Saturday, June 8, 2013

Kashu-do (歌手道): Understanding Falsetto vs Supported Soft Singing

Falsetto, head-voice, flute voice, mezza-voce, voix-mixte, etc are only a few terms that bring more confusion to the vocal discussion than necessary.  The reasoning for this confusion is that the conversation is usual based on sensations rather than function.  A definition of any of these terms must take into account the structure of the vocal vibration, which is often very different among the people having discussions about this subject.

There are many effects that resemble each other in the listener's ear, but when we are speaking about classical singing, particular in the context of the operatic stage, there must be some agreement about the acoustical signature of a well-sung tone.

Falsetto is distinguished by the fundamental frequency carrying most of the acoustical energy of the sung tone.  We can call that F0-dominant.  This means that the source tone (the fold vibration) is so weak that upper harmonics are not strong enough to be affected strongly by vowel formants or the Singer's Formant.

By contrast, when the source tone is strong, the "pockets of energy" (formants) in the vocal tract can have a notable influence on nearby harmonics.  In a strong source tone, the dominant formant will be the one that has the greatest influence on a nearby harmonic.  The frequency range of the formant and its proximity to the harmonic makes the difference as to whether the harmonic will be boosted strongly or not.  The frequencies of the formants of the vocal tract change as the shape of the vocal tract changes.  This means that the frequencies of formants change with vowels (shapes of the vocal tract).  This is where the principles of vowel modification come in to play

In short, if a singer is able to take advantage of these formants (particularly higher ones), it is a guarantee that the source tone is strong (i.e. the folds are meeting completely.  There is relatively little loss of energy).

View the clip below for acoustic acoustic analysis of Matti Talvela's pianissimo F4.  Falsetto or well supported tone?




Corelli's pianissimi are legendary.  It is interesting what I was able to find below.




I grew up listening to a fascinating singer.  Tino Rossi, a French of Corsican (ergo partially Italian) heritage, had the most touching voice I had ever heard.  I always taught his production was more or less of a falsetto nature, but in this recording which I heard often in my childhood, this is a well-supported example of soft singing that remains acoustically energized throughout.  See clip.





Tino Rossi had a very efficiently produced voice.  He never sang very loudly.  One always got the impression that his voice could not sustain the stress of great volume.  Gentle singing was his skill.  It is refreshing to know that the voice that schooled my childhood ears is actually properly produced.




This Di Stefano example is fascinating.  I would have imagined it was purely falsetto, but there is definitely some energy in the high harmonics.




Björling has the very best balance between low and high resonance found in any analysis of the voice. His soft singing is also very convincingly "not" falsetto.  Yet there is room for greater efficiency as demonstrated by Gigli below:






The point to be made in this series of spectrographic analyses is that the ideal as demonstrated by Gigli in the final note is difficult to maintain.  Even the very best in history do not keep it consistently. Maintaining strength in the Singer's Formant require efficiency of source tone (complete closure of the vibrating folds) and enough pharyngeal space to guarantee a 6:1 ratio between pharyngeal circumference and epi-laryngeal circumference. This also requires a stable fold posture that can produce a tone rich in overtones to begin with.  Strength in one of the lower vowel formants (F1 or F2) is also part and parcel of a balanced tone as demonstrated by Björling singing loudly.  Ideally, that kind of balance should be maintained in soft singing.

The presence of the Singer's Formant is not the only way.  Corelli's famous diminuendi show a strong F2 and little strength in the SF.

"The Threshold of Acceptance" (i.e. what the listener accepts as supported soft singing) requires that there be a strong enough source tone that makes "formant influence" viable.  In any source tone (without formant influence) the fundamental is the strongest harmonic. A formants will boost the strength of an harmonic near its field of influence. With a weak source tone, the formants would not be able to influence an upper harmonic enough to make it stronger than the fundamental.  That is the case in falsetto.  Therefore, if a singer singing softly is able to produce strength in one of the upper harmonics (anything beyond the fundamental), the sound will not be perceived as falsetto.  Yet a source tone is neither 100% weak or 100% strong.  There are tones that lie in between and will induce equivocal spectrograms (difficult to distinguish whether falsetto or supported soft tone).  In fact most singers singing softly are singing in a mode that lies between falsetto and supported soft singing.

The difference can be observed through spectrographic analysis.  That is obvious.  But the question I get all the time is how do you produce a supported soft tone.  The answer is simple:  The same way you produce a loud tone except softer.  It sounds simple, but as always it is not.  The kind of full voice tone that leads to good piano singing must be efficient enough to produce strong upper resonance.  The folds must be able to close fully without excessive medial pressure (pressed voice).  This is why deep fold posture is a part of Kashu-do training.  Still, the "flow" sensation experienced in falsetto singing is part and parcel of training soft singing.

