Understanding resonance with MRI scans

This blog post is all about resonance, and how singers manipulate it to their advantage. I recently discovered these fabulous videos on YouTube showing people singing inside an MRI scanner, meaning that we can see what is happening inside their mouth and throat when they are singing. The singers have done wonderfully, considering they had to lie on their backs, unable to move their heads inside the MRI machine! (This also explains the slightly strange sound quality of the recordings.)

I am writing this blog post on the assumption that most of my readers already have worked with me on understanding the following terminology: vocal tract, pharynx, layrnx, Formant 1, Formant 2, resonance, fundamental. My apologies to other readers, but I think it is still worth reading this and watching the videos even if you are not sure of all the terminology yet. I hope to add information about each of these terms in due course.

A man’s vocal tract

Michael Volle (baritone) singing ‘Lied an den Abendstern‘ (from Richard Wagner’s opera Tannhäuser, Act II sc iii, 1845). [O du, mein holder Abendstern, wohl grüsst’ ich immer dich so gern vom Herzen, das sie nie verriet, grüsse sie, wenn sie vorbei dir zieht, wenn sie entschwebt dem Tal der Erden, ein sel’ger Engel dort zu werden!]

Points to notice:

  • 0:06 – The ‘ah’ sound in Abend- has a relatively flat tongue, the back of the tongue near the back wall of the pharynx (back of the mouth), and the narrow vertical space of the pharynx. The flattened tongue keeps low away from the hard palate, which keeps Formant 2 (the ‘bright’ ‘treble’ end of end vowel) low, and less bright in sound – the highest harmonics are avoided. The narrow pharynx means that Formant 1 (the ‘dark’ ‘bass’ end of the vowel) is less dark or bass-y.
  • 0:12 – The vowel in –stern has a higher tongue than the ‘ah’, a little bit closer to the roof of the mouth – so Formant 2 is higher and brighter (more ‘treble-y’). The pharynx is also wider, so Formant 1 is lower, and more bass-y.
  • 0:13 – The letter ‘n’ at the end of -stern shows the tongue creating an airtight seal at the front of the mouth, preventing air or sound form escaping there. The soft palate, towards the back of the roof of the mouth, drops down towards the tongue, opening up the vertical space into the nasal cavity – this is how the air and sound escapes for a nasal consonant. You can see this happen on another nasal consonant ‘m’ in ‘immer’ at 0:20.
  • 0:21 – The ‘ee’ sound in dich shows a very high, fronted tongue (raising the treble-y Formant 2 for brightness), and the resultant huge horizontal and vertical pharyngeal space (lowering the bass-y Formant 1 for darkness / depth).
  • 0:23 – The ‘oh’ in so is given maximum ‘darkness’ and ’roundness’ by the lips going forwards. This helps of increase the size of the Formant 2 space, thereby lowering Formant 2 and reducing the treble-y-ness. The forward movement of the lips also lowers Formant 2 (increasing bass-y-ness), because the lips effectively lengthen / enlarge the whole resonating tube.

It’s interesting to see all that going on inside the singer’s vocal tract, invisible to the listener. Here is what it looks like from the outside, sung beautifully by Roman Tekel (begins at 2:00).

A woman’s vocal tract

A woman is singing about an octave higher than a man. This means that her fundamentals (the notes she sings) all sound higher and brighter, because they are a faster frequency of vibration. This also means that the harmonic series of each note is much higher and brighter. That makes for ringing, bright sounds. The downside is that, as she sings higher and higher, the strongest harmonics (those nearest her fundamental pitches) are so high that they start to sound horribly shrill and screechy. To offset this, a female singer (or anyone with a small vocal tract, including a child) has to maximise the size of their vocal tract, to pull Formant 1 and Formant 2 downwards. The problem is that a woman’s vocal tract is, on average, much smaller than a man’s; compare the size of Michael Volle’s mouth and pharynx, wit the woman’s in this next video. We’re going to see – and hear – that the woman will need to use very different strategies from the men in order to create a balanced sound – containing enough ‘darkness’ and ‘warmth’ to offset the inevitable brightness of her higher notes.

This soprano (no name given on the YouTube example, unfortunately) is singing O mio babbino caro (‘O my beloved papa’, from the opera Gianni Schicchi by Giacomo Puccini, 1918). [O mio babbino caro, mi piace, è bello, bello. Vo’andare in Porta Rossa a comperar l’anello!]

