Kashu-do (歌手道): Appoggio and Mask Resonance

I was very tempted to comment on Jack Livigni’s recent post, on appoggio and giro del fiato (turn of the breath), which generated a great deal of discussion. One of my most repeated statements here on Kashu-do (歌手道) is the following: Science is only the beginning! I make this statement because modern vocal research does not provide all the answers.  The science-based voice teacher’s job is to take all the proven elements into account and then fill in what is missing.  For this, we take advantage of the principles of the classical schools of singing.  Science is there to clarify the principles of the traditional schools and help make corrections particularly when traditional language is equivocal or wrong.

Singing is paradoxical.  I always speak of pressure and flow.  The two are in fact interdependent.and related to the subject of this post.

First, can one support by leaning the breath behind the mask?
Scientifically, no!  In proper singing, the breath does not arrive in the nasal cavity except when singing nasal vowels.  However, the sensations of intense vibration felt in the sinuses are directly related to what the singer experiences as support.

The efficient propagation of sound waves in a resonant vocal tract is felt in the sinuses through bone conduction. A resonant vocal tract depends not only on adequate adjustment of the vocal tract but on the correct timing of the vocal fold opening.  According to Donald Miller, the vocal folds must remain closed through the last maximum of the harmonic that dominates the spectrum of the sung tone.  This means that a singer producing an Ab4 on the vowel [a] (F1=c. 800 Hz, F2= c. 1200 Hz). The common setting of F2 (second formant) on H3 (third harmonic) requires that the folds remain close in excess of 75% of the cycle.

Tenor G4.bmp

The left side of the picture shows the third harmonic to be the carrier of most of the acoustic energy. The audio signal on the upper right lines up with the glottal signal below to show that acoustic energy is generated while the glottis is closed (the glottal signal peaks during glottal closure).  The audio signal shows that the glottis remains closed for nearly three complete peaks.  The glottal signal weakens during the drop-portion of the third peak.  The graph shows a CQ (contact quotient or closure quotient) of 73% as measured by the EGG device.

Tenor G4 spectrum.bmp

The spectrum view shows the green cursor on the third harmonic at 1202 Hz (the second formant value of the vowel [a]).  The strong peak between the orange lines is the singer’s formant.  This is an ideal spectrum for G4

 This long close cycle (completely appropriate) builds up appropriate subgottal pressure that the singer experiences as being connected with the body. In essence, the same pressure that gives a sense of support provides the conditions that produces the sensation of mask resonance.

If however the glottal closure quotient is too high, there will not be enough breath flow and therefore not enough sound pressure (i.e. resonance) to induce the intense bone conduction that is felt as mask resonance.  In effect, a balance in glottal flow/pressure must accompany vocal tract adjustments as well as the part of breath compression controlled by volume of air in the lungs and the global activity of core muscles, intercostal resistance and diaphragmatic activity.

In actuality, there is no column of air that goes from the lungs through the glottis and up and around to the mask. Still, the path of the breath does take a 90° turn from the laryngeal pharynx to the buccal pharynx. Many singers do not generate enough subglottal pressure in their lower range to sense such strong resonance in that region. Often, the sudden intensity that is felt in the mask in the male passaggio and upper range, or in the female upper middle range may be experienced as a sudden turn into the area where the vibrations are felt, or a giro del fiato (turn of the breath).

In the end, the singer’s experience of the voice is based on sensory feedback and indeed sensory memory.  The singer’s proprioceptive sense combines aural as well as vibratory feedback to complete the experience. Visualization and metaphors are important tools in the teacher’s kit.  But they must not be the only ones.  Teachers are sometimes frustrated when they are not able to pass on their sensations to the student.  Most often, the student has not yet achieved the muscular dynamics that would make such sensations possible. This is just as frustrating to the student.  When we have a scientific understanding of the conditions necessary to produce such sensory feedback, we can help the student develop these conditions over time. When the physical components are ready to produce the sensations, proprioception and metaphors become even more important. For that reason: Science is only the beginning! But I should add:

Without a beginning there is no end product!

© 10/28/2010

22 thoughts on “Kashu-do (歌手道): Appoggio and Mask Resonance

Add yours

  1. In the end, the singer's experience of the voice is based on sensory feedback and indeed sensory memory.

    I know that this is outside your concern in this article but I would add to that list of influences “verbal habits” and “social conditioning.” The relation between our proprioceptions and the expressions that we use in describing what we feel is one of mutual influence; and our choice of expressions is influenced by what we hear others say. We do not invent or select all our descriptive expressions on the basis of a solitary scrutiny of our experiences, but adopt them under the influence of what we hear from teachers and from other singers.


