Now record the plate tap tones using this Audacity software by pressing the red button with the mouse pointer. This button is top left in the picture on the right - click on the picture to see it. I usually tap the plate in the right place about 10 times in 5 seconds.
Select the 2 silences at the either end of the waveform with the mouse left hand button pressed and delete them using the delete key. This leaves a good waveform to work with: you may need to adjust the mic sensitivity down so the mic is not overloaded
Select all the recorded waveform left in the window using “Control + A” keys pressed simultaneously and go to the ‘Analyze’ option on the top line of the window and select the ‘Plot Spectrum’ menu item that drops down. This calculates the Fast Fourier Transform (the FFT) of the selected waveform: that is it shows all of the resonant frequencies or tap tones present in the sounds picked up my the microphone.
You will need to enlarge the ‘Frequency Analysis’ window, and at the bottom left of the screen select ‘Spectrum’, 16,384 (the sample window size), Hanning window, and finally ‘log’ (or logarithmic) display.
I’ve included 2 pictures (screen dumps) of typical Audacity windows, one above showing the waveform, and one right showing the FFT screen. Click on them for the full size pictures.
In this example of a front (belly) with a good bassbar fitted, Mode 2 peak is at 150 Hz, and, also shown selected with the cursor is the 340 Hz of Mode 5. What’s so good about Audacity here is that the cursor pops automatically to the nearest peak and shows its frequency under the display, under the purple area bottom left. Be careful to read the ‘peak’ frequency on the right and not the ‘cursor’ frequency on the left ! Now write it down - I keep a small transparent folder for each violin to record all progress, results and calculations.
You can find instructions for using Audacity software here, or use a 10-part tutorial on YouTube here.
Audacity also allows the export of data to MS Excel (or an OpenOffice spreadsheet) to calculate the energy at various frequencies.. It’s shown as an ‘Export’ button, bottom right of the Freq. Analysis window, and allows a graph of amplitude v. frequency to be drawn in the spreadsheet using the ‘X-Y plot’ option.
You might like to look at Vojtech Blahout’s website on violin making too as it has some help on using Audacity to measure plate mode frequencies.
Another possible and good waveform / FFT software program and audio sound generator is “Visual Analyser 2014” by Alfredo Accattatis.
It is free to use, and has a pretty good real-time spectrum analyser as well as a standard FFT function. Settings allow you to zoom in on a spectrum and also to set the spectrum range - for violin plates you only need the spectrum up to say 3 kHz.
For some reason I can record and measure Mode 5 at 300+ Hz well, but can’t get Mode 2 at 150+ Hz to be easily visible in the sound spectrum. Turn it on (button top left), then on ‘Main’ menu select FFT size 8192 and ‘mono’. Then under ‘Capture’ set 20 or more buffers. Then when the button on the right ‘Capture Spectrum’ is pressed about 5 seconds of the FFT of the sound just recorded is visible. You can set it for no phase display and log ‘x’ scaling, and zoom into the FFT display.
There is also a free version of Overtone Analyser from Sygyt software which can be downloaded from here. An example of a screen dump from Overtone Analyser (Free Edition) is here left: click on it.
This is the spectrum for a belly with no bassbar where Mode 5 is at 301.5 Hz. Pieter Swanepoel wrote in to say he uses it, but I find it a little confusing as the plate is tapped and the tones are not continuous: but each to his own. The paid-for versions are much more versatile. The ‘Live’ version is 99 Euros.
A company called Peterson make a range of strobe tuners, best described on this YouTube video. The latest is the ‘StroboSoft’ PC software for about $50. Strobe tuning techniques have been in use since Lloyd A. Loar’s outstanding work in the 1920’s on the Gibson F4 (and F5) mandolin plate tuning. They have a flat or rather carved front and back plate. His work has encouraged plate tuning on all kinds of instruments in the USA ever since, as the F4 and F5’s from that era are truly the ‘strads’ of mandolins! Roger Siminoff is the current US guru on plate tuning, and has published several books.
The advantage of ‘strobe tuners’ is they allow rapid visualisation of the tap tone’s pitch, and also its harmonics: i.e. a frequency and all of its octaves. This may require a compressor (as used with guitars) to stretch the tap tone out.
