Help reading flamenco tap tone data

[josswinn]

Hello,

I've made 4 flamenco guitars over the last couple of years and recorded tap tone data, which I could do with some help interpreting.

The data is here: https://docs.google.com/spreadsheets/d/ ... sp=sharing

You can see in the four worksheets that I've highlighted the numbers that I think are the T1/2/3 figures, but other than T1, I am not confident I am locating the right numbers. (I've summarised my readings at the bottom of this post).

What do you think?

I've used Audacity and more recently Adobe Audition to get the data. The recording is done on my laptop using the internal mic sat in front of me holding the guitar. Those graphs are produced Google Sheets.

The guitars are all of the same design, based on a Barbero plan, with cypress back and sides and spruce top. #2 and #3 were Engellman spruce and #4 and #5 were Euro spruce. The guitars weigh between 1119g and 1319g, mainly dependent on the neck and fingerboard wood used and whether they are geared or wooden pegs.

You'll see that #5 has a lower air resonance than the rest and I'm wondering why that is. I went thinner on the top for that one (1.7mm) but the bridge was a bit heavier (19g), compared to #4 which was about 2mm top and 15g bridge.

I've made 5 guitars in total (the first was a lattice classical) and I'd really like to start using this data more productively. I also have Trevor's books so can apply what is in them. His information on Flamenco guitars was really useful in setting the standard I tried to aim for with #2.

Thanks for any help with this.
Joss

#2
T1: 92.18
T2: 191.78
T3: 275.89

#3
T1: 90.17
T2: 191.78
T3: 279.25

#4
T1: 88.82
T2: 175.63
T3: 283.29

#5
T1: 82.76
T2: 182.34
T3: 291.34

[blackalex1952]

Computers have inherent noise, eg fans...so best to use an external microphone and soundcard. I had better, clearer results when I switched to external microphone. Chladni testing is also a useful tool. 1) can see the nodal points for each mode 2) the modes are easily recognised 3) resonant frequencies as read by the computer testing can be confirmed. I take iPhone shots of my Chladni results with a piece of paper in the shot which has the frequency of the mode written on it. I test both the back and the soundboard. Bear in mind that the guitar which sounds the nicest to you is the one which sounds nicest-to you! Cheers! Ross

[Dave M]

Hi Joss, you've certainly been busy since that first build with Roy Courtnall!

I won't comment on the resonant frequencies as there are people here much better qualified. There is quite a lot in the G&G book but I at least find it a bit hard to retain it all. One of the standout things is to avoid having a resonance exactly on a scale not. I tested the box I am currently working on yesterday and the main air is looking a bit close to G sharp. I'll be looking again after full assembly.

I thought I would flesh out Ross's note about Chladni patterns which are indeed very useful, particularly once past the first three main resonances I bought a pair of cheap car speakers and drive one through an old domestic stereo amp that like so many people I was reluctant to chuck out! I use Visual Analyser to generate a signal but there are plenty of other bits of free software around. So the whole thing has cost very little money, and the 'dancing tea leaves' have given friends some great amusement!

I would also agree strongly about a separate microphone. Cheers Dave M

[josswinn]

Thanks blackalex1952 and DaveM.

Chladni patterns, yes, but one step at a time! I'm going to get a microphone. I do have one, but an upgrade to MacOS seems to have stopped it being recognised. The cleanest data I've had so far as been with #5, which was the laptop mic and Adobe Audition. I don't know what went wrong with the recording for #4 but the data is full of noise.

I'm sure about T1 for all instruments. Fairly sure about T2, but really unsure about T3, so if you can interpret the data, it would really help to see if I'm reading it right.

Yes, I've been plugging away at the building since learning from Roy Courtnall. I'm hooked on flamenco blancas, trying to make them as light as I can, though they seem to be getting heavier! After building a laminated lattice classical with Roy, my first flamenco (#2) felt so light, there was no turning back. Any excess weight over 1200g is in the neck and I'm going to be more careful about wood selection for the neck and fretboard from hereon. Wooden pegs help, too, but not everyone wants them. I think the bodies, based on the same plan, are fairly close in weight.

Cheers
Joss

[Alain Lambert]

If you record a tap of the back this will give you an aproximation of where the T(1,1)3 is.

[josswinn]

Thanks Alan. I've still got guitars #2 and #5 so was able to tap the backs of those. I also put all the wav files through Adobe Audition and it's produced clearer data and charts, except for #4, which was clearly a poor recording. I've updated the online spreadsheet.

Anyway, I think the new readings are:

#2
T1: 92.18 F#2 -6
T2: 191.09 G3 -44
T3: 273.85 C#4 -21 (tap to back: 223Hz / A3 +23)

#3
T1: 90.83 F#2 -32
T2: 176.29 F3 +17
T3: 258.38 C4 -22

#4
T1: 88.82 F2 +30
T2: 175.63 F3 +10
T3: 273.20 C#4 -25

#5
T1: 82.76 E2 +7
T2: 182.34 F#3 -25
T3: 291.34 D4 -14 (tap to back: 241.1Hz / B3 -41)

At one point in his book (p. 3-39), Trevor indicates that the T3 for lightly built flamenco guitars might not appear on charts, which is somewhat reassuring given my difficulty locating it.

[Alain Lambert]

I have no idea how a Flamenco spectrum should look. But based on your numbers and the back tap, I would say the T(1,1)3 is that little bump around 217 Hz. It look like the back is not contributing much to the top resonance.

[josswinn]

Thanks for looking at the data Alain, and apologies for mis-reading your name earlier.

Trevor Gore includes a spectrum for a Conde guitar (Figure 3.2-11) and the peaks for it were: 80Hz, 170Hz, and 225Hz. He notes that it was an "excellent sounding flamenco guitar, but this example is rather extreme."

The average figures he gives across six vintage flamenco guitars, possessing the traditional sound, was T(1,1)1: 92.8Hz, T(1,1)2: 183.3Hz, T(1,1)3: 183.3Hz.

Cheers
Joss

[Dave M]

Joss, getting your head round section 2.3.7 could be useful. Trevor shows that you can often get the uncoupled back frequency by blocking the sound hole. This could possibly help to nail down the coupled back frequency if it is showing.

By the way I would gently point out that these sorts of discussions are usually in the Forum topic: Contemporary Acoustic etc. If you haven't browsed this topic there are loads of useful discussions on these subjects.

Cheers Dave

[josswinn]

Thanks Dave. Apologies for picking the wrong forum topic.

I've just spent time in the other forum topic and found Trevor's video course on all of this, which has opened doors. It's making much more sense now. I reckon I'll be building my Chladni set up in the next couple of weeks...

[TallDad71]

Cant help with frequencies I'm afraid but recording buzz free is easy.

FFT from the app store.

https://apps.apple.com/us/app/fft/id298840058

There's a google Play version too.

It generates lovely graphs, allows you to store them and crucially allows you to take average reading over multiple taps.

[josswinn]

Thanks Alan. I’ve just bought the app and can see it’s going to be very useful.

[TallDad71]

Excellent. The other ancillary benefit is that it is also a tone generator. You can plug it into your guitar amp and set it to your desired Hz level and blast away, perfect for Chladni patterns.