By John Shirley
On the Level
Now that we’ve taken some of the mystery out of metering, let’s look at setting levels. To do this properly, we will consider the equipment, the music, and the context within the recording process itself. While I will offer some starting points for level settings, note that these are just loose guidelines. There are way too many variables involved to consider these as “rules”. And remember, the ears, not the eyes, must ultimately judge the appropriateness of audio levels. Never rely solely on the meters!
Channel Input Levels
On an analog console or channel strip, either a 0 dBu or 0 VU level is a good place to start for your strongest signal. Listen carefully to your microphone preamps as you turn the trim up into the last quarter of gain. Often there is significant distortion as they are pushed into the top of their range.
On the input of a digital channel, a –12 dBFS level is considered a good compromise between leaving headroom and maximizing the signal-to-noise-ratio, especially for 16-bit recording.
Analog Recording Levels
When recording to analog tape, it is extra important to get a strong signal to tape in order to maximize S/N. While 0 VU is a good reference level, the appropriate level may range from –8 to +6 depending on the recorder and the audio source. On the VU meters so common to open-reel tape, it is important to listen carefully for distortion and use peak indicators wherever possible.
Unlike digital, with its absolute maximum of 0 dBFS, analog tape does have a bit more flexibility on the top end of it’s level range – it distorts gradually, and the distortion is mostly 2nd and 3rd harmonic (which is usually more acceptable to the ear than the harsh crunch of digital clipping). In addition analog tape has a natural tendency to compress the dynamics of the signal as it approaches saturation. Sometimes this is a highly sought-after sound.
Digital Recording Levels
The most common problem when setting record levels for digital recording is what I like to call “analog envy”. +3 VU just looks so cool, wavering up in the red on that open-reel multitrack. Digital recordists want to get that red color lighting up more… and get hotter levels! But remember, while positive numbers can sometimes be OK to analog tape, they are death in digital… and those pretty red peak lights are never good.
In fact, though I’ve already offered –12 dBFS as a guide for 16-bit recording, it may be wise to allow even greater headroom when making 24-bit multitracks, going as low as -20 dBFS. The inherent S/N of 16-bit recording, while quite good, still requires much attention be paid to setting a strong signal. Since 24 bits offer a 48 dB increase in the signal-to-noise as compared to 16, there is no need to worry about the extra 8 dB. (48 decibels is roughly the difference between the level of a normal conversation from about 2 feet away, and placing your head one foot away from a Fender Deluxe guitar amp, cranked!)
Mix levels
Especially in the digital realm, having a lot of headroom on your track as you begin mixdown is a blessing. This is because the signal level increases as the individual tracks are bussed together. For example, if 14 busy tracks, metering –6 dB each, are mixed together, the result could average as much as 16 dB greater for a combined level of + 10 dB at the mix bus. The resulting distortion can ruin a mix fast! Keep this in mind when mixing… and watch that master level! A digital stereo master must still remain under 0 dBFS. In fact, it is a good rule of thumb to mix to a maximum of –3 dBFS to leave headroom for the final mastering.
This is one reason, of many, to avoid the practice of normalizing your individual tracks. Normalizing finds the highest peak in a selected audio region and then increases the level of the entire region so that the peak is a full 0 dBFS. This leaves no headroom and will very likely cause clipping. Furthermore, it raises the original noise floor, adds more quantization noise… and, when you later lower the levels to avoid the inevitable clipping, will add even more (you guessed it) noise! Save normalizing for the very last step in mastering (if it turns out to still be necessary).
Mastering
Mastering is the last step undertaken before a final mix is ready to be pressed. It can entail anything from the simple preparation of a CD (putting songs in proper order, establishing track spacing, adjusting a song’s level here or there) to elaborate finessing of tonal and dynamic adjustments. The later is generally best left up to a dedicated professional mastering engineer, as it requires well-trained ears, experience and objectivity, and specialized equipment installed in a highly accurate listening environment.
