 The Clarinet BBoard
|
Author: Don Berger
Date: 2000-10-08 19:09
Thank you, Rene, for intro- and reintro-ducing many of us to the study of our horn's harmonics-overtones . We usually just accept that the clars have "odd" "character" giving us an extended range, and the makers the severe mechanical problems with the 12 ths. I attempted some study when I found out my oboe had more energy in the octave than in the fundemental, but didn't get very far. A review of Bonade etc may help me! I doubt I can add to your conclusions re: the spectra, but will try. Shouldn't energy distribution be helpful in tonal characterization of bright/dark, mellow/ etc terminology?? Wow!, Don
|
|
Reply To Message
|
|
Author: L. Omar Henderson
Date: 2000-10-08 19:56
Had to think twice about jumping into this one but could not help myself.
I think that spectral analysis using FFT can be a useful tool in characterizing the sound produced by instruments, including the clarinet. A few caveats:
1) The sampling rate, damping, smoothing, hardware, software, etc. can all contribute to the end spectra that are analyzed. Jim Pyne, as mentioned, would be a good source of help and also abstracts from the Acoustical Society can be found on the web, which deal with many of these factors.
2) The instrument has many variables (reed, mouthpiece, ligature, barrel) as well as different players and their consistency can contribute to the spectra.
3) The recording set up, including microphone position, distance, and angle as well as the acoustics of the recording room can affect the spectra.
4) Plus about a zillion other factors too numerous to enumerate here.
All of this aside, myself and a team of math geeks have been able to document certain patterns and construct mathematical search equations which can be reliably associated with "bright" and "dark" sounding clarinets. This is not perfect and in a few lines I cannot explain or document all the "what if's" to satisfy this expert audience. Suffice it to say that we are making significant progress on adding some quantitative parameters to some of the subjective assessments that we have all made about our instruments. Hopefully, using such a system one could evaluate their own sound and the contribution made by changing certain components of the clarinet "system - reed, mouthpiece, etc.". We are now using this system as a quality control tool in reed manufacture. This is presently a research tool.
|
|
Reply To Message
|
|
Author: Don Berger
Date: 2000-10-08 20:10
Glad you "jumped in" L O H , I had the same feeling of the deeps, much of my math + has disappeared with age, maybe this will help. My real reason for a second post is to correct to BEnade, Arthur H. , sorry, Bonade, I do like your ligs!! This will cause me to look in my old Broadhouse, "Musical Acoustics" [p b's] for HELP! Don
|
|
Reply To Message
|
|
Author: Rene
Date: 2000-10-09 09:59
Thanks for the hints. I have now given the thing a bit more thought. I do not believe that my recordings are bad and result in a wrong spectral analysis. The sampling rate is high enough (mathematically) and the results are seemingly too good to be errors. Of course, my own play or clarinet is definietely a factor. I'd love to have some samples (one tone, held long, solo, no vibrato) of professionals but did not find one.
It may be possible to use the spectral analysis as a qualtity criterion, but I have my doubts, if things work out that simple. I'd like to see more of your research, Omar. Will it be published? Or is it a trade secret?
Thanks for pointing me to the plots, Mark.
Fascinating subject,
Rene
|
|
Reply To Message
|
|
Author: L. Omar Henderson
Date: 2000-10-09 11:58
There is almost no such thing as bad data but we often do not take all the variables into consideration (or do not know what they are all the time) into the analysis of our data. Thoroughly read Benede's book as suggested by Don, I got the same suggestion when I posted a while back about some of my preliminary spectral observations and it has helped me greatly. Some of our work will be published after it is polished and peer reviewed by those that know a lot more about this than myself. I am using spectral analysis in some of my commercial endeavors but have little time to devote to publishing in this area since I am required to "publish or perish" in my day job. I hope that others will help me out and publish the answers so that I can just use their techniques.
|
|
Reply To Message
|
|
Author: L. Omar Henderson
Date: 2000-10-10 15:35
Found another web site with examples of the FFT patterns that you are seeing. http://hyperphysics.phy-astr.gsu.edu/hbase/music/clarinet.html
|
|
Reply To Message
|
|
The Clarinet Pages
|
 |
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
