Seeing Sound

I was hanging a mic on a drum kit today and I was aware once again that I was seeing sound.

Not literally, of course. That's called synesthesia (and is a fascinating subject on its own). No, this is more the ordinary quality of being a member of a visually-minded species that tends to organize information in visual/spacial terms.

I visualize the sound field. There is the pattern projected from musical instruments, speakers, and all sorts of other objects. There is the detection pattern of microphones. And there is the quality of the space; the room, and the air itself.

How to describe? The sound issuing from, say, a trumpet is a bit like a multi-colored squid. Most of the sound comes from the bell of the trumpet, but it is a narrow, blunted cone (this would be the semi-pointed mantle of the squid). The center of the cone is where the brightest sound is; as you move away from looking straight down the bell, the sound becomes less bright, with less highs. There are also little arms/tentacles coming off the trumpet at all angles; these are the bits of sound (also very directional) from valve noise, breath noise, and so forth. What a trumpet sounds like, then, depends on its angle to you and how far it is from you.

For some instruments this array of colors doesn't blend into the characteristic sound until you are some distance from the instrument. A flute vibrates itself, and sound also comes out of the holes, but the total sound forms several inches away. A violin even more so; from close-up you might hear the scratching of the bow, the distinct 800 Hz tone of the bridge, the booming sounds coming from the f-holes; or the more characteristic sound coming off the body, but it is from a foot or more away from the face that the violin begins to sound like a violin.

This, then is the art of placing a microphone; to determine what qualities are needed for a particular band, instrument, number, and to place the mic to emphasize those. For a jazz-fusion group I might want the mic a foot from the face, looking flat on. For an Celtic Rock fiddle sound go right up close to the bow. For a string quartet, back off to as much as six feet away looking over the violinist's shoulder and down!

And there is more; as that complex multi-colored shape of sound issues into the environment, it reflects and is absorbed by what is around it. The higher frequency sounds will be stopped more easily, but will also reflect more easily (especially off a hard surface). The lower frequency sounds will pass around smaller obstructions like a ripple in a lake passing around a rock. So this multi-colored, vaguely conical fountain issues from the trumpet into the world and cascades off music stand, floor, the body of the trombone player in front in a spray of color and complexity.




I was placing this mic over a drum set, and it was like navigating a field of multi-colored spheroids. I knew the drummer, and I knew I didn't want too much ride cymbal in the mic, so I tried to move the mic past the sphere of sound where the ride was loudest, and point the mic away from ride and hi-hat. The snare has a character that changes depending on your angle to the head and how close to the rim you are positioned. The toms, also, have a booming body that emanates from the sides and bottom, but a crisp attack coming off the top. A difference of a few inches in the placement of the mic, or an angle of a few degrees, changes the quality of the sound.

And mics are the same way, but in a sort of reverse. Your workhorse mic is a cardiod; that means the sensitivity pattern is roughly heart-shaped (but three dimensional). The mic "hears" loudest directly in front, with the sound level tapering along an Fibonacci curve as you angle more and more away from dead on, until at some point maybe 45 degrees behind it effectively drops to zero.

The mic also has a color. It hears different frequencies more efficiently. The old workhorse the SM57, for instance, has a pronounced "presence" peak; using this mic is the equivalent of taking the sliders around 8 kHz on a graphic equalizer and moving them up six db. What this mic "hears" is bright, in-your-face, with a bit of sizzle. The SM58 (using the same capsule but a different filter design) has a gentler rise closer to 4-6 kHz.

Except this also interacts with the pattern. No cardiod "hears" all frequencies evenly. Like the trumpet, it is more sensitive to the high frequencies that are right in front of it; things to one side of the mic will sound "duller," with more low-frequency content.

And like the trumpet, it has spikes; little lobes out at all angles including behind the microphone where it hears certain frequencies strongly.

Even more, because of the design of the mic capsule, low frequency sound interacts in a very funny way. When you are close to the source of a low frequency sound, the pressure difference between the front and back of the capsule takes over from the pressure wave that ordinarily drives the capsule. Within a few inches of a sound source, frequencies below 100 Hz are boosted sharply in something known as "bass tip-up" (or more technically, "proximity effect.")




