having a go at spectrums

[simonm]

I have two classical boxes closed up and will be putting the bridges on in the next day or so. Unfortunately, I don't have my "books" with me and with our current lock-down (Spain) I doubt if I will seem them again for a few months.

I made an attempt at spectrums just now and really don't know where to start. Yes they look like guitar spectrums. The planes roughly the Ramirez 1912 but both tops are still a bit thicker than what I associate with that plan. I reckoned I can always sand it down a bit to adjust the sand a bit after putting the bridge on and stringing up in the white.

The two guitars are an experiment using a commercial kit similar to something I used 10 years ago to see if my results will be substantially better than back then and see if I can get anything remotely like repeatability.

Both euro-spruce. The first one here is other stripy and colourful. The second quite wide grained and white. Both thicker around the bridge and towards the sound hole and thinner on the outer edges. Backs seem to be 209 and 277 respectively.

Any thoughts?

Thank you.

[johnparchem]

They look sort of noisy to me, like maybe your mic is clipping. You might turn down the gain. Because of the extra peaks I can not tell what the back might be.

[Trevor Gore]

I made an attempt at spectrums just now and really don't know where to start...

You can download the VA setup instructions from my website here. That should at least get you started!

There is definitely quite a bit of "grass" on those plots, possibly due to too much gain as John suggested, but also maybe because you are capturing more than one tap per buffer on some occasions. Just one double tap (out of ten taps) will cause quite a lot of extra noise on the plot. Getting the pots to look right is the first thing to do, so that you are sure you can rely on your measurements, as making decisions based on unreliable data is not to be recommended!

The T(1,1)3 frequency is usually just a bit higher than the B(1,1) which is the main back monopole frequency . You can get the B(1,1) by recording ten taps on the back, just like you would one the top. That should help you identify the T(1,1)3 peak, if there is one.

[simonm]

Thanks guys. The microphone is definitely crappy. I have mislaid the power supply for the pre-amp to go with my share microphone. I set up VA according to the instructions.

Having read your comments I tried the crappy mic on a mac (audacity) and a win10 (VA) along with there internal mics. It was quite interesting the see the differences and how the numbers were shifted a bit. I guess serious testing will have to wait till I can get a new power-supply of pre-amp. I will try tapping a bit slower and see if it makes a difference. I will try the gain settings in VA and see what happens.

[Dave M]

Simon I have had trouble getting to a satisfactory state on this. One issue is that you've got gain controls all over the place, on the preamp, on the PC sound card and on VA (at least). Even now I get better results on my desktop than a laptop, using the same (Shure) mic and USB preamp which I can't seem to bottom out.

I have certainly not got anywhere with the internal mic.

Averaging several runs in a spreadsheet does help to reduce the noise, though I don't normally do that.

Good Luck Dave

[Trevor Gore]

Depending on which version of Windows you have, there are different ways to control the gain. For VA 9.0.6 (which I use most of the time) the level controls in VA only work in Win XP. In later versions of Windows, you have to go to the Windows sound controller or other installed app (e.g. DTS Audio Control Panel or similar) or both, to control the level from there.

Win 10 and some versions of Win 7 have "audio enhancements", sometimes quite well hidden. Turn them off! Typically they are on by default and cut out low frequencies or echo cancel, which causes analysis problems. It can be a good idea to use a USB mic, because they typically bypass all the hard-to-find Windows sound card controller stuff, as the mic basically has its own sound card in-built. Also, make sure your mic is physically isolated from your computer as vibrations and noise pass easily through shared table tops and mic stands. Some Mac laptops and tablets have inbuilt mics that work well enough (if you can get VA to work on a Mac) but none of the PC laptops I've ever used has worked well using just its inbuilt mic.

If you still have problems, I covered a lot of stuff on collecting data using VA in the video series I did with Robbie O'Brien, Guitar Analysis and Testing with Trevor Gore, available from Robbie's website.

[Alain Lambert]

You can try my software Luthier Spectrum Analyser (free Demo). It is already setup for luthier usage.
There is a windows version here: https://drive.google.com/open?id=1X4FSJ ... uAeHgzvdnP
I have also a Mac version, but this is still a Beta version. It need more testing (looking for testers at the moment) I might release it in a few weeks.

[kiwigeo]

Some Mac laptops and tablets have inbuilt mics that work well enough (if you can get VA to work on a Mac) but none of the PC laptops I've ever used has worked well using just its inbuilt mic.

I ran VA on my Macbook Air via Windows running on top of Parallels. It works ok but moving data files in and out of the emulator environment is a pain. I ended up buying a cheap Windows laptop and run VA native on same.

[josswinn]

If you still have problems, I covered a lot of stuff on collecting data using VA in the video series I did with Robbie O'Brien, Guitar Analysis and Testing with Trevor Gore, available from Robbie's website.

I just bought this video course and it is a revelation. I've had your books for a while but this has given me a way in that I really needed. Thank you.

[simonm]

Got my microphone set up this afternoon and had another go at getting a spectrum. Do these look better?

My immediate reaction is that they are both quite similar. One is very recent, the other is about 2 years old. Construction is also quite different but both are clearly "guitar shaped objects" to a casual observer.

If the look useable, any suggestions as to what might be improved?

The first one is the same guitar as the first one in the first post. Using the tuner on my new tascam, it looks like the first is about G and the second one around F.

The 50hz peak is the power supply for my pre-amp. I will need to re-arrange the location of all my bits and pieces to get it out of range of the mic.

[Trevor Gore]

The extra "grass" looks like you managed to score two taps in one buffer. Also possibly a bounce (essentially the same as two taps per buffer). One tap per buffer definitely gives cleaner plots. It takes a bit of practice.

[simonm]

Thanks Trevor. I will practice a bit more and see if I can improve it.