Chapter 5

Mixers

In its simplest form an audio mixer combines several incoming signals into a single output signal. This cannot be achieved simply by connecting all the incoming signals in parallel and then feeding them into a single input because they may influence each other. The signals need to be isolated from each other. Individual control of at least the level of each signal is also required.

In practice, mixers also do rather more things than simply mix. They can provide phantom power for capacitor microphones (see ‘The capacitor or condenser microphone’, Chapter 3); pan control (whereby each signal can be placed in any desired position in a stereo image); filtering and equalisation; routing facilities; and monitoring facilities, whereby one of a number of sources can be routed to a pair of loudspeakers for listening, often without affecting the mixer’s main output.

A simple six-channel mixer

Overview

By way of example, a simple six-channel mixer will be considered, having six inputs and two outputs (for stereo). Figure 5.1 illustrates such a notional six-into-two mixer with basic facilities. It also illustrates the back panel. The inputs illustrated are via XLR-type three-pin latching connectors, and are of a balanced configuration. Separate inputs are provided for microphone and line level signals, although it is possible to encounter systems which simply use one socket switchable to be either mic or line. Many cheap mixers have unbalanced inputs via quarter-inch jack sockets, or even ‘phono’ sockets such as are found on hi-fi amplifiers. Some mixers employ balanced XLR inputs for microphones, but unbalanced jack or phono inputs for line level signals, since the higher-level line signal is less susceptible to noise and interference, and will probably have travelled a shorter distance.

On some larger mixers a relatively small number of multipin connectors are provided, and multicore cables link these to a large jackfield which consists of rows of jack sockets mounted in a rack, each being individually labelled. All inputs and outputs will appear on this jackfield, and patch cords of a metre or so in length with GPO-type jack plugs at each end enable the inputs and outputs to be interfaced with other equipment and tie-lines in any appropriate combination. (The jackfield is more fully described in ‘Patchfield or jackfield’, below, and ‘Jackfields (patchbays)’, Chapter 12.)

Figure 5.1   Front panel and rear connectors of a typical simple six-channel mixer

The outputs are also on three-pin XLR-type connectors. The convention for these audio connections is that inputs have sockets or holes, outputs have pins. This means that the pins of the connectors ‘point’ in the direction of the signal, and therefore one should never be confused as to which connectors are inputs and which are outputs. The microphone inputs also have a switch each for supplying 48 V phantom power to the microphones if required. Sometimes this is found on the input module itself, or sometimes on the power supply, switching 48 V for all the inputs at once.

Input channels

All the input channels in this example are identical, and so only one will be described. The first control in the signal chain is input gain or sensitivity. This control adjusts the degree of amplification provided by the input amplifier, and is often labelled in decibels, either in detented steps or continuously variable. Inputs are normally switchable between mic and line. In ‘mic’ position, depending on the output level of the microphone connected to the channel (see ‘Microphone performance’, Chapter 3), the input gain is adjusted to raise the signal to a suitable line level, and up to 80 dB or so of gain is usually available here (see ‘Miscellaneous features’, below). In ‘line’ position little amplification is used and the gain control normally provides adjustment either side of unity gain (0 dB), perhaps ±20 dB either way, allowing the connection of high-level signals from such devices as CD players, tape machines and musical keyboards.

The equalisation or EQ section which follows (see ‘Equaliser section’, below) has only two bands in this example – treble and bass – and these provide boost and cut of around ±12 dB over broad low and high-frequency bands (e.g.: centred on 100 Hz and 10 kHz). This section can be used like the tone controls on a hi-fi amplifier to adjust the spectral balance of the signal. The fader controls the overall level of the channel, usually offering a small amount of gain (up to 12 dB) and infinite attenuation. The law of the fader is specially designed for audio purposes (see Fact File 5.1). The pan control divides the mono input signal between left and right mixer outputs, in order to position the signal in a virtual stereo sound stage (see Fact File 5.2).

Output section

The two main output faders (left and right) control the overall level of all the channel signals which have been summed on the left and right mix buses, as shown in the block diagram (Figure 5.2). The outputs of these faders (often called the group outputs) feed the main output connectors on the rear panel, and an internal feed is taken from the main outputs to the monitor selector. The monitor selector on this simple example can be switched to route either the main outputs or the PFL bus (see Fact File 5.3) to the loudspeakers. The monitor gain control adjusts the loudspeaker output level without affecting the main line output level, but of course any changes made to the main fader gain will affect the monitor output.

The slate facility on this example allows for a small microphone mounted in the mixer to be routed to the main outputs, so that comments from the engineer (such as take numbers) can be recorded on a tape machine connected to the main outputs. A rotary control adjusts the slate level.

Miscellaneous features

Professional-quality microphones have an output impedance of around 200 ohms, and the balanced microphone inputs will have an input impedance of between 1000 and 2000 ohms (‘2 kΩ’, k = thousand). The outputs should have an impedance of around 200 ohms or lower. The headphone output impedance will typically be 100 ohms or so. Small mixers usually have a separate power supply which plugs into the mains. This typically contains a mains transformer, rectifiers and regulating circuitry, and it supplies the mixer with relatively low DC voltages. The main advantage of a separate power supply is that the mains transformer can be sited well away from the mixer, since the alternating 50 Hz mains field around the former can be induced into the audio circuits. This manifests itself as ‘mains hum’ which is only really effectively dealt with by increasing the distance between the mixer and the transformer. Large mixers usually have separate rack-mounting power supplies.

The above-described mixer is very simple, offering few facilities, but it provides a good basis for the understanding of more complex models. A typical commercial example of a compact mixer is shown in Figure 5.3.

Figure 5.2   Block diagram of a typical signal path from channel input to main output on a simple mixer

Figure 5.3   A compact stereo mixer: the Seemix ‘Seeport’. (Courtesy of Seemix Sound AS)

A multitrack mixer

Overview

The stereo mixer outlined in the previous section only forms half the story in a multitrack recording environment. Conventionally, popular music recording involves at least two distinct stages: the ‘track-laying’ phase, and the ‘mixdown’ phase. In the former, musical tracks are layed down on a multitrack tape recorder in stages, with backing tracks and rhythm tracks being recorded first, followed by lead tracks and vocals. In the mixdown phase, all the previously recorded tracks from the tape recorder are played back through the mixer and combined into stereo to form the finished product which goes to be made into a commercial release. More recently, with the widespread adoption of electronic instruments and MIDI equipment (see Chapter 14), the multitrack tape recorder has begun to play a smaller rôle in some recording studios, because MIDI-sequenced sound sources are now played directly into the mix in the second stage.

For these reasons, as well as requiring mixdown signal paths from many inputs to a stereo bus the mixer also requires signal paths for routing many input signals to a multitrack tape recorder. Often it will be necessary to perform both of these functions simultaneously – that is, recording microphone signals to multitrack tape whilst also mixing the return from tape into stereo, so that the engineer and producer can hear what the finished result will sound like, and so that any musicians who may be overdubbing additional tracks can be given a mixed feed of any previously recorded tracks in headphones. The latter is known as the monitor mix and this often forms the basis for the stereo mixdown when the tracklaying job is finished.

So there are two signal paths in this case: one from the microphone or line source to the multitrack tape recorder, and one from the multitrack recorder back to the stereo mix, as shown in Figure 5.4. The path from the microphone input which usually feeds the multitrack machine will be termed the channel path, whilst the path from the line input or tape return which usually feeds the stereo mix will be termed the monitor path.

Figure 5.4   In multitrack recording two signal paths are needed – one from mic or line input to the multitrack recorder, and one returning from the recorder to contribute to a ‘monitor’ mix

It is likely that some basic signal processing such as equalisation will be required in the feed to the multitrack recorder (see below), but the more comprehensive signal processing features are usually applied in the mixdown path. The situation is somewhat different in the American market where there is a greater tendency to record on multitrack ‘wet’, that is with all effects and EQ, rather than applying the effects on mixdown.

In-line and split configurations

As can be seen from Figure 5.4, there are two complete signal paths, two faders, two sets of EQ, and so on. This takes up space, and there are two ways of arranging this physically, one known as the split-monitoring, or European-style console, the other as the in-line console. The split console is the more obvious of the two, and its physical layout is shown in Figure 5.5. It contains the input channels on one side (usually the left), a master control section in the middle, and the monitor mixer on the other side. So it really is two consoles in one frame. It is necessary to have as many monitor channels as there are tracks on the tape, and these channels are likely to need some signal processing. The monitor mixer is used during track laying for mixing a stereo version of the material that is being recorded, so that everyone can hear a rough mix of what the end result will sound like, and then on mixdown every input to the console can be routed to the stereo mix bus so as to increase the number of inputs for outboard effects, etc. and so that the comprehensive facilities provided perhaps only on the left side of the console are available for the tape returns.

This layout has advantages in that it is easily assimilated in operation, and it makes the channel module less cluttered than the in-line design (described below), but it can make the console very large when a lot of tracks are involved. It can also increase the build cost of the console because of the near doubling in facilities and metalwork required, and it lacks flexibility, especially when switching over from track laying to remixing.

Figure 5.5   A typical ‘split’ or ‘European-style’ multitrack mixer has input modules on one side and monitor modules on the other: two separate mixers in effect

The in-line layout involves the translation of everything from the right-hand side of the split console (the monitor section) into the left side, rather as if the console were sawn in half and the right side merged with the left, as shown in Figure 5.6. In this process a complete monitor signal path is fitted into the same module as the same-numbered channel path, making it no more than a matter of a few switches to enable facilities to be shared between the two paths. In such a design each module will contain two faders (one for each signal path), but usually only one EQ section, one set of auxiliary sends (see below), one dynamics control section, and so on, with switches to swap facilities between paths. (A simple example showing only the switching needed to swap one block of processing is shown in Figure 5.7.) Usually this means that it is not possible to have EQ in both the multitrack recording path and the stereo mix path, but some more recent designs have made it possible to split the equaliser so that some frequency-band controls are in the channel path whilst others are in the monitor path. The band ranges are then made to overlap considerably which makes the arrangement quite flexible.

