In this installment I outline the basics of signal flow. There was a time when all signals in a studio passed through a big chunk of hardware called a mixer (a.k.a. desk, board, console). That’s becoming a thing of the past, especially in home studios, but you still need to have a good understanding of your signals, how they get from one place to another, and by which methods and technologies. The fundamentals are the same regardless of the type of system used: analog mixer and tape deck, or combinations of software, computer interfaces, hybrid workstations, control surfaces and separate channel strips.
Though many people now record straight to hard disk and mix in a virtual environment, all of the basic functions of the old “mixer and tape machine” model must still be accomplished. Acoustic sounds still need to be converted to an appropriate electrical signal to interface with the recording system; audio must be routed from one place to another; tracks must be mixed and panned; and dynamics and effects must be added and managed.
When musicians sing, play the guitar, or bang on drums, they create vibrations in the air, which our ears can sense. To be recorded, these airborne vibrations (acoustic energy) must be converted into electrical energy by a microphone. Mics don’t all work the same, due to differences in design. (Much more on microphones coming in TCRM 8.)
Just about all mics produce signals that are too weak for the rest of your studio to handle. The voltage levels of different microphone’s outputs (and their electrical resistance, called impedance) can vary greatly. A mic preamp must be used to ensure that the signal from the microphone is converted to the proper level and impedance for recording. Therefore, the preamp is one of the main input types into the traditional console channel (a dedicated signal path for a single input signal).
It’s important to start out with the right level at the input, before sending the signal further into the electronics of the console (or to the A/D converters). You usually have two tools to achieve that. A variable control knob, called the gain or trim, lets you control the amplification (called gain) given to the incoming signal at the input. By contrast, a switch called a pad can be used to reduce the level of the incoming signal by a fixed amount (often in the area of –12 to –26 decibels). The pad comes in handy for use with microphones that have higher output levels.
Together, the trim and the pad actually serve to increase the effective range of levels a preamp can handle. For instance, a preamp whose gain settings can boost low signals, between –60 dB and –16 dB, to 0 dB (which amounts to 16 to 60 dB of gain) can handle from –34 to +10 signals using a –26 dB pad (for a total of -10 to 34 dB of gain change). The combination of trim and pad allows this preamp to convert signals ranging from –60 dB, all the way to +10 dB, to a nice comfy 0 dB.
Some types of mics require electrical current to operate, called phantom power. Sometimes this current is supplied from a battery that the user places in a compartment on the mic itself, but most often it is +48 Volts DC sent from the mic preamp down the mic cable and can be switched on or off at the preamp.
Finally, preamps can also include phase switches and high-pass filters. The phase switch inverts the signal phase (or polarity) by 180 degrees. Negative becomes positive; positive becomes negative. This can be used to match the phase of mismatched gear or cabling. It can also be used to combat acoustic phase issues between two or more microphones. (More on this in TCRM 9, 10, 11, 35 and 36!)
The high-pass filter (HPF, sometimes called low-cut filter) does what its name implies: high frequencies are passed, low frequencies are cut. These unwanted low frequencies could come from a near-by bass amp on an open stage, from natural resonances of a room, or from physical vibrations in the microphone stand. This filter is also used to combat the bass boost (called proximity effect) exhibited by directional microphones when placed close to the sound source.
A mic preamp turns the weak microphone output into the more robust line-level signal. Some mic preamps do this in a transparent manner, meaning that the tone quality of the signal doesn’t change as the strength of the signal is increased. Other preamps impart a specific coloration to the sound, unique to that preamp model. Both approaches can sound “good” depending upon how they’re used. Therefore, which specific preamps you choose is a matter of both personal taste and style.
Line inputs are for audio sources that are already within the appropriate level range for recording and mixing (typically from .25 to 1.25 volts instead of the 1 to 2 millivolts of many mic signals). This includes devices like synthesizers, drum machines, CD players, and effects units.
Instrument Input (Hi-Z)
Some electric instruments don’t send a normal line-level signal. Typically these are instruments such as electric guitar and bass. Most notably, they have a much greater impedance. To properly interface these kinds of signals, many systems now use a dedicated instrument input. The label “Hi-Z” literally means high-impedance.
