The last five installments of this column have been focused on microphones and miking techniques. Now it’s time to reconsider some of the other methods of generating audio signals and converting them into appropriate forms for recording and mixing.
Why three different input types?
By way of recapping: The three input signal types are mic level, instrument level, and line level. While all analog mics, instruments, and line level sources supply electrical signals to represent sound, the nature of these signals varies greatly. Factors such as voltage, impedance, number of signal legs, and the kind of connectors used determine exactly what sort of interfacing is needed to convert signals to appropriate types for recording.
Mics can produce voltages ranging from mere microvolts to around a single volt, line-level signals (as found in such devices as consoles, CD players, tape machines, and pro-audio effects units) typically function from .003, to 12 or more volts. Instrument signals (from electric guitars, basses, keyboards, etc…) generally run within the expected voltage range of a line input, but are of a much greater impedance. In fact, the impedance is often greater than 100 times that of a line signal. For this reason, instrument signals are often called Hi-Z (as “Z” stands for impedance).
It is important to note that this drastic impedance differential means that instrument signals should not be plugged into line inputs, even though they may use the same 1/4-inch connector as the line-level sources do. The instrument output will be loaded down with excessive current, causing tonal shifts and/or distortion in the audio. In extreme cases, it could even cause damage to the equipment itself. What’s even more insidious is that all of these problems can occur even when the meter on the input signal level indicates that the audio is within an appropriate range. But it just looks as if everything is OK….
Fortunately, “Hi-Z” or “INST” inputs (specifically intended for signals of this kind) are becoming increasingly more common on such devices as mixers, interfaces, effects units, amp emulators, stand-alone preamps, and all-in-one desktop recorders.
Even if these specialized inputs are present, however, not all concerns are addressed. Instrument signals are commonly unbalanced, which makes them susceptible to signal loss, high-frequency attenuation, hum and noise: especially over longer distances. Audio devices which use unbalanced line-level signals suffer from these same problems and limitations. If only there was a way to eliminate these issues….
Introducing the DI
A DI (direct inject) box is designed to convert the signals from electronic instruments and/or amplifiers into ones more appropriate for mic inputs (yes, mic inputs, not line inputs - more in a moment). To accomplish this they attenuate the signal, convert unbalanced signals to balanced ones, and change high impedances to low. Most models can separate the input and output grounds, while some also add transformer isolation. Due to their great flexibility, DIs are an essential part of any studio or live performance rig. Traditional passive and active designs are available from $15 to $500 or more per channel.
Before exploring the basic elements and uses of the DI, let’s first consider an age-old question: If a DI turns an instrument signal into a mic-level signal that is subsequently plugged into a mic pre and amplified to a line level… why not just make the output of the DI a line level directly?
First of all, microphone preamps are much more flexible than line inputs. They are designed to handle a wider range of both input levels and impedances than line inputs. Since the outputs of various instruments (and amplifiers) also vary so greatly, it is much easier to match their converted signals within the useful operating range of a mic preamp. Additionally, since mic pres can also send phantom power, active DIs are often designed to run on this preexisting power supply, thereby reducing reliance on batteries.
To say that a DI can never be plugged directly into a line input would simply be false. In some circumstances this will work just fine, with an active DI capable of either boosting the signal (or just not reducing it). A DI can also work with a line input if the relationship between a specific instruments output signal and a DIs electrical conversion is just right. This happens mostly with very hot input signals and/or if a particular DI has an adjustable output level or input sensitivity. As more features are packed into new DI models and other forms of gear begin acting as multifunction hybrids, highly flexible output solutions are becoming common, especially on the more expensive models.
Passive versus active
The most basic passive DI design is simply an impedance matching transformer with an unbalanced 1/4-inch input and a balanced XLR output. A switch is often provided to connect/disconnect the input and output grounds to help eliminate hum caused by ground loops. The transformer design provides further electrical isolation between input and output. This can reduce noise significantly. A second output, often labeled thru, is sometimes included so that an instruments signal can be sent to the mic input while also continuing on its way to an amplifier unaltered. The circuitry does not require a power supply (battery, phantom, or AC) and so is quite reliable and portable.
