I started out as a bass player. Not my first instrument—that was piano, followed by acoustic guitar—but the first one I played well. Inspired by players like Ashley Hutchings, Dave Pegg, Paul McCartney, and Rick Danko, I fell in love with the sound of a good electric bass—growling or sweet, thunderous or lyrical, it provides the band’s foundation in a conversation with drums and lead instruments.
But how to record it? Most engineers try to use at least some of the sound from the amp and speakers, as the colorations they contribute are part of the instrument’s tone. They may mix this sound, miked, with the direct sound of the instrument, taken with a direct feed of some sort (hereinafter referred to as a “DI”). Taking this to its logical conclusion, some engineers try to record the amp speaker track out in the room, away from the sound source, to add air and room sound to the overall mix. And that’s where the trouble begins.
Who also stand and wave
Room acoustics is a messy subject, and I barely have space to scratch the surface here—for a comprehensive treatment see F. Alton Everest’s books, notably his Master Handbook of Acoustics. To hyper-simplify, all rooms resonate at certain frequencies; in rooms with parallel walls, the primary resonances happen at integer multiples of 565/d, where “d” is a dimension of the room, measured in feet.
So, for example, a room that is 10’ long will have a resonance at 565/10, or 56.5 Hz, and corresponding resonances at 2x this frequency (113 Hz), 3x (169.5 Hz), etc. These are known as axial resonances because they take place as the sound bounces back and forth along an axis of the room. They’re also known as standing waves because the hot and cold spots stay in one place.
Another type of standing wave is the tangential; this happens as a sound wave bounces around between four surfaces in a diamond pattern. The formula for calculating tangential resonances is more complex, and I won’t bore you with it. Tangential resonances are less intense than axial.
Finally, there are oblique resonances, caused when sound bounces in a complex pattern between all six surfaces of a rectangular room. Again the formula is more complex; oblique resonances are the weakest of the common types.
What do these resonances do? If a musical note falls on a resonance, that note is reinforced; its volume will be louder, its decay slower. In a good room, resonances support the sound, adding richness and character. Unfortunately, when it comes to bass frequencies most rooms are awful.
Setting a bad example
Table 1 lists the resonant frequencies of a hypothetical room, deliberately selected for badness. First of all, it’s small; as we’ll see in a moment, this leads to serious problems in the bottom octaves of the bass’s range.
Second, the length and width are nearly identical, and almost exactly 1.5x the height. This means resonances will occur at nearly identical frequencies. (Incidentally, I calculated these resonances using a handy bit of freeware called ROOMODE.BAS, written in QBasic for IBM-compatible computers. E-mail me at the magazine (email@example.com) and I’ll send you a copy.)
The most crucial resonances are the axial, which I have listed in italics. Note that several of these (45.20 & 46.12, 90.20 & 92.24) cluster close together, thanks to the room’s nearly identical length and width. The notes corresponding to these frequencies will be doubly reinforced, while the ones around them get little or no help. (Notice the large gap between 46.12 Hz and the next resonance, a weaker tangential one at 64.58 Hz). The frequency response of the room will resemble a rollercoaster, with certain notes dead and lifeless while others jump up and bite you.
How bad is it? Look at the third column, which lists musical notes corresponding to the resonant frequencies. This room will not support frequencies below F#—second fret on the E string. If you play a 5-string, this leaves the first six notes of the low string with no support, along with the first two of the E. The gap between 46.12 Hz and 64.58 skips all the notes between F# and C. D and D# get skipped, while E gets two c lose-together tangential modes to itself. G# and A get short shrift, while Bb gets a double whammy.
As the frequency gets higher, the gaps are filled in by multiple tangential and oblique resonances and the room smoothes out, although there are still pileups on certain notes (there’s a three-banger around 148 Hz, a D). The bottom octave and a half, however, is a standing (sorry) invitation to trouble.
Little room, big room
Similar gaps and pileups are endemic to any small room—the first few resonances will necessarily be at higher frequencies than we’d like, because the room dimensions are small. For comparison, Table 2 lists the resonances for a larger room more typical of a commercial studio. Note that the length and width are no longer identical, nor are they even multiples of the height.