For more traditional tape-based studio situations, mixer channels often also include separate tape inputs or tape returns. These are dedicated inputs for mixing the recorded signals that play back from the tape (or any recording device).
On modern software-based systems, input (recording) channels, tracks, and tape returns, are often linked in such a way that the channel toggles automatically between functions. If in record mode, the channel acts as a mic, line, or instrument input. When in playback mode, it acts as a tape return.
Moving Signal In, Out and Around
Now we need to talk about the concept of pre-fader versus post-fader. The main channel fader’s job is to control the signal’s level, or amplitude. (Often, this is also equated with volume, but that term should be reserved for discussions of perceived level (aka. loudness)). The channel fader (it could also be a knob) usually sits at the bottom of a channel strip. When a channel feature is said to be pre-fader (“pre” means before), that feature affects the signal before it has reached the channel fader, so the channel fader doesn’t affect the signal at that earlier stage.
By contrast, a channel feature that is post-fader (“post” means after) deals with the signal after it has been influenced by the setting of the channel fader. Keep that in mind as you read on, and come back to this paragraph if pre- and post- don’t seem to make sense.
A bus (sometimes spelled “buss”) is a signal pathway that can be used to route multiple signals from their individual origins to a common destination. The most frequently used of these is the mix bus (a.k.a. master, stereo, or surround bus). Here, signal from each mix channel is sent to a single master fader. There are many other types of busses including auxiliary sends, groups (submasters), and bus outs (track assign). Signals are sent to these by way of switches on each channel.
Signal flow starts with the channel inputs (mic, line or Hi-Z instrument) and ends with the main mix (stereo or surround masters). In between channels and final masters there may be submasters. These control busses that merge related sounds into groups, which helps in the organization and ease of mixdown. For instance, all of the drum channels can be assigned to a single stereo submaster. The relative balance of the instruments in the drum kit is determined by the channel faders; the level of the entire kit in the mix is controlled by a single submaster fader.
The busses for both master and submaster are post fader. This means that the amount of signal sent is controlled by the individual channel faders. This is not always true of the auxiliary sends.
(Note: submasters are sometimes called groups or subgroups. However, many DAWs use “group” to refer to linked channels. Here, moving one fader casues all linked faders to move along with it. Therefore, submasters and groups should now be considered as distinct features. Furthermore, some DAWs use “aux returns” to create submaster functions (more on aux returns later).
Auxiliary sends are often called Aux Sends or even just Auxes. An Aux bus takes signals from any number of channels and mixes them into one combined signal stream. On each channel the user can use the Aux Send knob to regulate how much signal is sent to the Aux bus. The Aux bus then has its own volume control for the entire mixed signal that’s being sent out of the console.
The idea here is that an independent mix can be sent to an effects unit or used for monitoring purposes. For example, let’s say that the lead and background vocals don’t sound like they’re in the same room. You send them all to an Aux bus to go to (and come back from) an effects device that gives them room ambience (reverb etc.). You set, say, aux 1 and 2 for a stereo send to the effects unit. Then you use the channel aux knobs on each channel of lead and backing vocals to regulate how much of each goes to the aux bus. Finally, you use the master aux send knobs (1+2) to regulate how much of the combined signal goes to the effects unit. To control how much of the effected signal returns to the console (and the mix) you can use aux returns (discussed in greater detail below).
Channel aux sends can be either pre or post channel fader. If a send is pre fader, then you can create two mixes that are entirely independent – the mix created from the combined aux sends, and the mix created at the master (usually stereo) section. This is necessary for creating separate monitor, cue, and record mixes.
If an aux is post fader, then it’s level is relative to the channel fader’s. This is often the best way to treat effects so that when you fade out the unaffected (dry) track, the effects also fade out.
It’s useful to note that auxiliary sends sometimes can send to either a physical (real) output, or to an internal effects processor. On a computer with its own software-based effects you’d expect that, but many dedicated hardware mixing consoles now come equipped with internal effects as well.
Bus outs (track assign)
There are three main methods for sending incoming signals from the input channels out to the multitrack recorder: bus outputs, direct outputs and insert points. Bus outputs (a.k.a. track assign) are sent out of the mix environment by way of either a real or virtual output. They are a flexible way of sending signals to the record tracks because they allow signal from any channel to be sent to any track. Signals can even be combined as they are sent out. Some mixers also allow these busses to be assigned to either pre or post channel fader.
