Why a wah pedal sounds something like a voice, and why it doesn't sound more like one
Copyright 1999 R.G. Keen. All rights reserved.
The thing that makes a wah pedal so immediately catchy is that it has a real "vocal" quality - it sounds like a a human voice making that noise. Everyone has done a quiet little "waaah" with their mouth to imitate it.
On the other hand, that's all it does. There's clearly something missing because it won't do anything but that "wah" sound. What's going on?
People who study the human voice have provided a number of answers. If you take common voice sounds and calculate the frequency spectrum - that is, how loud is it at each separate frequency - you find some interesting features. There is a basic frequency, which is kind of like the basic note you hit with your voice when singing. Above that, there are at least three and often more resonant peaks. In voice research and musicology, such resonant peaks are often called "formants". A saxphone sounds different from a clarinet at least in part because the resonances associated with its different physical form cause different formants to be audible.
To shorten a long story, us humans listen for and assign meaning to the relative spacing of the first three formants of the human vocal tract. We hear and notice the fundamental frequency, but that seems to be almost inconsequential in assigning meaning to vocal sounds. At most, we notice relative shifts of the fundamental frequency as denoting emotional states - as in someone's voice going up in pitch when they're under stress. The relative positioning of formants, and largely the first two formants, forms a kind of code that we interpret as vowel sounds. Here's one version of the decoding key, based on information from Bell Labs research into humal voice sounds early in the 1900's:
While this chart implies that there are discrete regions where a sound is more like an "ee" than an "i" sound, that is an oversimplification. There are no real boundaries, only a continuous shading from one vowel sound to another. This is one source of regional accents - everybody in a "local area" kind of learns from everyone else what they all agree will be an "ee" and an "ah", and they all use that "code" to talk to their friends. It's only when someone from somewhere else that uses a slightly different mix of formants comes in that they notice the differences.
We can immediately pick up some things about wah pedals from the chart. The "ah" sound has a first formant resonance from about 700 to 1200 Hz. This matches almost exactly the typical range of the resonant peak in the semi-standard wah circuit. We can also see what's missing - the second formant. A wah pedal has only one resonant peak.
However, it also is a lowpass filter with a resonant peak, not single frequency peaks like human vocal resonances. The presence of the frequencies below the resonance peak is an almost, kind of, maybe reminder of a "peak" lower than the resonance; at least it does not drop off much. As a result, we get a sound like the first and second formants are very close together - very much like the "ah" region.
Here's another way of looking at wah frequency response:
This shows the relative ranges of the first two formants along with the range of a typical wah pedal. The wah pedal range covers most of the frequencies where the first two formants overlap. Not surprisingly, a wah sounds something like a human voice, but not close enough to be really mistaken for a vowel sound.
This does point the way to even more "vocal" sounds from a wah. A second resonance, even a fixed one, should make for more vowel-ey sounds. To really make it "talk" you could do a second wah circuit that has a resonance that moves around in a different manner than the first. This is exactly the trick employed in the Electro-Harmonix Talking Pedal.
There's a pony here. Try using two wah pedals in series, with the first one rocked to one position and left sitting there, while you use the second one to modulate the sound. See what you think.
In a later update, I'll describe some things that can be done to get more vowel-ey sounds with specially designed circuits.