Figure 5.6   A typical ‘in-line’ mixer incorporates two signal paths in one module, providing two faders per module (one per path). This has the effect of reducing the size of the mixer for a given number of channels, when compared with a split design

Figure 5.7   The in-line design allows for sound processing facilities such as EQ and dynamics to be shared or switched between the signal paths

Further aspects of the in-line design

It has already been stated that there will be two main faders associated with each channel module in an in-line console: one to control the gain of each signal path. Sometimes the small fader is not a linear slider but a rotary knob. It is not uniformly agreed as to whether the large fader at the bottom of the channel module should normally control the monitor level of the like-numbered tape track or whether it should control the channel output level to multitrack tape. Convention has it that American consoles make the large fader ‘the monitor fader’ in normal operation, while British consoles tend to make it ‘the channel fader’. Normally their functions may be swapped over, depending on whether one is mixing down or track laying, either globally (for the whole console), in which case the fader swap will probably happen automatically when switching the console from ‘recording’ to ‘remix’ mode, or on individual channels, in which case the operation is usually performed using a control labelled something like ‘fader flip’, ‘fader reverse’ or ‘changeover’. The process of fader swapping is mostly used for convenience, since more precise control can be exercised over a large fader near the operator than over a small fader which is further away, and thus the large fader is assigned to the function which is being used most in the current operation. This is coupled with the fact that in an automated console, it is almost invariably the large fader which is automated, and the automation is required most in the mixdown process.

Confusion can arise when operating in-line mixers, such as when a microphone signal is fed into, say, mic input 1 and is routed to track 13 on the tape, because the operator will control the monitor level of that track (and therefore the level of that microphone’s signal in the stereo mix) on monitor fader 13, whilst the channel fader on module 1 will control the multitrack record level for that mic signal.

If a 24 track tape machine is in use with the mixer, then monitor faders higher than number 24 will not normally carry a tape return, but will be free for other sources. Remember that more than one microphone signal can be routed to each track on the tape, and so there will be a number of level controls which affect each source’s level in the monitor mix, each of which has a different purpose:

•  MIC LEVEL TRIM – adjusts the gain of the microphone pre-amplifier at the channel input. Usually located at the top of the module.

•  CHANNEL FADER – comes next in the chain and controls the individual level of the mic (or line) signal connected to that module’s input before it goes to tape. Located on the same-numbered module as the input. (May be switched to be either the large or small fader, depending on configuration.)

•  BUS TRIM or TRACK SUBGROUP – will affect the overall level of all signals routed to a particular tape track. Usually located with the track routing buttons at the top of the module. Sometimes a channel fader can be made to act as a subgroup master.

•  MONITOR FADER – is located in the return path from the multitrack recorder to the stereo mix. Does not affect the recorded level on the multitrack tape, but affects the level of this track in the mix. (May be switched to be either the large or small fader, depending on configuration.)

A typical in-line multitrack mixer is shown in the photograph in Figure 5.8.

Channel grouping

Grouping is a term which refers to the simultaneous control of more than one signal at a time. It usually means that one fader controls the levels of a number of slave channels. Two types of channel grouping are currently common: audio grouping and ‘control’ grouping. The latter is often called VCA grouping, but there are other means of control grouping that are not quite the same as the direct VCA control method. The two approaches have very different results, although initially they may appear to be very similar due to the fact that one fader appears to control a number of signal levels. The primary reason for adopting group faders of any kind is in order to reduce the number of faders which the engineer has to handle at a time, and is feasible in a situation where a number of channels are carrying audio signals which can be faded up and down together. These signals do not all have to be at the same initial level, and indeed one is still free to adjust levels individually within a group. A collection of channels carrying drum sounds, or carrying an orchestral string section, would be examples of suitable groups. The two approaches are described in Fact Files 5.4 and 5.5.

Figure 5.8   A typical in-line mixer: the Soundcraft ‘Sapphyre’. (Courtesy of Soundcraft Electronics Ltd)

An overview of typical mixer facilities

Most mixing consoles provide a degree of sound signal processing on board, as well as routing to external processing devices. The very least of these facilities is some form of equalisation (a means of controlling the gain at various frequencies), and there are few consoles which do not include this. As well as signal processing, there will be a number of switches which make changes to the signal path or operational mode of the console. These may operate on individual channels, or they may function globally (affecting the whole console at once). The following section is a guide to the facilities commonly found on multitrack consoles. Figure 5.9 shows the typical location of these sections on an in-line console module.

Input section

•  Input gain control

Sets the microphone or line input amplifier gain to match the level of the incoming signal. This control is often a coarse control in 10 dB steps, sometimes accompanied by a fine trim. Opinion varies as to whether this control should be in detented steps or continuous. Detented steps of 5 or 10 dB make for easy reset of the control to an exact gain setting, and precise gain matching of channels.

•  Phantom power

Many professional mics require 48 volts phantom powering (see ‘Microphone powering options’, Chapter 3). There is sometimes a switch on the module to turn it on or off, although most balanced mics which do not use phantom power will not be damaged if it is accidentally left on. Occasionally this switch is on the rear of the console, by the mic input socket, or it may be in a central assignable switch panel. Other methods exist: for example, one console requires that the mic gain control is pulled out to turn on the phantom power.

Figure 5.9   Typical layout of controls on an in-line mixer module (for description see text)

•  MIC/LINE switch

Switches between the channel’s mic input and line input. The line input could be the playback output from a tape machine, or another line level signal such as a synth or effects device.

•  PAD

Usually used for attenuating the mic input signal by something like 20 dB, for situations when the mic is in a field of high sound pressure. If the mic is in front of a kick drum, for example, its output may be so high as to cause the mic input to clip. Also, capacitor mics tend to produce a higher output level than dynamic mics, requiring that the pad be used on some occasions.

•  Phase reverse or ‘Φ’

Sometimes located after the mic input for reversing the phase of the signal, to compensate for a reversed directional mic, a mis-wired lead, or to create an effect. This is often left until later in the signal path.

•  HPF/LPF

Filters can sometimes be switched in at the input stage, which will usually just be basic high- and low-pass filters which are either in or out, with no frequency adjustment. These can be used to filter out unwanted rumble or perhaps hiss from noisy signals. Filtering rumble at this stage can be an advantage because it saves clipping later in the chain.

Routing section

•  Track routing switches

The number of routing switches depends on the console: some will have 24, some 32 and some 48. The switches route the channel path signal to the multitrack machine, and it is possible to route a signal to more than one track. The track assignment is often arranged as pairs of tracks, so that odd and even tracks can be assigned together, with a pan-pot used to pan between them as a stereo pair, e.g.: tracks 3 and 4 could be a stereo pair for background vocals, and each background vocal mic could be routed to 3 and 4, panned to the relevant place in the image. In an assignable console these controls may be removed to a central assignable routing section.

It is common for there to be fewer routing switches than there are tracks, so as to save space, resulting in a number of means of assigning tracks. Examples are rotary knobs to select the track, one button per pair of tracks with ‘odd/even/ both’ switch, and a ‘shift’ function to select tracks higher than a certain number. The multitrack routing may be used to route signals to effects devices during mixdown, when the track outputs are not being used for recording. In this case one would patch into the track output on the patchfield (see below) and take the relevant signal to an effects input somewhere else on the patchfield. In order to route monitor path signals to the track routing buses it may be necessary to use a switch which links the output of the monitor fader to the track assignment matrix.

In theatre sound mixers it is common for output routing to be changed very frequently, and thus routing switches may be located close to the channel fader, rather than at the top of the module as in a music mixer. On some recent mixers, track routing is carried out on a matrix which resides in the central section above the main faders. This removes unnecessary clutter from the channel modules and reduces the total number of switches required. It may also allow the storing of routing configurations in memory for later recall.

•  Mix routing switches

Sometimes there is a facility for routing the channel path output signal to the main monitor mix, or to one of perhaps four output groups, and these switches will often be located along with the track routing.

•  Channel pan

Used for panning channel signals between odd and even tracks of the multitrack, in conjunction with the routing switches.

•  Bus trim

Used for trimming the overall level of the send to multitrack for a particular bus. It will normally trim the level sent to the track which corresponds to the number of the module.

•  Odd/Even/Both

Occasionally found when fewer routing buttons are used than there are tracks. When one routing button is for a pair of tracks, this switch will determine whether the signal is sent to the odd channel only, the even channel only, or to both (in which case the pan control is operative).

•  DIRECT

Used for routing the channel output directly to the corresponding track on the multitrack machine without going via the summing buses. This can reduce the noise level from the console since the summing procedure used for combining a number of channel outputs to a track bus can add noise. If a channel is routed directly to a track, no other signals can be routed to that track.

Dynamics section

Some advanced consoles incorporate dynamics control on every module, so that each signal can be treated without resorting to external devices. The functions available on the best designs rival the best external devices, incorporating compressor and expander sections which can act as limiters and gates respectively if required. One system allows the EQ to be placed in the side-chain of the dynamics unit, providing frequency-sensitive limiting, among other things, and it is usually possible to link the action of one channel’s dynamics to the next in order to ‘gang’ stereo channels so that the image does not shift when one channel has a sudden change in level while the other does not.

When dynamics are used on stereo signals it is important that left and right channels have the same settings, otherwise the image may be affected. If dynamics control is not available on every module, it is sometimes offered on the central section with inputs and outputs on the patchbay. Dynamics control will not be covered further here, but is discussed in more detail in ‘The compressor/limiter’, below.

Equaliser section

The EQ section is usually split into three or four sections, each operating on a different frequency band. As each band tends to have similar functions these will be described in general. The principles of EQ are described in greater detail in ‘EQ explained’, below.

•  HF, MID 1, MID 2, LF

A high-frequency band, two mid-frequency bands, and a low-frequency band are often provided. If the EQ is parametric these bands will allow continuous variation of frequency (over a certain range), ‘Q’, and boost/cut. If it is not parametric, then there may be a few switched frequencies for the mid band, and perhaps a fixed frequency for the LF and HF bands.

•  Peaking/shelving or BELL

Often provided on the upper and lower bands for determining whether the filter will provide boost/cut over a fixed band (whose width will be determined by the Q), or whether it will act as a shelf, with the response rising or rolling off above or below a certain frequency (see Figure 5.14).

•  Q

The Q of a filter is defined as its centre frequency divided by its bandwidth (the distance between frequencies where the output of the filter is 3 dB lower than the peak output). In practice this affects the ‘sharpness’ of the filter peak or notch, high Q giving the sharpest response, and low Q giving a very broad response. Low Q would be used when boost or cut over a relatively wide range of frequencies is required, while high Q is used to boost or cut one specific region (see Fact File 5.6).

