Function of the Singing Voice

Last Fall, I attended a course, "Function of the Singing Voice", graciously hosted by Prof. Johan Sundberg and his wife, Dr Ulla Sundberg (phonetics & linguistics) at their lovely Summer home south of Stockholm.

It was a very interesting week, full of lectures and experiments. We could easily have spent another week doing more experiments, but I guess I'll have to sign up for the research program for that...

Since I'm an engineer by training and profession, this approach to singing feels absolutely natural, and I am convinced that a more systematic understanding of the science behind singing only helps me become a better singer.

Here are a few of my own observations from the course:

First of all, I felt that maestro Bengt has taught me well. :-)

In general, it was great fun to get an introduction to what we can describe scientifically about the voice. There are many things which we cannot easily measure, and there are obvious problems with just about all experimental setups. Still, there is a wealth of extremely useful knowledge, which I think every singer could benefit from. How to shape vowels (and the inherent problem for sopranos - see below), is a very obvious example.

(I found some slides here that seem to be based on Sundberg's material, for reference.)

The singer's formant is of great importance for male classical singers. The main reason is that it helps the singer to be heard above an orchestra. For rock and musical singers, it offers no such benefits, as the music usually has a different frequency spectrum, and the singers rely on amplification to make themselves heard. Sopranos cannot produce much of a singer's formant, but don't strictly need to, as they make themselves heard anyway.

A number of different techniques exist for creating a singer's formant, e.g. raising your tongue towards the palate will help modify the third formant (but of course also affect articulation), but the "natural" way of producing it is by lowering the larynx and enlarging the ventricle (epilarynx resonator) above the larynx. This is also described pretty well here.

The obvious main challenge for sopranos is making sure that the fundamental frequency, f0, doesn't rise above the first formant, f1. This is called "resonance tuning"), and one of the main techniques is raising f1 by increasing the jaw opening. Essentially, sopranos have little chance of getting the vowels right in their upper range, mainly because the fundamental frequency is well above the frequencies that shape the vowel sounds, and the formants which normally shape the vowels must be kept above the fundamental frequency for a pleasant sound.

The course helped clarify my understanding of how vowels are shaped, which has helped me in my practice (making me use the tongue more deliberately for articulation), and the notion of "tracheal pull" has helped me control air pressure and timbre - for example by noting that I can help the tracheal pull not only by lowering the diaphragm, but also by expanding or raising the chest, thereby either allowing the lungs to expand outward, or simply increasing the vertical room by "raising the ceiling", rather than "lowering the floor".

I did note with some disappointment, however, that there exist no good techniques for understanding the higher formants (it was only stated that they are generated in the larynx, and affect the "personality" of the voice). I noted with interest slides 29-32 from these seminar slides, where the spectral analysis charts for Corelli and Pavarotti show some similarity, as do the charts of Bjoerling and Domingo. I will draw no further conclusions.

Overall, it was a wonderful week, with great lectures, wonderful food, interesting labs and very good company.

Into the depths...

I turned 40 recently, and true to our tradition,
we invited friends and gave a concert.

I gave myself the opportunity to sing a passage out
of L'Elisir D'amore: the Nemorino/Dulcamara duet
(Voglio dire...), followed by the Nemorino/Adina duet
(Caro elisir...), followed by the Belcore/Nemorino/Adina
trio (Tran tran tran...)

My wonderful wife, Katarina sang the part of Adina, of course.
Peter Kajlinger did both Dulcamara and Belcore.
It was my
first time singing with Peter, and it was truly a joy. He has an
air of confidence and generosity about him, and he's a gifted
comedian as well as a consummate professional.

Coach Bengt made a valiant effort to make it despite first
attending an audition in Stockholm.
He arrived just in time
to fire off his contributions in rapid succession:
the Edgardo/Lucia duet (Verranno a te...), Cielo e mar,
and the Cavaradossi/Tosca duet (Mario, mario...), before
he was allowed to sit down and have a glass of wine.
He also helped spice up the party with his ukelele...

Bengt's singing reminds me of Franco Corelli.

Our Tosca was Maria Sloth, who has traveled a rocky road ever since
she was hailed as "the next Birgit Nilsson" at school.
She resolved the pressure by simply not singing at all for the longest
time, but a few years ago, she made a tentative comeback (in our
living room, if I recall...) She has now found her way back to her
first teacher (Jacqueline Delman), and has found her voice
again, after more than 20 years.