While it’s certainly not perfect, this imaginary room offers much more even support than the previous example. Above 46 Hz there are no significant gaps—and although there are a few pileups, they never involve pairs of axial resonances, so in practice they will be less of a problem. And a little tweaking of dimensions could smooth out this room even further; I pretty much pulled these numbers out of the air.
All well and good, but how many of us can record in a 19' x 31' room with 12' ceilings? Not me—you’d have to gut my house, and where would I shower? Most of us are perforce limited to smaller rooms—basements, bedrooms, and especially garages, which are often square and thus present the extra problem of identical length and width.
How can I leave this behind?
The obvious solution, of course, is to go direct; while you don’t get the amp sound, you never have to worry about room problems, since you’re not using the room.
But DI never sounds quite right, at least not to my ears; much of the instrument’s richness comes from the non-linearities of a speaker cone, the complex colorations of a good amp. So we want at least some miked sound; in a small room the only answer is to close-mic it.
Well, duh—most of you could tell me that. But having arrived at that point, let’s see what we can do to make the close-miked bass sound best.
The first question is mic choice and placement. The soundfield of a conical speaker is complex; different parts of the cone vibrate in different patterns, and the high frequencies get pretty directional. Most cardioid mikes have “proximity effect”—they get bassier when they get close to a sound source. A flat cardioid mic used up close on a bass amp can sound horrible, booming and woofing like a bad juke box.
Some mics have less proximity effect than others; the Electro-Voice RE20 and RE15 (the latter threatened with extinction—ack!) are designed to alter frequency response very little when placed close, and I’ve used them both to mike bass amps with great success. Sennheiser’s venerable MD441 is another good choice, with similar response, as is the Shure SM7.
Among condenser mics, some large-diaphragm mics such as the Neumann U87 have reduced proximity effect and can work well on bass speakers, although I confess to preferring the slight compression of a dynamic. The little Electro-Voice RE200 is lovely; its steep rolloff at the bottom nicely compensates for proximity effect, and the peak up top adds a bit of string bite to the sound.
Another tack is to avoid proximity effect entirely by using an omnidirectional mic—omnis typically maintain their frequency responses at any distance. A multi-pattern condenser like the Alesis/Groove Tubes AM52 or AM62 will work nicely in this spot, as will a good dynamic omni like the Electro-Voice 300 (these show up occasionally on eBay). The increased pickup of room sound won’t be a problem if you’re close enough to the cabinet—say, 1/4" from the speaker grille.
Which brings us to placement. You do want to be very close—the nearer you are, the louder the amp is compared with the room sound. Directly on-axis to the speaker cone gets you the brightest and (usually) the cleanest tone, which you may or may not prefer; I like the sound a bit thicker, so I often place my mic about 1/3 of the way off-center. (You’ll get all the top end you want from the DI track, which we’ll talk about in a moment.) Your ears are the final arbiter—but do your listening with the DI track potted up too, if you’re planning to mix them into a composite sound.
How loud should the amp be? My prejudice is toward tight bass sounds with well-defined starts and stops for each note. Amps tend to get looser as you turn them up, often becoming positively flabby at very high volumes; if that’s the effect you like, go ahead and get loud, but be careful to stay within your mic’s SPL limitations and to avoid clipping your mic preamp or mixer. Myself, I prefer to set the amp just a tad louder than normal conversation, say about 83 dB. This also avoids rattling shelves in the recording room.
Going through a phase
When double-recording an instrument through DI and amp, it’s important to be sure the two signals are in the same relative polarity. Simultaneously record a pair of test tracks: one from the DI, one via the miked amp. Now look at them in a digital editor—do both waveforms go up at the same time or does one go up while the other goes down? If the latter, something in the chain is inverting the polarity (sometimes wrongly called “phase”) of one of the signals.
If the DI and miked amp signal are recorded on separate tracks, this isn’t a hassle; you can reverse the polarity of one during mixdown, or load them into the computer and flip one of them cybernetically. If you’re mixing the DI and microphone direct to a single track, however, you’ll need to flip one signal in the mic preamp or mixer, or use a polarity-reversing mic cable; otherwise the two signals will cancel, leaving some very strange sounding residue. (A neat effect on occasion, but not for daily use.)
