(prior page: Nasality)

The vocal organ of birds, called a syrinx, functions very differently from the human vocal organ, called a larynx. The syrinx is located farther down, where the trachea splits into two bronchi to travel to the lungs. Researchers are still solving the mystery of exactly how the syrinx works, and it probably works differently in different species of bird. However, it is clear that in many birds, including most passerines, the syrinx has the ability to produce two separate sounds simultaneously, one from each lung.

Although many birds have “two voices” in theory, few birds actually use this ability in the wild. Some birds go through their entire lives using only one of their voices (that is, singing out of one side of their syrinx). Many birds use both sides, but not at the same time: they might alternate so as to be able to sing faster trills, or they might make low sounds with one side of the syrinx and high sounds with the other. Most Northern Cardinals do this: they switch from one side of the syrinx to the other in the middle of each note of their song, so quickly and smoothly that the transition can’t be detected even with spectrographic analysis.

When birds do vocalize simultaneously with both sides of their syrinx, odd sounds (and very odd spectrographic images) can result. The use of two voices is called polyphony, and sounds that are made with two voices are termed polyphonic. Polyphonic sounds can look extremely messy on the spectrogram, but our first example is cleaner than most:

Detail of alternate song of Prothonotary Warbler, Phillips County, Arkansas, 4/4/2006.
Detail of alternate song of Prothonotary Warbler, Phillips County, AR, 4/4/2006.

Does the tone quality of those three notes remind you of, say, a Pine Siskin call?  If so, that’s no coincidence, because the cardueline finches (along with the thrushes) are North America’s masters of polyphony, and their calls are often the first that come to mind when we hear the distinctive tone quality of a polyphonic sound, which I sometimes describe as “finchy” and which David Sibley sometimes calls “wiry.”

When you look at the spectrogram, the first thing you’ll notice about the partials is that they aren’t neat copies of one another stacked at regular intervals, like the partials in a nasal sound. Instead, they diverge and converge in odd patterns; some of them even cross each other. This entire complex pattern is comprised of three different kinds of partials:

  1. fundamental frequencies (two of them);
  2. harmonics of the fundamentals (three of them visible here); and
  3. heterodyne frequencies (all the rest of the traces).

The heterodyne frequencies are evidence of a fundamental concept of polyphony: when two sound waves mix, they combine to form additional sounds at frequencies equal to the sum of the original frequencies and the difference between them. Thus, if you mix frequencies A and B, in addition to the original sounds, you’ll get heterodyne frequencies A+B and A-B. If A and B have multiple harmonics apiece, then things get complicated, because there are lots of frequencies to add and subtract.  Here’s that last spectrogram again, this time with the frequencies labeled:

Same spectrogram as above, with labels showing fundamental frequencies (A & B), harmonics (integer multiples of A & B), and heterodyne frequencies (sums and differences of fundamentals and harmonics, in parentheses).
Same spectrogram as above, with labels showing fundamental frequencies (A & B), harmonics (integer multiples of A & B), and heterodyne frequencies (sums and differences of fundamentals and harmonics, in parentheses).

With so many heterodyne frequencies created by adding and subtracting each fundamental from each harmonic, it’s no wonder that some polyphonic sounds can produce spectrograms that look like plates of spaghetti. Note in the above spectrogram that the heterodyne frequencies, which are labeled in parentheses, are relatively faint, which is typical.

Polyphony can look a lot of different ways on the spectrogram, but one sure-fire clue is the simultaneous occurrence of rising and falling tones at the same exact time within a single sound. If you ever see that, you can be certain that you’re looking at a polyphonic sound. If you look carefully, you can see it in this American Goldfinch call:

American Goldfinch alarm call, Boulder, CO, 1/4/2009.
American Goldfinch, Boulder, CO, 1/4/2009.

Note that there appears to be a slight noisiness in the center of the spectrogram, associated with the call. Noise and polyphony are frequently found in association.  You can see a little bit of noise in this classic example of a polyphonic sound, the rising whine of the Hermit Thrush:

Hermit Thrush, Pinal County, AZ, 1/10/2008.
Hermit Thrush, Pinal County, AZ, 1/10/2008.

And here’s a much noiser sound that nonetheless derives a great deal of its characteristic tone quality from polyphony:

Mexican Jay, Santa Cruz County, AZ, 5/12/2009.
Mexican Jay, Santa Cruz County, AZ, 5/12/2009.


One type of polyphony produces distinctive images on the spectrogram, which are often referred to as sidebands. You can see this phenomenon in the middle section of this vocalization of Lesser Goldfinch:

Lesser Goldfinch, Pinal County, AZ, 1/10/2008.
Lesser Goldfinch, Pinal County, AZ, 1/10/2008.

Note that the spectrogram of this call looks like a teapot!  The first section of the call is “simple” polyphony: the two voices start together and split, the upper voice slurring down, the lower voice sharply underslurred, coming back up to meet the upper voice at the end of the teapot “handle.”  Then there’s the bizarre-looking middle section, which contains the sidebands.  The end of the call is a monotone whistle with a very brief downslur at the end.

The middle section is what I want to talk about.  If you listen carefully to the call, you can hear a distinctly different tone quality in that middle section — it sounds almost like the bird is going out of tune for a brief moment.  In fact, that might be exactly what it’s doing.  The bird is probably singing two notes at once that are almost but not quite identical — and the dissonance between them creates a distinctive tone quality and an odd band-painted pattern on the spectrogram.  It’s pretty much the same phenomenon as we saw above with the heterodyne frequencies — sidebands are essentially the same thing.  It’s just that here the heterodyne frequencies are much closer together since the difference between the two fundamentals is so small, and the whole thing starts to look like one wide banded call rather than a bunch of separate ones.

The brief moment of sideband dissonance in the goldfinch call can be hard to hear, so let’s listen to a nice long pure example from a champion sideband singer:

Varied Thrush, Humboldt County, CA, 3/28/2009.

In this ethereal Varied Thrush songtype, listen carefully for the beat frequencies, which come across as a kind of trilled texture to the sound.  From an auditory standpoint, it’s not quite as simple as the buzziness we’ve seen before; there’s that distinctive dissonance.  Looking at the spectrogram, you can see the characteristic “cross-hatching” of horizontal sidebands and vertical beats that is frequently seen in vocalizations of this type.  Birds that gain much of their distinctive tone quality from notes like this include the Yellow-headed Blackbird and the Bobolink (e.g., see the first note in the header at the top of this page).

Polyphony is in many ways a more complex subject than I’ve made it out to be on this page, but this should give you a good start on recognizing it when you see it and hear it in a bird vocalization.

P. S. — I will be the first to admit that I don’t fully understand all the acoustic terminology I’ve been using on this page, so I would love it if anybody out there with better training can set me straight on anything.  Email me with comments!