14 June 2010

What Singing Looks Like

I've been meaning to post this for several days. It is a real-time MRI of the vocal tract of two vocalists: a soprano and an emcee/beatboxer.

The video is posted at vimeo by Krishna Nayak, professor at the USC Viterbi School of Engineering, where he is associated with SPAN, the Speech Production and Articulation kNowledge Group.  (Good to know that my alma mater produces something besides corrupt sports teams and unemployed musicologists.)  This video was first presented at a conference in 2006.

the diva and the emcee from Krishna Nayak on Vimeo.

I find the video fascinating and enlightening, vastly more so than the static and rather poorly executed line drawings or blurry black and white photographs that one typically finds in books of vocal pedagogy. This video vividly demonstrates what voice teachers have long known: vowels that are readily distinguishable in a singer's lower register become essentially indistinguishable in the upper register.  Also note how far forward the tongue is on the Italian consonants—something one constantly tells English-speaking voice students, but a moving picture is worth at least a thousand words.

The SPAN site has (so far as I can see) only one other video clip of a singer: a much smaller and shorter one of a soprano singing "Ave Maria."

I note that the SPAN group seems not to include any musicians.  Hey guys, I'm available.

I haven't had the time to do more than a cursory search of the literature on MRIs of singers, but I did run across the following article: Matthias Echternach, et al. (2008), "Vocal tract and register changes analysed by real-time MRI in male professional singers—a pilot study," in Logopedics Phoniatrics Vocology. The article is behind a paywall....actually it's in a paybunker, where it costs (I am not making this up), €69.00 (around $83.68) from informaworld. The article is 7 pages (pp. 67-73), and the pdf is ~262 KB.  That comes to $11.95 per pdf page, or 32 cents per kilobyte.

In a separate post, I'll be starting a contest for more most outrageous article price, but I think I may already have won.

Bookmark and Share


  1. I meant to mention that I assume that the singer and the beatboxer were both lying down when the image was made, which is not the ideal position for singing. So I wonder what effect that had on the results.

  2. Great video my only compliant is they used a very mediocre emcee/beat-boxer :(. It would have been more interesting to see what happens with a skilled beat-boxer.


  3. I'm not a skilled critic of beat-boxers. But remember that he was lying down in an MRI with a weird apparatus over his face, so he may not have been at his best.