Audacity freeware: How does a 'typical' resonance spectrum of a good sounding violin look like?


Recommended Posts

Audacity just plots a response spectrum, much like any other spectral analysis program.  The important "generating" comes from how you generate the sound from the instrument, where you place the microphone, and plenty of other variables, which is a long discussion for later.

If you want to see some very good graphs and a discussion about them, see some recent posts on AD's bench thread.  Those are impact spectra comparisons... not Audacity, but like I said, that doesn't matter much.

I have been doing more with bowed semitone scales lately.  Owners of Cremonese instruments might not be too keen on whacking the bridge of their violin, so this might be a more convenient test.  Below is a comparison of the 1714 Ex-Jackson Strad and my violin that is played by Annelle Gregory.  Never mind the absolute dB scale, as they were taken with different computers that apparently had different gains... but the shape is what matters.  The Strad isn't terribly different in the signature mode area; the main differences are above that.  Around the 700 - 1200 Hz zone (transition hill), the Strad is more even, without big peaks.  In the "bridge/body hill" above that, the Strad is more "scrunched" into the 1300-4000 Hz zone, whereas the new instrument is not as even, and smeared out into the higher ranges.  (Resolution set to 2048 to smooth out the discreet notes)

1459988512_1714Stradvs21.thumb.jpg.89fc63a15068994326647bfb5e44b1b1.jpg

For reference, I had a cheap old Maggini model picked up on eBay that I repaired and regraduated for fun.  It would make a very bad violin, but was actually a fairly pleasant fiddle... good low end, very low on brightness and projection.  Strong transition hill, dead on the highs.  On the E string, you get the fundamental and maybe a bit of 1 overtone... completely different from the Strad where there would be several strong overtones.

387089093_CheapoldMaggini.jpg.cd04e2df40b5d3bab4214fc5219e41d0.jpg

 

Link to post
Share on other sites

I took semitone scale spectra of my old violin (red) and the one I just bought to replace it (red).  I took them in the same position, with the same mic, same bow, about 30 seconds apart.  I minimized as many variables as I could, so the dB scale is close to reality.  What you see is a better low end on the new one, which is evident with my ears and anyone else within earshot.

 

The old instrument was shrill but powerful.  The new one is deep and powerful.  Everyone who has heard both prefers the newer instrument.

 

Untitled.jpg

Link to post
Share on other sites
1 hour ago, dpappas said:

I took semitone scale spectra of my old violin (red) and the one I just bought to replace it (red).  I took them in the same position, with the same mic, same bow, about 30 seconds apart.  I minimized as many variables as I could, so the dB scale is close to reality.  What you see is a better low end on the new one, which is evident with my ears and anyone else within earshot.

 

The old instrument was shrill but powerful.  The new one is deep and powerful.  Everyone who has heard both prefers the newer instrument.

 

Untitled.jpg

I think you mean louder, not better. Whether it’s better depends on the shape of the wave and the makeup of the harmonics, as well as speed of attack and many other factors not visible on this kind of graph.

 

Link to post
Share on other sites
18 minutes ago, martin swan said:

I think you mean louder, not better. Whether it’s better depends on the shape of the wave and the makeup of the harmonics, as well as speed of attack and many other factors not visible on this kind of graph.

 

Martin,

 

I agree, 100%.  I should have clarified.  This is just one piece of the puzzle.  And I don't think these kinds of plots can predict performance (maybe for a maker they are a useful reference).  For me they are a curiosity.

Link to post
Share on other sites
2 hours ago, David Beard said:

Looks a lot like a bad sounding one.

B)

Just kidding.  

But more seriously, is there anyone here that believes they can look at a spectrum and decide by sight, 'Oh, that's a good sounding violin.'

Does anyone think they have that ability???

One of the problems is that not all spectra are taken the same way, so I think that relative characteristics between instruments taken with the same equipment and method are about all you can hope for.  If you have some reference "good" and "bad" instrument spectra for calibration, I think you could get some idea.

