CHAPTER 13

Plug-Ins and Outboard Equipment

 

Plug-Ins

The graphic equalizer

The compressor/limiter

The expander

Echo and reverb devices

Digital reverb

Multi-effects processors

Frequency shifter

Pitch shift

Digital delay

Miscellaneous devices

Connection of Outboard Devices

 

PLUG-INS

Plug-ins are now one of the fastest-moving areas of audio development, providing audio signal processing and effects that run either on a workstation’s CPU or on dedicated DSP. (The hardware aspects of this were described in Chapter 9.) Audio data can be routed from a sequencer or other audio application, via an API (application programming interface) to another software module called a ‘plug-in’ that does something to the audio and then returns it to the source application. In this sense it is rather like inserting an effect into an audio signal path, but done in software rather than using physical patch cords and rack-mounted effects units. Plugins can be written for the host processor in a language such as C+, using the software development toolkits (SDK) provided by the relevant parties. Plug-in processing introduces a delay that depends on the amount of processing and the type of plug-in architecture used. Clearly low latency architectures are highly desirable for most applications.

Many plug-ins are versions of previously external audio devices that have been modeled in DSP, in order to bring favorite EQs or reverbs into the workstation environment. The sound quality of these depends on the quality of the software modeling that has been done. Some host-based (native) plug-ins do not have as good a quality as dedicated DSP plug-ins as they may have been ‘cut to fit’ the processing power available, but as hosts become ever more powerful the quality of native plug-ins increases.

A number of proprietary architectures have been developed for plug-ins, including Microsoft’s DirectX, Steinberg’s VST, Digidesign’s TDM, Mark of the Unicorn’s MAS, TC Works’ PowerCore and EMagic’s host-based plug-in format, subsequently bought by Apple and re-named Logic Pro, part of their Logic Studio music application programs. Apple’s OS X Audio Units are a feature built into the OS that manages plug-ins without the need for third-party middleware solutions. It is usually necessary to specify for which system any software plug-in is intended, as the architectures are not compatible. As OS-based plug-in architectures for audio become more widely used, the need for proprietary approaches may diminish. Digidesign (now Avid), for example, has had a number of different plug-in approaches that are used variously in its products, as shown in Table 13.1. The oldest of these are being phased out, with the intention that AAX will become the default version.

Chapter 9 deals with audio processing for workstations, but a brief repetition covering some aspects of the Apple system is warranted here. Apple’s Core Audio provides plug-in facilities for audio signal processing and synthesis, as well as audio-to-MIDI synchronization. Its audio plug-ins are called Audio Units (AUs). A number of standard AUs are provided with the OS X operating system, offering a range of audio processing options to other Core Audio compatible software that runs on the platform. Audio workstation packages such as Logic, for example, work closely with Core Audio to implement aspects of their functionality, including plug-ins. Core Audio normally expects to work with audio represented as 32-bit floating point linear PCM, but there are means to translate between this and other PCM formats, as well as to coded formats such as MP3, AAC or Apple Lossless Audio Coding (ALAC).

Table 13.1 Avid/Digidesign Plug-in Formats

Plug-in architecture	Description
TDM	Uses dedicated DSP cards for signal processing. Does not affect the host CPU load and processing power can be expanded as required.
HTDM (Host TDM)	Uses the host processor for TDM plug-ins, instead of dedicated DSP.
RTAS (Real Time Audio Suite)	Uses host processor for plug-ins. Not as versatile as HTDM.
AudioSuite	Non-real-time processing that uses the host CPU to perform operations such as time-stretching that require the audio file to be rewritten.
AAX	Avid Audio eXtension. Most recent Avid plug-in format for Pro Tools. Comes in DSP (for Pro Tools HDX systems) and Native (runs on host processor) versions.