Since many people insist upon doing it themselves, however, I will make the following observations. All dynamics and eq processing should be done before the final level change. Dynamics processors will mean making yet another level adjustment. Equalization (wether boosting or cutting) can lead to higher peak levels. There should be enough headroom to allow for this. Only at the end should any final limiting/normalizing be done, if at all.
Hooking Up
All cabling used for audio is electrical in nature, with the exception of the optical digital formats. As such, it is subject to all of the basic elements and principles that govern electrical circuits including voltage, impedance, polarity, shielding, grounding, transmission loss, capacitance, and interference. Each of these factors influences performance and must be appropriate to the intended use.
Over the years, I have seen many musicians install their own home studios. Invariably, they buy some pretty cool, cutting-edge gear: fancy microphones, designer preamps, powerful multitracks…. Then they order the cheapest cabling they can find to hook it all up with.
Connecting an expensive condenser microphone to a state-of-the-art preamp and recording it in 24-bits at 192 kHz using el cheapo-brand cable is not a smart move. It’s like buying a racecar, feeding it regular unleaded 85-octane gas, and putting $20 tires on it. There should be little surprise when it fails to perform up to its potential.
With audio, both cabling and connectors can be sources of unwanted noise, hum, signal loss, filtering, and phase issues. These effects are both indiscriminate and cumulative. Indiscriminate in that all signals and all gear are affected (both pro and consumer, expensive and inexpensive); cumulative in that more of these unwanted side-effects occur as the amount of cabling and number of connections increases. Using high quality cabling, while not eliminating these problems, will significantly minimize their overall magnitude. Bad, or inappropriate, connectors or cabling can increase these audio problems exponentially.
Quality Factors
We will discuss the various types of cables and connections in a moment, but first let’s attend to the question of quality. Assuming the same type of cable with the same basic specifications, what could possibly make one brand or model of cable superior to another? Some of the factors include:
- Quality control: When quality control is low, the contacts between wires and connectors can be tenuous, phase can be reversed, and the wires or connectors themselves can be inconsistent in gauge or composition.
- Strength and durability: Many inexpensive cables use component materials that are not as robust. Their quality degrades quicker and their useful life is shorter. It’s sometimes shocking just how quickly some of these cables go bad.
- Gauge: Gauge is a way of referring to a wires thickness. Generally, the heavier the gauge (actually stated in decreasing units, e.g. 14-guage cable is thicker than 18-guage) the more current the cable can handle, all other factors being equal. Higher gauge numbers mean decreasing thickness, so less metal is used, thereby cutting costs.
- Conductivity: Both the wires and the connectors need to conduct electricity well. Not all metals or alloys are equally conductive. Unfortunately, the more conductive also tend to be the more expensive…. Gold plating is even used in some brands, as it is both highly conductive as well as resistant to oxidization.
- Capacitance: Capacitance refers to the ability of a medium to store a charge. The capacitance in a run of cable can cause a loss of higher frequencies. As the electrical capacitance increases, the more the higher frequencies are attenuated. The amount of capacitance a cable has depends upon its materials and length: the longer the cable, the higher the capacitance.
- Shielding: All shields are not created equal. They vary in design, materials, and overall effectiveness. Because they are towards the outside diameter of a cable, they need to be fairly flexible. Accomplishing good conductivity and flexibility while maintaining consistent coverage and effectiveness is not simple and can add to the cost of cabling.
While some of these factors are not easily determined when you go to purchase cable, I think they are worth noting as they do start to explain the differences between models and brands and why it is not always wise to buy the “bargain” stuff.
Unbalanced… or just disturbed?
Unbalanced connections are those that use a single conductor for the positive side of the electrical signal (called hot or send) but either share a conductor for the negative and shield (aka. cold/return and ground respectively) or simply have no shield. Because only these two conductors are used, the end connections require only two contacts (examples below).