An insert (short for insert point) is a place where a signal can be tapped into, diverted out of the mixer, and sent elsewhere… before returning to the channel and continuing on its merry way. Most often this is done to add a compressor or external effect to a single channels signal. Inserts are not considered busses because they send individual channel signals, not combined ones.
Example: You want to compress a snare drum sample before it even gets to a bus or to the multitracker, but you don’t want to disconnect the permanent cable from your sampler to the console’s channel marked “snare sample.” Instead, right at the channel input, you tap into the insert socket where you get the signal to go to and come back from a compressor.
Because it’s often a send-and-return situation on a single connector, the usual wiring scheme is that of a TRS (tip-ring-sleeve) 1/4” plug that branches out to two TS (tip-sleeve) 1/4” connectors (called a “Y” or “insert” cable). One of the two TS plugs goes into the input of the compressor (or whatever device you use), and the other connects to the compressor’s output. Just make sure that you know which (tip or ring) is the send, and which is the return, for your mixer – this may vary. (Commonly, the tip is the send and the ring is the return… but not always!)
It’s important to note that inserts can also be used to send audio to the multitrack recorder and back. If you do that, you can use the channel fader to set up a monitoring mix—the faders won’t affect the signal going to the multi-tracker because the insert is pre-fader.
Inserts have become just as popular, if not more so, in the virtual, software-based domain where many more are available per channel.
The final option for sending signal from the input channels to the multitrack is by direct outputs. Like inserts, a direct out can only send signal from one specific channel. Unlike inserts, these are outputs only and will not interrupt the signal’s progress down the channel strip. Direct outs may, or may not, have a separate level control like the aux sends.
The master section of a mixer is where the final master fader, subgroup master faders, and aux returns are located. What you do here does not affect the individual channels of instruments or voices, or their individual recorded tracks. This is the place where final levels of pre-combined signals of various kinds are controlled. By this stage, the individual components of the mix better be treated just right, so that nothing sticks out or gets lost.
We looked at using Aux Sends before - now it’s time to bring the sent signals back. Aux returns are dedicated inputs that bus directly to the main mix. Often, the return levels are controlled only by a rotary knob, called a pot (short for potentiometer), rather than a fader. The features attached to these inputs are generally much more limited than channel inputs in terms of bussing, eq and inserts. If you need some of these features on a returning effected signal, you can use a spare channel input as a return. Many inexpensive mixers don’t give you very many returns, expecting you to use channel inputs instead. But if you do use a channel input as an effects return, be careful to avoid feedback as the horrible noise can result in gear or hearing damage.
So how does such a feedback loop happen? If send 1 goes out to a reverb unit and comes back into a channel input, it is possible to send it back out to the same unit by bringing up send 1 on that same channel. It then comes back, goes out, comes back, goes out, comes back, and each time it gets amplified… becoming louder and louder (usually within a fraction of a second). If you don’t need to add the EQ or other features of a full input channel, it’s therefore safer to use dedicated aux returns when you have them available.
Note: Aux returns are sometimes called also known as FX (effects) returns. In some DAWs, however, the term “FX return” specifically refers to a return from a dedicated internal effects processor and cannot be used to return signal from an external source.
The monitor section controls what signal is going out the various mix outputs such as main, monitor, headphone, cue, and 2-track or tape. The tape, headphones, and monitor outs each have independent level controls that boost or cut the signal from the master fader.
If you have a choice, try not to use the Main L+R outputs to feed the monitor amp. The main outputs are taken directly off of the master fader and are best used to feed the two-track master recorder. If you were to use the Main L+R outputs to feed the monitor amp, it would mean that raising or lowering the control room listening level would also affect the signal going to the master and the headphones. (It’s also too easy to bump the fader and blow your speakers… assuming we’re not talkin’ virtual faders here.) Use the Monitor L+R out for monitoring.
Metering and Levels
Proper level settings are extremely important during recording, mixing, and mastering. Audio quality can easily be compromised if incorrect levels are used at any stage in the process. The main problems associated with levels are twofold: excessive noise caused by levels being too low at some point in the chain, or distortion caused by levels being too high. With this in mind, the basic principle of setting levels comes down to this: optimal levels are high enough as to minimize noise, but low enough so as not to cause distortion, or any unwanted misrepresentation, of the signal.