DirectX is a suite of multimedia extensions developed by Microsoft for the Windows platform. It includes an element called DirectShow that deals with real-time streaming of media data, together with the insertion of so-called ‘filters’ at different points. DirectX audio plug-ins work under DirectShow and are compatible with a wide range of Windows-based audio software. They operate at 32 bit resolution, using floating-point arithmetic and can run in real time or can render audio files in non-real time. They do not require dedicated signal processing hardware, running on the host CPU, and the number of concurrent plug-ins depends on CPU power and available memory. DirectX plug-ins are also scalable — in other words they can adapt to the processing resource available. They have the advantage of being compatible with the very wide range of DirectX-compatible software in the general computing marketplace. DXi, for example, is a software synthesizer plug-in architecture developed by Cakewalk, running under DirectX. DirectX 11.1, contained in Windows 8, includes XAudio2, an audio API (Application Programming Interface) for Windows and the Xbox360. XAudio2 succeeds DirectSound on Windows, and operates through the XAudio API on the Xbox 360. It operates through DirectSound on Windows XP, and through the low-level audio mixer WASAPI (Windows Audio Session API) on Windows Vista and higher. Features of XAudio2 include spatial processing and signal processing for high-level audio APIs such as XACT, a Cross-platform Audio Creation Tool which is a part of Microsoft’s DirectX SDK.

One example of a proprietary approach used quite widely is VST, Steinberg’s Virtual Studio Technology plug-in architecture. It runs on multiple platforms and works in a similar way to DirectX plug-ins. On Windows machines it operates as a DLL (dynamic link library) resource, whereas on Macs it runs as a raw Code resource. It can also run on BeOS and SGI systems, as a Library function. VST incorporates both virtual effects and virtual instruments such as samplers and synthesizers. There is a cross-platform GUI development tool that enables the appearance of the user interface to be ported between platforms without the need to rewrite it each time.

Plug-ins cover all the usual traditional outboard functions including graphic equalization, compressor/limiting, reverb and multi-effects processing, and a variety of ‘vintage’ examples mimic old guitar amplifiers and analog processors. Examples other than these more traditional types include plug-ins which will essentially transform one instrument into another; and others which are designed to create new and unfamiliar sounds. Fact File 13.1 gives some examples. Outboard equipment, the ‘hardware’ equivalent of the plug-in, which of course preceded it historically, is still much in use in studios and particularly in the live sound environment. Sometimes a ‘half-way house’ is encountered where a mixer will incorporate electronics from another manufacturer’s effects processor with front panel controls similar to the original unit. This offers useful on-board processing from elsewhere which will normally be a familiar and respected device.

The following describes a selection of devices and their functions, and it applies equally to plug-ins and their more traditional hardware equivalents. Plug-ins by their nature provide a vastly greater range and type of audio processing than can outboard devices, but there are still a number of functions common to both.

The graphic equalizer

The graphic equalizer, pictured in Figure 13.1, consists of a row of faders (or sometimes rotary controls), each of which can cut and boost a relatively narrow band of frequencies. Simple four- or five-band devices exist which are aimed at the electronic music market, these being multiband tone controls. They perform the useful function of expanding existing simple tone controls on guitars and amplifiers, and several of the latter incorporate them.

The professional rack-mounting graphic equalizer will have at least ten frequency bands, spaced at octave or one-third-octave intervals. The ISO (International Standards Organization) center frequencies for octave bands are 31Hz, 63Hz, 125Hz, 250Hz, 500Hz, 1kHz, 2kHz, 4kHz, 8kHz and 16kHz. Each fader can cut or boost its band by typically 12dB or more. Figure 13.2 shows two possible types of filter action. The 1kHz fader is chosen, and three levels of cut and boost are illustrated. Maximum cut and boost of both types produces very similar Q (see Fact File 5.6). A high Q result is obtained by both types when maximum cut or boost is applied. The action of the first type is rather gentler when less cut or boost is applied, and the Q varies according to the degree of deviation from the fader’s central position.

 

Many graphic equalizers conform to this type of action, and it has the disadvantage that a relatively broad band of frequencies is affected when moderate degrees of boost or cut are applied. The second type maintains a tight control of frequency bandwidth throughout the cut and boost range, and such filters are termed constant Q, the Q remaining virtually the same throughout the fader’s travel. This is particularly important in the closely spaced one-third-octave graphic equalizer which has 30 separate bands, so that adjacent bands do not interact with each other too much. The ISO center frequencies for 30 bands are 25Hz, 31Hz, 40Hz, 50Hz, 63Hz, 80Hz, 100Hz, 125Hz, 180Hz, 200Hz, 250Hz, 315Hz, 400Hz, 500Hz, 630Hz, 800Hz, 1kHz, 1k25Hz, 1k8Hz, 2kHz, 2k5Hz, 3k15Hz, 4kHz, 5kHz, 6k3Hz, 8kHz, 10kHz, 12k5Hz, 16kHz and 20kHz. The value of using standard center frequencies is that complementary equipment such as spectrum analyzers which will often be used in conjunction with graphic equalizers have their scales centered on the same frequencies.