Active designs require power but are generally capable of a greater range and number of inputs. Using electronic impedance-matching circuitry, they can function with much higher impedances than most passive DIs. Less expensive models skip the isolation transformer, as it is usually the single most expensive component. These transformerless designs can function well, but are not as versatile as designs using high-quality transformers, which are also noticeably cleaner under some circumstances. In many models the battery is automatically disconnected when nothing is plugged into the output and/or whether phantom power is available (to extend battery life).
Solutions for better tone
The first time you hear a guitar signal coming straight from a DI box, you'll be in for a surprise - it hardly sounds like the full-bodied guitar you're used to hearing from an amp. You can do one of two things to beef up the DI signal. First, you can record the "naked" DI signal and then use software on your DAW to spruce it up - either in real time so that the guitarist hears the more inspiring sound while playing, or after the recording has been completed. Second, you can record the "naked" DI signal and then send it out later to an amp that you'll mike and record (this is called re-amping). Re-amping lets you do the recording in peace and quiet, and later you can experiment with amps and mics. More on this in a moment.
These days we have another option to bypass a noisy amp and still get a realistic tone: you can connect the guitar directly to one of the many amp imulator boxes that are designed to accept a "dry" guitar signal and add the sound of an amplifier… without the actual amplifier. My suggestion here is not to get too fussy with the accuracy of specific emulation presets. While the patch for a Marshall JCM800 might not sound exactly like it’s namesake, it can be a convenient starting point to tweak settings and sculpt some great guitar tones. These simulator boxes are also available for bass as well.
If the particular sound you want is not attainable through a simulator, never fear: there are many more ways to deal with this problem. One of the most prevalent and convenient solutions comes in the form of the proliferation of DIs built into newer amplifier models. Since these most often capture the signal after the preamp (but before the power amp), they do capture the coloration of that circuitry as well as the nuances of its electrical interaction with the instrument's output.
Once again, purists will tell you that bypassing the power amp and speakers removes vital elements of tonal shaping. Amplifiers, especially tube ones, impart a distinctive sonic character to the audio. The speakers and cabinet in turn perform some fairly severe filtering functions and transient shaping. For this reason, some models of DI (whether a stand-alone box or the built-in kind) also offer speaker emulation. This may be as simple as a low-pass filter or as complex as a power amp / speaker emulator matrix selectable by make and model of your favorite classics.
So here we go again; speaker and amp emulation may not be your cup of tea. If you prefer to get the most timbral flavor out of your amp while still reaping the benefits of direct injection methods, there’s one last place in the signal chain to try: after the amp.
Some DIs are actually capable of converting speaker levels to balanced mic-level signals. These units have one or more attenuators (pads) linked in series to give multiple, stepped level ranges. They can be passive or active, though the later is most likely.
DIs with this ability can allow the sonic character of a guitarists or bass player’s amp to be captured without needing to use a microphone. This technique aids in isolating individual performers when an ensemble is tracked simultaneously. It can also be used to avoid any acoustic coloration by the room. Unfortunately, the influence of the speakers themselves is still left out of the equation. As before, emulators or EQ can be used to make up for this last, otherwise unavoidable, shortcoming.
In circumstances where the speakers themselves are disconnected, people's ears can be spared as well as your good relationships with the neighbors. Be aware, however, that when using a DI on the output of a power amplifier, the amp must still be connected to either the speakers or a specialized power soak. This is often accomplished by way of a speaker thru on the DI. This ensures that the amp sees an appropriate load impedance. Matching this electrical resistance is especially critical with tube amplifiers as they can be easily damaged when run into a load outside their intended range of operation.
A final word of warning: never use a DI with speaker levels unless you are certain it was specifically designed for that purpose. Fire and/or electrocution hazards can occur - a definite drawback.
More on the unnatural
As seen above, when any instrument is recorded using a DI, there are compromises made. There’s no influence of the speakers (or, possibly, amp or preamp). There’s also no little to no acoustic signature from the space: no reverb, reflections, or room modes. There’s no influence of the microphone: no frequency response curve, transient response, distortion, or proximity effect. Finally, if a guitar or bass is not subjected to the sound from the speakers, there’s no feedback: no extra sustain, comb filtering, or cool shrieks. So what can you do?