•  Frequency control

Sets the centre frequency of a peaking filter, or the turnover frequency of a shelf.

•  Boost/cut

Determines the amount of boost or cut applied to the selected band, usually up to a maximum of around ±15 dB.

•  HPF/LPF

Sometimes the high- and low-pass filters are located here instead of at the input, or perhaps in addition. They normally have a fixed frequency turnover point and a fixed roll-off of either 12 or 18 dB per octave. Often these will operate even if the EQ is switched out.

•  CHANNEL

The American convention is for the main equaliser to reside normally in the monitor path, but it can be switched so that it is in the channel path. Normally the whole EQ block is switched at once, but on some recent models a section of the EQ can be switched separately. This would be used to equalise the signal which is being recorded on multitrack tape. If the EQ is in the monitor path then it will only affect the replayed signal. The traditional European convention is for EQ to reside normally in the channel path, so as to allow recording with EQ.

•  IN/OUT

Switches the EQ in or out of circuit. Equalisation circuits can introduce noise and phase distortion, so they are best switched out when not required.

Channel and mix controls

•  Pan

See Fact File 5.2.

•  Fader reverse

Swaps the faders between mix and channel paths, so that the large fader can be made to control either the mix level or the channel level. Some systems defeat any fader automation when the large fader is put in the channel path. Fader reverse can often be switched globally, and may occur when the console mode is changed from recording to mixdown.

•  Line/Tape or Bus/Tape

Switches the source of the input to the monitor path between the line output of the same-numbered channel and the return from multitrack tape. Again it is possible that this may be switched globally. In ‘line’ or ‘bus’ mode the monitor paths are effectively ‘listening to’ the line output of the console’s track assignment buses, while in ‘tape’ mode the monitor paths are listening to the off-tape signal (unless the tape machine’s monitoring is switched to monitor the line input of the tape machine, in which case ‘line’ and ‘tape’ will effectively be the same thing!). If a problem is suspected with the tape machine, switching to monitor ‘line’ will bypass the tape machine entirely and allow the operator to check if the console is actually sending anything.

•  Broadcast, or ‘mic to mix’, or ‘simulcast’

Used for routing the mic signal to both the channel and monitor paths simultaneously, so that a multitrack recording can be made while a stereo mix is being recorded or broadcasted. The configuration means that any alterations made to the channel path will not affect the stereo mix, which is important when the mix output is live (see Figure 5.10).

•  BUS or ‘monitor-to-bus’

Routes the output of the monitor fader to the input of the channel path (or the channel fader) so that the channel path can be used as a post-fader effects send to any one of the multitrack buses (used in this case as aux sends), as shown in Figure 5.11. If a BUS TRIM control is provided on each multitrack output this can be used as the master effects-send level control.

•  DUMP

Incorporated (rarely) on some consoles to route the stereo panned mix output of a track (i.e.: after the monitor path pan-pot) to the multitrack assignment switches. In this way, the mixed version of a group of tracks can be ‘bounced down’ to two tracks on the multitrack, panned and level-set as in the monitor mix (see Figure 5.11).

Figure 5.10   A ‘broadcast mode’ switch in an in-line console allows the microphone input to be routed to both signal paths, such that a live stereo mix may be made independent of any changes to multitrack recording levels

Figure 5.11   Signal routings for ‘bounce’, ‘bus’ and ‘dump’ modes (see text)

•  BOUNCE

A facility for routing the output of the monitor fader to the multitrack assignment matrix, before the pan control, in order that tracks can be ‘bounced down’ so as to free tracks for more recording by mixing a group of tracks on to a lower number of tracks. BOUNCE is like a mono version of DUMP (see Figure 5.11).

•  MUTE or CUT

Cuts the selected track from the mix. There may be two of these switches, one for cutting the channel signal from the multitrack send, the other for cutting the mix signal from the mix.

•  PFL

See Fact File 5.3.

•  AFL

After fade listen is similar to PFL, except that it is taken from after the fader. This is sometimes referred to as SOLO, which routes a panned version of the track to the main monitors, cutting everything else. These functions are useful for isolating signals when setting up and spotting faults. On many consoles the AFL bus will be stereo. Solo functions are useful when applying effects and EQ, in order that one may hear the isolated sound and treat it individually without hearing the rest of the mix. Often a light is provided to show that a solo mode is selected, because there are times when nothing can be heard from the loudspeakers due to a solo button being down with no signal on that track. A solo safe control may be provided centrally, which prevents this feature from being activated.

•  In-place solo

On some consoles, solo functions as an ‘in-place’ solo, which means that it actually changes the mix output, muting all tracks which are not solo’ed and picking out all the solo’ed tracks. This may be preferable to AFL as it reproduces the exact contribution of each channel to the mix, at the presently set master mix level. Automation systems often allow the solo functions to be automated in groups, so that a whole section can be isolated in the mix. In certain designs, the function of the automated mute button on the monitor fader may be reversed so that it becomes solo.

Auxiliary sends

The number of aux sends depends on the console, but there can be up to ten on an ordinary console, and sometimes more on assignable models. Aux sends are ‘take-off points’ for signals from either the channel or mix paths, and they appear as outputs from the console which can be used for foldback to musicians, effects sends, cues, and so on. Each module will be able to send to auxiliaries, and each numbered auxiliary output is made up of all the signals routed to that aux send. So they are really additional mix buses. Each aux will have a master gain control, usually in the centre of the console for adjusting the overall gain of the signal sent from the console, and may have basic EQ. Aux sends are often a combination of mono and stereo buses. Mono sends are usually used as routes to effects, while stereo sends may have one level control and a pan control per channel for mixing a foldback source.

•  Aux sends 1–n

Controls for the level of each individual channel in the numbered aux mix.

•  Pre/post

Determines whether the send is taken off before or after the fader. If it is before then the send will still be live even when the fader is down. Generally, ‘cue’ feeds will be pre-fade, so that a mix can be sent to foldback which is independent of the monitor mix. Effects sends will normally be taken post-fade, in order that the effect follows a track’s mix level.

•  Mix/channel

Determines whether the send is taken from the mix or channel paths. It will often be sensible to take the send from the channel path when effects are to be recorded on to multitrack rather than on to the mix. This function has been labelled ‘WET’ on some designs.

•  MUTE

Cuts the numbered send from the aux mix.

Master control section

The master control section usually resides in the middle of the console, or near the right-hand end. It will contain some or all of the following facilities:

•  Monitor selection

A set of switches for selecting the source to be monitored. These will include tape machines (stereo), aux sends, the main stereo mix, and perhaps some miscellaneous external sources like CD players, cassette machines, etc. They only select the signal going to the loudspeakers, not the mix outputs. This may be duplicated to some extent for a set of additional studio loudspeakers, which will have a separate gain control.

•  DIM

Reduces the level sent to the monitor loudspeakers by a considerable amount (usually around 40 dB), for quick silencing of the room.

•  MONO

Sums the left and right outputs to the monitors into mono so that mono compatibility can be checked.

•  Monitor phase reverse

Phase reverses one channel of the monitoring so that a quick check on suspected phase reversals can be made.

•  TAPE/LINE

Usually a global facility for switching the inputs to the mix path between the tape returns and the console track outputs. Can be reversed individually on modules.

•  FADER REVERSE

Global swapping of small and large faders between mix and channel paths.

•  Record/Overdub/Mixdown

Usually globally configures mic/line input switching, large and small faders and auxiliary sends depending on mode of operation. (Can be overridden on individual channels.)

•  Auxiliary level controls

Master controls for setting the overall level of each aux send output.

•  Foldback and Talkback

There is often a facility for selecting which signals are routed to the stereo foldback which the musicians hear on their headphones. Sometimes this is as comprehensive as a cue mixer which allows mixing of aux sends in various amounts to various stereo cues, while often it is more a matter of selecting whether foldback consists of the stereo mix, or one of the aux sends. Foldback level is controllable, and it is sometimes possible to route left and right foldback signals from different sources. Talkback is usually achieved using a small microphone built into the console, which can be routed to a number of destinations. These destinations will often be aux sends, multitrack buses, mix bus, studio loudspeakers and foldback.

•  Oscillator

Built-in sine-wave oscillators vary in quality and sophistication, some providing only one or two fixed frequencies, while others allow the generation of a whole range. If the built-in oscillator is good it can be used for lining up the tape machine, as it normally can be routed to the mix bus or the multitrack outputs. The absolute minimum requirement is for accurate 1 kHz and 10 kHz tones, the 10 kHz being particularly important for setting the bias of an analogue tape machine. The oscillator will have an output level control.

•  Slate

Provides a feed from the console talkback mic to the stereo output, often superimposing a low-frequency tone (around 50 Hz) so that the slate points can be heard when winding a tape at high speed. Slate would be used for recording take information on to tape.

•  Master faders

There may be either one stereo fader or left and right faders to control the overall mix output level. Often the group master faders will reside in this section.

Effects returns

Effects returns are used as extra inputs to the mixer, supplied specifically for inputs from external devices such as reverberation units. These are often located in the central section of the console and may be laid out like reduced-facility input channels. Returns sometimes have EQ, perhaps more basic than on channels, and they may have aux sends. Normally they will feed the mix, although sometimes facilities are provided to feed one or more returns to the multitrack via assignment switches. A small fader or rotary level control is provided, as well as a pan-pot for a mono return. Occasionally, automated faders may be assigned to the return channels so as to allow automated control of their levels in the mix.

Patchfield or jackfield

Most large consoles employ a built-in jackfield or patchbay for routing signals in ways which the console switching does not allow, and for sending signals to and from external devices. Just about every input and output on every module in the console comes up on the patchbay, allowing signals to be cross-connected in virtually any configuration. The jackfield is usually arranged in horizontal rows, each row having an equal number of jacks. Vertically, it tries to follow the signal path of the console as closely as possible, so the mic inputs are at the top and the multitrack outputs are nearer the bottom. In between these there are often insert points which allow the engineer to ‘break into’ the signal path, often before or after the EQ, to insert an effects device, compressor, or other external signal processor. Insert points usually consist of two rows, one which physically breaks the signal chain when a jack is inserted, and one which does not. Normally it is the lower row which breaks the chain, and should be used as inputs. The upper row is used as an output or send. Normalling is usually applied at insert points, which means that unless a jack is inserted the signal will flow directly from the upper row to the lower.