Maria is a true dramatic soprano. Managing a dramatic voice is
no picnic (not that I would know...), but when it works, it's truly
something to experience. Thanks, Maria, for a wonderful
performance.

Katarina pulled off being a great hostess, singing Adina, Lucia
and the Bell Song from Lakmé. A perfect Birthday gift, and a
wonderful evening.

As for my own performance, I was actually quite pleased.
The acting was mostly acceptable, and I did attempt to go
all out (something that I've had trouble with, as most
amateurs do). The voice sounded stable and well grounded,
although the singer's formant can be improved. The legato
line was good.

A few weeks later I had a very interesting conversation
with Prof Johan Sundberg at the Dept of Speech, Music
and Hearing at KTH. Prof Sundberg's theoretical depth
is truly impressive, and just chatting with him for a couple
of hours taught me a great deal. Not that I can easily
account for what, exactly, but the combination of his
comments, the stuff that Bengt has been telling me for
years, and bits and pieces I've picked up elsewhere,
inspired me to go home and try a few things...

One item of discussion was a picture from Manén's
"Bel canto" book, showing a closed larynx. The picture
looked like the larynx was in a high position (not very
interesting from a singing perspective), but the text
indicated that it was in a low position. Closing the false
vocal cords, as indicated by the picture, in a low position
is not easy, apparently. Without an X-ray machine at
home, it's difficult to conduct your own experiments,
obviously...

But Manén mentions a "click" sound, "caused by ... a clash
of air rushing in from above and below the larynx"
(according to Manén - I can't judge whether this is a
correct description). I set out to attempt to reproduce
this effect. It took two days of hard work, and significantly
more downward pull of the larynx than I'm used to,
combined with a kind of squeezing (I imagined the larynx
as a "V", which needs to be squeezed together into an "I",
partly inspired by a drawing that Prof Sundberg made
during our conversation). After a while, I could do the
exercise suggested by Manén: a slow staccato scale
(apparently, from "Studio di canto", by A. Busti),
producing that faint "click" between each note, as the
larynx closes and re-opens. I don't know whether this
proves anything, but the tangible effect was a marked
improvement in my singer's formant (more "ringing",
and more power - basically, everything improved, as
far as I can tell; Katarina seems to agree).

This seems to be in line with Sundberg's observation
that the singers formant is formed in the larynx, and with
Bengt's insisting that the overtones are not produced
the way you'd expect. Not that I doubted this, but it's
always good to be able to internalise this knowledge,
so that your body agrees as well.

The added difficulty is exerting a significant downward
pull on the larynx while "freeing" the muscles shaping
the upper air passages. I would not have been able to
do this a year ago (much less four years ago, before I
started taking lessons for Bengt), even if I had
understood that it were needed for the kind of
singing that I aspired to.

I've been resting a few days now, to avoid over-excerting
myself. It's difficult, since I want to keep exploring this
newfound sensation. I'm looking forward to many more
years of continued discovery. This is obviously only the
beginning.

charting emotional content

It seems as if my thoughts on emotional indicators have been
answered scientifically long before I started thinking them.
Not surprising - the trick for a layman is of course to find all
the research. Much of it is not available via the web, unless
you belong to a participating research institution. If I were
to park myself in a library, I might have better luck...

However, I did find this:
Emotional Expression Code in Opera and Lied Singing

(Presented by Dr. Eliezer Rapoport at the 1996 Israel
Musicological Society Annual Meeting)

The author has run computer analyses of recordings
of great artists: Callas, Caballe, Margaret Price, Pavarotti,
Kraus, et al, and has performed a systematic breakdown
of how these singers vary their voice and musical expression
in order to express appropriate emotions. The article lists
some 50 different indicators. I have yet to read it thoroughly,
but here is an extract:

A higher degree in excitement than in the C or R modes is achieved by introducing a third element: pitch transition; a gradual increase in pitch in one or two stages from the onset to the sustained stage, mostly practiced by tenors. This is not singing off-tune but is a deliberate way of shaping the tone, endowing it with some extra qualities: openness, brightness, life, timbre embellishment, and expressive­ness. These are the qualities that bel canto tenors use in expressing love, exhilara­tion and happiness. (The tenor is the hero and the lover in Italian operas). T modes are perceived as a timbre effect. After becoming aware of it the trained listener can discern this gradual transition as a pitch effect. Figure 5 displays an example of the T1 mode taken from the aria "La rivedra nell'estasi..." from Un Ballo in Maschera by Verdi, sung by Luciano Pavarotti, expressing love (marked in Performance

[...]