Points to notice:

  • 0:02 The ‘ee’ sound in babbino has a lower tongue than Volle’s ‘ee’ at 0:21 in his video. Volle was trying to create a bright, rining ‘ee’. But the woman singing babbino is aiming for warmth and roundness of tone, so she lowers Formant 2.
  • 0:05 – The ‘ah’ in caro brings the tongue fairly flat. But notice that the horizontal space between tongue and hard palate is much smaller than in Volle’s ‘ah’ in the earlier video. Notice also that the soprano’s vertical pharyngeal space has almost disappeared, it is so narrow (Volle’s had more space than hers does).
  • 0:11 – The ‘o’ in bello is the interval of a tenth higher than the ‘ah’. This would be a very shrill sound unless the tongue could be moved further from the hard palate. To do this, the singer has dropped her jaw open very wide. To keep the muscles between jaw, tongue and larynx loose, she has dropped the jaw loosely, and straight down at the back; the horizontal space opens up, and so Formant 2 is lowered and less treble-y. Notice that this jaw release has also enabled the singer to create more space in her pharynx; critically, this helps lower Formant 1, thereby giving a little more ‘depth’ and warmth to the sound.
  • 1:13 to 1:25 – The soprano sings ‘la’ in the lower range of her voice, and the tongue is relatively flat, with the front of the mouth closed. As she moves into her upper range, she drops her jaw in order to get the tongue lower at the front. She also draws the tongue back, to increase the horizontal space and lower Formant 2.
  • 1:28 t0 1:42 – This is a nice demonstration in the lower pitches of the new tongue position needed to create ‘oo’ as distinct from ‘ah’; the tongue is raised at the back, creating roughly equally mouth (horizontal) and pharnygeal (vertical) spaces, so both Formants are reasonably low. As the woman sings higher, again, the jaw is dropped to accommodate the tongue.
  • 1:44 to 1:52 – A nice comparison of the clear differences of tongue position and sound between ‘ah’ and ‘oo’ on low pitches.
  • 1:53 to 1:57 – Top note ‘ah’ is rapidly compared to the top note ‘oo’. Notice the sound, jaw and tongue positions are the same for both! This is a clear demonstration that a woman has to opt for ‘ah’ or ‘uh’ on top notes, regardless of the vowel on the page. She must not sing the vowel the word requires, but sing the vowel that vocal acoustics and vocal mechanics require.

And it you want to see a truly extraordinary mastery of resonance, watch this (begins 0:38) – Amira Willighagen at 9 years old!

With such a tiny resonating space, it makes sense that she drops her jaw for the high notes. However, finally, here is a performance by the great Monserrat Caballé. She keeps her mouth much more closed at the top – this creates a wonderfully delicate high note. Caballé ensures that she does not put too much air pressure on the note, or it would come out loud and shrill. This is great resonance craftsmanship and artistic maturity. (Begins at 0:53)

Posted in Resonating
3 comments on “Understanding resonance with MRI scans
  1. Thank you, Alexander for illustrating what is going on. I’m amazed that singers could sing with all that noise & vibration of an MRI scan – generous offering for the rest of us to learn from.

  2. Jeff McCracken-Hewson says:

    My theory is that masque resonance or “ring” is produced by “pinching” or narrowing a part of the vocal tract with the tongue. For a back vowel – eg ah, oh – this is done by narrowing the throat with the back of the tongue. For a front vowel – eg ee, oo – it is done by bringing the front of the tongue up close to the hard palate. This is what I feel when I sing with ring and I think it is shown in this MRI – although it doesn’t seem to happen with the eh/air vowel, which puzzles me. I found an MRI of someone singing without ring in the following article, and sure enough the tongue was not used to pinch or narrow the vocal tract. http://rebrn.com/re/mri-of-an-opera-singer-2594872/ If I am right about this, it makes a bit of a nonsense of theories about the “open throat”.

    • Jeff, you don’t need to theorise about this. Tongue position is not a critical factor in producing ‘ring’ (the singer’s formant). Kenneth Bozeman, in ‘Practical Vocal Acoustics’ (p.106) says that is “generated by a low larynx, open laryngopharynx, and narrowed epilaryngeal tube exit, that results in an increase in power of the high frequency components of a sung sound.” Scott McCoy, in ‘Your Voice: an inside view’ (pp.47-49) gives a useful description of the phenomenon. He shares the now widely accepted view amongst voice scientists that ‘ring’ comes from a convergent resonator (small mouth opening, with wide and long pharynx), and a slight constriction of the epilayrnx, assuming the signal from the folds has strong harmonics in the region 2400Hz-3,200Hz. Bozeman also points out that when the second formant is high, it helps boost the cluster of harmonics found in the third formant (which is part of the singer’s formant) – so vowels with a higher second formant (especially EE for example), will most likely have more ring.

Leave a Reply

Your email address will not be published.