For example, I showe bowed spectra of a very good, and rather poor violin, and the difference is rather obvious in the spectra.  With dpappas' red and black spectra, the only thing I could tell by looking at it (assuming they were both taken the same way) is that the red line had a lot more power, mostly in the lower range.  That could be good or bad, as it is all in the balance as determined by the player or listener.

It is much easier to spot a horrid VSO from a spectrum.  Here's a good example (this is an impact spectrum) of an amateur-made fiddle, all mahogany:

Mahogany.jpg.65a0eb578082549f42936d4e78655046.jpg

Link to post
Share on other sites

Yeah. I can see how you can spot really bad.  I have trouble believing that one can see good.

Though, I have to admit your jackson spectrum shows things I tend to prefer.  So I see a good spread of energy in the lower range, not excess energy going crazy high, but plenty of energy going mid high for clarity.

Perhaps if I saw enough similar sprectra from good fiddles, then when I saw that again I might suspect a good fiddle.

But I sure wouldn't be surprised if some ugly sounding beast also had that distribution.

Link to post
Share on other sites
6 minutes ago, David Beard said:

But I sure wouldn't be surprised if some ugly sounding beast also had that distribution.

I haven't seen any spectrum yet that has looked all that close to the ex-Jackson.  After 15 years of looking.  But I must admit that I haven't had my hands on that many great violins.

I would be completely shocked if some awful-sounding violin had that kind of plot.

Link to post
Share on other sites

@Don Noon

thanks for all the information. 

All what I get from it right now is that I have first learn to read those graphs. For me it is pretty impressive that you could tell the difference between the graph taken from a new violin and an old violin. 

What I remember from some older research papers is that scientists look somehow on the spacing of some peaks as well as average loudness levels in certain areas 

looking at the graphs of the Jackson strad (thanks for posting it) and your instrument, the only thing I would note is that the strad seems to be stronger in the 3000- 6000 Hz area which, if I remember correctly, is a frequency range the human ear is very sensitive. (And singers need to learn how to produce overtones in this range to be heard against a full orchestra) 

Besides, if you play half tone scale, you just keep recording the spectrum all the way through? And if so how far do you play? (Supposedly not until you reach the end of the fingerboard on the e string?)

I looked at the page of AD as well and it seems I need first to calibrate my own testing setup.

Link to post
Share on other sites
47 minutes ago, Andreas Preuss said:

looking at the graphs of the Jackson strad (thanks for posting it) and your instrument, the only thing I would note is that the strad seems to be stronger in the 3000- 6000 Hz area

The solid strength from ~1400 Hz to 4000 Hz I think is the feature that gives it the "laser-like" Strad E string.  Not too long ago, some very major restoration and repairs were done on it, and for quite a while it was just a very nice violin, nowhere near what it had been.  The spectrum showed a marked attenuation of the high frequencies (compared to an earlier spectrum).  Fortunately it settled back into its normal self.

 

48 minutes ago, Andreas Preuss said:

Besides, if you play half tone scale, you just keep recording the spectrum all the way through? And if so how far do you play? (Supposedly not until you reach the end of the fingerboard on the e string?)

The way I do it is all 1st position, 7 semitones per string, except I go 1 more on the E string.  The overtones are primarily what fills in the higher frequencies of the plot. and if I played higher on the E string I might get spikes that don't get averaged out with the 2048 resolution.

The full spectrum is taken from the whole scale.  I can select smaller clips of the scale to look for various other things, or even single notes to see the balance of the overtones.  That can be useful if certain notes sound odd, to find out why.

Link to post
Share on other sites
9 hours ago, Michael Darnton said:

Oliver Rodgers was the King of the Violin Spectrum, and I learned a lot of things from his writings and from people who knew him. If you are really interested in the topic, you should look at things he's written, IMO. 

Actually I was digging out many old research papers recently.  In my reference books I found so far only one article he co authored with Carleen Hutchins.

