(prior page: Trills and Beats)
Nasality is a feature of bird sounds that are tonally complex. This means that the bird simultaneously produces several sounds that our brains interpret as one. Most sounds that we hear are tonally complex, including most bird sounds, all human voices, and the sound of almost every musical instrument. On a bird sound spectrogram, tonal complexity often looks like a set of stacked copies of one original sound. We call the vertically stacked copies partials because they are only partial components of the sound.
Here is a classic example of a bird sound with many vertically stacked partials:
Our brains hear one sound with a nasal tone quality, but in reality each of the partials in these notes is itself a pure-toned whistle. To illustrate this fact, I copied one note from the above recording and used a frequency filter to isolate one partial at a time. The unaltered call appears at the beginning and end for comparison:
Extra credit: Earlier I referred to the partials in the Black-capped Chickadee’s song as harmonics. Why am I calling them partials now? In the science of acoustics, partials are any components of a complex sound, while harmonics are a specific kind of partial, the kind produced by a single vibrating source. In a true harmonic series, the fundamental frequency is typically the lowest and loudest partial (meaning it is the lowest and darkest band on the spectrogram). In the case of the Red-breasted Nuthatch, you’ll notice that the darkest bands are in the neighborhood of 3 kHz, with the partials getting fainter both above and below that. This means that Red-breasted Nuthatches probably produce their nasal notes by using both sides of their syrinx simultaneously, but not independently: instead the two vibrating sources are coupled to create what looks like a harmonic series, but technically isn’t (see Nowicki & Capranica 1986).
Extra extra credit: Those of you with musical training may notice that the first interval between the nuthatch’s partials is a perfect fifth, and the second interval is a perfect fourth, meaning the third individual note is an octave higher than the first. From there on out the intervals continue to decrease as you would expect in a harmonic series. However, the first note in the series (the fundamental) is missing. It should appear at roughly 0.5 kHz, at half the frequency of the lowest note in the stack and therefore an octave lower.
More energy in higher partials = more nasal
Nasality can be difficult to interpret on a spectrogram, because not every sound that contains partials has a nasal tone quality. Here’s a good example of a bird sound with a pretty significant set of partials, but no nasal tone:
Why no nasality? Because the fundamental is dominant. That is, the lowest note in the stack is also by far the loudest (darkest), and therefore the other notes don’t contribute much to the sound.
Here, then, is the key concept: nasality is determined primarily by which partials are loudest (that is, darkest on the spectrogram). If the fundamental (the lowest partial) is loudest, then the tone quality will be perceived as whistled, not nasal at all. The higher the energy maximum climbs up the stack of partials, the more nasal the resulting tone quality.
Bird sounds that are loudest in the second partial constitute something of a special class. They do not sound nasal in the same way as a Red-breasted Nuthatch, but they can sound piercing, like this Killdeer call:
At a lower pitch, this type of vocalization comes across as a little more mellow, like this Sora’s call.
One of the reasons why the second partial doesn’t contribute greatly to nasality is that the second partial is usually exactly an octave above the fundamental, so it “blends” with it very well. As you add higher partials that come at ever-smaller musical intervals, they tend to blend less well and provide more of the “clash” between sound components that we interpret as nasality.
Here’s a bird sound that is strongest in the second partial, but also has some decent energy higher up, especially in partials 6-8, which makes this sound slightly more nasal:
The Pinyon Jay’s call is loudest in the second partial too, but not by much; its energy is more evenly spread among higher frequencies, making it slightly more nasal than the quail:
Note that when nasal notes are upslurred or downslurred, the spacing between the partials doesn’t stay constant over time. Instead, the distance between partials in the higher-pitched parts of the sound is always greater than the distance between partials in the lower-pitched parts of the sound. Thus, in each higher partial, the slope of slanted lines on the spectrogram gets progressively steeper. You can see this pattern clearly in the second half of the Pinyon Jay’s first note.