This configuration is inexpensive to implement, but bypasses one of the main benefits of having a shield in the first place: to capture and remove electrical and magnetic interference from external sources by running them to ground. Since the negative side of the audio travels through the shield, noise or hum cannot be removed by either grounding or cancellation.
Most consumer audio gear (and some professional gear) uses unbalanced connections. Keep unbalanced cabling shorter than 10 feet to minimize interference – the shorter the better. In cases where a signal must be run further, it should first be balanced (see description below). Direct boxes (DIs) or special line amplifiers can be used for this purpose.
The Balancing Act
Balanced signals use two conductors, one for positive and one for negative, surrounded by a separate and dedicated shield. Because of these three elements, the connectors used must have three contacts (sometimes more). Since the shield (ground) is separate from the audio signal, any interference noise that it picks up can be kept out of the program material. Furthermore, when interference does make its way through the shield and into the signal, it can be cancelled out in a process called common mode rejection.
Most interference picked up will be equal and in the same phase relationship on both the positive and negative lines. Since the audio signal itself is the same, but phase inverted, between the two conductors, if one side is inverted the two sides will now be in phase, but the noise will be out of phase. The result will boost the audio and cancel the noise. This not only makes for cleaner audio signals, but also allows for much greater cable runs than in unbalanced lines.
Now that’s twisted…
One way some manufacturers have improved a balanced line’s common mode rejection is to twist the two inner wires. This way, any interference that gets through the shield is picked up more evenly between the two. This is because they each share equally in their relative positions along the length of the cable. Each is equally as often on the left or the right (or up or down…) as the other. For obvious reasons, this kind of cable is called twisted pair. It is slightly more expensive than the non-twisted design as it is more complicated to manufacture and requires more wire per foot.
Another approach to equaling out the distribution of interference between positive and negative is to use four inner conductors, rather than two. Two wires, on opposite sides from one another, carry the positive and the other two carry the negative. This way, neither signal can be relegated to the opposite side of the cable relative to the interference source. These cables are known as quad. Can you say “more money?” And when quad cables are also twisted…
Interference
The type of interference described above, which is picked up by audio cabling, can be generated from many sources. Some common culprits include fluorescent lights, dimmers, amplifiers, speakers, video monitors, power lines and electrical outlets, power supplies, electrical motors, phones, even radio stations. I have twice picked up a taxi radio on a recording, and once even got to listen in on a neighbor’s conversation on her cordless phone. While this can certainly be amusing, it does hinder the recording process.
Some of the culprits listed above can simply be avoided in the studio or while recording. For others, sending properly shielded and balanced signals, from properly shielded gear, is the only way to combat these interference problems.
Keep all audio lines as far away as possible from sources of electronic or magnetic radiation. Do not run microphone or line-level audio cabling alongside power cables or cables carrying signals to speakers or headphones. Also, do not leave live cables coiled up – the coil can amplify any interference.
Connectors
Each end of an audio cable is either terminated in a special connector or, in some circumstances, is left raw. Connectors with two contacts can only function as unbalanced. Connectors with three contacts can be used as either balanced or unbalanced depending on how they are wired and the equipment to which they are hooked up.
Using a three-conductor cable on the unbalanced output of an effects unit will not make the signal balanced. Likewise, any three-conductor cable in which either end terminates in a two-conductor connector can only send an unbalanced signal. While it is a simple matter to unbalance a balanced signal, a special circuit (as found in DIs as mentioned above) must be used to balance an unbalanced signal.
Below is a list of some common connectors used in audio cabling:
2-conductor
- RCA – also called phono: mostly used for –10 dBV “consumer” line-level signals, but also as a cost or space-saving measure on some pro gear, especially for DI use. RCA isn’t necessarily a sub-standard connection format, but it’s worth taking a bit more care to use quality cabling and make sure your cables plug into their jacks firmly – of all the connector types, RCA is most prone to developing loose or intermittent connections due to strain on the plug.