The quirky idiosyncrasies of recording with a DI should not be taken blindly as negatives. On the contrary, they should be seen as a chance to be creative, getting down and dirty with the tone itself. A sense of space can be crafted totally uniquely using artificial reverb, delay, phase, EQ, and panning. The interactive feel of the microphone can be found using mic emulators… even recreated manually or expounded upon with EQ, distortion/overdrive, and fast-attack dynamic expansion. Some of the tonal coloration and sustain of feedback can be revisited with EQ, delays (with feedback) and compression. The proper compressor threshold settings, along with a long release time, can create a convincing illusion of increased sustain.
But if this illusion is proving too hard to conjure, or too easy to see through, maybe real feedback is called for….
The art of re-amping
If a DI track has been captured and found lacking in any of the areas mentioned above, and artificial means of addressing these shortcomings are not making the grade, never fear, there’s always re-amping! This is where the direct-recorded signal is sent back to an amplifier and then re-recorded with a mic (or another flavor of DI, I suppose).
Nowhere is the true flexibility of the direct methodology more obvious. To begin with, you can send the player home and do this whenever circumstances (like sleeping offspring or nervous neighbors) allow. In fact, the original track can now be played through various combinations of amps, preamp settings, speakers, spaces, and microphones. The acoustic context is not only back, but more control is available than ever before! By editing the track first, a definitive version can be assembled even before the natural acoustic environment and final tone are assigned. (Maybe man was not meant to have this much power.)
As you may have already surmised from earlier discussions on level and impedance, re-amping is not as simple as plugging an aux send (at line level) into your amp input. The signal must be changed back to a Hi-z instrument level by a sort of reverse-DI, called (appropriately) a re-amplification box. We should not here that the name Reamp is a trademark of REAMP, Inc., but that many other devices exist to perform this operation, including little labs Redeye and the Radial X-amp. Some offer extra features as well.
Since re-ampboxes convert any line-level source to an instrument signal, a cool alternative use of these devices is with virtual instruments. A realistic acoustic context can now be stamped onto computer-generated sounds or samples. Sounds from the popular program Reason can be run through your Mesa Boogie Rectifier for some extra zip! Way cool.
Next time we will begin a three-chapter look at EQ, how it works, and how it’s used. (That’s right… three chapters! EQ deserves it!)
John Shirley is a recording engineer, composer, programmer and producer. He’s also a Professor in the Sound Recording Technology program at the University of Massachusetts Lowell and chairman of their music department. Check out his wacky electronic music CD, Sonic Ninjutsu, at http://cycling74.com/products/c74music/c74009/
Main uses and benefits of DIs:
To convert instrument signals to the proper level and impedance for a mic input.
To lift an audio ground, breaking ground loops and alleviating hum problems.
To help isolate sounds from one another. This is great for both recording and live sound applications.
To balance unbalanced signals allowing for greater cable lengths and increased sonic accuracy while reducing noise and hum.
To add flexibility when mixing (greater possibilities with effects treatments, emulations, and re-amping).
Supplemental Media Examples
It is very common to record the electric bass using a DI. Below are a number of recordings of the same bass using various Direct Inject techniques.
First, let's hear how the bass sounds when recorded using a microphone. Here, a Shure SM57 was used on-axis straight in front of the speaker. TCRM13_1.wav
Now, the same performance through a DI (in this case a Countryman Type 85). TCRM13_2.wav
Now, another take of the same bass, player, amp and part… but using a Behringer DI100 Ultra-DI. TCRM13_3.wav
That same performance through a Whirlwind Director. TCRM13_4.wav
With more gain on the amp, the same passage is now recorded using a PZM microphone taped to the floor in front of the cabinet. TCRM13_5.wav
Finally, that last take is also recorded using the built-in DI on the amp. TCRM13_6.wav
* Special thanks to Connor Smith for playing bass and helping set up for the bass session, as well as to Scott Cabana for the generous use of his bass rig.
Now, some seriously old-school licks on the guitar.
First, the recording straight from the guitar cabinet using a microphone (again a Shure SM57 on-axis straight in front of the speaker). TCRM13_7.wav
Now listen to how different it sounds straight into a Behringer DI100 Ultra-DI. TCRM13_8.wav
These amp simulations can be layered as well. Here, the recording is run through three instances of the Amp Farm. TCRM13_11.wav
Now four. TCRM13_12.wav
A dry signal, recorded using a DI, can also be sent back out to a real, physical amplifier. Here, it is sent to a Fender mini practice amp. TCRM13_13.wav