Collecting Spectrographic data

[Dominic]

Thanks Trevor, I have been dropping the data into notepad then excel. Its definitely as I described, the highest peak is T(1,1)2 in the chart but reverts to T(1,1)1 if I touch the data or when I export it. I am using the latest version of VA and it seems a bit buggy. Won't save my settings and if I try to open the saved ini file with my settings it crashes. I'll try version 9.
And I’ll try pointing the mic at the bridge area.
Cheers
Dom

[Tphil]

I don't suppose anyone knows of a MAC analog of VA??

[Trevor Gore]

Sorry, Tphil. I'm one of Bill's slaves.

You might want to try this on the main forum, as somebody will have come across a solution.

[J.F. Custom]

Yeah... the solution being "Parallels" !!

Sorry Tphil, I didn't even check as I had parallels set up already. If it helps, VA works well within windows in parallels.

I'd be interested if anyone has found a Mac equivalent. Quite sure it would exist.

Jeremy.

[kiwigeo]

I had a good look around a few months ago for decent Mac spec analysis software and couldn't find much at all.

[kiwigeo]

I had a good look around a few months ago for decent Mac spec analysis software and couldn't find much at all.

Many months later and Ive given up looking for a Mac spectrum analyser as good and as cheap as VA. Ive given in and today fired up VA on the mac running via Parallels. Seems to run ok but is a p in the a having to crank up snotty old windows just to run one application.

[Allen]

I had to set up Bootcamp and Windows 7 to run some CAD software on my Macs. I'll be absolutely stuffed trying to figure out why anyone would continue to use Windows. There is no possible way I'd use it unless it was the only option.....and unfortunately for the time being it is for a couple of programs. I feel your pain Martin.

[Bob Connor]

I rest my case your honour.

[J.F. Custom]

Yep Martin, that's the way I went at first too.

It works ok - The latest Parallels is at least without the issues some of the earlier versions had.

But I agree, very irritating to boot that up just for the one thing...

So in the end I just bought a cheap old PC (less than $100) to keep in the workshop. Loaded it up with VA, Strobosoft and trial frequency generators thus far.

Means the Mac is no longer burdened by Windows, nor exposed to any potential dust... Just as it should be.

Jeremy.

[kiwigeo]

The latest version of Parallels actually runs surprisingly smoothy and Windows 7 seems to run at an acceptable pace with only moderate allocation of Mac resources.

Jeremy....I will most likely do what you did and get a hold of a cheapo windows laptop and run all the software on that machine.

[Graham Long]

Hi Trev,
I haven't built any guitars since attending your course in Canberra last year.
I've just finished building a new shed and now I'm back building.

Just a question about tap testing, which peak do I record. I can find peaks that line up with the sample data on you spreadsheet, but they aren't the largest peaks or a fundamental.

Cheers
Graham

[Dominic]

Hey, remember from the course, we know generally where the peaks for long, cross, and twisting should fall. Its in the book and the course material. And then move your finger hold and tapping around so you accentuate the peak that falls in that range. Light tapping I found to work best at isolating the peak you are looking for.
Cheers
Dom

[Graham Long]

Hi Dom,
If the typical frequencies weren't in the book how would you know which one to use in the modulus calculation?
Cheers
Graham

[Trevor Gore]

Graham,

The general rule when tapping wood for material properties is to look for the largest amplitude lowest frequency peaks for each of the modes you're looking at. If you're using "standard sized wood" e.g. a typical spruce half top blank ~4mm thick, these peaks should correspond to the ranges noted in the book. If they don't, it is indicative of some sort of problem. Possibilities are

1) Your frequency meter/spectrum analyser not set up properly
2) Your microphone not interfacing properly
3) Your microphone not having a very good low frequency response
4) Your holding, tapping and mic positions not commensurate with the mode being measured
5) Occasionally, interference of one mode with another when all the above are correct. This is due to there being a particular relationship between the aspect ratio of the wood and ratio of the long and cross Young's moduli.

The easiest to get right is the long grain marimba bar tap mode. If this isn't coming out in the ball park you need to check through points 1 to 4 above. If that doesn't sort things, get back with more detail of what you're doing and what's going wrong. However, maybe before you do that, try tapping, say, a fretboard blank and see if you can get anything useful out of that.

[Nick Payne]

I had a good look around a few months ago for decent Mac spec analysis software and couldn't find much at all.