Even with tight constant Q filters, the conventional analog graphic equalizer still suffers from adjacent filter interaction. If, say, 12 dB of boost is applied to one frequency and 12 dB of cut applied to the next, the result will be more like a 6dB boost and cut, the response merging in between to produce an ill-defined Q value. Such extreme settings are, however, unlikely. The digital graphic equalizer applies cut and boost in the digital domain, and extreme settings of adjacent bands can be successfully accomplished without interaction if required.

Some graphic equalizers are single channel, some are stereo. All will have an overall level control, a bypass switch, and many also sport separate steep-cut LF filters. A useful facility is an overload indicator — usually an LED which flashes just before the signal is clipped — which indicates signal clipping anywhere along the circuit path within the unit. Large degrees of boost can sometimes provoke this. Some feature frequency cut only, these being useful as notch filters for getting rid of feedback frequencies in PA/ microphone combinations. Additional dedicated frequency-sweepable notch filters can be incorporated for this purpose. Some can be switched between cut/boost, or cut only. It is quite possible that the graphic equalizer will be asked to drive very long lines, if it is placed between mixer outputs and power amplifiers for example, and so it must be capable of doing this. The ‘+20dBu into 600 ohms’ specification should be looked for as is the case with mixers. It will be more usual though to patch the graphic equalizer into mixer output inserts, so that the mixer’s output level meters display the effect on level the graphic equalizer is having. Signal-to-noise ratio should be at least 100dB.

The graphic equalizer can be used purely as a creative tool, providing tone control to taste. It will frequently be used to provide overall frequency balance correction for PA rigs. It has formerly been used to equalize control room speakers, but poor results are frequently obtained due to the fact that a spectrum analyzer’s microphone samples the complete room frequency response whereas the perceived frequency balance is a complex combination of direct and reflected sound arriving at different times. The graphic equalizer can also change the phase response of signals, and there has been a trend away from their use in the control room for monitor EQ, adjustments being made to the control room acoustics instead.

The parametric equalizer was fully described in ‘Equalizer section’, Chapter 5.

The compressor/l imiter

The compressor/limiter (see Fact File 13.2) is used in applications such as dynamics control and as a guard against signal clipping. Such a device is pictured in Figure 13.3. The three main variable parameters are attack, release and threshold. The attack time, in microseconds (μs) and milliseconds (ms), is the time taken for a limiter to react to a signal. A very fast attack time of 10 μs can be used to avoid signal clipping, any high-level transients being rapidly brought under control. A fast release time will rapidly restore the gain so that only very short-duration peaks will be truncated. A ducking effect can be produced by using rapid attack plus a release of around 200–300ms. A threshold level is chosen which causes the limiting to come in at a moderate signal level so that peaks are pushed down before the gain is quickly reinstated. Such a ducking effect is ugly on speech, but is useful for overhead cymbal mics for example.

A slow release time of several seconds, coupled with a moderate threshold, will compress the signal dynamics into a narrower window, allowing a higher mean signal level to be produced. Such a technique is often used in vocals to obtain consistent vocal level from a singer. AM radio is compressed in this way so as to squeeze wide dynamic range material into this narrow dynamic range medium. It is also used on FM radio to a lesser extent, although very bad examples of its application are frequently heard on pop stations. An oppressive, raspy sound is the result, and in the pauses in between items or speech one hears the system gain creeping back up, causing pumping noises. Background noise rapidly ducks back down when the presenter again speaks. Many units offer separate limiting and compressing sections, their attack, release and threshold controls in each section having values appropriate to the two applications. Some also include gates (see ‘Noise gates’, Chapter 7) with variable threshold and ratio, rather like an upside-down limiter, such that below a certain level the level drops faster than would be expected. This is called ‘expansion’, described below. ‘Gain make-up’ is often available to compensate for the overall level-reducing effect of compression. Meters may indicate the amount of level reduction occurring.

One advantage of in-the-box compressor-limiting as opposed to separate outboard processors is that the latter react to incoming signals in real time, and so are unable to anticipate the state of signals before action can be taken. Processors contained within computer workstations can preview a signal, once recorded, on an ongoing basis, anticipating action to be taken beforehand, and by this means processes such as compression and limiting of signal peaks can be handled less abruptly and therefore less obtrusively.

