Wood or metal end? Test results...

[Chris Ghent]

A chance remark in a conversation between myself and Jurgen Suttner at Willie Clancy Week a couple of months ago led me to perform this test. It is a handful of notes played twice on the same concertina. Each pair of notes consists of a note played with a wooden end (composite ebony/2mm plywood/ebony) fitted and a note played with a metal end (.65 nickel silver) of the same pattern fitted. I'll leave it to you to sort out which note is which.

[Chris Ghent]

endmaterialtest.MP3 Edited November 15, 2012 by Chris Ghent

[Chris Ghent]

Had trouble uploading the file and can't get the editor to work but as you can see the file is in the second post...

 

Chris

[Greg Jowaisas]

Dear Chris,

As a repairman I was so distracted by the valve noise that I forgot to listen to the notes. Could you please play them again?

 

I'd venture that the first note was metal end and the second wood. First note seemed to have more overtones especially higher ones.

 

Of course I'm only guessing and it may be your turn to remind me of why I repair rather than play for a living!

 

Keep up the good work!

 

Greg

[Chris Ghent]

Dear Chris,

As a repairman I was so distracted by the valve noise that I forgot to listen to the notes. Could you please play them again?

 

Well, you know what they say, never buy a mechanic's own car There is an underlying hum on the recording, it did not seem worth redoing the recording for as it does not interfere with the ability to listen to the notes.

 

I'd venture that the first note was metal end and the second wood. First note seemed to have more overtones especially higher ones.

 

Of course I'm only guessing and it may be your turn to remind me of why I repair rather than play for a living!

 

I think I'll let a few others have a listen before I say. I feel it makes you listen from first principles rather than listening for confirmation...

 

Chris

[Geoff Wooff]

That is a fairly difficult test for the old ears but I think I detect a little more 'top' with the first note, but not in all the groups.

[Dirge]

That is a fairly difficult test for the old ears but I think I detect a little more 'top' with the first note, but not in all the groups.

Yes, me too, not much on the lower notes and more pronounced in the higher pitches, so I would also expect it to be metal first, wood second.

[Anglo-Irishman]

Yes, me too, not much on the lower notes and more pronounced in the higher pitches, so I would also expect it to be metal first, wood second.

 

I second that. Still just a very slight difference, though, even on the high notes.

It would be interesting to hear a similar commarison with nice, full chords. I seldom play a single note by itself!

 

Cheers,

John

[ttonon]

Hi Chris, thanks for the data for an interesting experiment. Apart from exploring the consequences of different materials used for endplates, this experiment can perhaps also evaluate the widely held concept that “brighter” sounds are associated with higher harmonics in the frequency spectrum.

 

I thus plotted the frequency spectrum of each of the notes you played, using the open source free software Audacity, which is a marvelous accomplishment by generous individuals. For those interested, I attached a Word.doc file with the results. There’s a graph for each of the six notes, plotting the normalized response level vs frequency, showing the peak overtones. The dark markers are for the first sound of each note (A), and the red markers are for the second sound (. I’d be glad to furnish more explanation for anyone so interested.

 

But having a frequency spectra of the wave forms is only a starting point, because one must still resort to subjective convictions on how to evaluate them. There’s of course a possibility that two spectra can show large enough differences so that one can confidently decide for instance that one spectrum has a larger contribution from upper harmonics than another. We can also, of course, resort to mathematical algorithms as a way to attempt a more objective evaluation, but in the end, it’s my feeling that subjective decisions will have to be made.

 

With all that said, going by these frequency spectra and assuming the sound files are an accurate representation of the true audio sound, my guess is that the first sound for Note 1 and the first sound for Note 5 should sound directly “brighter” to the human ear, if indeed higher harmonics are the determining factor in how “bright” the sound is to our ears. For these notes, markers A have a decidedly larger response than markers B, at the high end of the spectrum. Such a conclusion is marginally the same for notes 3 and 4. The story with the other notes, for me, is not as clear. Concerning my ear response, yes, I can convince myself that I hear in the files what the frequency spectrum is suggesting, but only really for Notes 1 and 5.