Score No. 3). The aria "De miei bollenti spiriti" from La Traviata by Verdi, and the preceding recitative, express great happiness. Pavarotti and Alfredo Kraus use the T1 and T2 modes extensively in this aria. Further on in the aria at the climax of happiness, the phrase "io vivo quasi in ciel" is repeated five times, each time leading to a climax - sung by Kraus in the T1 mode.

This sort of thing appeals to my engineering brain. (-:

A technocratic, optimistic spin on this might be that, given
such wonderful analysis tools, and the ability to describe
exactly what is going on in these wonderful performances,
we should be able to resurrect, and perhaps even improve
the old magic.

The only remaining problem is of course to attract singers
that are talented enough, and devoted enough, to subject
themselves to the years of training that will still be required.

(Actually, I think that isn't a problem, because lots of talented
and devoted singers do this today. As with many other fields,
the trick is to integrate research into the teaching and wide
practice of the art. Most singers I know will probably be
slightly put off by discussions on Fast Fourier Transforms,
vowel formants, vibrato periods and unit pulses...)

apologies to RSS subscribers

I've been republishing the last posts several times.
If anyone is subscribing to this blog, I apologize.

I've switched to the new Blogger version, and it
was playing tricks on me. The word wrap wasn't
working right, and while the preview looked okay,
the published version looked horrible.

... at least using the Opera web browser, but honestly,
when writing a blog on Bel canto, using another browser
is surely blasphemy?

neural response

I've been out fishing lately.

I started thinking when listening to a concert recently. It was
 an accomplished opera singer, who performed, by all accounts,
 very well. I was quite pleased that I could actually enjoy myself, 
and not just analyse her singing to death. (:

However, there was time for some analysis too. I couldn't help
noticing how she sang with basically the same timbre
throughout. This is not unusual. It is what most singers seem to 
do.

She used a fairly wide range of "indicators" to signal emotion, 
and we, attentive listeners, were eager to comply and respond
with the appropriate emotion.  I wouldn't have reacted to this 
a few years ago, but having realised how much a singer can 
actually vary his/her vocal colour, you can't help not longing
to hear that on stage. 

Actually, the thinking started long before this. A couple of years 
ago, I listened to a live rendition of "Nessun dorma", with a 
reasonably skilled tenor. His middle range was uninteresting,
though, which obviously becomes problematic in "Nessun
dorma". To exaggerate a bit, he stumbled through the aria and
made a nice recovery on the high notes. To my surprise, my
friends hadn't noticed his relative failure in the 90% of the
aria that doesn't consist of "money notes". I guess that's why
they call them "money notes" - if a tenor can only deliver the
high Cs, all is forgiven (in Nessun dorma, it's a H).

So here's the theory. Nessun dorma is all about anticipation,
but since everyone in the audience is so full of anticipation of
that last high H, the singer can probably get away with just
indicating the rest of the song. After all, none of it is real.
The artist builds an illusion, and we, the audience fill in the
blanks.

So why go through all that trouble to deliver real emotion,
when it's obviously possible to just indicate, and let the
audience do the work? Shouldn't there be a difference, after all?

I started googling for neurological papers to find out if we are
actually able to detect what goes on in the brain when e.g.
listening to singing.

I turns out that we can - almost.

Robert Zatorre in Montrëal seems to be on the forefront of
neurological studies related to music. One paper has the
impressive title

"Emotional responses to pleasant and unpleasant music correlate with activity in paralimbic brain regions".

Unfortunately for me, trying to prove my theory, we seem to
have some way to go before we fully understand how the brain
processes emotional content in music.

It turns out that listening to music triggers a lot more brain 
activity than does e.g. speech. Several different parts of the 
brain (even the "reptile brain") are engaged, and the result 
is collated into a unified experience.

This paper does claim to support the hypothesis that "there
may be a dissociation between perceptual and emotional
responses to music".

Furthermore, the paper states that "it is possible that music
that induces different types of emotions would recruit
different neural substrates. This may be especially likely
if emotion is elicited through memory or association,
rather than spontaneously."

That last bit seemed to relate to what I was digging for. I was
thinking that if an artist uses physical or vocal "gestures"
to indicate to the audience what they should feel, this may
have the effect that the audience relates to the music in a
slightly more intellectual way, having to draw on their
memory for the proper understanding of the gestures. 