  2. Without a doubt Miles! Your commentary is relevant not only to this post but probably to all posts on this blog. Our social conditionings (which influence our verbal habits as well)are always front and center. When they are good we encourage them and when they are bad we replace them. The influence of the teacher is of course of primary importance to the singer's development and future. Therefore, it is incumbent upon the teacher to be equipped with knowledge and indeed with a diversity of potential approaches dependent upon the students' needs.

    Yet as artists we must eventually become selective in allowing ourselves to be influenced. With experience comes real knowledge, which should be used to inform our decision making, relative to what we are being given as information.


  3. I'm not sure I understand what you're saying about the CQ.

    First off, just as a detail, the post says Ab but the images are for G. G is in fact just under 400 Hz so I assume this is what you mean.
    Second, in an H3 dominated note, the tallest peaks should divide the cycle into thirds, so wouldn't the CQ need to be above 67% not 75%?
    Third, this is the first I've heard about the CQ being higher than the peak of the dominant harmonic, but it kind of makes sense. Does this mean that the only way to make H4 dominant is to exceed 75% and the only way to make H5 dominant is to exceed 80%? What about singers where the singer's formant is dominant? Does the rule not apply to those strategies? Do they achieve CQs in the mid to high 80's? Or do those singers, despite achieving high singer's formants, not actually produce a tone where the SF is higher than all the other harmonics?
    Does this mean switching to a lower dominant harmonic makes the note require less breath pressure? Could a tenor singing high C at 523 Hz on a vowel around /a/ or /u/ (F2 around 1050) use F2 to resonate H2 and only need a CQ of 50%?

    Also, I thought your view was that a CQ of around 55% +/- 5% was the goal. Has D.G. Miller changed your mind, or are you advocating a different resonance strategy than what the spectrogram you posted shows?

    Thanks for an interesting post and for any light you can shed on these questions for me!



  4. Lots of good questions Klaus!

    It shows how much you really keep up with the discussion here!

    1. I think there is value to Miller's work relative to the top of the male voice and middle of the female voice. When I listen to Del Monaco and Corelli or even Giocomini or Galouzine, I hear this intensity. We can even extrapolate from the audio signals to come up with an approximate CQ value.

    2. This is one ideal however. Perhaps an old school ideal. I think there are questions not yet answered. For example, how does supraglottal inertia affect these values? Would the inertial component register as a closed glottis or would the EGG show only the vocal fold dynamics without the effect of the sluggish air above the glottis?

    3. Why does Björling sound so fluid when his acoustic strategy matches that of Corelli or Del Monaco (in terms of the dominant harmonic in the upper range)?

    To your questions: When the 3rd harmonic dominates, the expected CQ according to the Miller theory would be around 83% (i.e. the entirity of the first two peaks plus the maximum of the third peak. In essence 5/6 or 83.33%. What I observe is that it is sufficient if the folds are closed through the beginning of the third peak, not even at the top of the maximum. So somewhere beyond 2/3 seems to be enough to excite the 3rd harmonic. Or a little more than 75% is enough to carry the 4th harmonic and so on.



  5. I always love your bloggs Jean-Ronald, but am way too lazy to get into online discussion. One of these days I will track you down in the hope of having a very long conversation. This particular blogg entry made me even more determined for this to happen. 🙂


  6. To Klaus, Cont…

    With respect to singer's formant singers like Domingo, I suggest there are other things at play. Fold closure is not the only thing that determines strength in the singer's formant. We know that that formant is highly affected by the ratio between the circumference of the Epylarynx and the pharynx, somewhere around 1:6, 1:7, 1:8…(no lower than 1:6). That influence plus a vowel formant within the SF region could make the SF the dominant energy on the fith or sixth harmonic. The narrowing of the Epilarynx is (insterestingly enough) accomplished by the oblique inter-arytenoids, which are also necessary for medial compression. I would submit that the SF strategy might not be as linear as we would like to think.

    The strategy of C5 (523Hz) with F2 on H2 is not a foreign concept. I believe Gioacchini Livigni has referred to this at some point referring to the strong 6th harmonic that results instead of the 5th as is the case with post Lauri-Volpi tenors. Worth investigating.

    After spending a couple of days with Don Miller, I believe there is relevance to his theory. How far it extends is yet to be confirmed. He, himself feels that research needs to be done on this.

    Sorry about the G scale. I mention an Ab scale but the scale is actually a G scale. In a sense a venial sin on my part. Since G or Ab would have the same strategy (i.e. F2 on H3 for the [a] vowel), it does not make much difference.

    Thank you for your wonderful observations as usual. I wish I had time to make all the corrections of the blog. I must do this eventually if this is to remain useful over the long haul!

    Warm Greetings,



  7. JR, you said that Bjoerling's strategy matched that of Corelli and Del Monaco when it comes to the way he tunes the harmonics. This is not the cast. Bjoerling's top notes are similar instead to Caruso's, Lauri Volpi's and other old school tenors. The dominant harmonic is the 5th.