 

If I might make some general comments, the higher pitch notes (4, 5, and 6) have a low frequency artifact, easily discernible as an annoying buzz, and I believe this is due to the compression of the sound file, to MP3 format. This buzz does interfere with my audio evaluation. I suggest for such experiments that high-resolution sound files (e.g., wav) be used. Also, it may be better to make mono recordings, instead of stereo, because there can be unnecessary complicating factors when two microphones are used instead of one.

 

Perhaps most importantly, the response of the microphone and amplifying circuit is part of what we hear (and do calculations with). I assume you did not use a calibrated microphone/amplifier system, and of the two, it’s most likely that the microphone would be the weaker link, and certainly also, cheap computer speakers add their own problems. For instance, if the response of the microphone is weaker for the higher frequencies, we will then not hear, or measure, in the sound files, the true response, and the contribution from the high frequencies may not appear as large as they really are, when you hear the instrument in the room. Thus, such unfortunate outcome would invalidate such an experiment. In other words, without knowing the frequency response of the microphone you used, we really can’t say for sure how accurate the sound files are. I find that some electret mic cartridges have flat response – which is what you want - over the entire human ear response range. And they are in general no more expensive than others. Such microphone elements are available from electronics stores such as Digikey for only a couple/three dollars each, and their frequency response can be viewed online, before purchase.

 

Best regards,

Tom

www.bluesbox.biz

Chris Ghent Experiment, spectra.doc

[Chris Ghent]

Tom,

 

thanks for the response, and to everyone else too.

 

Tom I accept your description of the limitations of the audio setup in the "experiment". The recording was with an Edirol R09 in MP3 format. The buzz I heard, but as it did not seem to impact on the sound overall I accepted it. If I seem over willing to accept the low sound quality I have to say it is because the sound I am hearing in the recording matches my perception when playing the actual concertina. Your description of a better microphone makes sense.

 

The two ends (wood and nickel silver, not steel as you say in the heading to your spectrum analysis) were played in the order 1) wood 2) metal. I appreciate this will be a surprice to those who guessed in the opposite order, and can I say it was a surprise to me when I performed the test. The concertina concerned is very loud with a nice "cut" to the sound, perfect for an Irish session but perhaps a little loud for a reverberant kitchen. It is not harsh, I would describe it as semi sweet. When I fitted the metal end I was amazed to discover to my ear the higher partials which aid the cut had been tidied up. The sound was more controlled.

 

I don't claim this to be a test of all instruments, metal versus wood. I think it shows what happens to this instrument when you fit one or the other end.

 

One thing I think could come out of this is a curbing of the tendency for everyone to describe metal ends as intrinsically being louder or more penetrating. I accept there is a difference; we have not yet learned to describe it.

[Greg Jowaisas]

Greg Jowaisas, on 14 November 2012 - 09:15 PM, said:

 

 

I'd venture that the first note was metal end and the second wood. First note seemed to have more overtones especially higher ones.

 

Of course I'm only guessing and it may be your turn to remind me of why I repair rather than play for a living!

 

 

Although the first inclination is: "The older I get the better I was", I've decided a better working credo is: "The older I am the more practice I have being humbled."

 

Thank you for another reminder!

 

Greg

[Dirge]

So do you think this is an odd result from this particular instrument? Or are we all deluding ourselves generally?

 

I WOULD have said quite confidently that wood ended instruments (and I'm usually talking big duets here; thicker ends maybe? You were talking quite thin ply I think?) had a much softer, rounder tone with more emphasis on the bass, and that it was quite clear and unmistakeable. Now you have me totally confused.

[Daniel Hersh]

One thing I think could come out of this is a curbing of the tendency for everyone to describe metal ends as intrinsically being louder or more penetrating. I accept there is a difference; we have not yet learned to describe it.

