Spectra question

Talk about musical instrument construction, setup and repair.

Moderators: kiwigeo, Jeremy D

Post Reply
Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Spectra question

Post by Jim watts » Sun Jan 20, 2019 6:47 am

I'm currently working on a guitar and have a questions concerning the dip in the spectrum at 201Hz. I was anticipating the top to be there. The guitar is nearing completion, It's a fan braced classical, without a bridge at this point. I'd like to pull everything down a bit.
So, what is it and why does it do that?
mb sprecta plot.jpg

User avatar
kiwigeo
Admin
Posts: 9588
Joined: Sat Sep 29, 2007 5:57 pm
Location: Adelaide, Sth Australia

Re: Spectra question

Post by kiwigeo » Sun Jan 20, 2019 3:21 pm

Hi Jim,

Could you fill us in on the equipment you're using to record your spectra and the procedure you use to tap test (eg hittin device, where youre hitting the top and where the mic is located). Some info on the software and settings would be useful too.
Martin

blackalex1952
Blackwood
Posts: 605
Joined: Tue Aug 27, 2013 6:36 pm
Location: North East Victoria

Re: Spectra question

Post by blackalex1952 » Sun Jan 20, 2019 5:07 pm

Where is the main back resonance relative to the top? Have you established the uncoupled back resonance and uncoupled top resonances? If the back and top are too close to each other, these kind of things can appear. I am finding that the decision is often best left until the bridge has been glued on, then I look at my data and other tests, eg Chladni testing with the tea leaves or whatever, and then think about whether to adjust the back bracing alone, the soundboard bracing or both. I tend to go for a light bridge so as to keep the inertial mass of the soundboard low, but there is the option of adding mass inside the guitar at the bridge location.
Cheers! Ross
"Everything I say on the topic is based solely upon inexperience and assumption!"

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Sun Jan 20, 2019 5:12 pm

Hey martin,
some addition information;

- I used a Zoom h4 recorder for the mic. It actually has a pair of mics.
- Recorded into Va, fft size of 16384 and sample rate of 11025.
- The guitar was tapped with an eraser on a stick, lightly in the bridge area.
- Averaged 10 samples.
- The mic was little less than a meter in front of the lower bout.
- I blocked the sound hole after I posted this, and re-tapped, the top is at 222 hz. I had stated earlier that I expected this top to be around 200 Hz, that was a different guitar. There's no surprises here on this one, other than the funny spectra.

I'm thinking this is related to my tapping somehow, but I'm not sure. I would like to know what causes this.
Thanks for considering it.
Jim

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Sun Jan 20, 2019 5:15 pm

Ross,
It looks we posted at the same time.
The uncoupled top is at 222 Hz. I had planed to do some top thinning on this one, but this is probably more than I wanted.

blackalex1952
Blackwood
Posts: 605
Joined: Tue Aug 27, 2013 6:36 pm
Location: North East Victoria

Re: Spectra question

Post by blackalex1952 » Sun Jan 20, 2019 6:11 pm

Jim...I am by no means an expert here, but I recall some of my own taps with a similar dip in the spectral data. The dip disappeared after I fitted the bridge. My feeling is that the top, air and back need to end up with a relationship which works (without hitting any scale tones, ie in between the notes frequency wise, to avoid admittance problems ) so that an impulse delivered to the soundboard from the strings causes the guitar to "throw"(my terminology) the air out from the guitar in the low frequencies. This can be felt with the hand, and has a sustained quality about it. After that, the top takes over as a sound radiator, which is where the higher soundboard resonances can be adjusted to achieve that certain "magic"!!!????!!!*** #ahem....Do you have "The GG Books?" There is a section about fan brace adjustments and also reference to the spectral data when top and back are too close together.
Looking at the picture you posted, here is my speculation: Top 222hz which is pretty close to the note A3 (220Hz). Back (240Hz) looks like it is less than a tone difference, ie too close. Adding the bridge along with its extra mass should lower the Top resonance somewhat, thus creating larger difference between it and the back resonance . If you can get your top down just shy of a G there will be four semitones difference.The dip will most likely take care of itself, not sure why either, so a good question in the first place. Then look at the back and aim for the four semitones difference.
Refer to the Gore Gilet build volume section 2.3.10 last paragraph in reference to a classical build.That's if you have those books or access to them. Hopefully Trevor Gore will chime in here! Watch this space!..cheers! Ross
"Everything I say on the topic is based solely upon inexperience and assumption!"

