 The Clarinet BBoard
|
Author: Audi-Graph
Date: 2010-01-11 18:00
Hi all. I’ve lurked on the boards for a while but decided to post and see what everyone thought of using acoustics and physics of sound as an educational approach to teaching sound quality.
I bring this up because of the difficulty we all have in explaining what sound is, and how to produce a good sound. Plus all those posts on recommendations on a new reed or upgrade the clarinet or ligature to make a good sound when a majority of it comes from player, not the horn. This came from some acoustics research I was doing when I worked for NASA.
This isn’t intended to be a product plug, but me sharing my teaching approach that seems to work with the tech savvy students.
Without getting too technical, it analyzes your sound so you can see the harmonic frequencies generated when you play. It shows timbre and tone quality. The software contains the recorded harmonic frequency patterns of some local symphony musicians across their chromatic scale. You practice to make the picture of your sound match the picture of their sound, and you sound the same.
http://www.audi-graph.com
It's basically like playing into a tuner, but it helps you develop a core fundamental sound against a professional example.
Does anyone have any thoughts?
------------------------------------------
Anthony Vasser - audi-graph@audi-graph.com
http://www.audi-graph.com
Post Edited (2010-01-11 18:07)
|
|
Reply To Message
|
|
Author: L. Omar Henderson
Date: 2010-01-11 19:54
(Disclaimer - this represents research and product offerings not yet commercially available)
Thanks so much for sharing your new product. Some of us have been fooling around with similar concepts about analyzing sound and developing algorithms from FFT spectral analysis that by objective analysis describe the subjective descriptions of "bright", "dark", "mellow", "warm", etc. My own approach is to use algorithms to pick out certain sound descriptions e.g. "warm" and put a color value - perhaps red to violet - versus "bright" which might be blue to white and using sets of blended concentric circles with the major harmonics centered in the middle which also acting as a tuning scale with expanding circles of color to illustrate changes in sound quality. This too would use a microphone but with a set of 3D glasses corded to a computer worn by the player.
Our approach may be a little more specific than trying to imitate the spectral characteristics using FFT of a particular player's note or particular sound characteristics and would allow the individual to explore their own sound and make adjustments to trend toward a particular sound spectrum. The clarinet and other woodwinds have particular and characteristic sound patterns with odd harmonic features which add specific characteristics to the sound.
I am also not sure that any group clarinetists could reach a consensus opinion of what is the "best" sound or even if a particular sound was better than the next out of the context of a piece of music expressing various notes with a similar sound quality?
L. Omar Henderson
www.doctorsprod.com
|
|
Reply To Message
|
|
Author: johng ★2017
Date: 2010-01-11 20:23
This is an interesting idea that I would like to explore a bit more. The idea of providing a range of tonal types and full chromatic scales for comparison of a student's sound seems like a good one. As Omar said, this isn't really about "music", but I think it is true that one can spend lots of time trying to improve a student's tone quality apart from playing music. It appears to me that we are becoming ever more visually oriented and so a visual teaching concept may reach students.
There is one thing I wonder about, though. In the web site, it is noted that what is being studied is the sound after the attack. There are times when it is very difficult to tell the difference between instruments when you eliminate the attack. Would it be possible to take a "snap shot" of the attack for comparison as well? That is something else I spend a lot of time on with students.
John Gibson, Founder of JB Linear Music, www.music4woodwinds.com
|
|
Reply To Message
|
|
Author: DavidBlumberg
Date: 2010-01-11 20:52
Placement of the recommended Microphone for Clarinet - what kind of stand do I need for it? (as it's not a clip on)
I'm curious about that one.
Also, is there a 2 instrument "doubler" discount? (instead of the $100 - I know that's already a discount from regular, but is there a combo besides the all instrument one?)
http://www.SkypeClarinetLessons.com
|
|
Reply To Message
|
|
Author: Audi-Graph
Date: 2010-01-11 22:32
L. Omar:
Your approach sounds extremely fascinating! Do you have anything presentable that you could share (either publicly or privately)?
JohnG:
As it stands right now, the characteristic of the sound changes significantly during the attack time, and it happens very quickly. At the moment, Audi-Graph only contains samples of the professionals at the stabilized tone.
You can certainly use Audi-Graph to help look into attack time. A staccato would be on screen for about 3/4 of a second or so... which should give you enough time to at least see what is happening so to discuss with your students. I'll look at putting in something for attack time.