Very interesting. I too had guessed in the wrong direction. I wonder if part of this tendency is that we may be, perhaps unconsciously, comparing instruments that are not otherwise equal. Most Lachenals have wooden ends, and most Jeffries have metal ones. Of course, I know better than this intellectually - but when I think "wooden ends", my mind calls up the wooden ended instruments I've actually heard or played. In any case, i will be more careful about what I say about this in the future.

[Chris Ghent]

One thing I think could come out of this is a curbing of the tendency for everyone to describe metal ends as intrinsically being louder or more penetrating. I accept there is a difference; we have not yet learned to describe it.

Very interesting. I too had guessed in the wrong direction. I wonder if part of this tendency is that we may be, perhaps unconsciously, comparing instruments that are not otherwise equal. Most Lachenals have wooden ends, and most Jeffries have metal ones. Of course, I know better than this intellectually - but when I think "wooden ends", my mind calls up the wooden ended instruments I've actually heard or played. In any case, i will be more careful about what I say about this in the future.

 

Daniel,

 

Well put...

 

Chris

[Chris Ghent]

So do you think this is an odd result from this particular instrument? Or are we all deluding ourselves generally?

 

I am reluctant to draw too many conclusions from this test other than the one I have stated, that this is what happens to this instrument in this case. By that I don't mean it is misleading or odd, but sufficiently against expectations that careful consideration of the test would be prudent.

 

It is worth noting the particularly "penetrating" sound I find difficult in some Jeffries concertinas does not record as such. Also I find it impossible to tell whether a concertina on a Cd is wood or metal.

 

I WOULD have said quite confidently that wood ended instruments (and I'm usually talking big duets here; thicker ends maybe? You were talking quite thin ply I think?) had a much softer, rounder tone with more emphasis on the bass, and that it was quite clear and unmistakeable. Now you have me totally confused.

My experience of larger Wheatstone instruments is they are very lightly built. The ply in this case is a little under 4mm, made up of two layers of ebony and a very thin ply of 2mm.

 

Chris

[Chris Ghent]

Although the first inclination is: "The older I get the better I was", I've decided a better working credo is: "The older I am the more practice I have being humbled."

 

Thank you for another reminder!

 

Greg

 

Greg,

 

my apologies for the setup, I suspected people would listen closer if they needed to decide for thenselves. If it is any consolation I would have failed this test myself.

 

Cheers

 

Chris

[Chris Ghent]

Apart from exploring the consequences of different materials used for endplates, this experiment can perhaps also evaluate the widely held concept that “brighter” sounds are associated with higher harmonics in the frequency spectrum.

Tom,

 

re-reading your post I realise this sentence encapsulates a thought I was trying to formulate. Bright sounds suggest high partials, but could they be a relationship between lower partials. What about partials a third apart which are disproportionately loud?

 

Cheers

 

Chris

[DaveM]

Hi Chris, thanks for the data for an interesting experiment. Apart from exploring the consequences of different materials used for endplates, this experiment can perhaps also evaluate the widely held concept that “brighter” sounds are associated with higher harmonics in the frequency spectrum.

 

I thus plotted the frequency spectrum of each of the notes you played, using the open source free software Audacity, which is a marvelous accomplishment by generous individuals. For those interested, I attached a Word.doc file with the results. There’s a graph for each of the six notes, plotting the normalized response level vs frequency, showing the peak overtones. The dark markers are for the first sound of each note (A), and the red markers are for the second sound (. I’d be glad to furnish more explanation for anyone so interested.

 

[snip]

 

Best regards,

Tom

www.bluesbox.biz

 

<pedantic-on>

It seems to me that you identified the fundamental and then plotted the spectral amplitudes at the exact harmonics of this frequency.

I'm under the impression that inharmonicities between the upper partials are an important part of an instrument's timbre.

I don't expect that it invalidates any of the conclusions that have been drawn so far, but is there the possibility your plots

don't reflect the maximal values of the peaks, since they may be slightly offset from their ideal frequencies?

I'd also believe it if someone could confirm that free-reeds have particlarily harmonic overtones, which would be consistent

with my understanding of how the sound is produced, but I'm not 100% about this.

</pedantic-on>