- 1/4-inch TS (tip/sleeve) – sometimes called guitar plug: found on a wide array of cabling including instrument cables, line-levels audio cable, and speaker cable. Note: When you buy a cord with TS plugs, label it according to its intended use. These various types are not be used interchangeably as they have different impedances, gauge (therefore power handling), and capacitance. While connecting an external effects unit using speaker cable may “only” result in a crummy signal, using a guitar cable to connect speakers to an amplifier can be dangerous!
- Mini TS (1/8-inch tip/sleeve) – used on some older analog synthesizers as well as some portable audio equipment.
- Banana – sometimes used for speaker cables, also used in hobby electronics projects and on some analog modular synthesizers. Most banana connectors have a plug on one end and a jack on the other, with the cable attached in the middle. This way multiple plugs, and therefore multiple cables, can be stacked on a single jack, making for a built-in signal splitter or mult. (We’ll talk more about mults in TCRM 7 when we discuss patch bays.)
- BNC (British Naval Connector, or Bayonet Nut Connector, or Bayonet Neill Concelman – take your pick) – does not usually carry audio, but instead is used for video connections and digital word clock. They twist and lock in place when inserted properly.
- Speakon – Made by Neutrik for speaker cables, these have a high power handling capability as well as an easy method to lock firmly in place. They’re common on large speaker arrays, but some passive studio monitors use them as well.
3-conductor (or greater)
- XLR – also called cannon, most commonly found on microphone cables as well as +4 dBm “professional” line-level audio cables and other professional applications. One end of an XLRcable is male (with three pins) and the other end is female (with three sockets). The signal flows from the female end to the male end. Correspondingly, XLR input jacks are always female and outputs are male. Most XLR cables lock in place with a spring-loaded catch that prevents accidental disconnects.
- ¼-inch TRS (tip/ring/sleeve) – often used for +4 dBm line-level signals, but also commonly used for making unbalanced connections such as used for stereo headphones, and for switching jacks. Switching jacks allow an unbalanced signal to be input, but uses the third conductor inside the jack to make or break the circuit. For example: automatically disconnecting the XLR input of a DAW channel when a guitar is plugged into its Hi-Z jack. Another popular unbalanced use is on insert cables (aka Y or send/return cables) where one end has a single TRS connector and the other end is split off into two TS connectors. Here, as in the mini TRS, two unbalanced signals are carried. The difference is that they are going in opposite directions: one out of the console (the send) and one back into the console (the return). Unfortunately, there’s no reliable standard as to which of the connectors, tip or ring, corresponds to the send or return on console insert jacks (see TCRM 5). You’ll have to read your manual to be certain.
- Tiny Telephone (TT) or Bantam – a TRS plug with a smaller size than a standard ¼-inch often used in pro studios’ balanced patchbays because you can fit many more patch points in the same amount of rack space.
- Combo – this is a special type of space-saving jack made by Neutrik for inputs that can accommodate either XLR or ¼-inch TRS plugs. It has holes for 3 XLR pins as well as a central ¼-inch jack. It is used on many DAWs, preamps/channel strips, and computer audio interfaces.
- Mini TRS (tip/ring/sleeve) – also called 1/8-inch or 3.5mm stereo: Though a three-conductor connector, this is most often used to carry two unbalanced signals for headphones, computer audio, portable recording gear, mp3 players, or anywhere else that miniaturized stereo connectors are needed. You should also note that there are also specialized four and five conductor versions of the 3.5mm connectors. The four conductor variant is found on some laptops and camcorders are well as on the iPod Shuffle (2nd generation).
- 2.5mm stereo (TRS) – Also called 3/32-inch stereo, these are most commonly found on cell phones to carry one mono output and a mono mic input. These can also be found in a four-conductor variety on some cell phones to carry stereo sound plus a mono mic input. 2-conductor versions also exist, but are very rare.