  8. Dear Jack,

    The Björling et al comparison was made relative to the dominant harmonic (H3) in reference to the CQ question only. Elsewhere I refer to your commentary regarding the upper part of the spectrum (i.e. H5 and H6). It is important to figure out the relevance of the tuning at the SF region. I am interested in your commentary regarding this difference. With examples so we can determine whether we are talking about vowel formants or indeed the SF.


  9. @Jean Thank you for your answers. I'm particularly interested in the CQ question because my voice has a tendency to crack in the passaggio when I'm not feeling 100% (even a mild cold) and I'm trying to figure out if it's because of a lack of adductive strength which time will fix because I originally learned those notes in a loose form of production, or if I need to change strategy.

    @Jack How would you describe the difference between the Bjorling and the Corelli high note strategy?


  10. @Klaus: I think Jack's new post on covering or not covering the passaggio might have some answers for you. In my own transition to tenor, the most difficult part has been dealing with the passaggio. It is a very delicate muscular balance to achieve what Fisichella refers to paradoxically as the “open-covered” approach to F, F# and G. The idea is to maitain the “lean” efficiency of phonation as the voice goes to F2 dominance. I believe that the tendency in the upper range is to sing thinly (i.e. too little fold mass). To achieve the acoustic turn more fold mass is needed. Often we go too far resulting in loose phonation. I would recommend Jack's idea of singing “open and high” at first to achieve efficiency and then gradually add the adequate mass without losing the appropriate closure.

    @Jack: Waiting for your commentary on the pre vs. post LauriVolpi tenors.



  11. So I played around with VV this morning and sang the same G scale. My spectrogram is clearly H2 dominant on the G, with F0 and H3 and the SF cluster all at about the same level as each other.

    I didn't have the CQ hooked up but I'm usually in the 55-58% range.

    It seems to me I can interpret this one of two ways.
    1. I have a more lyric instrument and my strategy on a G is different
    2. I need to work on my adductive strength in order to get a strong enough closure to allow H3 to become dominant.



  12. Hey Klaus,

    F1 on H2 is not appropriate on G4 regardless of vocal weight. This is an acoustic issue. This means the larynx is climbing to track F1. This is also confirmation of a slightly thin phonation and hearing you, I imagine the arytenoid end of your folds is relaxed. Adduction is one way of putting it. But as you know, I advocate fold depth as a part of the fold posture package.

    My thoughts for remedy: Maintain the fullness from the lower range and/or work for better adduction. Begin the tone with an “emotionally charged” pianissimo and then cresendo. This is one way of working out the adduction part.



  13. I'm sure that my larynx isn't riding up, both by feeling it with my hand and by observing it. And I can definitely tell when it comes up when I'm singing. Is there some other way to have F1 be reinforcing H2? Or could it be that F2 isn't getting up high enough to hit H3 and is partially reinforcing H2? H2 is only a few dB higher than H1 and H3 on my G4, and the SF cluster is also almost at that same level.

    I thought that maybe it's a CQ thing. If you need to hit around 65% to get H3 to be the dominant harmonic, then perhaps I'm lining up the articulators properly but not giving it the proper “weight” at the vocal fold source to ring. Is this possible?


  14. Thanks for your previous reply. I went at it again this morning with a full night's sleep and I can get the appropriate H3 ring to happen now.

    It definitely feels more energetic and fully engaged. Probably after a day of rehearsing Handel yesterday I was just physically done.



  15. Hi JR! How's your Radames coming along!

    I disagree with you concerning your comment about no air passing through the nasal cavity in proper singing. I believe that to create the most effective nasal resonance some air does pass through the nasal cavity.



  16. Dear Rebecca,

    Sorry for the late response. Jetlagged and teaching a lot. The sensations in the mask are so important to singers that it is difficult to believe that no air needs to pass through the nose. Not everyone is going for the same sounds of course, but it is a common understanding that air passing though the nose is tantamount to nasality. One of the few experiments by the great pedagogue Vennard which remains uncontrovertible is his experiment of stuffing cotton and milk in the sinuses to test the participation of the sinuses in optimum resonance. It was found that the nasal cavity does not contribute to the vocal tract resonance at all and if it participates it produces inefficient results. The experiment has been reproduced by other scientists over the years with the same results.

    I have found a slight nasality to be often symptomatic of “getting close” to ideal fold posture for balanced resonance. But this adjustment is not the final product as it reveals a mild glottal squeeze leading to high larynx and low soft palate if produced at higher volumes.

    The sensations are very deceptive. The vibrations we feel in the mask/sinuses are products of bone conduction of the vocal tract resonances and not actual air going through the nose.