Audacity is available for Mac. I haven't used it there (don't have a Mac), but I've used the Linux version of Audacity for plotting the spectrum of a tap test. I use either a Zoom H2 recorder or (lately) a smartphone to record the taps, then load the file into Audacity and choose Plot Spectrum from the Analyze menu.

[Graham Long]

Hi Trev,
Thanks for the reply, I've been using VA and the microphone in a Logitech webcam, maybe not the best quality, but i did about a dozen or so soundboards the results all looked credible until I did a set where one half had already been thicknessed down probably too far. When I did the tap testing the target thickness can out even thinner, and the target thickness of the other half of the set coming out at believable value.
I was thinking about thicknessing the thicker half and testing as periodically, should I be able to the get the point where the target thickness and the actual thickness are the same?
Cheers
Graham

[jeffhigh]

Some microphones just won't pickup low frequencies well and if you use them on a plate that is already thin you will not register the fundamental note.
This will obviously throw out your modulus and plate thickness calculations.

Paul (ozwood ) had this issue at first with his mic, and plugging in my kickdrum mic made a huge difference.

[kiwigeo]

I do my tap tests using a Shure SM57 microphone running via a preamp into my Mac. I'm running VA on the Mac via Windows and Parallels.

I tried Audacity but my preference is for VA.

[Trevor Gore]

I've been using VA and the microphone in a Logitech webcam, maybe not the best quality, but i did about a dozen or so soundboards the results all looked credible until I did a set where one half had already been thicknessed down probably too far. When I did the tap testing the target thickness can out even thinner, and the target thickness of the other half of the set coming out at believable value.

Graham, that sounds like it could be a mic problem along the lines that Jeff mentioned.

I was thinking about thicknessing the thicker half and testing as periodically, should I be able to the get the point where the target thickness and the actual thickness are the same?

I've never tried that, because I want to get to the target thickness after I've joined the two halves and then cleaned up the joint. I usually loose a bit of thickness doing that, never mind more thickness loss cleaning up the rosette and general finish sanding. However, if you have all the equations set up right, it should work. The reason for this is that the first part of the calc is working out the elastic constants (Elong, Ecross and G) which should come out right irrespective of the thickness of the wood when you tap it. Then using those material constants (which should always come out the same) you should get the same target thickness. So even if you are too thin already, if you can measure the frequencies accurately and therefore get the elastic constants right, the second half of the calc should tell you that you should have left it thicker!

I've measured a number of half plates on different occasions as I've been working them down and always got elastic constants that closely agreed (and hence the same target thickness), but I've never taken the wood close to final thickness for the reasons above.

So, all in all, I'd be suspecting the mic. I'm using a Shure PG 57 (cheap version of the SM 57) plugged straight into the mic input of my Toshiba laptop. The input is "biased" for a electret mic, which gives a much higher output than the PG 57, which is a dynamic mic, but if you use the mic boost feature in the Volume Control set up, it works fine.

[kiwigeo]

So, all in all, I'd be suspecting the mic. I'm using a Shure PG 57 (cheap version of the SM 57) plugged straight into the mic input of my Toshiba laptop. The input is "biased" for a electret mic, which gives a much higher output than the PG 57, which is a dynamic mic, but if you use the mic boost feature in the Volume Control set up, it works fine.

For those of you using a Mac the signal produced by going straight into the mic input wont be high enough...youll need a preamp in most cases. The other alternative is a usb mic.

[jeffhigh]

I use a dynamic mic through a xlr to usb adaptor cable

[Graham Long]

Hi Trev,
Thanks, I'll upgrade the mic and try again.
Cheers
Graham

[kiwigeo]

I use a dynamic mic through a xlr to usb adaptor cable

Jeff are you running a Mac?

[jeffhigh]

No- a windows machine

[Barry Daniels]

I am trying to get my setup working properly. I have tried both versions 9 and 11 of Visual Analyzer. They seem to be working OK and I am getting repeatable and reasonable results so that makes me fairly confident. But there is one aspect I am not really understanding concerning the buffer. I have set all the settings to the recommendations of the book, Trevor's supplemental document on his website, and several postings here. But it seems that the recording length (if that is what you can call it, or perhaps the buffer) is only set by the setting the Capture Scope. I first set this to 10 (as well as the Capture Spectrum Buffer set to 10). So the total length I get is only 10 seconds which is insufficent to do 10 taps if I wait 1.67 seconds between taps. Also, I do not see anything "refreshing" to time my taps to. I just end up counting "one thousand one, one thousand two", and then tap again.

Any further explanations on this issue would be helpful. Thanks,

Barry