The expander

The expander, or expander/gate as it is sometimes labeled, is the antithesis of the compressor/limiter, offering upward and downward expansion of dynamic range. Upward expansion will increase the maximum level of an existing signal, and is rarely called for, particularly as it can lead to overload difficulties. Downward expansion, reducing the level of the signal when it drops below a certain user-defined threshold, is a far more commonly used facility, and the term ‘expander’ normally refers to this process. It can be regarded as a more sophisticated form of noise gate; with the latter the gate closes when the signal reaches the user-defined muting threshold, but the expander enables the user to define the rate at which the gain is reduced, tailing the sound gently into silence. This is termed ‘slope’. Below the threshold level, a 1:2 slope ratio for instance doubles the attenuation of the signal compared with the un-processed signal. Thus, a signal level that was 3 dB below the threshold would now be reduced to 6 dB below the threshold, and signals 10dB below would now be reduced a further 10dB. The release time defines the speed at which the slope activates. The ‘hold’ control can be used to delay the onset of the expansion for a length of time before it takes action. The attack time defines the time it takes the processor to revert to unity gain once the signal rises above the threshold, and this can be used creatively to alter the sound of the transient attack of short, percussive sounds, for instance.

Echo and reverb devices

Before the advent of electronic reverb and echo processing, somewhat more basic, ‘physical’ means were used to generate the effects. The echo chamber was literally that, a fairly large reverberant room being equipped with a speaker and at least two spaced microphones. Signal was sent to the speaker and the reverb generated in the room was picked up by the two microphones which constituted the ‘stereo return’. The echo plate was a large thin resonant plate of several meters in area suspended in a frame. A driving transducer excited vibrations in the plate, and these were picked up by several transducers placed in various positions on its surface. Some quite high-quality reverb effects were possible. The spring reverb, made popular by the Hammond organ company many decades ago, consists literally of a series of coil springs about the diameter of a pencil and of varying lengths (about 10–30cm) and tensions depending on the model. A driving transducer excites torsional vibrations in one end of the springs, and these are picked up by transducers at the other end. Quite a pleasing effect can be obtained, and it is still popular for guitar amplifiers. The tape echo consisted of a short tape loop together with a record head followed by several replay heads spaced a few centimeters apart, then lastly an erase head. The output levels from the replay heads could be adjusted as could the speed of the tape, generating a variety of repeat-echo effects. Control of the erase head could generate a huge build-up of multi-echoes.

When one hears real reverb, one hears ‘pre-delay’ in effects processing parlance: a sound from the source travels to the room boundaries and then back to the listener, so there is a delay of several tens of milliseconds between hearing the direct sound and hearing the reverberation. This plays a large part in generating realistic reverb effects, and Fact File 13.3 explains the requirements in more detail.

Digital reverb

The present-day digital reverb and processor, such as that pictured in Figure 13.4, can be quite a sophisticated device. Research into path lengths, boundary and atmospheric absorption, and the physical volume and dimensions of real halls, have been taken into account when algorithms have been designed. Typical front panel controls will include selection of internal pre-programmed effects, labeled as ‘large hall’, ‘medium hall’, ‘cathedral’, ‘church’, etc., and parameters such as degree of pre-delay, decay time, frequency balance of delay, dry-to-wet ratio (how much direct untreated sound appears with the effect signal on the output), stereo width, and relative phase between the stereo outputs can often be additionally altered by the user. A small display gives information about the various parameters.

Memory stores generally contain a volatile and a non-volatile section. The non-volatile section contains factory preset effects, and although the parameters can be varied to taste the alterations cannot be stored in that memory. Settings can be stored in the volatile section, and it is usual to adjust an internal preset to taste and then transfer and store this in the volatile section. For example, a unit may contain 300 presets. The first 150 are non-volatile, and cannot be permanently altered. The last 150 can store settings arrived at by the user by transferring existing settings to this section. The method of doing this varies between models, but is usually a simple two- or three-button procedure, for example by pressing 1, 5 and 1 and then ‘store’. This means that a setting in the non-volatile memory which has been adjusted to taste will be stored in memory 151. Additional adjustments can then be made later if required.

Several units provide a lock facility so that stored effects can be made safe against accidental overwriting. An internal battery backup protects the memory contents when the unit is switched off. Various unique settings can be stored in the memories, although it is surprising how a particular model will frequently stamp its own character on the sound however it is altered. This can be either good or bad of course, and operators may have a preference for a particular system or house ‘sound’. Sometimes the bandwidth of a processor’s output reduces with increasing reverberation or echo decay times. Such a shortcoming is sometimes hard to find in a unit’s manual.