User avatar
lamanoditrento
Blackwood
Posts: 256
Joined: Thu May 05, 2016 9:50 am

Re: Spectra question

Post by lamanoditrento » Mon Jan 21, 2019 10:11 am

I was of the understanding that the dip is a frequency where the top is not excited. So if you were to do clandi patterns, no tea leaves would be hopping at 201hz. I may be wrong on this but the reason it is so low is two shouldering peaks are coupling to push the dip lower. On the higher side the the harmonic of 105hz at 210hz not sure what would be before it at 195Hz.
Trent

User avatar
kiwigeo
Admin
Posts: 9588
Joined: Sat Sep 29, 2007 5:57 pm
Location: Adelaide, Sth Australia

Re: Spectra question

Post by kiwigeo » Mon Jan 21, 2019 9:28 pm

Yes I'd certainly get some tea leaves on the job.
lamanoditrento wrote:
Mon Jan 21, 2019 10:11 am
I was of the understanding that the dip is a frequency where the top is not excited. So if you were to do clandi patterns, no tea leaves would be hopping at 201hz. I may be wrong on this but the reason it is so low is two shouldering peaks are coupling to push the dip lower. On the higher side the the harmonic of 105hz at 210hz not sure what would be before it at 195Hz.
Martin

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Tue Jan 22, 2019 5:09 am

Thanks for the replies.
This guitar is being built by a student over a long time period, a few years. I was just going over my notes, it appears we wanted to leave the top thick so we could bring down the resonances. The idea being he'd learn something from that exercise (me too of course) and he'd potentially end up with a better guitar as a result. It's just going to be more thinning than I wanted.
I do have Trevor's books and they are a great resource. I can't say how much I appreciate the work Trevor and Gerard Gilet put into these books.
So... it seems like perhaps there's a dipole being set up and causing the dip if I'm tracking with Trent's post? Maybe due to my tapping skills, or lack of. Perhaps I was too far off the center line of the guitar. I'm pretty sure I was only a couple of inches off though.
I currently don't have a signal generator to run the Cladni patterns, as I'd have to haul my desktop out to the shop. I need to get that set up again. I believe Trevor had some plans to build one, I need to do that or get a new shop computer or something...
Jim

User avatar
lamanoditrento
Blackwood
Posts: 256
Joined: Thu May 05, 2016 9:50 am

Re: Spectra question

Post by lamanoditrento » Tue Jan 22, 2019 8:50 am

You can use a online generator on your phone like this http://www.szynalski.com/tone-generator/ or one of the million apps available to a blue toothspeaker as a stop gap. I find they work ok if you have a loud BT speaker and or on the lower frequencies like you are looking at.

I don't believe it the dip is because of your tapping. I don't have my books with me but I think it is something to do with coupled phase of two resounaces. Each peak has a positive and negative side and if you get two negative together they drag down the interveening frequency. I think it expains it in the black book where Trevor is modelling the FRC.
Trent

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Tue Jan 22, 2019 10:16 am

Thanks Trent, I'm one of those very few anti smart phone guys.
As I was thinking about it a little more this seems more than likely a phase thing with the back. I'll check the black book for some more info.
I'll try retapping in the next day or two and see if I can get anything different.
Jim

blackalex1952
Blackwood
Posts: 605
Joined: Tue Aug 27, 2013 6:36 pm
Location: North East Victoria

Re: Spectra question

Post by blackalex1952 » Tue Jan 22, 2019 1:06 pm

If you are using Visual Analyser it has a signal generator in it! You just need to get the signal in to an amplifier via a lead and use a cheap 4 inch speaker in a piece of poly downpipe with a cap to seal the end.
Cheers! Ross
"Everything I say on the topic is based solely upon inexperience and assumption!"