Falsetto and supported soft singing can feel very much alike.  But there are different kinds of falsetto.  The one we are most concerned with is full-closure falsetto.  In full closure falsetto, the folds close completely during the close phase of vibration.  However, the arytenoids are relaxed and allow air to flow freely.  The relaxed arytenoids are a response to the raised sub-glottal pressure that occurs when the folds are too shallow, necessitating greater medial pressure to maintain the length of the vibration cycle.  In other words, the arytenoids relax to avoid the pressure of pressed voice.

By contrast, when the folds are deep enough, the length of the cycle is achieved without pressing.  Therefore, the folds can be set into vibration with very little breath pressure (i.e. soft singing).

Full-closure Falsetto can be a good beginning, because the folds close fully and some breath pressure is built.  However the increased pressure necessary for louder singing is destabilized by the fact that air pressure is lost through the open arytenoid juncture.  From full-closure falsetto to full-voice, there is usually a sudden change in mode.  The sudden closure of the arytenoids causes a sudden rise in pressure that the pressed folds cannot resist.  The result is a "break" (cracking).  The break is only avoided when the soft mode is not leaking air as is the case in full-closure falsetto. That said, there are singers who have excellent control of the arytenoid juncture and can gradually open and close it as needed to create an effect of crescendo-diminuendo.

Soft singing must be practiced and each singer comes to it from a difference history.  Even a singer who has appropriate fold structure for supported soft singing, controlling the air pressure is also a question of coordination and practice.  Breathing technique becomes part of the equation.

In my baritone days I had excellent full-closure falsetto in the tenor high range (should have been sign of my tenor nature).  Converting that set-up to supported singing took time.  Developing supported soft singing requires the sensation of falsetto.  Warming up the voice with falsetto from G5 down has helped  tremendously in developing flow-phonation in a deep fold posture vibration.

The key in all of this is not to resort to an "either/or" mentality but rather adopt an "and" philosophy.  There is much too be learned from a falsetto sensation.  Most singers can crescendo diminuendo from falsetto to full voice and back in the comfortable middle and lower range.  The question is not to learn a technique for doing that in the difficult upper range, but rather to develop an upper voice that has the characteristics of the balanced middle and/or lower voice.

As for acoustic analysis, one needs to have done many hours of analysis to be able to make an assessment of what is actually occurring.  A trained opera singer will most likely produce some energy at least in the lower formants even when there is leakage of air at the arytenoid level.  An opera singer is used to singing with fully closing vocal folds.  Even when attempting falsetto there may be enough energy in the source tone to create some energy in the lower vowel formants.  Therefore, I would make the Threshold of Acceptance for supported soft singing to be at least activity at the second formant level.  The goal should be activity at the SF level, which would guarantee full closure including at the arytenoid level.

In closing, I would recommend analyzing Dietrich Fischer-Dieskau.  Dieskau's Pianissimi almost always included strong SF content. The SF was also relatively high pitched for a baritone.  More on Dieskau in upcoming posts.

© 06/08/2013

10 comments:

Banjo said...

Great post and lots of interesting clips, but I could not listen to the one with Matti Talvela. The clip was "private". Is it something that you could unlock?

I have some fundamental lacks of understanding.

Is medial pressure the same as air pressure below the vocal folds?

What is deep fold posture? Is it about the mode of the vocal folds or about the location of the larynx?

Benjamin

Kashu-Do said...

Thanks Banjo, I will unlock the Talvela recording. Medial pressure is how much the vocal folds press against each other to keep the glottis closed. Too much medial pressure builds up pressure below the folds (sub-glottal pressure). But sub-glottal pressure is dependent not only upon the folds closing but also upon the contractions of the muscles of exhalation.

deep fold posture refers to the amount of vertical contact area of the folds during phonation. A strong source tone creating strong harmonics depends on "deep fold posture" (appropriately greater contact area).

Kashu-Do said...

Talvela clip released.

Benjamin Laur said...

Great post, per usual. Hey, did you do those spectral analyses yourself, and if so, with what program? They look great. Haha, I really need to buff up on some voice science so I can foray back into the debates!

-Ben
http://belcantochronicle.blogspot.com/

Lloyd W. Hanson said...

