to a dampening of the movement of the top by adding mass and the movement of the tube?

Really interesting stuff. Thanks for posting it.

The current setup, the bottle tube weighs next to nothing. It's about an inch from the bottom of the uke, no contact. I don't think it's having a significant effect on the top, but... I guess it doesn't take much to make a difference. I think that being so thin and fairly rigid, it does introduce some of its own vibration/resonance at certain frequencies. (and I doubt that anyone but the player can hear that... and only when playing softly, as I do) Otherwise, it's mostly Helmholtz body resonance at work.

The silicone hose will have more mass, but still won't contact anything but the soundboard hole.

In either case, I'm going for friction fit to the sound hole. It's just simpler that way.

Part of me wants to buy an even cheaper uke... plug the sound hole, and put like a 1" port on the side with a long piece of PVC pipe to lengthen the resonator port. Did Blue Man Group ever do a stringed "tubulum" instrument? Would be fun to take things to extremes just to see what happens.

Okay, let's try to get all experimental here.

I just put a new compensated saddle on my uke, and the intonation is nearly perfect. E string is slightly sharper than the rest, but we're talking a couple cents at the first fret, and maybe 4-5 cents at the 12th. Maybe I'll revisit that after I put new strings on. I bought a set of 5 pairs of spare nuts and saddles, I can play.

But, that's not what we're here for... I'm all tuned up and I want to make some baseline observations. I'm just going to play each note, and also play some chords and see if I can identify any repeatable areas of concern. Then I'll have a "hit list" of things to check with the cheap Tornavoz in place.

Hey, I've got a sound meter app on my phone! I can get dB readings from a fixed distance. And it also shows me a running graph of the sound level, so I can measure sustain to some degree.

Here we go!

Ambient noise was about 33db. What I'm calling sustain is what I could readily see above the noise on the graph. Audible sustain was longer than that in all cases, but I think this is less subjective, and there are obvious differences between notes. I decided to also include a subjective "quality" for each note, which is more about how the note resonated and sounded.

I'll attach the spreadsheet when I'm done for anyone who's interested. I'm having a hard time deciding what's important enough to put on a chart with just the baseline data. Maybe with before/after data, I can do a comparison chart of something useful.

Some interesting info about my uke from the baseline data:

Measuring from maybe 18" from the soundhole and off to the side. (I'm holding the uke seated, phone is 3" from the middle of my left thigh)
Peak sound from a healthy pluck of each string ranges between 57 and 70 dB. 60 dB is considered "conversational level", 70 is close to your vacuum cleaner. A change of 3 dB is said to be the smallest difference we can perceive, 10 dB will be perceived as roughly twice as loud.
D# up to G were the only notes louder than 65 dB.
Anything below 62 dB was generally on the A string, or above the 5th fret of the E string.
So, MOST of the notes are in the 62-65 dB range.

Sustain was measured roughly in seconds by looking at the graph. (it gave me bars at 5 second intervals, so I could estimate 1, 2, 2.5, 3... or a measly half) You can probably double all of those values as to what was PERCEPTIBLE, this is just a rough measurement from what was visible.

Sustain of 2 seconds or more happened on almost all notes at or below the 7th fret. The exceptions all being on the E string.

Really poor sustain of 1 second or less was mostly above the 9th fret. Again, except the E string.

Really great sustain of 3 seconds or more was on open strings (except E) and G string up to fret 5, C string up to fret 9.

Subjectively, my "quality" grade mostly followed the sustain. With only 3 exceptions, anything with 2 seconds or more sustain had a good resonant sound. Below that tended to sound somewhere between "dull" and "thunky". The only open string that really sounded great was the A string. E was a dud. C seemed harsh. G was boomy.

Subjectively, common chords that left the G string open, or fretted at the first or 2nd fret had a boomy low string sound. A bit less so on the 3rd fret, almost gone at the 4th fret. Running the C7 shape (or better, a C7Sus4 shape... beautiful sound) down the fretboard is a good test to show this.

Same experiment, medicine bottle Tornavoz installed. It's roughly 1-5/8" long, 60mm diameter. About 1/8" or so sticking out of the sound hole.

Doing the chord comparison first, because I JUST did it without the Tornavoz, and things sound quite different. First thing I notice is that the overall sound in general is lower. I think the Low G boom is still there, but it's more subdued. I'd say it seems to have moved up about two frets. Better on the open string, still there on the first fret, pretty much gone at the 2nd fret. (anecdotally, I remember reading reviews of the O-port that said it dropped the range of acoustic guitars by 2-3 semitones, too -- so, that seems about right)

Let's see what the notes do...

Ambient noise is down to around 28. I guess the AC is off and the fridge isn't running. Soooo quiet.

And the results are in!

Definitely quieter, 58 to 66 dB. (one string was 55, one string was 67) The majority were in the 61-65 dB range. I'd say what it lost in volume, it picked up in sustain. (which makes sense... the energy has to go somewhere)
The notes above 64 dB were C# and above.
Anything below 61 dB was generally on the A string, or above the 5th fret of the E string, but also some G this time. (again, more sustain, less volume)

Sustain. Wow! Didn't really put it together until I started measuring and really listening. The first note I hit was the Low G, and it rang for FIVE SECONDS on the graph... much longer by ear. It ended up being the best by far. But, WOW!