I've taught and been taught if you have the embouchure to produce a good sound in long tones, that if you set your embouchure before you even start playing, your note attack will be good.
David:
Regarding microphone placement. I recorded everything with (and recommend) the Samson CO1U condenser mic which comes with little tripod stand. I recommend placement about 3 feet from the middle of the horn (as opposed to on the floor). Obviously better microphones are better, the Samson gets you range and response without a huge investment. Audi-Graph was made to remain accurate while being used in practice rooms or at home, so nothing needs to be studio quality. What kind of mic do you use?
As for multiple instruments, send me an email and let me know what you play. audi-graph@audi-graph.com
------------------------------------------
Anthony Vasser - audi-graph@audi-graph.com
http://www.audi-graph.com
Post Edited (2010-01-14 17:37)
|
|
Reply To Message
|
|
Author: mrn
Date: 2010-01-11 22:34
Howdy fellow Ag! The idea behind your program is an interesting one. I think visual feedback regarding one's tone quality could be valuable if placed in the right format and used in the right way.
John's point above about attacks touches on one of my thoughts regarding this sort of thing. Your program, functioning as essentially as a real-time spectrum analyzer, displays frequency-domain information for an instant in time (and then the next one and then the next one, thus creating a real-time animation). However, much of what we actually perceive about musical tone really exists in the time domain, not the frequency domain. Now, to the extent that you animate your spectral plot, you do convey some of that to the person looking at the screen, but it's really hard to step back and do the same kind of comparison you do with the spectral plots to see how things change over time. I think it would be helpful to be able to view, at a glance, how the frequency spectrum changes over time, since that would give someone a better representation of what the ear actually hears.
To illustrate my point, consider an electric guitar with a "wah-wah" pedal. The "wah" sound is caused by varying the parameters of a filter over time, so you have a frequency spectrum that changes from one instant to the next. You can't represent a complex sound like a "wah" in a really meaningful way with just an FFT, because the FFT (or the DFT, I should say) is based on the fiction that you are viewing a steady-state periodic waveform (and calculating the DFT based on one period of that waveform). There's no really good way within the confines of a simple DFT to model those things that we perceive with our ears to be transient in nature.
Also, from a pedagogical viewpoint, I think it is more valuable to learn how to create a palette of different tone colors than it is to duplicate a single tone color, as Omar suggested. I think your program could potentially be a valuable tool for this purpose, but obviously, you'd have to provide a pretty wide variety of sounds in the sound library to do this.
|
|
Reply To Message
|
|
Author: rgames
Date: 2010-01-11 23:07
I think we always need to be careful when using visual tools to teach auditory perception. As a tool, they can be useful (e.g. tuners) but training a student to recognize auditory stimuli via listening should be the first concern.
Also, a significant portion of the recorded sound is due to the acoustic response of and position within the room in which the recording is made. Unless there's a way to account for that fact (which is quite tough) then I'm not certain the tool is providing results that can be acted upon. Furthermore, microphones can also vary widely in their frequency response - they will also color the measurement.
In short, you can take the same player and record him/her in different locations using different microphones and get very different measurements. How does the tool account for those differences? I would guess it's extremely difficult to create an algorithm to do that.
A well-trained ear, however, can do it readily
rgames
____________________________
Richard G. Ames
Composer - Arranger - Producer
www.rgamesmusic.com
|
|
Reply To Message
|
|
Author: Mark Charette
Date: 2010-01-12 00:05
Dear BBoarders - by the time I got to this there were already a few followups, so please note:
Normally this would not be approved for posting since, as per our posted rules, this is self-promotion, albeit an interesting product. Since there are already a number of followups, I'm going to leave the thread up, but it is an exception, not the rule, and should not have made it to the BBoard.
|
|
Reply To Message
|
|
Author: Wes
Date: 2010-01-12 00:38
Hi!
This is very interesting to me as I've used spectrum analyzers of all kinds for many years, even in the office of the CEO of Citicorp. I've planned to buy an audio spectrum analyzer for my own use.
Perhaps I did not look carefully enough, but I did not see technical specifications such as the vertical and horizontal scale information, frequency calibration information, repitition rates, bandwidth, etc. Could it be used for instruments such as the contrabassoon, bass clarinet, piccolo, or cymbals?
Good luck on this project!
|
|
Reply To Message
|
|
Author: L. Omar Henderson
Date: 2010-01-12 01:22
Mark, thanks for bending the rules to facilitate a very interesting discussion!