I read your explanations with great interest. As with all attempts to define or analyze vocal production the descriptions can, at time, become troublesome.
According to the voice scientist, Ingo Titze, complete closure of the vocal folds is only one part of the equation describing the development of a strong fundamental and a rich harmonic spectrum. A more important consideration is the amount of time the vocal folds remain closed during each oscillating cycle. If the folds remain closed for about 30% of the time of each cycle, the resultant tone will be lacking in upper harmonic presence. The ideal is a closure of about 50% to 60%. If the vocal folds display a closure of 70% or more the tone will be observed as a "pressed" tone. The only way to access this closure rate is through the use of an Electroglottalgraph (EGG). Don Miller with his use of VoceVista had developed a strong basis of study of the effect of glottal closure time.
In pianissimo singing, which os often described as a well supported pianissimo singing, the vocal folds not only completely close during each oscillating cycle (as you have well described) but also stay closed for the desired 50% of the cycle.
However, the vocal folds are thin, not as thick as they are for forte singing. Consequently the vocal folds do not display the typical vertical wave opening of the folds from the bottom to the top which is characteristic of their function in forte singing. Instead, the folds show only a minuscule opening from bottom to top because of their thinner configuration.
The traditional way for the singer to learn this thinner configuration of the vocal folds is through the studied use of a balanced onset of tone. Such onset practice teaches the singer how to best balance breath pressure with thinner vocal folds such that a clear but gentle or soft tone is produced; soft yet rich in harmonic content.
I admire the spectrographic display that you are using, especially its apparent ability to identify the frequencies of each emphasized harmonic. Very useful. And I also admire your direct and most accurate descriptions of what is occurring during each analyzed example

Kashu-Do said...

Benjamin, I did the spectrograms myself using a very cheap but sophisticated program called Voice Analyzer Pro for the MAC. Available at the Apple app store. It has two windows Spectrum and Spectrogram and uses the computer's internal mic. So by reducing the size of my browser window, I fit the youtube screen around the Spectrum and Spectrogram display. The program also calculates formants and displays the dominant formants. I will be doing a tutorial on this software soon.

Kashu-Do said...

Dear Lloyd,

It is always an honor to have your commentary. Your explanation above is traditionally what had been observed. But as is often the case, the samples observed in any vocal study are not broad enough to reflect all the possibilities.

Don Miller (the same Voce Vista creator you mention above) has done well in expanding our views of what we accept as healthy singing. Don contends that not all high CQs (close quotients) mean pressed voice. He did EGG examination on many respected professional singers including several singers at the Lindemann Program at the MET and concluded that "robust" singers (e.g. dramatic voices) display higher CQs without sounding pressed. He has examples of two tenors on the Voce Vista website executing "messa di voce" successfully with different CQs. The lower CQ displayed some falsetto patterns half way through the diminuendo.

The data supports the theory I have had about greater contact area. There have been many studies since Titze/Sunberg that shows a more efficient pattern where the fold body is stiffer--that is when the stretching action of CT is strongly opposed by Vocalis (internal TA). When the fold body is stiff, the vibration is carried by the fold cover, which is precisely what we want. It is not the thinness of the folds that produce an easy vibration but that the vibration is being carried by the soft flexible tissue (i.e. cover) requiring less sub-glottal pressure.

See this post: http://kashu-do.blogspot.com/search?q=fold+vibration

There are many in-between patterns. The lighter tenor in Miller's examples shows a proper modal pattern when the singer sings forte but increasingly falsetto as he diminishes. Such a singer would gradually open the arytenoid junction and allow the pressure to decrease to give the effect of a diminuendo.

The other model is to maintain the base sub-glottal pressure that sustains the oscillation of the folds and increase and decrease additional pressure to create the diminuendo. This reflects Corelli's famous Celeste Aida Bb.

Kashu-Do said...

I mean to create the crescendo-diminuendo

Hemichromis said...

It seems that Gigli doesn't turn his voice either. he sings F#-A on a second harmonic dominance and even a B on a recording of nessun dorma.
I say second harmonic as it probably is the second formant influencing the second harmonic on the B but certainly not on F#-Ab.

Here is the high B:
http://www.youtube.com/watch?feature=player_detailpage&v=ru8Lf_SAPIo#t=154

To me it sounds as though he nearly acheives a 2nd formant dominance but then goes back to 1st formant.

I would love to hear your commentary on this

Kashu-Do said...

Sorry for the late reply, Hemichromis. I uploaded an acoustic analysis of the Gigli clip you sent me. The F# at tramontate o (STE)lle is definitely F2 dominant on the E vowel. The dominant peak is around 1485 on the fourth harmonic as it is on the third harmonic for the top B. This fourth relationship between also occurs at the final "pensier" of La donna è mobile. Textbook! Gigli might sound open because he is able to sing a perfectly clear open [E] vowel. But it is precisely the vowel that is needed here to access the second formant. I hope this helps.