Unfortunately, this is not quite a simple as it sounds. Determining optimal levels can depend a lot on what type of gear your using, how it’s calibrated, whether it’s analog or digital, and where you are in the production process. An acceptable recording level is not necessarily the same as a desirable mixdown level. Before discussing the specifics of level setting, however, we first need to look at measuring levels, in a process known as metering.
Metering is one of the most misunderstood and overlooked aspects of any mixer/recorder system. Meters give a visual representation of a signal’s level and allows the user to see whether there is too much (causing distortion) or too little (introducing extra noise) for a particular situation. Not all meters work the same, however, nor do they all even use the same units of measurement.
VU meters are, traditionally, those cool-looking meters with a moving needle (ol’ skool). Now, they can also come with bargraph LEDs, either real or virtual. They display level in Volume Units (VUs), an early broadcasting standard intended to approximate perceived loudness. To do this, the VU is based on average levels rather than instantaneous peaks. As a result, very quick spikes in the signal (called transients) may not be registered by the meter at all. Though unseen by the meter, these spikes may go high enough to audibly distort. In this case, the levels must be brought down to allow more headroom, despite the meter readings. This is most common when recording instruments with sharp attacks such as drums/percussion, piano and brass. VUs are not equally as useful on all sounds or in all situations.
You need to establish what the strength of your signals is at the console’s outputs while the VU meter shows 0 VU. That signal strength will be the nominal level, and it matters greatly because it has to relate to the devices to which you’re sending the signals: multitracker, two-track recorder, monitors. In the analog world and with balanced equipment, you’ll find this—more often than not—to be +4 dBu (reference 1.23 Volts RMS). When feeding digital recording equipment, where 0 dBFS (zero dB full scale) is the absolute allowable maximum, 0 VU might typically correspond to 20 dB below 0 dBFS.
Exceeding 0 VU on the console (referred to as “going into the red”) is only acceptable when 0 VU allows for that safety margin called headroom; this is around 10 dB or so for analog tape due to its natural compression characteristics, and just under 20 dB for digital in the case mentioned above.
There is much more to the math behind these levels and calibrations. Within the scope of this installment of our series, suffice it to say that VU meters give a reasonable representation of changes in volume, but not with enough accuracy to relieve you from two chores: You have to research the exact relationships between your console, your recording equipment, and your monitoring system AND never stop using your ears. If it sounds distorted, it is distorted, never mind what the meters say. In TCRM 6 we’ll take a more detailed look at these relationships.
Analog meters which use small lights or liquid crystal data displays often reference levels in dBu, a unit of measurement where 0 dBu is equal to .775 Volts RMS. Though dBu are a different scale than VU, the analog LED meter often also displays an average level. Since there is no physical needle to move, however, they may be capable of showing instantaneous level changes. Some even allow the user to switch between averaging and instantaneous or peak (see below): very handy.
A consistent drawback to meters of this type is the rather large steps in the range of the display, creating gaps in its resolution. Since a limited number of lights are used, often these gaps are 3 dB or greater. That means values between, say, –6 dB and –3 dB will all just register as -6.
That’s not terribly accurate, and if you take away enough lights you could end up, as some budget consoles do, with only two LEDs per channel, one for “signal present” (some audible but low number like -40 dBu) and one for “overload” (usually just under 0 dBu for safety’s sake – see below).
Digital LED/LCDs, like their analog counterparts, can be used to display either average or peak. Most often they relate their readings to 0 dBFS, and since 0 dBFS is the absolute allowable maximum for digital recording, all readings will be in the minus range. Any signal that exceeds this threshold of 0 dBFS is digitally represented only as 0 dBFS (there can be no signal higher than 0 dBFS), and clipping occurs (as described in the TCRM 3).
To add functionality to averaging meters, overload indicators are sometimes added. These are single LED lights that flash when the signal has gone into distortion levels, even if just for a moment. These indicators are extremely handy for determining if fast transient attacks are too hot.
In digital this ceiling is hard and fast. In analog, the overload light may come on long after significant distortion is present, or when only a little is audible. Regardless, if the overload lights are coming on, the signal level should be brought down until they no longer do.