In some of the above devices it should be noted that the input is mono and the output stereo. In this way ‘stereo space’ can be added to a mono signal, there being a degree of decorrelation between the outputs. A reverb device may have stereo inputs, so that the source can be assumed to be other than a point.

Multi-effects processors

Digital multi-effects processors such as that shown in Figure 13.5 can offer a great variety of features. Parametric equalization is available, offering variations in Q, frequency and degree of cut and boost. Short memory capacity can store a sample, the unit being able to process this and reproduce it according to the incoming signal’s command. MIDI interfacing (see Chapter 14) has become popular for the selection of effects under remote control, as has the RS 232 computer interface, and a USB port or RAM card slot is sometimes encountered for loading and storing information. Some now have Firewire or USB ports for digital audio streaming, or an alternative real-time digital interface (see Chapter 10). Repeat echo, autopan, phase, modulation, flange, high and low filters, straight signal delay, pitch change, gating and added harmony may all be available in the presets, various multifunction nudge buttons being provided for overall control. Many units are only capable of offering one type of effect at a time. Several have software update options so that a basic unit can be purchased and updates later incorporated internally to provide, say, longer delay times, higher sample storage capacity, and new types of effect as funds allow and as the manufacturer develops them. This helps to keep obsolescence at bay in an area which is always rapidly developing.

Frequency shifter

The frequency shifter shifts an incoming signal by a few hertz. It is used for acoustic feedback control in sound reinforcement work, and operates as follows. Feedback is caused by sound from a speaker re-entering a microphone to be reamplified and reproduced again by the speaker, forming a positive feedback loop which builds up to a continuous loud howling noise at a particular frequency. The frequency shifter is placed in the signal path such that the frequencies reproduced by the speakers are displaced by several hertz compared with the sound entering the microphone, preventing additive effects when the sound is recycled, so the positive feedback loop is broken. The very small frequency shift has minimal effect on the perceived pitch of the primary sound.

Pitch shift

Pitch shifting and time stretching, described in Chapter 8, is normally accomplished in computer recording and editing systems and outboard processors using techniques such as altering sampling frequency, re-sampling at a different frequency, or using other more sophisticated pitch alteration algorithms. Another, more basic way of achieving pitch change has been used in effects units and ‘stomp boxes’ for the music industry to provide effects such as flange and chorus which involve small and varying degrees of pitch shift, as well as for more straightforward pitch change duties in harmonizers.

Downward pitch shift can be achieved by introducing a continuously increasing delay to the signal. A continuously decreasing delay is applied to achieve upward shift. Figure 13.6 illustrates the process for reducing pitch. A square wave has been chosen as the signal for clarity, and the horizontal time axis is calibrated in milliseconds. The top trace is the original signal, and the bottom trace shows how it has been pitched down by one octave, its total duration being increased from 8 ms to 16 ms. At A and A’, the original and processed signals begin. At B, 2 ms has elapsed. At B’, a delay of 2 ms has been applied to the signal, and 4 ms in all has therefore elapsed since the start of the lower trace. At C, 4 ms after the original signal has begun, a delay of a further 4 ms has been introduced to obtain point C’; and for point D, 6 ms along the time axis, point D’ is obtained by introducing 6 ms of delay. This describes what goes on at several specific points during a continuously increasing delay process which therefore achieves downward pitch shift, the degree of shift being a function of the rate of increase of the delay.

A problem with the process as described is that continuously increasing the delay will soon require a prohibitively large amount of digital storage as the program continues, and one way of avoiding this is for the processor to re-start the delay operation after a short period of time, a smooth transition between the two states being managed by the software. The rate of increase of delay determines the degree of pitch shift, this being independent of the length of time the process has continued. Another technique for avoiding prohibitive levels of data storage is for a second processor to begin just before the first processor is due to re-start, and a cross-fade is performed between the two.

Upward pitch shift is the exact opposite of that just described, a continuously decreasing delay of the original signal now being applied. Figure 13.6 can now be read from right to left, the bottom trace being regarded as the original signal, the top trace the upward-pitched output. It will immediately be noticed that the processed signal does not start until some time after the original signal has begun. This is because the processor first has to store up a certain amount of the incoming audio before it can begin to apply the continuously decreasing delay. An inherent latency is therefore present in the upward-pitch process, and a delay of up to about 50 ms can be accepted before the ear begins to perceive the emerging signal as having a definite time lag.