May we please have a name - I hate the Dear Audi-Graph logo. Yes, off the BB I can share the stage of development of our product. As others have observed the first major peaks of the FFT real time graph are the major harmonics and the right hand portion of the graph are the minor or odd harmonic series that are so important at least in woodwind sound characterization.
Unfortunately, without some characterization of spectral graphs (i.e. algorithms to interpolate groups of spectral data) it is hard to know how twist and turn your embouchure or get your hardware adjusted to change the odd harmonic series characteristic of the "dark" or "warm", or "mellow" sounds (spectra) which seem to be the desired sounds most coveted by (some) clarinet players. Certain players (not to get into a naming game) have distinctive sound signatures that they use to add distinction to their playing and interpretation of music which would not fit your "ideal" symphonic player's mold. As a teaching tool it might not be a bad idea to alter some neophyte player’s sound but to set a standard for more mature players could be folly!
I think also that the "attack" portion of creating sound is part of the process of playing but your approach is real time and just requires interpretation of the segments from the beginning to a stable sound spectrum. These are two separate questions or facets of the total sound.
All of this follows the tenants of learning - visual or auditory or both. Many players will benefit from a visual approach since auditory is subject to too many variables in interpretation.
We wish you luck with this new endeavor and continued improvement in the product.
L. Omar Henderson
www.doctorsprod.com
|
|
Reply To Message
|
|
Author: Audi-Graph
Date: 2010-01-12 04:01
Mark: Thanks for bending the rules a little. I meant the post to more facilitate a discussion on if others are teaching tone using technology, with my approach as an example.
Mrn: Howdy! I was in the Aggie Band, which explains why I’m using everything at my disposal to revitalize the elegance in my playing! You are correct in your statement of the importance of the time domain in the perception of sound. The guitar’s whammy bar “wah wah,” or vibrato in general, is not something that can be represented in a snapshot.
Vibrato is just one of many effects musicians apply to their tone (like bending or glissing) to add expression to a note or phrase. However, if the musicians underlying sound is of poor quality, the listener has trouble getting past the poor sound … like playing beautiful vibrato on a horn way out of tune. The intent is that this can help students develop a core sound foundation to which to apply these effects.
David: I’m not sure I understand your question, so I’ll take a quick guess at it. I’m trying to include as many examples of musicians as possible because each plays with a different sound. Audi-Graph comes with all the musicians I’ve currently recorded on that instrument class (right now just Eb clarinet, but later it’d include bass clarinet, etc). It gives you the opportunity to learn against a “dark” sound or a “bright” sound. If you find the sound you like, you just practice to achieve that sound, and ignore the other recordings.
Rgames: I agree, room acoustics are very important. I recommend a particular mic, and suggest placement should be close to the instrument (about 3 feet). Audi-Graph is made to discount a lot of the environmental acoustics in favor of just the acoustics emanating from the instrument. Environmental resonance does add to the overall acoustics, but I’ve done everything I can to minimize it.
Audi-Graph is more geared towards helping student musicians (like 6-12th grade) in school practice rooms or at home, where studio level acoustics aren’t available. Or those like me coming back after learning bad habits and can’t figure out the subtle balance in embouchure to emulate that professional’s good sound.
Wes: Each Audi-Graph version is calibrated for that particular instruments range. Sure, you could play clarinet into the tuba version, but it’ll be applying the wrong math and highlighting the wrong data (and you won’t have the right example to compare to).
Yes you can play bass clarinet into the clarinet version and get roughly the same data, since the reach of their harmonics ranges are roughly the same, but I don’t yet have professional recordings included for bass clarinet (I record new people weekly for my research but not all are always included). When I do, bass clarinet will be included with clarinet.
I refrain from putting details like scaling, sampling rates, etc because for the most part that level of detail is irrelevant and distracting to all but the most tech savvy.
L. Omar: I agree, this is not setting a standard for matured players, but providing an example to learn with, similar to listening to recorded musicians and emulating their sound. I'd love to hear more about your spectrum analyzer development.
Does anyone have questions or insights to share on clarinet harmonics?
Anthony
------------------------------------------
Anthony Vasser - audi-graph@audi-graph.com
http://www.audi-graph.com
|
|
Reply To Message
|
|
The Clarinet Pages
|
 |
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
.jpg){kind=small}