It is especially important to trust all of your overload indicators when recording. I have seen a number of people record with the overload lights coming on their interface who still believe that there’s no clipping present. They often say “It’s OK. The clipping indicator is not lighting up on the DAW channel.” This can be a misleading aspect of digital metering, so let’s get this straight….
When you record a strong input signal with peaks above 0 dBFS, it is clipped at the A/D converter, which displays that fact on the overs indicator. When the channel meter sees the signal, nothing exceeds the 0 dB mark (as it has already been cut off), so it may not display an over (depending on the DAW). This does not change the fact that it is clipped! Zooming in on the recorded signal in the edit window will show that the tops of the waveforms have been flattened.
To compound their folly when recording, some people simply bring down the DAW channel fader slightly when the clip light trips there until it is not displayed. Again, this does not change the fact that clipping has occurred (at the A/D input), it merely gets the channel light not to show it.
This brings us to the discussing on pre-fader versus post-fader metering. On record, your channel meters should be set to pre-fader. Since the amount of signal recorded is not affected by the channel fader, it is better to view pre-fader to determine actual record levels. The post-fader levels display how much of the recorded signal is being sent to the mix bus, and so is more useful when mixing. These are very different concerns; how much level you want to record is determined by technical factors such as S/N ratio, while how much you want to send to the mix bus is a matter of musical balance.
Of course, it is possible to be too conservative with record level settings…. Keeping transients and distortion in mind, many recording musicians are tempted to record levels that are too low in order to be safe. This leads to excessive noise. In both analog and digital audio gear, there is a noise floor whose level is basically constant. The difference between the signal level and the noise is known as the signal-to-noise (S/N) ratio. The greater this difference, the better.
Signal A is recorded at –28 dB on a system whose noise floor is at –78 dB, there is only 50 dB separating the signal and the noise.
Signal B is recorded at –8 dB on that same system, so there’s 70 dB separating the signal and the noise.
When each track is brought up to an appropriate mixdown level (let’s say -6 dB) the noise introduced to the mix from A is at -56 dB, while B adds much less, at only -76 dB.
Peak metering and peak hold
As an alternative to the averaging style metering of the classic VU, many current digital audio devices allow metering of peak levels. These allow you to see just how hot the signal is getting and how close even the fast transients are coming to clipping. They work by indicating the absolute highest peak occurring during a short time interval (typically between 50 and 150 milliseconds).
(Beware that this type of peak metering differs drastically from the peak (overload) indicator presented above, which only indicate if a peak has exceeded the distortion threshold. For this reason, care should be taken to keep the concepts of peak metering and peak indicators (preferably called overload or clip indicators) distinct.)
While most peak meters use LED displays, some traditional physical (or virtual) needle types can also be found. Due to the physical limitations of a moving needle, the way the needle responds to fast transients is still somewhat limited.
To make the quicker transient peaks even more visually obvious many peak meters now also offer a peak hold function. This allows the top LED to stay lit for a while through subsequent reporting intervals if those peaks have dropped to a lower level.
Other Mixer Features
EQ and dynamics
Approaches to the use of equalization (tone controls) will be covered in great detail in TCRM 14, 15 and 16. Suffice to say that most individual mixer channels and also some master channels have some sort of equalization. Dedicated stereo channels often do not, because the idea is that such stereo inputs are often used for synths, submixers, and the like, where you have control over the timbre and EQ right at the source.
Eq is common on most mixers, but built-in dynamics processors are not—except for the newer digital mixers, or expensive analog consoles.
Solo & Mute
Most good mixers have both solo and mute (sometimes just called “on”) buttons on every channel. Engaging the mute button will stop any audio from getting past the channel fader to the post fader busses. It may or may not stop the audio from going out any pre-fader sends, inserts, or direct outs. Consult your manual to determine this.
The solo button is useful for quickly isolating and checking the audio of a specific channel. When a channel is soloed, only it will be heard through the monitors. If the mixer has multiple solo, then you can select several channels to be heard together while everything that is not soloed stops sounding. It’s easier to hit one or a few Solo buttons than a whole slew of Mute buttons.
The solo function can be either PFL (pre fader listening: handy for checking the sound source) or AFL (after fader listening: good for focusing on specific elements in a mix). Some mixers even let you select between these two.