Digital delay

The digital delay line is incorporated into any substantial sound reinforcement installation as a matter of course. Consider a typical SR setup which may consist of main speakers each side of the stage (or against the proscenium arch or ‘prosc’ in a theater); front-fill speakers across the front of the stage; additional speakers covering extreme right- and left-hand sides under galleries; a line of speakers under a gallery covering the rear of the stalls; and a flown cluster covering an upper balcony, these latter speakers being rigged somewhat forward of the stage. Arrival times of the sounds from the various speakers to a particular location in the auditorium will of course vary due to the different physical path lengths, and the sound can be quite blurred as a result. Comb-filtering effects — abrupt attenuation of sound at certain frequencies as sound from one speaker is canceled by anti-phase sound coming from another due to the different path lengths — can also be encountered at some locations, and a slow walk from one side of an auditorium to the other whilst listening to pink noise will quite often make these problems apparent.

Digital delay lines are patched in between the mixer’s outputs and the power amplifiers to alleviate the problem (digital mixers often incorporate delays on their outputs for this sort of application) and are set up as follows. First, the sound of sharp clicks, rather like those of a large ticking clock — about one tick per second — is played through the main speakers each side of the stage with the other speakers switched off. Whilst standing close to the front of the stage, the front-fill speakers can then be added in. The clicks from these will reach the listener sooner than those from the main speakers, and the clicks will sound blurred, or perhaps even a double click will be heard each time. Adding a delay of perhaps 10 ms or so to the front-fills will bring the sound back into sharp focus, the exact value of delay needed depending upon the distance between the listener and the two sets of speakers. The front-fills are then switched off, and the listener moves to the side, under the balcony, and the speakers covering these areas are switched on. Again, indistinct clicks will be heard, and a delay to the side speakers of perhaps 25 ms will be required here to bring the sound into focus. For the line of speakers covering the rear stalls, a delay of perhaps 50 ms will be found to be needed. A rule of thumb when setting initial values is that sound travels just over a foot per millisecond, and so if, say, the line of rear stalls speakers is about 50 feet in front of the main stage speakers, an initial setting of 50–55 ms can be set. Many delay devices will also display the delay in terms of feet or meters, which is very useful. Moving up to the balcony, the flown cluster can now be switched on. A delay of perhaps 120 ms will be needed here to time-align the flown cluster with the main speakers each side of the stage. As well as giving a much cleaner sound, the use of delays in this manner also has the effect of making the speakers forward of the stage ‘disappear’, and the sound appears to come just from the stage itself.

As air temperature rises, sound travels faster, and delay settings obtained in a fairly cold auditorium during the day may well be a little high for the evening concert when the air is somewhat warmer. Some digital delay devices have an input for a temperature sensor, and if this is used the delay settings will automatically adjust to compensate for temperature changes.

Several computer-based systems are available which in conjunction with measuring microphones placed in the auditorium will display required delay settings for the various speaker locations when special test tones are played through the system. Additionally, using pink noise to drive the speakers the EQ curve requirements for flat frequency responses can be displayed, and parametric equalization can be used to mirror the display curves. A word of caution concerning EQ settings — air absorbs high frequencies to a rather greater extent than low frequencies, and sounds coming from a distance naturally sound duller. At a distance of 50 meters, assuming a 20°C temperature and 20% relative humidity, there is about a 10dB loss at 8kHz, and about a 35dB loss at 16kHz. Measuring microphones placed some distance from flown speakers will therefore register treble loss as a natural consequence of air absorption, and one must add treble boost with caution. An unnaturally bright sound can result, and excessive power can be fed to high-frequency horns.

The cathedral, or large conference hall, is a particularly difficult environment in which to install sound reinforcement. Speech intelligibility is of the highest priority, and a large, reverberant space requires special consideration. The usual approach is to place a number of small loudspeakers discreetly upon pillars along a nave or throughout an auditorium, each reproducing the sound of the person speaking at a low level, providing coverage only for its immediate surroundings. This helps to give a clean, clear sound without exciting much of the building’s natural reverberation. One might consider adding progressive delay to the speakers as one approaches the rear of a large venue, and this is usually timed in such a way that the direct sound and that from loudspeakers further toward the front arrive before that from the closest loudspeaker, so that the perceived direction is from the front. (See Fact File 2.4 on the precedence effect.) But the intelligibility of speech is impaired if the sound from a loudspeaker, which may be some distance from the person speaking, is out of synchronization with his lips. Using no delay at all connects the sound one hears intimately with the subject, and this is a special case in which the use of delay would be inappropriate.