So just what are these Modes?
What is Mode 2 or the X-Mode? and what is Mode 5 or the Ring-Mode? Here are the ‘tea leaf’ patterns that show up when the plates are vibrated using a loudspeaker underneath at a particular frequency - the ‘tap tone’. It is a resonant frequency of the plate as a whole.
On the left here is one of the excellent photos on Joseph Curtin’s website and shows the most important Mode in a belly plate, Mode 5, the ring tone. The black lines are the nodal lines on a violin belly vibrating in Mode 5 at about 350 Hz. The pictures below show Mode 2 in a front plate at about 170 Hz. Have a look too at a BBC web page here that uses “holographic interferometry” (!) to do the same thing - only even better!
These pictures show that when a plate, either the front or back of an instrument is vibrated at one constant frequency then fine sand, tea leaves, sawdust or glitter placed on it will move to areas of the plate that are not moving i.e. not vibrating. This only happens at particular resonant frequencies, and finding those resonant frequencies is what this is all about. These lines of no motion are called the ‘nodes’, or in this case the ‘nodal lines’. Anti-nodes are where the motion is at a maximum.
In summary: The “nodal lines” are the black lines in the photos and videos where the plate does not vibrate at that particular frequency. So the plate can and must be held somewhere on a nodal line without affecting the vibrating pattern of that Mode.
There is a good YouTube video by Jonathan McKinley (click on the picture left) that clearly shows plate modes 1, 2 and 5 vibrating as the tea leaves migrate to the lines of zero-vibration - the nodal lines of the violin plate as the vibrating frequency is gradually increased. Jonathan McKinley has photographed his violin parts here too: but his belly: Mode 2 is high, and Mode 5 low!
Understanding these Modes and their shapes is fundamental to understanding tap tones. You will need to know how to get these Modes vibrating. It was Carleen who first used a speaker placed under a plate at an anti-node, and fed the ‘speaker at various frequencies until a glitter pattern showed up.
These patterns are well illustrated by other website pages to be found in the “Really Useful Links” page. If you click on either these two glitter patterns you can see the nodal lines on a back plate that Alan Jhones from Brazil sent me (thanks!). Mode 5 is the one above, and Mode 2 is on the right.
We can also get the plates vibrating by holding at a point on a nodal line, and tapping at an ‘anti-node’, or a point where the vibration is or will be a maximum: it works well, but there is information lost about the detailed shape of the nodal lines: Carleeen thought that is important.
Animations of plates vibrating
Borman Violins also shows some excellent and fascinating animations of violin and other plates here vibrating in situ on a violin and viola, made possible by the work of instrument maker George Stoppani.
This helps us understand how the vibration of the belly and back plates vibration modes while free (as we measure them) are linked to how they behave actually in a completed instrument.
Vibration modes of the completed violin
Getting the tap tones right for the plates will give you a very very good chance that the resonance modes of the completed violin body will be right - or at least similar to the world’s greatest violins! There is a direct link from the plate resonances to the resonances of the completed instrument. Have a look here for the resonances of the completed violin body ......
...but what tap tones don’t tell us
Well you wont make a violin that sounds like a £15,000 violin using tap tones like this unless you are very, very, very, very lucky and at least prepared to let an expert choose the (costly) wood to start with. In particular tap tones are not good at reflecting what the edges of a violin plate are like. They are a whole different subject.
There is also another and much higher level of the art, experience and science that the master craftsman has that us mere mortals can only dream of.
You can glimpse some of it in the work some people are doing to actually listen openly to what violins sound like: there’s a discussion here on maestronet for instance, which I’ve edited to a .pdf document here to make more of a narrative (and remove the occasional name calling these forums suffer from).
Others for instance have tried area tuning (see this fascinating site of Keith Hill’s and one here too - figure right) and the ‘Vigdorchik strip’ tap tuning (see David Langsather’s example here) which I think has great possibilities, but you will need a lot of faith, patience, and a very, very good ear to do it: which I do not have.
And much of a lifetime.