Here’s a bird sound that is loudest in the fourth partial, which makes it sound slightly more nasal still:
And here’s one of the grand prize winners of nasality, a vocalization whose energies are strongest in roughly the ninth or tenth partial:
Effects of pitch on nasality
Just as pitch affects the tone quality of whistled sounds, so it affects the tone quality of nasal sounds. In basic terms, the higher-pitched the vocalization, the less nasal it can possibly sound. Think about it: the partials are usually at integer multiples of the fundamental. In a harmonic series, if your fundamental is at 2 kHz, that means your second partial will be at 4 kHz, your third at 6 kHz, and so on. In theory, you’ve got room for about 10 partials before you reach the upper limit of human hearing, which could potentially allow for lots of nasality. But if your fundamental frequency is at 7 kHz, then your second partial will be at 14 kHz, and the vast majority of humans won’t be able to hear the third partial at 21 kHz no matter how much energy it contains. So there’s no real way that such a high-pitched vocalization can sound nasal.
The phenomenon I just explained is the same one we saw above, in the Pinyon Jay spectrogram: the higher-pitched the sound, the more widely spaced the partials. But in learning to read spectrograms, it may be even more useful to think of this phenomenon from the other way around: the more widely-spaced the partials, the higher-pitched the sound. The California Gnatcatcher spectrogram demonstrates this nicely. The energy maximum is mostly between 6 and 8 kHz, which is pretty darn high-pitched — significantly higher than the Mountain Chickadee’s song we saw a couple of pages ago and not too much lower than the stratospheric alarm call of the robin. But the gnatcatcher sounds much lower-pitched than either of those, because it is composed of partials that are extremely close together.
This recording of the alarm call of the Northern Goshawk also helps demonstrate this point. The individual notes of the call are quite brief, so I’ve zoomed in the time axis somewhat.
Note that the first four notes of the call sound very low-pitched — significantly lower-pitched, for example, than the Pinyon Jay above, even though the maximum energies of both sounds are concentrated between 2 and 3 kHz. Starting at the 2-second mark, though, a second syllable begins to follow the first, and you’ll note that it sounds much, much higher-pitched than the first part of the call, even though it too has its energy maximum centered squarely at 2 kHz. Looking at the spectrogram, you can tell it will sound higher-pitched (and also less nasal) because its partials are so much farther apart.
Many birds with nasal voices, particularly gulls and hawks, often have voice breaks in their calls. You’ll hear them towards the end of this Laughing Gull recording:
In the last six notes of the above spectrogram, notice the “vertical fault lines” along which the zebra stripes don’t match up. They coincide with what your ear will hear as a “break” in the voice of the bird: an abrupt jump to a different pitch. In this case all the jumps are to a higher pitch, because the partials after the break are more widely spaced than those before the break.
A somewhat confusing phenomenon can be seen in the second call of this Sinaloa Crow:
|Sinaloa Crow, Alamos, Sonora, 3/25/2005.||Same individual. Note period doubling.|
Note that the second call is almost exactly identical to the first, except that it has fainter partials added in between all the partials that were already there. The loudest partial hasn’t really changed or moved — it’s just acquired an additional “shadow” copy beneath it, as have all the other partials. This phenomenon is called period doubling, and the “extra” partials are what acousticians call subharmonics. Exactly how and why they show up in bird vocalizations is not well understood. Your ear treats them quite differently than it treats the other partials. If they were “normal” harmonics, then in theory their appearance should cut the pitch of the call roughly in half, as they cut in half the average distance between the partials. But instead, subharmonics merely add a subtle quality to the call: a very slight “roughness” or “hoarseness” that is audible in direct comparison.
Putting it all together
This Red-shouldered Hawk brings together in one vocalization all the phenomena we’ve seen on this page: nasality, voice breaks, and period doubling.
Note the following:
- Period doubling does not occur in the first note, but occurs throughout the second note, and then “flickers” on and off up to three times per note thereafter. The subharmonics are the “partials between partials” that aren’t always present. When present, the subharmonics add a subtle “rough” or “hoarse” quality that is particularly noticeable in the second note.
- You have to ignore the subharmonics when evaluating where the energy is to predict how nasal the call will sound. When they are discounted, it’s apparent that the maximum energy is concentrated in the second partial, with a relatively even scattering of energies in the partials above it, meaning that nasality will be present, but not strong.
- The voice breaks downward towards the end of most of the individual notes. You can tell the pitch will be lower after the voice break because (ignoring subharmonics) the partials are always more closely spaced after the break.
- The pitch drops despite the fact that the maximum energy after the voice break is concentrated in the third partial, not the second. (This is most readily noticeable in the tenth and final note on the spectrogram, and in subsequent notes on the recording.) Switching energy to a higher partial, as you will recall, doesn’t affect pitch so much as tone quality: each note gets more nasal after the voice break.
(next page: Polyphony)