My friend who scans brains for a living needed a series of voiceovers recorded. 80 innocuous words, spoken without inflection, nothing fancy. My recording room wasn’t finished being built, but for a down-and-dirty job like this, I figured I could set up a mic on a desk stand at my workstation with a pair of headphones for monitoring. Simple, right
Ha. When I put on the cans and started saying innocuous words, I heard weird sounds—like a bad room opening into an atrocious echo chamber. Taking off the cans, I listened to myself talking; lo and behold, there was the coloration, sounding like a single note resonating (for three seconds!) through the world’s cheesiest spring reverb.
Being on deadline, I didn’t have time to trace down the source of the echo, although it was clearly coming from the area of my computer. Instead, muttering a few not-quite-innocuous words, I found a spot across the room and finished the job.
Afterwards, I wondered, What else is polluting the sound as I monitor? For that matter, what is polluting the sound I record? And what can be done about it?
So this is the first of a series of articles on keeping unwanted stuff out of your recordings. I’ll talk mostly about resonances this time—not the room resonances we covered in the January issue, but rather the rings, rattles, and roars of walls, furniture, and objects.
First, a note to serious technical readers: I am about to engage in an act of deliberate oversimplification. Live with it, okay?
Intuitively, we know what happens when an object resonates mechanically: it vibrates at one frequency more readily than at others. When stimulated, it will ring, and continue ringing for a time after the stimulus is removed; the object stores energy, which is then released into the air as sound.
The frequency at which a particular item rings depends on several factors, including size and shape (a rectangular panel rings quite differently from a round one, for example). For a given size and shape, an object’s resonant behavior is determined by three factors: mass, elasticity, and damping.
The mass and elasticity interact to dictate the frequency of resonance; the damping dictates how long the ringing continues. Engineers refer to resonant systems as possessing an attribute called “Q,” which is the inverse of damping. A resonance with a high Q is poorly damped and rings for a long time.
High-Q resonances are narrow-band—they ring at one particular frequency, even if excited by a broadband sound, and they are exceptionally liable to be excited by sounds at that preferred frequency. Very low-Q resonances, because they cover a wider band, are excited more easily, but the higher damping means they won’t ring for very long. Medium-Q resonances—wide enough to be easily excited, narrow enough to ring for a while—can add a surprising amount of coloration to a room.
Looking for trouble with a whoop and a hey
First, we’ll find the gross problems. Go into the studio or control room, with everything turned off (including the furnace or air conditioner), and let out a loud “Whoop!”, letting your voice go into falsetto. When the whoop cuts off, listen for ringing. Walk around the room, whooping loudly at suspicious objects and taking notes.
Repeat the process, but this time shout “Hey!” instead of “Whoop!”, making sure the cutoff is equally abrupt. Do this with your voice at mid pitch, and again near the bottom of its range. (Borrow a basso if necessary; don’t try to explain what you’re doing, just feed him a beer for his trouble.)
Technically, what you’re doing is stimulating the room with gated broadband signals, biased toward high, medium, and low frequencies, and also biased toward low cost. The highest-Q resonances will be painfully obvious.
In whooping and heying around my living room (in a figure-8 pattern), I found a double whammy resonance from my grandmother’s lovely old floor lamp. The obvious ringing was from the glass tulip-shaped diffuser, but when I removed that I found a second, stronger resonance from the large Mogul light bulb. Both are now banished from recording sessions.
Sweeps and beeps
The next set of tests requires more equipment. If you own a sine wave oscillator with continuously variable frequency, patch it through a power amp and speaker and sweep slowly through the spectrum, from 20 Hz up to where you can’t stand it any more (usually around 5 kHz). Keep the volume loud enough to preclude comfortable conversation.
As you sweep, listen for sudden reinforcements of the sound at particular frequencies, localized at certain places or objects in the room. Listen especially for buzzes and rattles. (It’s useful to walk around the room and feel for sympathetic vibrations with your fingertips—sorry if this sounds New Agey.) You may have some unpleasant surprises, including odd noises from your speaker cabinets themselves. Write them all down.
If you don’t own a variable frequency oscillator, there are several test CDs on the market containing swept sine waves. These are less convenient to use than an oscillator, since you can’t stop the sweep at a particular spot, but they’re cheap and they do the job.
