Chapter 6

Analogue recording

A short history of analogue recording

Early recording machines

When Edison and Berliner first developed recording machines in the last years of the nineteenth century they involved little or no electrical apparatus. Certainly the recording and reproduction process itself was completely mechanical or ‘acoustic’, the system making use of a small horn terminated in a stretched, flexible diaphragm attached to a stylus which cut a groove of varying depth into the malleable tin foil on Edison’s ‘phonograph’ cylinder or of varying lateral deviation in the wax on Berliner’s ‘gramphone’ disc (see Figure 6.1). On replay, the undulations of the groove caused the stylus and diaphragm to vibrate, thus causing the air in the horn to move in sympathy, thus reproducing the sound – albeit with a very limited frequency range and very distorted.

Cylinders for the phonograph could be recorded by the user, but they were difficult to duplicate for mass production, whereas discs for the gramophone were normally replay only, but they could be duplicated readily for mass production. For this reason disks fairly quickly won the day as the mass-market prerecorded music medium. There was no such thing as magnetic recording tape at the time, so recordings were made directly on to a master disc, lasting for the duration of the side of the disc – a maximum of around 4 minutes – with no possibility for editing. Recordings containing errors were either remade or they were passed with mistakes intact. A long item of music would be recorded in short sections with gaps to change the disc, and possibilities arose for discontinuities between the sections as well as variations in pitch and tempo. Owing to the deficiencies of the acoustic recording process, instruments had to be grouped quite tightly around the pick-up horn in order for them to be heard on the recording, and often louder instruments were substituted for quieter ones (the double bass was replaced by the tuba, for example) in order to correct for the poor frequency balance. It is perhaps partly because of this that much of the recorded music of the time consisted of vocal soloists and small ensembles, since these were easier to record than large orchestras.

Figure 6.1   The earliest phonograph used a rotating foil-covered cylinder and a stylus attached to a flexible diaphragm. The recordist spoke or sang into the horn causing the stylus to vibrate, thus inscribing a modulated groove into the surface of the soft foil. On replay the modulated groove would cause the stylus and diaphragm to vibrate, resulting in a sound wave being emitted from the horn

Electrical recording

During the 1920s, when broadcasting was in its infancy, electrical recording became more widely used, based on the principles of electromagnetic transduction (see Chapter 3). The possibility for a microphone to be connected remotely to a recording machine meant that microphones could be positioned in more suitable places, connected by wires to a complementary transducer at the other end of the wire, which drove the stylus to cut the disc. Even more usefully, the outputs of microphones could be mixed together before being fed to the disc cutter, allowing greater flexibility in the balance. Basic variable resistors could be inserted into the signal chain in order to control the levels from each microphone, and valve amplifiers would be used to increase the electrical level so that it would be suitable to drive the cutting stylus.

The sound quality of electrical recordings shows a marked improvement over acoustic recordings, with a wider frequency range and a greater dynamic range. Experimental work took place both in Europe and the USA on stereo recording and reproduction, but it was not to be until much later that stereo took its place as a common consumer format, nearly all records and broadcasts being in mono at that time.

Later developments

During the 1930s work progressed on the development of magnetic recording equipment, and examples of experimental wire recorders and tape recorders began to appear, based on the principle of using a current flowing through a coil to create a magnetic field which would in turn magnetise a moving metal wire or tape coated with magnetic material. The 1940s, during wartime, saw the introduction of the first AC-biased tape recorders, which brought with them good sound quality and the possibility for editing. Tape itself, though, was first made of paper coated with metal oxide which tended to deteriorate rather quickly, and only later of plastics which proved longer lasting and easier to handle. In the 1950s the microgroove LP record appeared, with markedly lower surface noise and improved frequency response, having a playing time of around 25 minutes per side. This was an ideal medium for distribution of commercial stereo recordings, which began to appear in the late 1950s, although it was not until the 1960s that stereo really took hold. In the early 1960s the first multitrack tape recorders appeared, the Beatles making use of an early four-track recorder for their ‘Sergeant Pepper’s Lonely Hearts Club Band’ album. The machine offered the unprecedented flexibility of allowing sources to be recorded separately, and the results in the stereo mix are panned very crudely to left and right in somewhat ‘gimmicky’ stereo. Mixing equipment in the 1950s and 1960s was often quite basic, compared with today’s sophisticated consoles, and rotary faders were the norm. There simply was not the quantity of tracks involved as exists today.

Magnetic tape

Structure

Magnetic tape consists of a length of plastic material which is given a surface coating capable of retaining magnetic flux rather in the manner that, say, an iron rod is capable of being magnetised (see Figure 6.2). The earliest recorders actually used a length of iron wire as the recording medium. In practice all modern tape has a polyester base which was chosen, after various trials with other formulations which proved either too brittle (they snapped easily) or too plastic (they stretched), for its good strength and dimensional stability. It is used throughout the tape industry from the dictation microcassette to the 2 inch (5 cm) multitrack variety. The coating is of a metal oxide, or metal alloy particles.

