How a Voice Beats an Orchestra

An opera singer stands in front of an orchestra of eighty players, with no amplification, and is audible. Not faintly: clearly, at the back of a two-thousand-seat house, over a full tutti.

The obvious explanation is that they are loud. The obvious explanation is wrong, because the orchestra is louder. Eighty instruments against one set of vocal folds is not a contest about power, and no amount of training makes a human body louder than a brass section.

The actual explanation was worked out by Johan Sundberg in the 1970s, using long-term average spectra: measure the energy of a sound at every frequency, averaged over minutes rather than instants, and look at the shape.

Do that to an orchestra and the energy is concentrated low. The peak sits somewhere around five hundred hertz, and above that the curve falls away steadily. An orchestra is a loud object in the lower and middle part of the spectrum and a comparatively quiet one higher up.

Do it to a trained operatic voice and there is a distinct extra peak, broad, somewhere around two and a half to three kilohertz. It is produced by the third, fourth and fifth resonances of the vocal tract moving close together and reinforcing each other, which happens when the pharynx is widened relative to the opening of the larynx. It has a name: the singer's formant.

Put the two spectra on top of each other and the mechanism is visible. The voice has a strong peak precisely where the orchestra has almost nothing. The singer is not competing with the orchestra; they are occupying a band it has abandoned.

Which reframes what the technique is. The classical vocal training that produces this sound is not a method for being louder. It is a method for redistributing the energy you already have into a region where it will be heard, and that is why Sundberg's own framing is about vocal economy: loud sound at minimum effort.

This also explains two things that are otherwise puzzling about the sound.

It explains why the trained operatic voice has that particular ring, which listeners either love or find artificial. The ring is the formant cluster. It is not decoration added on top of a natural voice; it is the acoustic signature of the configuration that makes the voice audible at all.

And it explains why imitating it fails so characteristically. An untrained singer trying to be heard over an orchestra pushes harder, which raises pressure, which does not move the resonances. You get effort and no ring. The adjustment needed is a shape, not a force, which is exactly the lesson of the previous chapter arriving in a performance situation.

Two caveats, since the popular version of this is tidier than the research.

The cluster's centre moves with voice type, lower for basses and higher for tenors, with baritones between, so there is no single frequency to quote. And the phenomenon is best documented in male voices; the literature on how and whether it works the same way across all voice types is less settled than summaries suggest, partly because very high sopranos have a different problem to solve and appear to solve it differently, by tuning a resonance to the note itself.

What generalises is the principle, and it is the same one as the harmonics lesson in the first book. A resonance does not create energy. It moves it. Find the gap in what everybody else is doing and put your energy there.