Perhaps other components in the music speak directly to 
our more intuitive systems, triggering a direct emotional 
response which we will have great difficulty suppressing?

Of course, tons of complications arise if one would try to
measure this in a scientific study.

Of course, it could be that our willingness to be swept away by
the artist is the most important factor in enjoying a musical
performance. Perhaps our brains are clever enough to
provide us with the necessary emotions if we just don't get
in its way? Call me stubborn, but I don't think that's the
whole truth.

Something to look forward to

My beloved wife chastised me for writing that I didn't have much
to look forward to (here). I admit that that was only partly true.
I meant that I didn't have anything like a full-scale production in
front of me.

A few fun things have happened, though.

Katarina and I were engaged to sing on a Tallinn cruise. I was
really supposed to ride along, but since we enjoy singing, we
promptly decided that I should be part of the program. Such
"coups" are always a bit scary, since they can backfire if the
customer ends up thinking that I didn't contribute to the
enjoyment. Or maybe that's just me thinking that I'm still
not good enough for professional engagements... As it was,
I only participated in a duet ("Rosen in Tirol" by Zeller) and
one aria ("O Paradis", by Meyerbeer). It went quite well,
though, and I felt like a real pro... almost.

The next occasion was our traditional Christmas concert,
now in a new church (since we have moved). It was well
visited, and people seemed pleased. I sang some Christmas
hymns, and some arias from Händel's Messiah (Comfort ye
and Every Valley). We recorded it, and listening to the
tape, I was horrified! I was singing out of tune on several
occasions. My wife insisted that I wasn't, and she has a
good ear...

Enter Lucy Manén's Bel canto book. She observes (pg 50) that

The high harmonics of the imposto mechanism, not only
give to all the vocal registers the beauty and brilliance
of Bel Canto; they also ensure that the voice will sound
in tune to a distant listener (without adequate brilliance,
a distant voice will often sound flat even when, heard
from near by, the notes seem correctly pitched).

When I checked the parts of the tape where I recalled emphasizing
the higher harmonics, the voice didn't sound flat. I had focused too
much on creating a full voice, and forgetting about the brilliance.
Oh, well. Let's hope the one's in the back didn't have such good
ears...

(If you want to know what imposto is, read the book, or go see
your local Bel Canto teacher.)

Later, Katarina and I gave a recital in Oxelösund - Church coffee
for the senior members of the congregation. Katarina grew up in
Oxelösund, so she has a tendency to sell out when she visits. So
it was this time too. The church was full and we had a great
time. I sang Ah, fuyez... from Manon - a terribly difficult aria,
but I never did have the good sense to go for the easy wins.
The challenge with Ah, Fuyez is that you have to be able to
express several different and conflicting emotions. De Grieux
prays to be rid of the obsession with Manon, and curses the
world and his misfortune, but really, he doesn't want anything
other than to be with her - but he does sincerely long for the
peace of God too, and he keeps shifting in and out of these
emotions. To add some complexity, it's in French, which I
don't know, and the tessitura is terribly high.

Bel canto to the rescue. If you're able to produce the sound
with a lowered larynx, whatever you feel will be strangely
amplified (because the emotional colorings of the voice
emanate from there.) So if you take the time to really
understand what you're singing, you will be able to portray
the proper emotions (obviously there is more to a great
interpretation than that, but just the basics will get you
surprisingly far). The 'imposto' thingy helps anchor the high
voice especially, which saves you when De Grieux first
utters his "ah fuyez" in a piano G. Also, you can bring on
the crying in the ff exclamations - as long as you keep your
position and let the body provide (as Bengt keeps telling me).

The truly amazing feeling is that you can go all out, both with voice and emotion, as long as you keep directing the force the right way.

For fun, I later decided to try singing it unsupported (actually,
still much more supported than I ever sang before I went to
Bengt three years ago). I found that I could produce a cool-
sounding timbre, and sing loudly enough, but when trying to
convey emotions, the voice suddenly felt like a wall between
me and the (imaginary) audience. I started realising why many
singers are left with just varying the volume, phrasing and
attack. What else can you do?

I decided to quit experimenting and return to practising the
real thing. I don't master it well enough to goof around.

Then we did some Nemorino/Adina, "Caro Elisir". Fun stuff.
I need to find more opportunities to sing on stage.