At the bottom of the patchfield will be all the master inputs and outputs, playback returns, perhaps some parallel jacks, and sometimes some spare rows for connection of one’s own devices. Some consoles bring the microphone signals up to the patchbay, but there are some manufacturers who would rather not do this unless absolutely necessary as it is more likely to introduce noise, and phantom power may be present on the jackfield. Jackfields are covered in further detail in Jackfields (patchbays), Chapter 12.

Digital mixers

Much of what has been said applies equally to both analogue and digital mixers. Features which are most commonly found in digital mixers will now be looked at.

In a digital mixer incoming analogue signals are converted to a digital signal as early as possible so that all the functions are performed entirely in the digital domain. Digital inputs and outputs can be provided to connect recording devices and other digital equipment without conversion to analogue. The advantage of this is that once the signal is in the digital domain it is inherently more robust than its analogue counterpart: it is virtually immune from crosstalk, and is unaffected by lead capacitance, electromagnetic fields from mains wiring, additional circuit distortion and noise, and other forms of interference.

Functions such as gain, EQ, delay, phase, routing, and effects such as echo, reverb, compression and limiting, can all be carried out in the digital domain precisely and repeatably using digital signal processing as described in Chapter 8. The microphone amplifiers can be sited remotely from the mixer, close to the microphones themselves, the gain still being adjusted from the mixer, so that their line level outputs can now be converted to digital format before being input to the mixer. Operationally a digital mixer can remain similar to its analogue counterpart, although the first commercial examples have tended at least partially to follow the assignable route described above. The more fully assignable digital mixing console is ergonomically quite different from its analogue counterpart, and a brief description follows.

Because many of the controls of the traditional console such as pan, eq, aux send and group assign are either absent entirely from the assignable console’s control surface or present only as single assignable sections or multi-function controls, many facilities can be packed into a unit of modest dimensions and quite modest cost, the signal remaining in the digital domain with as much as 32-bit internal processing resolution to cope with extremes of signal level, eq settings and other effects. Inputs can be a mixture of analogue and digital (the latter configurable via plug-in modules for Tascam, ADAT, Yamaha and AES/EBU formats, for example) with digital and analogue main and monitoring outputs.

Typically, the control surface consists of many input channel faders and/or rotary knobs, each channel having ‘active’ and ‘select’ buttons. Much smaller areas of the control surface are given over to single sections of eq, routing (aux and group) and processing: these sections are automatically assigned to one particular channel when its ‘select’ button is active before adjustments can take place. Thus most processes which are affecting the signals are not continuously on view or at the fingertips of the operator as is the case with the traditional analogue desk. The assignable console is therefore more suitable to recording work (particularly post-session mix-downs) where desk states can be built up gradually and saved to scene memories, rather than to live performance and primary recording work where continuous visual indication of and access to controls remains desirable.

Facilities such as channel delay, effects processing, moving fader automation and fader ganging (see Figure 5.15), scene memories offering total recall of all settings, MIDI (including memory load and dump via separate MIDI data filers), and timecode interface are typically offered, and a display screen shows the status of all controls: either in simple global formats for the whole console for parameters such as routing, channel delay, scene memory details and the like, or in much greater detail for each individual channel. Metering can also be shown. Cursors facilitate both navigation around the screen displays and adjustments of the various parameters.

Digital mixing has now reached the point where it can be implemented cost effectively, and Yamaha has released a number of low cost digital mixers with full automation. An example is pictured in Figure 5.12. At the other end of the scale companies are manufacturing large-scale studio mixers with an emphasis on ultra-high sound quality and an ergonomically appropriate control interface. An example of such a mixer with a multi-purpose assignable control surface is pictured in Figure 5.13, and the alternative knob per function ‘in-line’ design of Solid State Logic’s MT Production console is shown in Figure 5.14. At the low cost end of the scale, digital mixers are implemented within computer-based workstations and represented graphically on the computer display. Faders and other controls are moved using a mouse.

EQ explained

The tone control or EQ (= equalisation) section provides mid-frequency controls in addition to bass and treble. A typical comprehensive EQ section may have firstly an HF (High-Frequency) control similar to a treble control but operating only at the highest frequencies. Next would come a hi-mid control, affecting frequencies from around 1 kHz to 10 kHz, the centre frequency being adjusted by a separate control. Lo-mid controls would come next, similar to the hi-mid but operating over a range of say 200 Hz to 2 kHz. Then would come an LF (Low-Frequency) control. Additionally, high- and low-frequency filters can be provided. The complete EQ section looks something like Figure 5.16. An EQ section takes up quite a bit of space, and so it is quite common for dual concentric controls to be used. For instance, the cut/boost controls of the hi- and lo-mid sections can be surrounded by annular skirts which select the frequency. Console area is therefore saved.

Figure 5.12   Yamaha 02R digital mixing console. (Courtesy of Sound PR)

Figure 5.13   Sony OXF-R3. (Courtesy of Sony Broadcast and Professional Europe)

Figure 5.14   Solid State Logic OmniMix. (Courtesy Solid State Logic)

Figure 5.15   Fader automation can be dispalyed as ‘streams’ against time. (Digico D5T, courtesy of the RSC)

Figure 5.16   Typical layout of an EQ section

Principal EQ bands

The HF section affects the highest frequencies and provides up to 12 dB of boost or cut. This type of curve is called a shelving curve because it gently boosts or cuts the frequency range towards a shelf where the level remains relatively constant (see Figure 5.17(a)). Next comes the hi-mid section. Two controls are provided here, one to give cut or boost, the other to select the desired centre frequency. The latter is commonly referred to as a ‘swept mid’ because one can sweep across the frequency range.

Figure 5.17(b) shows the result produced when the frequency setting is at the 1 kHz position, termed the centre frequency. Maximum boost and cut affects this frequency the most, and the slopes of the curve are considerably steeper than those of the previous shelving curves. This is often referred to as a ‘bell’ curve due to the upper portion’s resemblance to the shape of a bell. It has a fairly high ‘Q’, that is its sides are steep. Q is defined as:

Q = centre frequency ÷ bandwidth

where the bandwidth is the distance in hertz between the two points at which the response of the filter is 3 dB lower than that at the centre frequency. In the example shown the centre frequency is 1 kHz and the bandwidth is 400 Hz, giving Q = 2.5.

MF EQ controls are often used to hunt for trouble-spots; if a particular instrument (or microphone) has an emphasis in its spectrum somewhere which does not sound very nice, some mid cut can be introduced, and the frequency control can be used to search for the precise area in the frequency spectrum where the trouble lies. Similarly, a dull sound can be given a lift in an appropriate part of the spectrum which will bring it to life in the overall mix. Figure 5.17(c) shows the maximum cut and boost curves obtained with the frequency selector at either of the three settings of 1, 5 and 10 kHz. The high Q of the filters enables relatively narrow bands to be affected. Q may be varied in some cases, as described in Fact File 5.6.

Figure 5.17   (a) Typical HF and LF shelf EQ characteristics shown at maximum settings

Figure 5.17   (b) Typical MF peaking filter characteristic

The lo-mid section is the same as the hi-mid section except that it covers a lower band of frequencies. Note though that the highest frequency setting overlaps the lowest setting of the hi-mid section. This is quite common, and ensures that no ‘gaps’ in the frequency spectrum are left uncovered.

Filters

High- and low-cut filters provide fixed attenuation slopes at various frequencies. Figure 5.17(d) shows the responses at LF settings of 80, 65, 50, 35 and 20 Hz. The slopes are somewhat steeper than is the case with the HF and LF shelving curves, and slope rates of 18 or 24 dB per octave are typical. This enables just the lowest, or highest, frequencies to be rapidly attenuated with minimal effect on the mid band. Very low traffic rumble could be removed by selecting the 20 or 35 Hz setting. More serious low-frequency noise may require the use of one of the higher turnover frequencies. High-frequency hiss from, say, a noisy guitar amplifier or air escaping from a pipe organ bellows can be dealt with by selecting the turnover frequency of the HF section which attenuates just sufficient HF noise without unduly curtailing the HF content of the wanted sound.

Figure 5.17   (c) MF peaking filter characteristics at 1, 5 and 10 kHz

Figure 5.17   (d) High-pass filters with various turnover frequencies

Stereo line input modules

In broadcast situations it is common to require a number of inputs to be dedicated to stereo line level sources, such as CD players, tapes, etc. Such modules are sometimes offered as an option for multitrack consoles, acting as replacements for conventional I/O modules and allowing two signals to be faded up and down together with one fader. Often the EQ on such modules is more limited, but the module may provide for the selection of more than one stereo source, and routing to the main mix as well as the multitrack. It is common to require that stereo modules always reside in special slots on the console, as they may require special wiring. Such modules may also provide facilities for handling LP turntable outputs, offering RIAA equalisation (see ‘RIAA equalisation’, Appendix 2).

With the advent of stereo television, the need for stereo microphone inputs is also becoming important, with the option for MS (middle and side) format signals as well as AB (conventional left and right) format (see ‘Stereo microphones’, Chapter 3).

Dedicated monitor mixer

A dedicated monitor mixer is often used in live sound reinforcement work to provide a separate monitor mix for each musician, in order that each artist may specify his or her precise monitoring requirements. A comprehensive design will have, say, 24 inputs containing similar facilities to a conventional mixer, except that below the EQ section there will be a row of rotary or short-throw faders which individually send the signal from that channel to the group outputs, in any combination of relative levels. Each group output will then provide a separate monitor mix to be fed to headphones or amplifier racks.

Introduction to mixing approaches

Acoustic sources will be picked up by microphones and fed into the mic inputs of a mixer (which incorporates amplifiers to raise the low-voltage output from microphones), whilst other sources usually produce so-called ‘line level’ outputs, which can be connected to the mixer without extra amplification. In the mixer, sources are combined in proportions controlled by the engineer and recorded. In ‘straight-to-stereo’ (or surround) techniques, such as a classical music recording, microphone sources are often mixed ‘live’ without recording to a multitrack medium, creating a session master which is the collection of original recordings, often consisting of a number of takes of the musical material. The balance between the sources must be correct at this stage, and often only a small number of carefully positioned microphones are used. The session master recordings will then proceed to the editing stage where takes are assembled in an artistically satisfactory manner, under the control of the producer, to create a final master which will be transmitted or made into a commercial release. This final master could be made into a number of production masters which will be used to make different release formats. In this case the mixing console used may be a simpler affair than that used for multitrack recording, since the mixer’s job is to take multiple inputs and combine them to a single stereo output, perhaps including processing such as equalisation. This method of production is clearly cheaper and less time consuming than multitrack recording, but requires skill to achieve a usable balance quickly. It also limits flexibility in post-production. Occasionally, classical music is recorded in a multitrack form, especially in the case of complex operas or large-force orchestral music with a choir and soloists, where to get a correct balance at the time of the session could be costly and time consuming. In such a case, the production process becomes more similar to the pop recording situation described below.

‘Pop’ music is rarely recorded live, except at live events such as concerts, but is created in the recording studio. Acoustic and electrical sources are fed into a mixer and recorded on to a multitrack medium, often a few tracks at a time, gradually building up a montage of sounds. The resulting recording then contains a collection of individual sources on multiple tracks which must subsequently be mixed into the final release format. Individual songs or titles are recorded in separate places on the tape, to be compiled later. It is not so common these days to record multitrack pop titles in ‘takes’ for later editing, as with classical music, since mixer automation allows the engineer to work on a song in sections for automatic execution in sequence by a computer. In any case, multitrack machines have comprehensive ‘drop-in’ facilities for recording short inserted sections on individual tracks without introducing clicks, and a pop-music master is usually built up by laying down backing tracks for a complete song (drums, keyboards, rhythm guitars, etc.) after which lead lines are overdubbed using drop-in facilities. Occasionally multitrack recordings are edited or compiled early on during a recording session to compile an acceptable backing track from a number of takes, after which further layers are added. Considerable use may be made of computer-sequenced electronic instruments, under MIDI control, often in conjunction with multitrack disk recording. The computer controlling the electronic instruments is synchronised to the recording machine using time code and the outputs of the instruments are fed to the mixer to be combined with the non-sequenced sources.

Once the session is completed, the multitrack recording is mixed down. This is often done somewhere different from the original session, and involves feeding the outputs of each track into individual inputs of the mixer, treating each track as if it were an original source. The balance between the tracks, and the positioning of the tracks in the stereo image, can then be carried out at leisure (within the budget constraints of the project!), often without all the musicians present, under control of the producer. During the mixdown, further post-production takes place such as the addition of effects from outboard equipment to enhance the mix. An automation system is often used to memorise fader and mute movements on the console, since the large number of channels involved in modern recording makes it difficult if not impossible for the engineer to mix a whole song correctly in one go. Following mixdown, the master that results will be edited very basically, in order to compile titles in the correct order for the production master. The compiled tape will then be mastered for the various distribution media.

Basic operational techniques

Level setting

If one is using a microphone to record speech or classical music then normally a fairly high input gain setting will be required. If the microphone is placed up against a guitar amplifier then the mic’s output will be high and a much lower input gain setting can be used. There are essentially three ways of setting the gain control to the optimum position. Firstly, using PFL or prefade listen (see Fact File 5.3).

PFL is pressed, or the fader overpressed (i.e.: pressed beyond the bottom of its travel against a sprung microswitch), on the input module concerned and the level read on either a separate PFL meter or with the main meters switched to monitor the PFL bus. The channel input gain should be adjusted to give a meter reading of, say, PPM 5, or 0 VU on older analogue desks, and a meter reading of perhaps 6–10 dB below maximum on a digital desk. This gain-setting procedure must be carried out at a realistic input level from the source. It is frequently the case during rehearsals that vocalists and guitarists will produce a level that is rather lower than that which they will use when they actually begin to play.

The pan control should be set next (see Fact File 5.2) to place the source in the stereo image. The main output faders will normally be set to 0 dB on their calibration, which is usually at the top. The channel faders can then be set to give both the desired subjective sound balance and appropriate output meter readings.

The second way of setting the gain is a good way in its own right, and it has to be used if PFL facilities are not provided. First of all both the channel fader and the output faders need to be positioned to the 0 dB point. This will be either at the top of the faders’ travels or at a position about a quarter of the way down from the top of their travel. If no 0 dB position is indicated then the latter position should be set. After the pan control and faders have been positioned, the input gain may then be adjusted to give the desired reading on the output level meters. When several incoming signals need to be balanced the gain controls should all be positioned to give both the desired sound balance between them and the appropriate meter readings – normally PPM 6 or just over 0 VU during the loudest passages. See ‘Correlation between different metering standards’ on page 139 for more information regarding appropriate output meter readings.

These two gain-setting methods differ in that with the former method the channel fader positions will show a correspondence to the subjective contribution each channel is making towards the overall mix, whereas the latter method places all the channel faders at roughly the same level.

The third way is similar to the second way, but one channel at a time is set up, placing channel and output faders at 0 dB and adjusting the gain for a peak meter reading. That channel fader is then turned completely down and the next channel is set up in a similar way. When all the channels which are to be used have been set up, the channel faders can then be advanced to give both the desired subjective balance and peak meter readings.

Use of the EQ controls often necessitates the resetting of the channel’s input gain. For example, if a particular instrument requires a bit of bass boost, applying this will also increase the level of signal and so the gain will often need to be reduced a little to compensate. Applying bass or treble cut will sometimes require a small gain increase.

Using auxiliary sends

Aux facilities were described in ‘Auxiliary sends’, above. The auxiliaries are configured either ‘pre-fade’ or ‘post-fade’. Pre-fade aux sends are useful for providing a monitor mix for musicians, since this balance will be unaffected by movements of the faders which control the main mix. The engineer then retains the freedom to experiment in the control room without disturbing the continuity of feed to the musicians.

Post-fade sends are affected by the channel fader position. These are used to send signals to effects devices and other destinations where it is desirable to have the aux level under the overall control of the channel fader. For example, the engineer may wish to add a little echo to a voice. Aux 2, set to post-fade, is used to send the signal to an echo device, probably positioning the aux 2 control around the number 6 position and the aux 2 master at maximum. The output of the echo device is returned to another input channel or an echo return channel, and this fader can be adjusted to set the amount of echo. The level of echo will then rise and fall with the fader setting for the voice.

The post-fade aux could also be used simply as an additional output to drive separate amplifiers and speakers in another part of a hall, for example.

Using audio groups

The group outputs (see ‘Channel grouping’, above) or multitrack routing buses (see ‘Routing section’, above) can be used for overall control of various separate groups of instruments, depending on whether mixing down or track laying. For example, a drum kit may have eight microphones on it. These eight input channels can be routed to groups 1 and 2 with appropriate stereo pan settings. Groups 1 and 2 would then be routed to stereo outputs left and right respectively. Overall control of the drum kit level is now achieved simply by moving group faders 1 and 2.

When feeding a multitrack tape machine it is normally desirable to use the highest possible recording level on every track regardless of the final required balance, in order to achieve the best noise performance, and each multitrack group output will usually have an output level meter to facilitate this.

Technical specifications

Input noise

The output from a microphone is in the millivolt range, and so needs considerable amplification to bring it up to line level. Amplification of the signal also brings with it amplification of the microphone’s own noise output (discussed in ‘Microphone noise in practice’, Chapter 3), which one can do nothing about, and amplification of the mixer’s own input noise. The latter must therefore be as low as possible so as not to compromise the noise performance unduly. A 200 ohm source resistance on its own generates 0.26 μV of noise (20 kHz bandwidth). Referred to the standard line level of 775 mV (0 dBu) this is −129.6 dBu. A microphone amplifier will add its own noise to this, and so manufacturers quote an ‘equivalent input noise’ (EIN) value which should be measured with a 200 ohm source resistance across the input.

An amplifier with a noise contribution equal to that of the 200 ohm resistor will degrade the theoretically ‘perfect’ noise level by 3 dB, and so the quoted equivalent input noise will be −129.6 + 3 = −126.6 dBm. (Because noise contributions from various sources sum according to their power content, not their voltage levels, dBm is traditionally used to express input noise level.) This value is quite respectable, and good-quality mixers should not be noisier than this. Values of around −128 dBm are sometimes encountered which are excellent, indicating that the input resistance is generating more noise than the amplifier. Make sure that the EIN is quoted with a 200 ohm source, and a bandwidth up to 20 kHz, unweighted. A 150 ohm source, sometimes specified, will give an apparently better EIN simply because this resistor is itself quieter than a 200 ohm one, resistor noise being proportional to ohmic value. Also, weighting gives a flattering result, so one always has to check the measuring conditions. Make sure that EIN is quoted in dBm or dBu. Some manufacturers quote EIN in dBV (i.e.: ref. 1 volt) which gives a result 2.2 dB better.

An input should have high common mode rejection as well as low noise, as discussed in Fact File 5.7. Digital inputs can be of a variety of types, including AES/EBU on XLR sockets, SP/DIF on RCA phono sockets, multipin D connectors for interfacing with digital multitrack machines of a particular manufacturer’s standards, or via an optical link (see Chapter 10). Plug-in modules which provide these options are usually offered. Parameters such as signal level, noise and input impedance do not need to be considered here, but it is worth saying that appropriate signal levels are just as important in the digital domain as in the analogue domain. With analogue, signal levels need to be kept high enough to maximise signal to noise ratio whilst maintaining adequate headroom for a safe overload margin, and the 0 dBu standard (=775 mV) ensures this. With digital, levels must be chosen with equal care. Digital signals coming in to the mixer will already have a clearly defined maximum level, and no surprises should be encountered. A digital input gain control may not even be provided. Once in the mixer, signal levels must be controlled so as to provide meter reading levels along the signal chain which produce adequately high output meter readings, not only for optimum signal to noise ratio but also for low distortion, this latter requirement being in contrast with the analogue mixer. ‘Normalisation’ of signal levels afterward (i.e.: bringing inadequate digital signal levels up to full level at a later stage) cannot lower the distortion nor improve the signal to noise ratio of the programme.

When digital devices are interfaced with one another, it is important that their sampling frequencies continue exactly in step. If they do not, one can experience ticking noises and even drop-outs every so often as the interconnected units drift in and out of sync. Just as analogue tape machines do not run at precisely the same speed, digital machines do not sample at precisely the same rate. For this reason, digital mixers and some other devices are equipped with word-clock sync input and output BNC coax sockets. The output socket gives a continuous stream of data at the chosen sample rate, ‘ticking’ once per sample with no audio data present, which tells receiving devices exactly when each digital word begins. If the receiving device, such as a multitrack recording machine, is equipped with a work-clock sync input then this can be used to receive the sync signal from the mixer. It must then be set to ignore its own internal clock.