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Tue Jan 22, 2019 3:12 pm

Ross, I just trashed my shop computer and my old amplifier a few months back, cleaning up the shop! Oh well :cry:

User avatar
Trevor Gore
Blackwood
Posts: 1408
Joined: Mon Jun 20, 2011 8:11 pm

Re: Spectra question

Post by Trevor Gore » Tue Jan 22, 2019 9:03 pm

Jim,

Check out Figs 2.4-13 and 2.4-14 and captions. Then tell me what you think!

Cheers, T.

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Fri Jan 25, 2019 4:25 pm

Trevor, Thanks for the references. I've been pretty tied up with some other work stuff, so I'm still chewing on this. I'll get closer to having this figured out in a few days, I think.
Jim

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Sun Jan 27, 2019 4:24 pm

Well, I spent part of the day trying figure this out and I believe I made some progress. Looking at Trevor's references of section 2.4.2 it appears that frequencies can interact with one another depending on their phase and strength. I won't bore you with a bunch of spectra plots unless you want, but I took a the spectra with the sound hole plugged and I still had the deep trough. I was a little confused, I should note that I can be easily confused at times!
Long story short it appears to be a reflection off a back wall interacting with the top. When I originally took the spectra the mic was about 18"or so from the wall. I reoriented everything so it was about 2M from the wall behind it and I placed some pillows behind and on the sides of the mic.
After verifying this multiple time going back and forth this what I got with the new orientation.
MB guitar reoriented spectra.jpg
The trough to the left of the top (T1,2) is now gone, but oddly enough everything is a little higher in frequency too. I'm not sure why that is, ( see more confusion :) )
If I appear to be off track on this, please speak up!
Moral of the story, watch your environment and look out for reflections. It changes your spectra.

User avatar
Trevor Gore
Blackwood
Posts: 1408
Joined: Mon Jun 20, 2011 8:11 pm

Re: Spectra question

Post by Trevor Gore » Tue Jan 29, 2019 12:18 pm

That's good progress, Jim and the FRC looks a lot more real than before.

What Figs 2.4-13 and 2.4-14 show is how the FRC changes when the amplitude of different "sources", in this case the sound hole and the top, change in relative magnitude. Normally, the T(1,1)1 is lower in amplitude that the T(1,1)2. Your original plots showed the T(1,1)1 as larger in amplitude and I assumed that this was due to positioning the mic too close to the sound hole. This would give the deep trough effect. But reflections could also cause that. However, the original plot also does not look very much like a guitar plot and could well be because the signal amplitude is too much (mic placement again, or just too much gain on the mic). One of the other "standard" problems is picking up too much extraneous noise from the environment, either electrical or acoustical. Computer fans can be a large contributor to this. Having the mic acoustically isolated from the computer helps a lot, like just placing the mic on a nearby chair rather than on the desk with the computer.

I would not worry too much about the changing frequencies you mention. I suspect that either too much gain or extraneous noise was distorting the signal and that what you have now is much closer to reality. To be sure, check your mic gain (record e.g. some music and listen to how it plays back, or just knock the mic gain right down and bring it up slowly over a series of taps and see when you have enough level to get a reliable plot).

Good luck!

Jim watts
Myrtle
Posts: 64
Joined: Thu Oct 10, 2013 3:37 pm

Re: Spectra question

Post by Jim watts » Tue Jan 29, 2019 3:55 pm

Trevor,
The amplitude difference between the t1,1 and t1,2 was my first thought too after rereading the section. With the set up in the original configuration, I held a book (your book BTW) 5-6" in front of the sound hole to know the amplitude of the t1,1 down. But I still had the deep trough, so I went searching some more. It's encouraging to hear my first thought was also on track.
Here's the spectra with the sound hole obstructed.
sound hole obstructed.jpg
sound hole obstructed.jpg (105.43 KiB) Viewed 486 times

Post Reply

Who is online

Users browsing this forum: No registered users and 24 guests