Among CDs with sweep frequencies on them are Taking It To The Road (intended for testing car stereos, and produced by Kenwood); Denon Hi-Fi Check CD; Test Disc III (produced by Hi-Fi News and Record Review) and Audiosource Utility CD. (Look for the first of these at a Kenwood autosound dealer; the rest are available from Old Colony Sound Labs, PO Box 243, Peterborough NH 03458-0243, USA; phone 603-924-6371. The Audiosource disc also cleans your CD player.
A synthesizer with a pitch bender will work too; make sure you can sweep continuously through a wide band of frequencies.
Bust your knuckles
Let’s get tactile. Walk around the room knocking on everything in the room—and I mean everything. Such as:
- walls (every foot or so)
- empty shelves
- equipment racks, occupied or
- equipment (tops and bottoms);
- this includes computers
- table tops
- doors (especially closet doors)
- heating/cooling ducts
- storage cabinets
- speaker cabinets
- light fixtures (tap bulbs lightly
- with a fingernail)
- power amp heat sinks
- and anything else that doesn’t move or meow.
As you knock, listen to the sound. A short “plock” is good, a “clunk” or “clank” is bad, and a “clang” or “boom” is awful.
For some objects, you may want to repeat the test by tapping with a dowel rod or drumstick. (I found a third small resonance in the column of Grandma’s lamp that way.)
Don’t overlook the obvious. I once struggled for an hour with a difficult room before I realized that one wall held a large Chinese gong—very decorative. I removed it. I’ve also done concerts in a music store’s display room, sharing the space with dozens of banjos and dulcimers. When I gave a hey, they rang for a long time.
Particularly in recording spaces, loudspeakers themselves can be a problem. A boxed speaker (or even an open-backed combo amp) is a resonant system, tuned to a low note. (Two notes for a vented box.) If not damped, this resonance can color the room’s sound to a surprising degree.
Soak your knuckles and contemplate your list.
Walls and ceilings are tough. Luckily, drywall (Sheetrock™) panels probably don’t need fixing; by letting sound escape, they may well be keeping the room from becoming excessively boomy.
Thin wooden panels, on the other hand, can sound terrible. Using thicker panels will help, as will bracing them. Better yet, replace thin wooden paneling with thicker, better damped drywall. (If you’re really fond of that “kitchen of a mobile home” look, glue paneling onto the drywall. Make sure you use a thick, even coating of glue, and press it thoroughly into place as it dries to avoid trading new buzzes for old.)
Rattling windows should be grouted or caulked. A short term fix is to apply Mortite™ or 3M Strip-Calk™, two “rope caulk” products that can be molded into place. However, note that Mortite in particular exerts a mysterious influence on cats, who will pry it from the window frame and stash it in the sofa.
Damping large, rectangular panels of any sort follows certain rules. Many materials, especially wood and wood products, are better damped at higher frequencies; pushing panel resonances to higher frequencies can help in its own right.
A good rule of thumb is that the larger the circle that can be drawn on an unsupported section of the panel, the lower the resonant frequency. Thus, a 2' x 3' panel can have a 24" circle inscribed on it; bracing it crosswise limits the circle to 18" (good), but bracing it lengthwise limits the circle to 12" (better).
In Peter Tappan’s tests (see suggested readings, below), a 12" x 18" panel with a resonant frequency of 60 Hz had its resonance raised to 100 Hz by a crosswise brace and 115 Hz by a diagonal brace, but 160 Hz by a lengthwise brace. Moving the brace slightly off-center creates two slightly different resonant frequencies, rather than two identical ones that will reinforce each other.
Bracing can help with many resonant panel problems, including monitor cabinets, doors, table tops, shelves, etc. I find 1 x 2s ideal for braces; they’re stiff enough to be effective, but small enough to stay out of the way when mounted on the bottoms of horizontal surfaces. (I’ve used 2 x 4s inside monster speaker cabinets and under large table tops.)
Attach braces with both glue and screws (aliphatic wood glues are excellent), mounting them along their narrow side. Inside cabinets, in addition to lengthwise braces on the sides, it’s often effective to add a cross brace between sides, cutting across the space inside the cabinet.
Dig my pad, man
Bracing can help, but stubborn resonances need stronger remedies. The clangor of equipment cases and cabinets, in particular, can require heroic measures.