Do you recommend any specific paper he wrote?

Link to post
Share on other sites

Nearly 30 years ago Duennwald had an article in the Catgut Acoutical Society journal showing spectra of Old Italian, modern master, and factory-made instruments. See Figure 3 of this article.  It pretty much reflects the same thing I show in my plots above of the Strad, my contemporary violin, and a bad violin.  It's kindof amazing that this information has been around for so long.

Link to post
Share on other sites
1 hour ago, Don Noon said:

Nearly 30 years ago Duennwald had an article in the Catgut Acoutical Society journal showing spectra of Old Italian, modern master, and factory-made instruments. See Figure 3 of this article.  It pretty much reflects the same thing I show in my plots above of the Strad, my contemporary violin, and a bad violin.  It's kindof amazing that this information has been around for so long.

Unfortunately he didn't group them as good, mediocre, and bad from large group blind listening and player tests.  He also unfortunately normalized all their sound outputs so  weak instruments might look the same as powerful ones.

 It is kind of amazing this misleading information has been around for so long.

Link to post
Share on other sites
On October 14, 2020 at 11:19 PM, Don Noon said:

The way I do it is all 1st position, 7 semitones per string, except I go 1 more on the E string.  The overtones are primarily what fills in the higher frequencies of the plot. and if I played higher on the E string I might get spikes that don't get averaged out with the 2048 resolution.

The full spectrum is taken from the whole scale.  I can select smaller clips of the scale to look for various other things, or even single notes to see the balance of the overtones.  That can be useful if certain notes sound odd, to find out why.

So just as a quick try I made a plot of a violin. I played half notes as you described. Can you see if it is good or bad?

image.jpeg

Link to post
Share on other sites
50 minutes ago, Andreas Preuss said:

Can you see if it is good or bad?

As I said before, due to differences in the experimenter and personal taste, I can't say too much about good or bad.  Really, really bad can be obvious, but this is not one of those.

On the low end, it is curious that the CBR is stronger than the A0, and the B1- is barely visible.  Unusual, but not necessarily "bad". Is this violin construction abnormal in any way?

The huge peak at ~G#5 on the E string could be annoying, both from an overly-strong fundamental on the E string and an overly strong overtone on the note an octave below.  I would suspect that the soundpost is too far behind the bridge foot, which makes this resonance stand out.  But if the player likes it like that, then there's no problem.

The high end is much more operator and equipment-dependent, so any analysis would be suspect.  If I had takent the data, a plot like that would look a bit weak in the broad range of 900 - 1600 Hz... but not bad.

Link to post
Share on other sites
1 hour ago, Don Noon said:

As I said before, due to differences in the experimenter and personal taste, I can't say too much about good or bad.  Really, really bad can be obvious, but this is not one of those.

On the low end, it is curious that the CBR is stronger than the A0, and the B1- is barely visible.  Unusual, but not necessarily "bad". Is this violin construction abnormal in any way?

The huge peak at ~G#5 on the E string could be annoying, both from an overly-strong fundamental on the E string and an overly strong overtone on the note an octave below.  I would suspect that the soundpost is too far behind the bridge foot, which makes this resonance stand out.  But if the player likes it like that, then there's no problem.

The high end is much more operator and equipment-dependent, so any analysis would be suspect.  If I had takent the data, a plot like that would look a bit weak in the broad range of 900 - 1600 Hz... but not bad.

Interesting.

The violin is by no means unusual. It is a Guad copy i made a few years ago. It has a rather high arch and the wood used was steam treated. This was probably the main reason that the total weight in playing condition without chinrest came down to 370g. (with no intention from my side)

On a subjective scale it didnt sound good at all at the beginning and only after 2-3 month developed its sound

The soundpost for this plot was at 4mm distance. I moved it a bit farer away because the sound felt a bit tense under the bow. 

 

Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...
  • Recently Browsing   0 members

    No registered users viewing this page.