Delay can also be a useful creative and corrective tool. If one is recording an electric guitar for instance, one might wish to place a microphone in front of the amplifier’s loudspeaker and also DI (direct inject) the signal straight from the guitar, or from its amplifier’s pre-amp output. The distance the microphone is from the loudspeaker will cause a slight delay with respect to the DI signal, and this produces frequency-selective reinforcements and cancellations, the ‘toothcomb’ effect. It can be corrected for by introducing a delay into the DI feed, time-aligning the two signals. Alternatively, different values of delay can be experimented with to modify the tonal quality of the instrument.

Miscellaneous devices

The de-esser cleans up closely miced vocals. Sibilant sounds can produce a rasping quality, and the de-esser dynamically filters the high-level, high-frequency component of the sound to produce a more natural vocal quality.

The Aphex company of America introduced a unit called the Aural Exciter in the 1970s, and for a time the mechanisms by which it achieved its effect were shrouded in a certain amount of mystery. The unit made a signal ‘sparkle’, enhancing its overall presence and life, and it was usually applied to individual sounds in a mix such as solo instruments and voices, but sometimes also to the complete stereo signal. Such devices succeed entirely by their subjective effect, and several companies later produced similar units. They achieve their psycho-acoustic effect by techniques such as comb filtering, selective boosting of certain frequencies, small increments of pitch change, introducing relatively narrow-band phase shifts between stereo channels, and other such processes which achieve the desired subjective effect.

Effects such as these go back a long way, and in many cases will be long obsolete and unavailable. But the value of such things as old valve (tube) compressors and limiters, distortion devices, tape and plate echoes, room simulators and other vintage-sounding devices such as certain guitar amplifiers is reflected in the market for ‘plug-ins’ for computer-based digital workstations: computer programs which have been developed to simulate the sounds of such old and well-loved devices. These things succeed purely on their own subjective terms, and are very much part of the creative process. Plug-ins do of course also provide up-to-the-minute effects and processing.

CONNECTION OF OUTBOARD DEVICES

A distinction needs to be made between processors which need to interrupt a signal for treatment (series connection), and those which basically add something to an existing signal (parallel connection). Graphic and parametric equalizers, compressors, de-essers and gates need to be placed in the signal path. One would not normally wish to mix, say, an uncompressed signal with its compressed version, or an ungated with the gated sound. Such processors will generally be patched in via the mixer’s channel insertion send and returns (see Figure 13.7), or patched in ahead of the incoming signal or immediately after an output. Devices such as echo, reverb, chorus and flange are generally used to add something to an existing signal and usually a channel aux send will be used to drive them. Their outputs will be brought back to additional input channels and these signals mixed to taste with the existing dry signal (see Figure 13.8).

Sometimes just the effects signal will be required, in which case either the aux send will be switched to pre-fade and that channel’s fader closed, or the channel will simply be de-routed from the outputs. The channel is then used merely to send the signal to the effects unit via the aux. The returns will often contain a degree of dry signal anyway, the ratio of dry to effect being adjusted on the processor.

MIDI control for selecting a program has already been mentioned. Additionally, MIDI can be used in a musical way with some devices. For instance, a ‘harmonizer’ device, designed to add harmony to a vocal or instrumental line, is normally set to add appropriate diatonic harmonies to the incoming line in the appropriate key with the desired number of voices above and/or below it. Results are thereafter in the hands of the machine. Alternatively, a MIDI keyboard can be used to control the device so that the harmonizer adds the notes which are being held down. Composition of the harmonies and voice lines is then under the control of the musician. This can be used in recording for adding harmonies to an existing line, or in a live situation where a keyboard player plays along with a soloist to generate the required harmonies.

RECOMMENDED FURTHER READING

Case, A., 2007. Sound FX: Unlocking the Creative Potential of Recording Studio Effects. Focal Press.

White, P., 2003. Basic Effects and Processors. New Amsterdam Books.

USEFUL WEBSITES

www.gersic.com/plugins: A website listing a large number and variety of plug-ins which are available, including regular updates of the latest releases.