  17. Hope it's OK to chime in a little late in the discussion…

    I would like to go back to the statements in your post:
    According to Donald Miller, the vocal folds must remain closed through the last maximum of the harmonic that dominates the spectrum of the sung tone.
    Is this a conclusion based on experience that Miller has gathered while looking at a lot of graphs? I cannot fathom any obvious acoustic reason why that should be the case. It makes sense that a larger closed quotient should generate a source sound richer in harmonics, but I seriously doubt that there is any “rule” for “timings” of the opening in relation to the dominating harmonic. Generalizing such a rule to different harmonics would lead to absurd consequences as KG makes clear in one of the commentaries. Note that it is not the case that a wave traveling towards the glottis somehow will be “swallowed” during an open phase and reflected during the closed phase! Also, I cannot see much support in the waveform for your claim that that acoustic energy is generated while the glottis is closed and that The glottal signal weakens during the drop-portion of the third peak .
    In fact, there are large peaks in the beginning and in the end of the marked period. (How easy is it to line up the plots in time, btw?)

    Sorry about just having negative comments today… Love your blog anyway!


  18. Dear Martin,

    Miller has observed this by exploring EGGs with professional singers, some at the Metropolitan Opera a little while back. He is not 100% sure about his theory and is still investigating. In fact he wanted your input.

    As per his book, it is not a matter of the waves getting swallowed by the open glottis but rather the timing of breath stroke to reinforce the wave precisely at the time that it changes direction from toward the glottis to away from it–as in pushing a swing at the right time to make it oscillate higher. This is consistant with resonance theory as I understand it, but you are the expert and I am ready to learn.

    I do not have the reasons for this, but as I observe the clips he collected with EGG included, there seems to be a correlation. As I implied, I am not sure it is necessary that the glottis remains closed through the final maximum. I think rather that it should be closed beyond the next to last maximum. As for the situations when the SF is the dominant wave, as in the Domingo model, we wonder why there is such extreme subglottal tension involved. This would lead me to lean in agreement with Miller's hypothesis or at least to be curious about it.

    I will check the spectrograph I posted against Miller's data. I may not have lined up the EGG and microphone signals adequately.

    At any rate, we were able to see a corrolation between the mic signal and the EGG as to where precisely the dip occurs, which normally corrolates with glottal opening unless the folds are close enough to each other to cause another small rise during the dip.

    Don't ever apologize for correcting me or to question me. We learn a lot here from your expertise. I wish I had time to catch up and understand this in a deeper level.



  19. You said but rather the timing of breath stroke to reinforce the wave precisely at the time that it changes direction from toward the glottis to away from it–as in pushing a swing at the right time to make it oscillate higher.

    First of all, what we see in the microphone signal is the pressure variation as recorded by the microphone. If the microphone is say 20 cm from the glottis, we will get a time delay of about 60 ms, to be compared with the period for a G4, which is about 2.5 ms. How would you go about to line up the EEG and microphone signals? To say something about the timings I would prefer to have the microphone close to the glottis.

    Moreover, I am not convinced by your reasoning above. One important complicating issue is the time delay between the “breath stroke” and the build-up of pressure, which is an important mechanism to enable a self-sustained oscillation.

    I would rather say that a lot of action is taking place between opening and closing of the glottis. When the glottis opens, a pressure buildup occurs due to inertia, and a corresponding pressure dip occurs just after the closing. The smaller the open quotient, the closer in time these peaks will be, which corresponds to higher harmonics.

    Well, well, it is easy to go wrong with this kind of reasoning without a lot of data or simulations to back it up…


  20. Voce Vista has a mechanism for determining the time delay based on the microphone distance, and this particular clip comes with the delay value. It is one of Don's clips. I may have lined it up wrong. I will double check the values.

    At any rate, I think you need to have this discussion with Don. I would be interested what you two come up with. The timing of the breath stroke is crucial I would think whether by the complex propagation of waves relative to th contours of the vocal tract or by the timing of the glottal oscillation or both.

    I think there is something worth investigating here. I am not presenting it as a conclusion but as a hypothesis, which begs to be studied.


  21. I find the relationship between CQ and enahnced harmonics interesting.

    I have recently got rid of a loose high voice and become a little pressed and thin(High CQ). This has led to my 4th Harmonic being very much enhanced, by as much as 40db at lower dynamics!

    My 3rd however is very difficult to enhanced at anything below A4. High B and C are heavily 3rd harmonic dominant with a corresponding relaxation (lower CQ?) though i can enhance the 5th alongside.

    Previously i thought this purely a Formant issue however it now looks as though there is more to the vocal puzzle!

    One thing is clear, Pressed phonation MUST GO!!!

    Would a comparison add Credence to your theory?
    compare Giacomini and Corelli, Giacomini far more pressurized with dominant 4th harmonic and even more powerful 5-7th. Corelli, far more relaxed production and 3rd harmonic dominant.

    Their approaches may be too dissimilar to make this comparison though!


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Website Powered by WordPress.com.

Up ↑

%d bloggers like this: