Opera singers never cease to amaze spectators as they fill auditoriums with songs — without any amplification. Our voice is inconstant, changing by year, by day, and even by hour. So what allows opera singers to change octaves, stay even throughout the range, and carry sound night after night?

In a real-time MRI scan, scientists at the Freiburg Institute for Musician's Medicine captured jaw-dropping footage of baritone Michael Volle singing an aria from Richard Wagner's 1845 opera Tannhäuser. Dr. Matthias Echternach of the University of Klinikum in Germany approached the singer about the test because he wanted to learn more about how the vocal tract acts during a performance. Our voice source, the sound that comes from our vocal folds, passes through the vocal tract resonator, which shapes it acoustically and gives it pitch. These sounds, in turn, can be shaped into recognizable speech by articulators like the tongue and lips.

Opera singers need to eliminate anything that may interfere with the voice. For example, muscle tension, even in the legs, uses oxygen, which leads to breathing problems and stress reactions. The ability to support the voice properly, forcefully, and without tension becomes a challenge even for the most experienced performers. Opera singers require a loose, agile tongue, and a relaxed jaw to reach their intended pitch.

They do this with a special technique that allows them to shape their throats in a way that produces an extra frequency when they sing. Called the singer’s formant, according to the National Center for Voice and Speech, it registers in the range of 3,000 hertz.

When an orchestra plays, all the instruments create sounds that have frequencies within a certain range, mostly below the 800 hertz range. Since an opera singer’s frequency is much higher than that, the audience can hear it over the instruments.

This MRI scan shows us there are some subtle and not-so-subtle movements spanning everywhere from the vocal cords to the windpipes that play a role in the grandiose opera voice.