The easy (but expensive) solution is to go to an auto customizers’ shop and look for damping pads and other products from Dynamat™ and similar manufacturers. These are usually made from felt that’s been impregnated with resins or other substances; they’re self-adhesive, they add mass and damping to the surface, and they do a bang-up job. Unfortunately, they cost a lot.
A lower-tech, proletarian alternative may be found at large hardware stores. Look for a product variously known as “sidewalk separator” or “expansion joint”; it’s intended to be placed between sections of concrete, and its “give” allows the sidewalk to expand and contract with temperature changes. I’ve seen it sold as 1/2" x 4" x 5' strips, and also in 4 x 8 sheets; it’s made of asphalt-impregnated felt, essentially the same stuff the BBC uses to damp the cabinets of their fine studio monitors. (The Brits call it “bitumenized felt.”)
Expansion joint isn’t hard to work with; a utility knife will cut it if you apply elbow grease. I don’t suggest using your table saw unless you have an old blade, ready for the scrap heap; asphalt felt gunks up the teeth.
I attach expansion joint to wooden or metal surfaces using Liquid Nails. Use a lot of the adhesive, and spread it evenly with a flat bladed tool until the entire surface of the asphalt felt is generously buttered. Clamp the felt firmly in place or weigh it down with a heavy object, and let it dry for at least 48 hours. (Make sure the room is well ventilated; fumes from the adhesive are plentiful and unpleasant.)
Objects on shelves can be a problem; a good way to damp their buzzing is to place them on a thin, soft felt pad. A mouse pad (available with chintzy promotional messages from your local computer store) is ideal.
Instruments and such
Ideally, the only instruments in your recording room should be those being recorded. If you can move unused acoustic instruments out, do it. Unfortunately, this isn’t easy when the offender is a piano.
When I was married, my wife sewed “snakes” for me—long, skinny things made from denim and pillow stuffing, about 1" around and 30" long. (She added beady little eyes, too.) They twisted nicely across the strings of a piano, damping them more thoroughly than the instrument’s own felts, but they came into their own when I had a hammered dulcimer in the room.
Unlike a piano, a dulcimer’s strings have little or no damping, and they ring with fervor. Many dulcimers are also tricky to keep in tune, and don’t like to be moved, so it seemed wise to leave the instrument in place throughout the session. A couple of snakes under the strings kept them quiet.
If you don’t have snakes, a couple of old pillows will shut a piano up effectively, if less elegantly. Drums are often a problem; small throw pillows on the toms can dampen their resonances nicely, but often cymbals and snares will need to be removed when not in use.
In that music store filled with instruments, we used to spend half an hour before concerts weaving index cards between the strings of banjos and mountain dulcimers. (Push the cards up toward the nut of the instrument so they fit tightly, unless you fancy a comb and tissue paper kazoo effect.)
When possible, keep inactive loudspeakers out of the recording area; otherwise, throw a mover’s blanket over them when they’re not in use. A speaker’s resonance will be worst when it’s not connected to anything; speakers that aren’t plugged into turned-on amplifiers should have their terminals shorted.
Why bother with all this? Is there any real advantage to chasing down and squelching these little demons? Yes, although it’s sometimes subtle. In a control room, you want to hear the speakers and the space—nothing else. A booming panel, a resonant table top, a zinging light fixture…these color the sound in a way that can fool you into a bad mix. A control room should be as neutral as possible.
In the recording space, of course the same principle applies, but there is an additional factor I call “articulation.” In a good recording room, the musical details become better defined: pitches are more definite, especially in the bass; the subtleties of vocal sounds are more audible, both to the singer and the microphone; individual notes are more cleanly delineated. Musicians, hearing the difference, play better; singers sing better. And perforce, the recordings sound better.
Oh yes, what about my workstation? I’ve had to punt on that one for the moment; the “cheesy echo chamber” turned out to be the ringing of the glass bell forming the display tube in my computer’s monitor. That’s something even I am unwilling to futz with.
Does anyone know where I can get a cheap flat screen monitor?
Tappan, Peter W. “Loudspeaker Enclosure Walls” Journal of the Audio Engineering Society 10:3 (July 1962), p. 224.
Benade, Arthur: Horns, Strings and Harmony, Anchor Books, Doubleday & Co., 1960.
Paul J. Stamler is a producer/engineer and session player in the St. Louis area.