Speaking of phase, if you record to separate tracks, you have the chance to compensate for the slight delay inherent in the miked amp. Measure the distance from the speaker cone to the microphone in inches; divide this by 13.56 to get the number of milliseconds the signal is delayed. So, for example, if the mic is 5" from the speaker cone, the delay would be about 0.369 ms; multiply this by the sampling rate in kHz to get the number of samples, about 16 at 44.1 kHz. Sliding the miked track forward by this amount eliminates the comb-filtering effect over 4 kHz.
Fixing to DI
While we’re here, let’s talk for a moment about direct boxes, or DIs. Most are passive devices, using a transformer to step the level and impedance of an electric instrument down to mic level, balanced so it can be carried for long distances without picking up noise. Some of them include attenuator networks, which allow them to be hooked to the output of guitar and bass amps. And a few use active circuits—in effect adding a preamp—that can be made from transistors or tubes, the latter devices costing an arm and a leg. (They sound great, though.)
Transformer-coupled passive DIs can be a bit of a problem on electric instruments, which usually don’t have active outputs. (If your bass is active, you know it, especially when it’s time to replace batteries. This paragraph doesn’t apply to you.)
A typical DI transformer will have a 12:1 voltage stepdown ratio, which translates to a 144:1 impedance ratio. This means that a typical preamp input impedance of 1000 ohms will be seen as a load of 144,000 ohms, or 144K, by the pickups and controls of the bass. This isn’t quite enough; most electric instruments perform best into an impedance of 470K or higher.
So running a bass directly into the DI may not produce as good a tone as you’d like—although your mileage certainly may vary; I’ve had some instruments that sounded quite gorgeous through transformer-coupled DIs.
You have a few alternatives. One, the high-priced spread, is to use an active direct box, which will provide the high input impedance your instrument likes to see. (In effect you’ve turned your bass into an active unit, albeit externally.)
These start at about $130 and go up from there; unfortunately, in my experience the cheaper ones tend to be noisy and pick up radio stations. Some of the better units take phantom power from the mixer or mike preamp, which saves batteries. The Rolls-Royces of this field are probably the Manley and Retrospec boxes, which use vacuum tubes and high-grade transformers. (See the 7/98 issue for a look at several models.)
You can also use an “everything box”—combined preamp and processors—as a DI in the studio, running its output directly to the multitrack at line level. (If you remember my 4/99 review of the Focusrite ToneFactory, you may remember that I found it particularly copacetic on bass.) Similarly, you can use an instrument preamp that incorporates a passive DI; Aguilar and LA Audio both make preamps designed specifically for use with basses. Again, not cheap.
Finally, there’s the amp itself. Modern bass amps often have a balanced XLR output at mic level, effectively a built-in active DI. Usually this takes its signal from the output of the preamplifier and tone control section, giving you the advantage (or disadvantage, depending on your preferences) of a similar signal to the one feeding the speakers.
Or you can use a passive direct box with an attenuator and tap into the speaker output of the amp, in parallel with the speaker cabinet. (The DI’s impedance is sufficiently high that the amp won’t know it’s there.) This gets you a signal with the same level of distortion that the speakers are receiving, which you may or may not want. (I usually like it on guitar amps, but not on basses.)
A simple Radio Shack Y-connector lets you connect the DI to the speaker jack if the amp doesn’t have one to spare.
Room—give me room!
So how do you make a space for the bass? With a reverb usually, but I’d like to call your attention to a handy bit of software called Soundstage. It’s a DirectX plug-in from the makers of Cakewalk, and it’s not to be confused with the old DOS editing program from Turtle Beach (which I still use, by the way). Bill Stunt took a quick look at it 8/99, but it’s worth another mention here.
Soundstage lets you simulate a room of any size and shape, and to place an instrument (taken from a track in the PC) and a pair of virtual mics anywhere within that room. If you like you can “mic” your bass amp from 10 feet away with a pair of crossed cardioids in a room the size and shape of Winchester Cathedral. (You may want to process your DI similarly, or at least slide it to avoid phase problems.)
Certainly other reverb programs can add space to a track, but none works in quite the same way. Check it out.
I mentioned Rick Danko as one of my inspirations. As we went to press I learned that he has just died. This article is dedicated to the memory of a fine musician and a gracious man.