The most common coating used is of gamma-ferric oxide, a kind of purified rust with specially shaped particles (signified by ‘gamma’). This formulation is used in cassettes along with the alternative formulations of chromium dioxide and its substitutes, and metal particles. It is also almost exclusively used for open-reel analogue master tapes of all widths – quarter inch, half inch, 1 inch, 2 inch – although a particular brand and formulation may not always be available in all widths.

Figure 6.2   Cross-section through a magnetic recording tape

Recent history

The BASF company of Germany introduced a chromium dioxide formulation in the early 1970s, claiming better high-frequency performance and improved signal-to-noise ratios. The patent on this formulation spurred other manufacturers to experiment with alternatives, and a substitute which consisted of cobalt-enriched ferric oxide emerged. Then so-called ‘ferri-chrome’ tape was introduced which consisted of a layer of ferric oxide over which was applied a layer of chromium dioxide or a substitute. High frequencies tend to be recorded close to the surface of an oxide layer, lower frequencies being recorded more deeply. The dual layer was therefore intended to exploit this, taking advantage of the top layer’s good performance at high frequencies together with the lower layer’s strengths in the areas of distortion and good lower-frequency output capability. This type of formulation is infrequently used today, partially because it tended to display appalling middle-frequency distortion characteristics, and also because of improvements in standard ferric and chrome formulations.

During the 1980s pure metal tape was developed for the cassette medium. Initially this posed great difficulties for the manufacturers due to the fact that very tiny particles of pure iron or iron alloy of the size required for recording tape tended to oxidise rapidly – they caught fire! This problem was, however, overcome and metal cassettes were launched which were rather more expensive than the oxide types and additionally required specially equipped cassette machines with record heads and circuitry capable of effectively magnetising the high-coercivity metal tape. Many cheaper cassette machines, although intended to cope with this tape and labelled as such, could not fully exploit it. The coercivity of the tape – its willingness to accept and retain magnetic flux – was such that cheap tape heads often saturated magnetically before the tape did, due to the extremely high bias currents required to magnetise it (see bias requirements, below). Improvements in distortion, output level and signal-to-noise ratios were obtainable with metal tape from good cassette machines, at a cost.

The metal formulation came into its own in a modified form when its high recording density capability proved ideal for digital tape recording, in special cassette form for R-DAT (see Chapter 9) and also for hand-held domestic camcorder machines.

Cassette tape

Cassette tape comes in various lengths to give appropriate playing times for given applications – ‘C5’ gives 2.5 minutes playing time per side, C90 gives 45 minutes per side and so on. The longer-playing tapes are thinner in order that more tape can be loaded on to the spools. The C120 cassette has very thin tape which causes problems with many machines since their transports are unable to handle the delicate and very flimsy tape adequately. The tape will tend to stick on to the rubber pinch roller as it passes between the roller and the capstan, causing the transport to chew the tape up in bad cases, wrapping it around the capstan and pinch roller. Excessively thin tape does not sit comfortably in the tape guides in the cassette housing, and it is also not pulled across the tape heads very evenly, causing poor head-to-tape contact with loss of output and frequency response. It would be difficult to find a cassette machine manufacturer who actually sanctioned the use of C120 cassettes.

Open-reel tape

Open-reel quarter-inch tape intended for analogue recorders has been available in a variety of thicknesses. Standard Play tape has an overall thickness of 50 microns (micrometres), and a playing time (at 15 inches (38 cm) per second) of 33 minutes is obtained from a 10 inch (25 cm) reel. Long Play tape has an overall thickness of 35 microns giving a corresponding 48 minutes of playing time, which is very useful for live recording work. In the past ‘Double Play’ and even ‘Triple Play’ thicknesses have been available, these being aimed at the domestic open-reel market. These formulations are prone to snapping or stretching, as well as offering slightly poorer sound quality, and should not really be considered for professional use.

Standard Play tape is almost always ‘back coated’. A rough coating is applied to the back of the tape during manufacture which produces neater and more even winding on a tape machine, by providing a certain amount of friction between layers which holds the tape in place. Also, the rough surface helps prevent air being trapped between layers during fast spooling which can contribute to uneven winding. Long Play tape is also available with a back coating, but as often as not it will be absent. It is worth noting that the flanges of a tape spool should only serve to protect the tape from damage. The ‘pancake’ of tape on the spool should not touch these flanges. Metal spools are better than plastic spools because they are more rigid and they do not warp. Professional open-reel tape can be purchased either on spools or in ‘pancake’ form on hubs without flanges. The latter is of course cheaper, but considerable care is needed in its handling so that spillage of the unprotected tape does not occur. Such pancakes are either spooled on to empty reels before use, or they can be placed on top of a special reel with only a lower flange. Professional tape machines are invariably operated with their decks horizontal. Half inch, 1 inch and 2 inch tape intended for multitrack recorders always comes on spools, is always of Standard Play thickness, and is always back coated.