Solo-in-place (SIP) is a feature that keeps a soloed track in its panned position rather than making it move into the mono (equally left and right), centered position.
Note: Be sure you know exactly to which outputs the soloed signals will be sent to on your particular system. A groovy take will quickly screech to a halt if the entire band suddenly hears only the singer in the headphones. Again… read your manuals!
Computer Interfaces and Channel Strips
I began this column by saying that traditional mixers are no longer necessary parts of the studio. But no matter how digital/virtual the recording platforms get, there’s always a need for basic analog functions such as preamplifiers, line inputs and monitor mixes. To achieve these basic functions in a more flexible package, many manufacturers are now making stand-alone channel strips, offering all of the basic functions of a console channel without the bulk. These are then connected to whichever computer interface or hybrid workstation the user desires. Additionally, these types of units allow people to mix and match gear to taste, and to purchase only what they need for their own particular studio situation.
Similarly, more and more interfaces (devices used to get audio in and out of the computer) also incorporate elements of the mixing desk. Some, like the MOTU 828mk2 are even capable of acting as a stand-alone mixer. You can take it to gigs and use it to mix… even without a computer!
Though the dedicated hardware mixer is becoming less necessary in this age of digital audio workstations, its spirit will live forever. Like the multitrack recorder itself, the mixer has simply transcended the constraints of this physical world. It now exists in both the physical and virtual planes and can take many forms. A good knowledge of the basics of signal flow and mix functions makes it much easier to understand the great variety of possible configurations.
TCRM 6 will address how to best connect your gear, looking at cabling, grounding, hum, noise, patch bays, connectors, and both level and impedance matching – how the different signals and connections come together properly (and sometimes improperly).
John Shirley is an audio engineer, composer, author, programmer and producer. He’s also on faculty in the Sound Recording Technology Program at the University of Massachusetts Lowell. Check out his wacky electronic music CD, Sonic Ninjutsu, at http://www.cycling74.com/c74music/009.
Supplemental Media Examples
Audio: Setting Levels Too High
Following are sample recordings of solo acoustic guitar and solo flute that demonstrate various distortions caused by excessive recording and mixdown levels. Proper calibration, level settings and metering are necessary to avoid unwanted distortion and loss of clarity.
TCRM5_1a.wav – original sample (digitally recorded at close to 0dBfs)
TCRM5_1c.wav – the original guitar recording sent through a tube preamp calibrated for 12dB past unity. Here, the overload light only blinked more regularly during the loudest parts and the VU meter even exceeded unity in a few spots.
TCRM5_1d.wav – the original guitar recording sent through an analog solid state gain stage calibrated for 6dB past unity. The peak light was more active than in TCRM5_1b.
TCRM5_1e.wav – the original guitar recording sent through an analog solid state gain stage calibrated for 12dB past unity. The peak light was quite active.
TCRM5_1f.wav – the original guitar recording digitally clipped by 6dB.
TCRM5_1g.wav – the original guitar recording digitally clipped by 12dB.
TCRM5_1h.wav – the original guitar recording saturating 2-inch tape.
(note: to keep guitar and flute labeling the same, there is no 2b as there is no tube up 6dB recording.)
TCRM5_2a.wav – original flute sample (digitally recorded at close to 0dBfs)
TCRM5_2c.wav – the original flute recording sent through a tube preamp calibrated for 12dB past unity.
TCRM5_2d.wav – the original flute recording sent through an analog solid state gain stage calibrated for 6dB past unity.
TCRM5_2e.wav – the original flute recording sent through an analog solid state gain stage calibrated for 12dB past unity.
TCRM5_2f.wav – the original flute recording digitally clipped by 6dB
TCRM5_2g.wav – the original flute recording digitally clipped by 12dB
TCRM5_2h.wav – the original flute recording saturating 2-inch tape
Note: TCRM5_1a, TCRM5_1h, TCRM5_2a, and TCRM5_2h were originally recorded at the University of Massachusetts Lowell by graduate students Gavin Paddock and Tim Brault.
The following is a program to demonstrate the effects of inverting phase. It is offered for both Mac OS and Windows. As it is freeware, there’s no support or warranty of any kind (either stated or implied). That said,… enjoy!