Using the 2-second metric, the dividing line is closer to the 9th fret now. Most everything 9th or below is there, and the exceptions seem random.

The differences show up more in the formerly poor sustaining notes. There are much fewer of those. And for the most part, even the lesser-sustaining notes (looking at the graph) still SOUNDED better.

Which brings us to the subjective "Quality". The audible sustain was much greater than the measurable sustain. I think that's what the characteristics of this particular Helmholtz resonator should do. We changed it such that the "q factor" went from less than 1 to over 20. It's related to that.

As before, measured sustain of more than 2 pretty much guaranteed a good, resonant sound. But, a lot of the notes with really poor measured sustain actually had a good resonant audible sound, and only one sounded downright "thunky".

Where I rated 29 notes as "resonant" or better before, that number is now 39.

Bottom line: It seems to even things out. It takes away some overall volume, lowers to overall tone, and greatly improves overall resonance. As long as your goal isn't to be "as loud as possible", these all seem like good things to me... for someone who wants to play Low G.

Looking for something to plot between the two datasets. Looking at dB, the trends are the same. The peaks and valleys are in the same places, nothing to see there.

Graphing the measured Sustain is surprisingly similar, too. The trends are almost identical, there are just a number of peaks that are higher with the Tornavoz. Not much to see there.

Graphing the subjective Quality showed the most, and I don't like that. It wasn't all that dramatic, either, so I'm not going to bother saving the chart as an image to show you. Again, the peaks and valleys are more or less in the same places. But, with the Tornavoz, the peaks are wider and more stable, and the valleys aren't as low. Interestingly, not only did it make the low end lower... it seems to have also made the high end higher! It's like it spread out the resonant range of the instrument. Same valleys in the response, just a little narrower and spread slightly further apart.

I guess I shouldn't be surprised. So much of what we hear is subjective and qualitative. That's why nobody can tell you what strings to put on your ukulele, and everybody has such trouble DESCRIBING the qualities of sounds. "Mellow", "Harsh", "Warm", "Soft", "Sweet", etc, etc.

Will be interesting to see what Tornavoz Version 2.0 does. But, unless it's dramatically different, I'm not sure I'll bother going through all of this testing again!

Spreadsheet attached, if anyone wants to dig into it. It's an Excel 2000 spreadsheet. Yes, I'm still using Office 2000. Amazingly, it still works.

Okay, I'm going to wrap this up for now.

My silicone part came in today. If you forgot, it's an automotive intake coupler, 51mm to 76mm ID, 5mm wall thickness.

It had a lot of excess length, being longer than the depth of my ukulele. So, I started by trimming down the narrow end of it to where there was about 1/4" before it started flaring out to the larger size. Thinking that I can always cut more, but never cut less... I tried it like that, even though it was still quite long. There was at least 1/4" of open space to the back of the uke.

Before shoving it inside, I checked fit from the outside. It went in without much trouble and was the nice friction fit that I expected. So, I folded it and stuck it in the hole. Then spent the next 15 minutes fighting with it trying to get the end of it to pull through the hole from the inside! It just didn't want to go. And part of the reason was that the bottom of it was hitting a back brace. So, I tried to pull it back out. But, there's no way to get enough fingers in there to fold the thing to pull it back out the way it went in. It's soft and pliable, but still a bit thick and wants to hold its shape.

Ended up sticking my scissors down the hole and cutting it down the side. That allowed me to easily roll it up and pull it out.

Then, I trimmed the fat end off all the way up to where it starts to "bell" to the smaller end. The resulting piece is about 1-3/8" long. Rolled it up, stuck it in the hole, pulled the narrow end through, and manipulated the cut ends back together... and it all fell into place. Still a nice friction fit, and it sort of holds itself together. It ain't going anywhere, and the end being bound together by the hole holds the whole cut seam together nicely.

But, how does it sound? Well, I messed around with it for most of the evening. It's got a lot of the same characteristics of the other version. Lots of sustain, but even less volume. I didn't take any measurements, it just seemed very flat. Sort of like an old AM transistor radio in a way. Some things sounded pretty good, a lot of things didn't. Not sure if it's because of the extra weight hanging on the sound board, or just the nature of a thick piece of silicone acting like a damper.

Pulled it out, and confirmed that yeah, it really sounds a LOT better without it.

Put the other one back in. And it sounds nice, too.

I'm going to deduce that there's a reason the Tornavoz in a vintage guitar is often made of something like brass. It needs to be something rigid and light. My rigid plastic isn't too bad. Would probably be better if it were made of thin metal. Maybe someday I'll experiment with that idea. But, that will take some fabrication.

Overall, fun experiment, and I'm going to live with my medicine bottle Tornavoz for a while... maybe pull it out from time to time to compare. It's easy to add/remove. I may even play with the "damping" silicone version some more. It doesn't sound great... but, probably better than a full sound hole cover. It is nice to be able to play while my wife is watching TV and not really bug her.

Thanks for posting your updates. I scanned it but haven't had time to properly read it.