Moving from the small scale of the mic capsule to the large scale of the performance space, the entire room can be treated as a musical instrument. Just as every woodwind generates sound through a column of air within which travel pressure waves that are harmonic multiples of the length of the tube, a room itself has compression waves bouncing from wall to wall; standing waves at every harmonic multiple of the distance between the walls.

As the air temperature and humidity change, these characteristic frequencies change. And depending on materials in the walls, placement of walls in relationship to each other, and of course sound sources within the space, some of these modes may be driven strongly. Just as a trumpet player concentrates to make the column of air in the trumpet vibrate strongly at the desired note, the air in the room itself will begin to generate a strong tone.

All of these multiple and intersecting pressure waves interact, of course. Depending on frequency (worse, on the harmonic relationship!) and phase, they may positively interact to make one frequency stand out, or destructively interact to make one frequency quieter.

Take the simple case of a singer in front of a music stand. Her voice enters the microphone, but it also bounces off the stand and arrives out of phase at the same microphone. If the mic "sees" the stand too well, the two signals will mix, and some frequencies will be in phase and boosted, and others out of phase and squashed. The resulting sound is as if you took your graphic equalizer and randomly moved some of the sliders to the top and some to the bottom -- hence the descriptive name of this effect, "comb filtering."

To combat this, you tip the stand, or you angle the mic. If the side of the mic "sees" the stand, not only is the level lower -- and the destructive interaction less -- but the lack of sensitivity to higher-frequency sounds at the side of the mic means the comb filtering mostly happens out of the frequency range of her voice.

In other cases, you might chose to move the mic very close to an obstruction -- as I sometimes have by pointing a mic directly at the lid of a grand piano from under an inch away. The length of the pressure wave at 8 kHz is 4 centimeters; that means if the mic capsule is only two centimeters away from the reflective surface, destructive phase interference is well above the fundamental and up where it doesn't really hurt the sound.



So this is how I see the room within which I try to work the art and science (more art than science, and more guesswork and compromise than art!) of sound design. Loudspeakers beaming a range of frequencies into the room, with boomy sound issuing off the back end and gliding around obstructions and through doorways to fill the space; high frequency sounds squirted out in a line of sight, to hit turbulent air and be diffused, mid-range sounds to strike a wall and bounce back setting up a powerful room node at thirty cents flat of Concert A.

Into this room, voices and acoustic instruments also send their sounds, to reverberate and combine and mix. And microphones attempt to navigate the mess, with what they hear being selectively tailored then sent back into the confusion in hopes of enhancing certain needed elements in frequency or time domain.

It is a constant battle between the different needs of relative volume, placement (the sense of where a sound is coming from), frequency content, and intelligibility (the quality of vocal material that makes it possible for human ears to extract understandable words from speech or lyrics). In this mass of compromises, it might be necessary to make an upright bass too "loud" because the indirect sound is too low and too late to give the defined rhythm necessary to support the musical material -- so you mic it just to pop the crisp sound of the attack and define that all-important beat. Or you might end up putting it too loud because the dancers need to hear it to keep in time!



A final thought. The microphone has a huge disadvantage over the human ear, a disadvantage that is at the heart of the difficulty in making a mic'd instrument sound like a "live" instrument. And that is that the mic is in one place. There's only one of it, and it is fixed in position. When we are in an acoustic space ourselves, our ears frequently find themselves in a room node or some other place where frequencies are destructively combined. But we move. Every tiny unconscious motion of our heads shifts the position slightly, and changes the combination of waves mixing and entering at the ear. And we have two ears, as well, each hearing something different, and each making nice little calculations about distance and position that also help us to unconsciously compensate. We sum over time, and we compensate through years of instinct, so a live trumpet in a live space will usually sound like a trumpet. It is the microphone, that lacks any of this subjective adjustment but can only listen objectively, that finds in each set-up, each different night of performance, each note of each tune, some different and incomplete picture of the complete sound.