A number of digital devices, including many processors, do not have word-clock sync sockets. Here, the AES/EBU or SP/DIF inputs will themselves lock onto the word-clock which is present with the incoming digital signals, and no problems should be encountered.

Output noise

The output residual noise of a mixer, with all faders at minimum, should be at most −90 dBu. There is no point in having a very quiet microphone amplifier if a noisy output stage ruins it. With all channels routed to the output, and all faders at the ‘zero’ position, output noise (or ‘mixing’ noise) should be at least −80 dBu with the channel inputs switched to ‘line’ and set for unity gain. Switching these to ‘mic’ inevitably increases noise levels because this increases the gain of the input amplifier. It underlines the reason why all unused channels should be switched out, and their faders brought down to a minimum. Digital mixers with ‘scene’ memories tend to be programmed by copying a particular scene to another vacant scene, then modifying it for the new requirements. When doing this, one needs to ensure that all unwanted inputs and routing from the copied scene are removed so as to maintain the cleanest possible signal. Make sure that the aux outputs have a similarly good output noise level.

Impedance

A microphone input should have a minimum impedance of 1 kΩ. A lower value than this degrades the performance of many microphones. A line level input should have a minimum impedance of 10 kΩ. Whether it is balanced or unbalanced should be clearly stated, and consideration of the type of line level equipment that the mixer will be partnered with will determine the importance of balanced line inputs. All outputs should have a low impedance, below 200 ohms, balanced (600 ohms sounds nice and professional, but it is much too high, as described in 600 ohms, Chapter 12). Check that the aux outputs are also of very low impedance. Sometimes they are not. If insert points are provided on the input channels and/or outputs, these also should have very low output and high input impedances.

Frequency response

A frequency response which is within 0.2 dB between 20 Hz and 20 kHz for all combinations of input and output is desirable. The performance of audio transformers varies slightly with different source and load impedances, and a specification should state the range of loads between which a ‘flat’ frequency response will be obtained. Above 20 kHz, and probably below 15 Hz or so, the frequency response should fall away so that unwanted out-of-band frequencies are not amplified, for example radio-frequency breakthrough or subsonic interference.

Distortion

In digital mixers, 24 bit, 96 kHz sampling rates are becoming common for input and output converters, and it is also common for mixers and computer-based digital editors to operate using 32 bits internally so as to meet adequate internal headroom and distortion requirements over a fairly wide range of signal levels.

With an analogue mixer, distortion should be quoted at maximum gain through the mixer and a healthy output level of, say, +10 dBu or more. This will produce a typical worst case, and should normally be less than 0.1 per cent THD (Total Harmonic Distortion). The distortion of the low-gain line level inputs to outputs can be expected to be lower: around 0.01 per cent. The outputs should be loaded with a fairly low impedance which will require more current from the output stages than a high impedance will, this helping to reveal any shortcomings. A typical value is 600 ohms.

Clipping and overload margins are discussed in Fact File 5.8

Crosstalk

Crosstalk (signal from an input, an output, or from an internal section of a mixer bleeding through to another section of a mixer) should not be a problem in digital mixers because of the nature of digital signal processing. Interfering data at very low levels which may stray across from adjacent wiring or PCB tracks are ignored by digital circuitry as it is usually below the threshold at which it operates. In analogue mixers, a signal from one input may induce a small signal in another channel, and this is termed ‘crosstalk’. Crosstalk from adjacent channels should be well below the level of the legitimate output signal, and a figure of −80 dB or more should be looked for. Crosstalk performance tends to deteriorate at high frequencies due to capacitive coupling in wiring harnesses for instance, but a crosstalk of at least −60 dB at 15 kHz should still be sought. Similarly, very low-frequency crosstalk often deteriorates due to the power supply source impedance rising here, and a figure of −50 dB at 20 Hz is reasonable.

Ensure that crosstalk between all combinations of input and output is of a similarly good level. Sometimes crosstalk between channel auxiliaries is rather poorer than that between the main outputs.

Metering systems

Metering systems are provided on audio mixers to indicate the levels of audio signals entering and leaving the mixer. Careful use of metering is vital for optimising noise and distortion, and to the recording of the correct audio level on tape. In this section the merits of different metering systems are examined.

Mechanical metering

Two primary types of mechanical meters are in existence today: the VU (Volume Unit) meter (Figure 5.18) and the PPM (Peak Program Meter), as shown in Figure 5.19. These are very different to each other, the only real similarity being that they both have swinging needles. The British, or BBC-type, PPM is distinctive in styling in that it is black with numbers ranging from 1 to 7 equally spaced across its scale, there being 4 dB level difference between each gradation, except between 1 and 2 where there is usually a 6 dB change in level. The EBU PPM (also shown in Figure 5.19) has a scale calibrated in decibels. The VU, on the other hand, is usually white or cream, with a scale running from −20 dB up to +3 dB, ranged around a zero point which is usually the studio’s electrical reference level.

Figure 5.18   Typical VU meter scale

Although this section is not intended to be a tutorial on tape machine line-up and reference levels, it is impossible to cover the subject of metering without reference to such topics, as they are inextricably intertwined. It is important to know how meter readings relate to the line-up standard in use in a particular environment, and to understand that these standards may vary between establishments. Fact File 5.9 discusses the relationship between meter indication and recording level on a tape recorder.

Problems with mechanical meters

PPMs respond well to signal peaks, that is they have a fast rise-time, whereas VUs are quite the opposite: they have a very slow rise-time. This means that VUs do not give a true representation of the peak level going on to tape, especially in cases when a signal with a high transient content, such as a harpsichord, is being recorded, often showing as much as 10–15 dB lower than a peak-reading meter. This can result in overmodulation of the recording, especially with digital recorders where the system is very sensitive to peak overload. None the less, many people are used to working with VUs, and have learned to interpret them. They are good for measuring continuous signals such as tones, but their value for monitoring programme material is dubious in the age of digital recording.

Figure 5.19   (Left) BBC-type peak programme meter (PPM). (Right) European-type PPM

VUs have no control over the fall-time of the needle, which is much the same as the rise-time, whereas PPMs are engineered to have a fast rise-time and a longer fall-time, which tends to be more subjectively useful. The PPM was designed to indicate peaks which would cause audible distortion, but does not measure the absolute peak level of a signal. Mechanical meters take up a lot of space on a console, and it can be impossible to find space for one meter per channel in the case of a multitrack console. In this case there are often only meters on the main outputs, and perhaps measuring some auxiliary signals, these being complemented on more expensive consoles by electronic bargraph meters, usually consisting of LED or liquid crystal displays, or some form of ‘plasma’ display.

Electronic bargraph metering

Unlike mechanical meters, electronic bargraphs have no mechanical inertia to be overcome, so they can effectively have an infinitely fast rise-time although this may not be the ideal in practice. Cheaper bargraphs are made out of a row of LEDs (Light Emitting Diodes), and the resolution accuracy depends on the number of LEDs used. This type of display is sometimes adequate, but unless there are a lot of gradations it is difficult to use them for line-up purposes. Plasma and liquid crystal displays look almost continuous from top to bottom, and do not tend to have the glare of LEDs, being thus more comfortable to work with for any period of time. Such displays often cover a dynamic range far greater than any mechanical meter, perhaps from −50 dB up to +12 dB, and so can be very useful in showing the presence of signals which would not show up on a mechanical PPM. Such a meter is illustrated in Figure 5.20.

There may be a facility provided to switch the peak response of these meters from PEAK to VU mode, where they will imitate the scale and ballistic response of a VU meter. On more up-market designs it may be possible to use the multitrack bargraphs as a spectrum analyser display, indicating perhaps a one-third octave frequency-band analysis of the signal fed to it. Occasionally, bargraph displays incorporate a peak-hold facility. A major advantage of these vertical bargraphs is that they take up very little horizontal space on a meter bridge and can thus be used for providing one meter for every channel of the console: useful for monitoring the record levels on a multitrack tape machine. In this case, the feed to the meter is usually taken off at the input to the monitor path of an in-line module.

Miscellaneous meters may also be provided on the aux send outputs for giving some indication of the level being sent to auxiliary devices such as effects. These are commonly smaller than the main meters, or may consist of LED bargraphs with lower resolution. A phase meter or correlation meter is another option often available, this usually being connected between the left and right main monitor outputs to indicate the degree of phase correlation between these signals. This can be either mechanical or electronic. In broadcast environments, sum and difference (or M and S) meters may be provided to show the level of the monocompatible and stereo difference signals in stereo broadcasting. These often reside alongside a stereo meter for left and right output levels.

Figure 5.20   Typical peak-reading bargraph meter with optional VU scale

Correlation between different metering standards

The correlation between meter indication and electrical output level varies depending on the type of meter and the part of the world concerned. Figure 5.21 shows a number of common meter scales and the relationship between these scales and electrical output level of the mixer. As introduced in Fact File 5.9, there is a further correlation to be concerned with, this being the relationship between the electrical output level of the mixer and the recording level on an analogue or digital tape machine. This is discussed in greater detail in Fact File 6.5.

In Figure 5.20, the digital meter on the right has no indication of a reference level, e.g.: dBu or a VU scaling. When such an output meter is encountered one must be aware of what the zero at the top of the scale means in terms of output signal level. For a digital output, this indicates the maximum level just before digital clipping, and one would normally operate the mixer so as to produce peak readings close to zero when recording digitally in order to achieve maximum signal to noise ratio and lowest distortion from both the mixer and the recorder. If the mixer has analogue outputs which one wishes to use to drive power amplifiers or other analogue devices, the zero meter reading also normally corresponds to maximum output level just before analogue circuit clipping, and this may well be as high as +26 dBu for a balanced output stage. Such a level would be rather too high for many applications and so the gain structure through the mixer would need to be adjusted, usually at the master output faders, to give an operating level in the region of –20 dB if the analogue outputs are to be used. This will produce an output level rather closer to 0 dBu.