- D-Sub – a widely used multipin connector, most often in either 25 or 50-pin varieties. These can carry analog or digital audio to multitrack decks. Exactly which pin corresponds to what signal can vary widely, so be sure to use the right cable for the right box (again, check your manual).
- Elco – a venerable multipin connector made famous by the ADAT standard in which it was a 56-pin format.
One thing to remember when reading technical literature: in North America a “jack” generally refers to a socket, but in Europe it commonly refers to a plug, usually of the ¼-inch variety.
Making it match
Whenever connecting audio recording equipment, it is best to use the least amount of cable with the least number of connections necessary. This means cables should be just the right length, and that individual connections should be completed using only two connectors whenever possible. Using more than one cable for a single signal can degrade the audio, as can the use of adapters.
That being said, it is still a good idea for a recording studio to have a well-rounded collection of adapters (or better yet… odd cable types) for quick fixes and emergencies. It is also a good idea to have a supply of phase reverse adapters. You can easily make short cable runs with one end wired to swap positive for negative if your willing to break out the old soldering iron. These should be marked very clearly as phase reversal adapters. Otherwise, they might be confused with regular cables.
Though there are generally accepted conventions, in the United States anyway, regarding exactly which conductor of each cable type is positive and negative, these sometimes get reversed by mistake. Reverse wiring schemes are also not uncommon on equipment from abroad, or just older gear.
For reference, XLR audio connections are supposed to have pin 2 is Hot (positive), pin 3 Cold (negative) and pin 1 Ground. Balanced TRS lines are supposed to be Tip Hot, Ring Cold and Sleeve Ground. Be sure to check all equipment and cabling to determine if it matches these conventions. In fact, a cable tester is also a must have for any studio. There are a few on the market these days that will check continuity, and wiring schemes… even for intermittency (loose connections).
It is also imperative that a studio be equipped with some decent DI boxes. As stated earlier, these are often necessary to balance a signal that must be run longer than 7 feet or so. They can also help prevent ground loops.
Well, I guess this is a good place to stop for now. TCRM 7 will discuss both ground and power issues, along with patch bays, tie-lines, and digital audio connections.
Supplemental Media Examples
Following are recordings of “nothing”. In fact, these recordings of “silence” are intended to illustrate the noise levels created when recording signals through long cable runs within various environmental scenarios. Basically, they are examples of what to avoid!
Each of these recordings is of an extremely low-level signal sent out from, and returned to, the line-level I/O of an M-Audio FireWire 410. The record levels remained constant for each of these so as to allow the most direct comparison between the various recordings. All of the final examples here have been boosted by exactly 36dB to aid in listening. Please note that this makes the last two examples very hot.
The first example was recorded through a 100-foot long run of coiled, balanced cable. Compare this to all subsequent examples.
The second example is made through a 100-foot long run of coiled, unbalanced cable. Likewise, use this example as a base-line comparison for all subsequent examples as each is a variation on this unbalanced run.
This third example is still through the 100-foot unbalanced run, but now adds a cheap, old adapter at the end.
For the fourth example I removed the adapter used in example 3, but placed a coiled 14-foot AC extension cable on top of my audio cable.
Now I removed the extension cable and turned on a nearby fluorescent light.
Now instead of the fluorescent light I turned on an incandescent light fixture with a dimmer.
The noise gets worse when I decease the amount of light (make the room dimmer).
Just for fun lets add the extension coil back while the lights are still dimmed.
Now let’s try something crazy…. What happens if we plug in one of those “wall-wart” transformers to the extension cable? Be careful this one gets LOUD!
Finally, I use my fingers as an “adapter” by running the positive leg of my audio signal into one finger and out of the other, thereby inserting my hand into the signal path. Again, be careful this one gets REALLY LOUD!
John Shirley is a recording engineer, composer, programmer and producer. He’s also on faculty in the Sound Recording Technology Program at the University of Massachusetts Lowell.