Just thought I'd add a little bit of discussion about the frequency response of mics since I posted these here.
A frequency response curve is a 2-dimensional representation of how a mic responds to various musical pitches (and beyond). The X axis (from left to right) represents frequency going from low to high. The Y axis (vertical) represents how loud the output of the mic would be if that frequency were played acoustically through the air at a known volume. So in effect, if the curve is high at a particular frequency, it means that the mic is more sensitive at that pitch; if it is low, it is less sensitive at that frequency.
A theoretically ideal (but not necessarily practically ideal) mic would be a straight line. In practice, the ideal mic depends on the instrument or voice. Having bumps and dips in certain places can accentuate or diminish certain aspects of the source that may make it more or less pleasant.
To give you an idea of what the frequencies actually mean, the bottom A on a piano is just shy of 30 Hz. The top note on a piano is just shy of 4.2kHz. That's for the fundamental tone, though.
The harmonics of any instrument or voice go up from the fundamental, so a 4kHz piano will have audible overtones at about 8.4kHz and 16.8 kHz (that would be a C the octave above the piano, and the G an octave above that). Your ear perceives those frequencies as being part of what makes a piano sound like a piano and not, for example, a guitar.
For example, because it is a struck stringed instrument, the harmonic series on a piano becomes progressively sharp at higher harmonics. The exact amount of this inharmonicity depends primarily on the length of the strings. This is true to a lesser degree even for guitar. (Oddly enough, this does not occur for bowed strings, nor for brass/woodwinds.)
Okay, going way off track here. The point is that these subtle differences in the overtone series give each instrument its unique character. For example, the clarinet has predominantly the even harmonics (by the musical definition) while the saxophone emphasizes the odd harmonics, resulting in a more complex tone. In fact, the clarinet as an instrument cannot play the odd harmonics (again, by the musical definition).
Note that I said "by the musical definition". This is because musicians and mathematicians have the definition of even and odd backwards. For musicians, the fundamental is not considered a harmonic. It's the fundamental, and the octave above that is the first harmonic. For mathematicians, they go based on multiples of the fundamental frequency, so the first harmonic is "fundamental x 1" (the fundamental), the second harmonic is "fundamental x 2" (which musicians would call the first harmonic), etc. You can see why this is confusing. This also means that if you read any text on the subject, they will likely use the mathematical version, but I'm using the musical version here because it is more understandable (not that any of this is understandable
).
This is all only for closed tube instruments like clarinet or sax. Then, you have to take into account the whole difference between a closed tube (almost all wind instruments) and an open tube (pretty much the flute, piccolo, pan flute, and other similar instruments). Open tubes allow all the harmonics to be present, but emphasizes lower order harmonics. They also have a fundamental that is an octave higher than with a closed tube for a given length of tubing, though that's really more of an aside.
Again, going off track a bit, my apologies. The point is that those harmonics make the tone quality of an instrument or voice what it is.
Vocals are even more interesting. The vocal tract behaves like a closed tube resonator. Unlike instruments, however, it is very nonuniform---that is, it has all sorts of curves and twists that alter the resulting sound. It also has additional closed resonator cavities such as the sinuses that alter the sound in various ways.
And, of course, as you raise and lower the soft palate, you change the character of the sound significantly. You also are slightly changing the effective length of the tube, IIRC. Since the pitch is defined by the vocal cord tension and not the tube length (as it would be in wind instruments), this results in changes to which overtones are emphasized.
As a result of these complexities in the shape of this tube, the voice has very nonstandard frequency ranges in which overtones are emphasized significantly. These are referred to as formant regions.
Just to be clear, instruments have formants as well, but you'll hear it talked about a lot more in the context of voice because the formants change so much from one moment to the next depending on the sound being sung and are not entirely driven by the pitch being sung.
Formants of instruments
http://ccrma-www.stanford.edu/~jmccarty/formant.htm
Formants of voice
http://en.wikipedia.org/wiki/Formant
The point is that fricative sounds (e.g. the letters "s", "f", etc.) produce frequencies that have nothing whatsoever to do with the pitch being sung at the time. These sounds extend outside the human hearing range (above 22 kHz). This is one reason that people with high frequency hearing loss often have trouble understanding consonants (and similarly, people trying to do crossword puzzles over cell phones). Of course, the information above about 10 or 12 kHz is not strictly necessary, but it does improve understandability.
So the very top end mainly has an impact on fricative consonant sounds, cymbals, etc. A presence peak in the 2-5 kHz range can enhance understanding by making these consonants cut through the mix better.
A word of caution is in order, however. If you go too far in this range, you will get a very painfully harsh quality to the sound because certain vowel sounds have formant regions up in that region.
Also, some microphone circuits end up emphasizing too much of certain harmonics, resulting in a harsh quality to the sound. I find that many mics with a really uneven high frequency response also exhibit this characteristic. They also tend to really make sibilance (the sound of those fricatives I mentioned) stick out like a sore thumb.
I've probably massively overloaded you and everyone else who reads this thread, but hopefully that will shed some light on what these graphs mean as far as what the mics should sound like. I'll leave with some lists of interesting frequencies for voice:
80-1100 Hz - fundamental pitches of speech or singing.
200-1000 Hz - first formants of vowel sounds.
800-1500 Hz - round vowel second formants (ah, oh, oo, aw, and so on).
1400-2500 Hz - non-round vowel second formants (ee, eh, a as in apple).
50-8000 Hz - plosive sounds (p, b, t, and so on).
200 -22,000 Hz - fricative consonants (s, f, wh, sh, and so on).