Figure 5.21   Graphical comparison of commonly encountered meter scalings and electrical levels in dBu. (After David Pope, with permission)

Meter take-off point

Output level meter-driving circuits should normally be connected directly across the outputs so that they register the real output levels of the mixer. This may seem self-evident but there are certain models in which this is not the case, the meter circuit taking its drive from a place in the circuit just before the output amplifiers. In such configurations, if a faulty lead or piece of equipment the mixer is connected to places, say, a short-circuit across the output the meter will nevertheless read normal levels, and lack of signal reaching a destination will be attributed to other causes. The schematic circuit diagrams of the mixer can be consulted to ascertain whether such an arrangement has been employed. If it is not clear, a steady test tone can be sent to the mixer’s output, giving a high meter reading. Then a short-circuit can be deliberately applied across the output (the output amplifier will not normally be harmed by several seconds of short-circuit) and the meter watched. If the indicated level drastically reduces then the meter is correctly registering the real output. If it stays high then the meter is taking its feed from elsewhere.

Automation

Background

The original, and still most common form of mixer automation is a means of storing fader positions dynamically against time for reiteration at a later point in time, synchronous with recorded material. The aim of automation has been to assist an engineer in mixdown when the number of faders that need to be handled at once become too great for one person. Fader automation has resulted in engineers being able to concentrate on sub-areas of a mix at each pass, gradually building up the finished product and refining it.

MCI first introduced VCA (voltage controlled amplifier) automation for their JH500 series of mixing consoles in the mid-seventies, and this was soon followed by imitations with various changes from other manufacturers. Moving fader automation systems, such as Neve’s NECAM, were introduced slightly later and tended to be more expensive than VCA systems. During the mid 1980s, largely because of the falling cost of microprocessor hardware, console automation enjoyed further advances resulting in developments such as snapshot storage, total dynamic automation, retrofit automation packages, and MIDI-based automation. It is now possible to install basic fader automation on a console for only a few thousand pounds, whereas previously one might have been contemplating tens of thousands. The rise of digital mixers and digitally controlled analogue mixers with integral automation is likely to continue the trend towards total automation of most mixer controls as a standard feature of new products.

In the following sections a number of different approaches to console automation will be presented and discussed.

Fader automation

There are two common means of memorising and controlling the gain of a channel: one which stores the positions of the fader and uses this data to control the gain of a VCA or digital attenuator, the other which also stores fader movements but uses this information actually to drive the fader’s position using a motor. The former is cheaper to implement than the latter, but is not so ergonomically satisfactory because the fader’s physical position may not always correspond to the gain of the channel.

It is possible to combine elements of the two approaches in order that gain control can be performed by a VCA but with the fader being moved mechanically to display the gain. This allows for rapid changes in level which might be impossible using physical fader movements, and also allows for dynamic gain offsets of a stored mix whilst retaining the previous gain profile (see below). In the following discussion the term ‘VCA faders’ may be taken to refer to any approach where indirect gain control of the channel is employed.

With VCA faders it is possible to break the connection between a fader and the corresponding VCA, as was described in Fact File 5.5. It is across this break point that an automation system will normally be connected. The automation processor then reads a digital value corresponding to the position of the fader and can return a value to the VCA to control the gain of the channel (see Figure 5.22). The information sent back to the VCA would depend on the operational mode of the system at the time, and might or might not correspond directly to the fader position. Common operational modes are:

•  WRITE: VCA gain corresponds directly to the fader position

•  READ: VCA gain controlled by data derived from a previously stored mix

•  UPDATE: VCA gain controlled by a combination of previously stored mix data and current fader position

•  GROUP: VCA gain controlled by a combination of the channel fader’s position and that of a group master

Figure 5.22   Fader position is encoded so that it can be read by an automation computer. Data returned from the computer is used to control a VCA through which the audio signal flows

The fader position is measured by an analogue-to-digital convertor (see Chapter 8), which turns the DC value from the fader into a binary number (usually eight or ten bits) which the microprocessor can read. An eight bit value suggests that the fader’s position can be represented by one of 256 discrete values, which is usually enough to give the impression of continuous movements, although professional systems tend to use ten bit representation for more precise control (1024 steps). The automation computer ‘scans’ the faders many times a second and reads their values. Each fader has a unique address and the information obtained from each address is stored in a different temporary memory location by the computer. A generalised block diagram of a typical system is shown in Figure 5.23.

The disadvantage of such a system is that it is not easy to see what the level of the channel is. During a read or update pass the automation computer is in control of the channel gain, rather than the fader. The fader could be half way to the bottom of its travel while the gain of the VCA was near the top. Sometimes a mixer’s bargraph meters can be used to display the value of the DC control voltage which is being fed from the automation to the VCA, and a switch is sometimes provided to change their function to this mode. Alternatively a separate display is provided for the automation computer, indicating fader position with one marker and channel gain with another.

VCA faders are commonly provided with ‘null’ LEDs: little lights on the fader package which point in the direction that the fader must be moved to make its position correspond to the gain of the VCA. When the lights go out (or when they are both on), the fader position is correct. This can sometimes be necessary when modifying a section of the mix by writing over the original data. If the data fed to the VCA from the automation is different to the position of the fader, then when the mode is switched from read to write there will be a jump in level as the fader position takes over from the stored data. The null lights allow the user to move the fader towards the position dictated by the stored data, and most systems only switch from read to write when the null point is crossed, to ensure a smooth transition. The same procedure is followed when coming out of rewrite mode, although it can be bypassed in favour of a sudden jump in level.

Figure 5.23   Generalised block diagram of a mixer automation system handling switches and fader positions. The fader interfaces incorporate a multiplexer (MPX) and demultiplexer (Demux) to allow one convertor to be shared between a number of faders. RAM is used for temporary mix data storage, ROM may hold the operating software program. The CPU is the controlling microprocessor

Update mode involves using the relative position of the fader to modify the stored data. In this mode, the fader’s absolute position is not important because the system assumes that its starting position is a point of unity gain, thereafter adding the changes in the fader’s position to the stored data. So if a channel was placed in update mode and the fader moved up by 3 dB, the overall level of the updated passage would be increased by 3 dB (see Figure 5.24). For fine changes in gain the fader can be preset near the top of its range before entering update mode, whereas larger changes can be introduced nearer the bottom (because of the gain law of typical faders).

Some systems make these modes relatively invisible, anticipating which mode is most appropriate in certain situations. For example, WRITE mode is required for the first pass of a new mix, where the absolute fader positions are stored, whereas subsequent passes might require all the faders to be in UPDATE.

Figure 5.24   Graphical illustration of stages involved in entering and leaving an UPDATE or RELATIVE mode on an automated VCA fader

A moving fader system works in a similar fashion, except that the data which is returned to the fader is used to set the position of a drive mechanism which physically moves the fader to the position in which it was when the mix was written. This has the advantage that the fader is its own means of visual feedback from the automation system and will always represent the gain of the channel.

If the fader was permanently driven, there would be a problem when both the engineer and the automation system wanted to control the gain. Clutches or other forms of control are employed to remove the danger of a fight between fader and engineer in such a situation, and the fader is usually made touch-sensitive to detect the presence of a hand on it.

Such faders are, in effect, permanently in update mode, as they can at any time be touched and the channel gain modified, but there is usually some form of relative mode which can be used for offsetting a complete section by a certain amount. The problem with relative offsets and moving faders is that if there is a sudden change in the stored mix data while the engineer is holding the fader, it will not be executed. The engineer must let go for the system to take control again. This is where a combination of moving fader and VCA-type control comes into its own.

Grouping automated faders

Conventional control grouping (Fact File 5.5) is normally achieved by using dedicated VCA master faders. In an automated console it may be possible to do things differently. The automation computer has access to data representing the positions of all the main faders on the console, so it may allow any fader to be designated a group master for a group of faders assigned to it. It can do this by allowing the user to set up a fader as a group master (either by pressing a button on the fader panel, or from a central control panel). It will then use the level from this fader to modify the data sent back to all the other VCAs in that group, taking into account their individual positions as well. This idea means that a master fader can reside physically within the group of faders to which it applies, although this may not always be the most desirable way of working.

Sometimes the computer will store automation data relating to groups in terms of the motions of the individual channels in the group, without storing the fact that a certain fader was the master, whereas other systems will store the data from the master fader, remembering the fact that it was a master originally.

Mute automation

Mutes are easier to automate than faders because they only have two states. Mute switches associated with each fader are also scanned by the automation computer, although only a single bit of data is required to represent the state of each switch. A simple electronic switch can be used to effect the mute, and this often takes the form of a FET (field effect transistor) in the signal path, which has very high attenuation in its ‘closed’ position (see Figure 5.25). Alternatively, some more basic systems effect mutes by a sudden change in VCA gain, pulling it down to maximum attenuation.

Storing the automation data

Early systems converted the data representing the fader positions and mute switches into a modulated serial data stream which could be recorded alongside the audio to which it related on a multitrack tape. In order to allow updates of the data, at least two tracks were required: one to play back the old data, and one to record the updated data, these usually being the two outside tracks of the tape (1 and 24 in the case of a 24 track machine). This was limiting, in that only the two most recent mixes were ever available for comparison (unless more tracks were set aside for automation), whole tracks had to be mixed at a time (because otherwise the updated track would be incomplete), and at least two audio tracks were lost on the tape. Yet it meant that the mix data was always available alongside the music, eliminating the possibility of losing a disk with the mix data stored separately.

Figure 5.25   Typical implementation of a FET mute switch

More recent systems use computer hardware to store mix data, in RAM and on disks (either hard or floppy). Data is synchronised to the audio by recording time-code on one track of the tape which uniquely identifies any point in time, this being read by the automation system and used to relate tape position to stored data. This method gives almost limitless flexibility in the modification of a mix, allowing one to store many versions, of which sections can be joined together ‘off-line’ (that is, without the tape running) or on-line, to form the finished product. The finished mix can be dumped to a disk for more permanent storage, and this disk could contain a number of versions of the mix.

It is becoming quite common for cheaper automation systems to use MIDI for the transmission of fader data. A basic automation computer associated with the mixer converts fader positions into MIDI information using a device known as a UART which generates and decodes serial data at the appropriate rate for the MIDI standard, as shown in Figure 5.26. MIDI data can then be stored on a conventional sequencer or using dedicated software. This proves adequate for a small number of channels or for uncomplicated mixes, especially if a dedicated sequencer is used for the storage of automation data, but it is possible that a single MIDI interface would be overloaded with information if large mixes with much action were attempted. One recent system overcomes these limitations by using a multiport MIDI interface (see Chapter 14) and a non-standard implementation of MIDI to carry ten bit fader data for a large number of channels.

Figure 5.26   A UART is used to route MIDI data to and from the automation computer

Integrating machine control

Control of a tape machine or machines is a common feature of modern mixers. It may only involve transport remotes being mounted in the centre panel somewhere, or it may involve a totally integrated autolocator/synchroniser associated with the rest of the automation system. On top-flight desks, controls are provided on the channel modules for putting the relevant tape track into record-ready mode, coupled with the record function of the transport remotes. This requires careful interfacing between the console and the tape machine, but means that it is not necessary to work with a separate tape machine remote unit by the console.

It is very useful to be able to address the automation in terms of the mix in progress: in other words, ‘go back to the second chorus’, should mean something to the system, even if abbreviated. The alternative is to have to address the system in terms of timecode locations. Often, keys are provided which allow the engineer to return to various points in the mix, both from a mix data point-of-view and from the tape machines’ point-of-view, so that the automation system locates the tape machine to the position described in the command, ready to play. This facility is often provided using an integral synchroniser with which the automation computer can communicate, and this can be a semi-customised commercial synchroniser from another manufacturer.

MIDI Machine Control (MMC) is also becoming a popular means of remote control for modular multitrack recording equipment, and is another way of interfacing a tape recorder to an automation system.

Retrofitting automation

Automation can usually be retrofitted into existing consoles which do not have any automation. These systems usually control only the faders and the mutes, as anything else requires considerable modification of the console’s electronics, but the relatively low price of some systems makes them attractive, even on a modest budget. Fitting normally involves a modification or replacement of the fader package, to incorporate VCAs in consoles which don’t have them, or to break into the control path between fader and VCA in systems which do. This job can normally be achieved in a day. It is also possible to retrofit moving fader automation.

A separate control panel may be provided, with buttons to control the modes of operation, as well as some form of display to show things like VCA gains, editing modes, and set up data. The faders will be interfaced to a processor rack which would reside either in a remote bay, or under the console, and this will normally contain a disk drive to store the final mixes. Alternatively a standard desktop computer will be used as the control interface.

Total automation systems

This title means different things to different people. SSL originally coined the term ‘Total Recall’ for its system, which was not in fact a means of completely resetting every control on the console; rather it was a means of telling the operator where the controls should be and leaving him to reset them himself. None the less, this saved an enormous amount of time in the resetting of the console in between sessions, because it saved having to write down the positions of every knob and button.

True Total Reset is quite a different proposition and requires an interface between the automation system and every control on the console, with some means of measuring the position of the control, some means of resetting it, and some means of displaying what is going on. A number of options exist, for example one could:

•  motorise all the rotary pots

•  make all the pots continuously rotating and provide a display

•  make the pots into up/down-type incrementers with display

•  provide assignable controls with larger displays

Of these, the first is impractical in most cases due to the space that motorised pots would take up, the reliability problem, and the cost, although it does solve the problem of display. The second would work, but again there is the problem that a continuously rotating pot would not have a pointer because it would merely be a means of incrementing the level from wherever it was at the time, so extra display would be required and this takes up space. None the less some ingenious solutions have been developed, including incorporating the display in the head of rotary controls (see Figure 5.27). The third is not ergonomically very desirable, as the human prefers analogue interfaces rather than digital ones, and there is no room on a conventional console for all the controls to be of this type with their associated displays. Most of the designs which have implemented total automation have adopted a version of the fourth option: that is to use fewer controls than there are functions, and to provide larger displays.

The concept of total automation is inherent in the principles of an assignable mixing console. In such a console, few of the controls carry audio directly as they are only interfaces to the control system, so one knob may control the HF EQ for any channel to which it is assigned, for example. Because of this indirect control, usually via a microprocessor, it is relatively easy to implement a means of storing the switch closures and settings in memory for reiteration at a later date.

Figure 5.27   Two possible options for positional display with continuously rotating knobs in automated systems. (a) Lights around the rim of the knob itself; (b) lights around the knob’s base

Dynamic and static systems

Many analogue assignable consoles use the modern equivalent of a VCA: the digitally controlled attenuator, to control the levels of various functions such as EQ, aux sends, and so on. If every function on the desk is to be dynamically memorised (that is, changes stored and replayed in real time, and continuously) then a lot of data is produced and must be dealt with quickly to ensure suitably seamless audible effects. Complex software and fast processors are required, along with a greater requirement for memory space.

Static systems exist which do not aim to store the continuous changes of all the functions, but they will store ‘snapshots’ of the positions of controls which can be recalled either manually or with respect to timecode. This can often be performed quite regularly (many times a second) and in these cases we approach the dynamic situation, but in others the reset may take a second or two which precludes the use of it during mixing. Changes must be silent to be useful during mixing.

Other snapshot systems merely store the settings of switch positions, without storing the variable controls, and this uses much less processing time and memory. Automated routing is of particular use in theatre work where sound effects may need to be routed to a complex combination of destinations. A static memory of the required information is employed so that a single command from the operator will reset all the routing ready for the next set of sound cues.

Digital mixers

The difficulty and expense of implementing true ‘total recall’ of an analogue mixer, that is automated resetting of all surface controls, has already been discussed. Digital mixers can incorporate such a feature routinely, and the Digidesign D5T illustrated in Figure 5.28 is a typical example of a console in which all setup parameters including such things as input gain and phantom power switching are recallable in seconds when a particular project or show is loaded into the mixer or recalled from its memory store. Such mixers are essentially versions of computer mixing systems but with a hardware control surface to provide a more traditional mode of hands-on operation, still an essential feature for live mixing work and many types of recording and broadcast session. Ergonomically, the mixer combines traditional ‘analogue’ facilities of channel faders, aux send and EQ knobs, VCA and group faders, with a considerable degree of assignability using usefully large touch sensitive screens and several selectable ‘layers’ across which banks of inputs can be displayed and accessed (Figure 5.29). A master output screen can display a variety of things such as group outputs, automation parameters, scene memory information, matrix settings and the like. A console such as this can offer 96 input channels, 20 aux sends, 24 group sends, and in a theatre version a 32-output matrix section. It is not difficult to appreciate the huge size and cost of an analogue console which offered comparable features. A QWERTY key pad facilitates the labelling of all sections. Features typical of such mixers will be briefly described.

Figure 5.28   Digico D5T. (Courtesy of the RSC)

Inputs and outputs, digital and analogue, are often provided by a series of outboard rack-mounting units which incorporate the D/A and A/D convertors, microphone amplifiers and phantom power, and these can be positioned where needed; in a recording studio one could be by the control surface itself, one or more in the recording studio area, and one by the recording machines. In a theatre, units would be placed next to power amplifiers, and in off-stage areas where musicians play. These units are connected in a daisy-chain loop to the main control surface via coax BNC cabling, MADI interface, or fibre optic links, the latter being preferable for longer distances. Patching of tie-lines into the console is thus largely avoided, and so there is a reduction in jackfield area needed.

Typically, adjacent to each bank of channel faders will be a row of buttons for the accessing of different control layers. Layer 1 could be input channels 1 to 8, the accompanying screen display showing such things as input gain, phantom power, routing, aux send levels, and EQ. Touching the appropriate area of the display expands that area for ease of viewing and adjustment, e.g.: touching the EQ section of a channel displays the settings in much more detail and assigns the EQ controls adjacent to the screen to that channel. Layer 2 could display channels 25 to 30 (channels 9 to 24 being provided by adjacent banks of faders). Layer 3 of all the fader banks could give fader control of all the matrix outputs, or all the group outputs, or all the aux master outputs, or a combination. All of these things are chosen by the operator and set up to his or her requirements, and the top layer would be assigned to inputs which normally need to be continuously on view, e.g.: musicians’ microphones and vocal mics, radio mics, and DI inputs. The lower layers would be assigned to things such as CD players, sampler and other replay machine outputs, and probably to some of the effects returns. These inputs do not normally need to be accessed quickly. Other features such as digital delay and EQ on inputs and outputs (the latter particularly useful in live sound work), compressors and limiters, and internal effects processors, are routinely available. This reduces the number of outboard effects processors needed. The settings for these are all programmable and recordable along with the rest of the console settings.

Figure 5.29   Detail of input channels display. (Digico D5T, courtesy of the RSC)

From the foregoing two main observations can be made regarding the operation of such consoles compared with their analogue counterparts. Firstly, a good deal of initial setting up, assigning and labelling needs to be carried out before a session can begin. Input and output channels need to be assigned to appropriate sockets on the outboard units around the building; the various layers have to be assigned to inputs/outputs/auxs/ VCAs as appropriate, and labelled; and a series of scene memories has to be created in anticipation of what will be required for the show or recording session. Secondly, the operation of the console often requires a two-stage thinking process. Although channel faders and some other facilities for a particular layer will be instantly available for adjustment, many other facilities will need to be accessed either on a different layer or by touching an area of a screen before adjustments can be made. Additionally, adjustments need to be stored in a scene memory. Normally, storing changes such as input gain, EQ, and aux send levels in a particular scene will automatically store those changes to the other scene memories. Channel fader adjustments will be stored only to that particular scene. Just what adjustments are stored to the present scene, and which ones are automatically stored to a bank of scenes, can be chosen by the operator. The complete project then needs to be stored to the mixer’s hard disk drive, and preferably also to an external backup. This all needs an operator who is familiar with that particular console and its software quirks. Digital consoles necessarily have many common features, but manufacturers have their own proprietary ways of doing things. The analogue console, in contrast, will be fairly familiar to a user after ten or fifteen minutes.

A digitally controlled analogue console will be looked at briefly next. The Midas Heritage 3000 shown in Figure 5.30 is a good example of such a mixer. Its control surface is analogue, and the signals remain in the analogue domain throughout. Digital control gives such things as mute and mute group automation, VCA assign, and virtual fader automation (a row of LEDs adjacent to each fader displays the audio level of the fader regardless of its physical position; moving the fader to the top lit LED gives the operator manual control). Scene memories can thus be programmed into the desk giving the appropriate fader positions and channel mutes, these being of great value in the live mixing situations for which such consoles are designed. Other consoles also provide automation of EQ in/out, insert, aux send enable, group assign, and moving fader automation, albeit at a somewhat higher cost, and such consoles undoubtedly prove their worth in the live sound market where visiting and freelance sound engineers need to become quickly familiar with a console which does not have too much automation. Such consoles are likely to be in use until well into the second decade of the present century.

Figure 5.30   The Midas Heritage 3000. (Courtesy of Klark Teknik)

Recommended further reading

See General further reading at the end of this book.