Open-reel tape should have a batch number printed on the box. This number ensures that a given delivery of tape was all manufactured at the same time and so will have virtually identical magnetic properties throughout the batch. Different batches can have slightly different properties and a studio may wish to realign its machines when a new batch is started. It can be said, however, that variations from batch to batch are minimal these days.

The magnetic recording process

Introduction

Since tape is magnetic, the recording process must convert an electrical audio signal into a magnetic form. On replay the recorded magnetic signal must be converted back into electrical form. The process is outlined in Fact File 6.1. Normally a professional tape recorder has three heads, as shown in Figure 6.3, in the order erase–record–replay. This allows for the tape to be first erased, then re-recorded, and then monitored by the third head. The structure of the three heads is similar, but the gap of the replay head is normally smaller than that of the record head. It is possible to use the same head for both purposes, but usually with a compromise in performance. Such a two-head arrangement is often found in cheaper cassette machines which do not allow off-tape monitoring whilst recording. A simplified block diagram of a typical tape recorder is shown in Figure 6.4.

Figure 6.3   Order of heads on a professional analogue tape recorder

The magnetisation characteristics of tape are by no means linear, and therefore a high-frequency signal known as bias is added to the audio signal at the record head, generally a sine wave of between 100 and 200 kHz, which biases the tape towards a more linear part of its operating range. Without bias the tape retains very little magnetisation and distortion is excessive. The bias signal is of too high a frequency to be retained by the tape, so does not appear on the output during replay. Different types of tape require different levels of bias for optimum recording conditions to be achieved, and this will be discussed in bias requirements, below.

Equalisation

‘Pre-equalisation’ is applied to the audio signal before recording. This equalisation is set in such a way that the replayed short-circuit flux in an ideal head follows a standard frequency response curve (see Figure 6.5). A number of standards exist for different tape speeds, whose time constants are the same as those quoted for replay EQ in Table 6.1. Although the replayed flux level must conform to these curves, the electrical pre-EQ may be very different, since this depends on the individual head and tape characteristics. Replay equalisation (see Figure 6.6) is used to ensure that a flat response is available at the tape machine’s output. It compensates for losses incurred in the magnetic recording/replay process, the rising output of the replay head with frequency, the recorded flux characteristic, and the fall-off in HF response where the recorded wavelength approaches the head gap width (see Fact File 6.2). Table 6.1 shows the time constants corresponding to the turnover frequencies of replay equalisers at a number of tape speeds. Again a number of standards exist. Time constant (normally quoted in microseconds) is the product of resistance and capacitance (RC) in the equivalent equalising filter, and the turnover frequency corresponding to a particular time constant can be calculated using:

Figure 6.4   Simplified block diagram of a typical analogue tape recorder. The bias trap is a filter which prevents the HF bias signal feeding back into an earlier stage

Figure 6.5   Examples of standardised recording characteristics for short-circuit flux. (N.B: this is not equivalent to the electrical equalisation required in the record chain, but represents the resulting flux level replayed from tape, measured using an ideal head)

Table 6.1   Replay equalisation time constants

Figure 6.6   Examples of replay equalisation required to correct for the recording characteristic (see Figure 6.5), replay-head losses, and the rising output of the replay head with frequency

f = 1/(2πRC)

The LF time constant of 3180 μs was introduced in the American NAB standard to reduce hum in early tape recorders, and has remained. HF time constants resulting in low turnover frequencies tend to result in greater replay noise, since HF is boosted over a wider band on replay, thus amplifying tape noise considerably. This is mainly why Type I cassette tapes (120 μs EQ) sound noisier than Type II tapes (70 μs EQ). Most professional tape recorders have switchable EQ to allow the replay of NAB- and IEC/CCIR-recorded tapes. EQ switches automatically with tape speed in most machines.

Additional adjustable HF and LF EQ is provided on many tape machines, so that the recorder’s frequency response may be optimised for a variety of operational conditions, bias levels and tape types.

Bias requirements

Higher bias levels enable the audio signal to be recorded more deeply into the oxide layer than would be the case with lower bias levels. High-coercivity tapes require a higher bias than low-coercivity tapes in order to magnetise the tape suitably. Such tapes can, however, display slightly poorer print-through performance (see below). The correct setting of bias level is vital for obtaining optimum performance from an analogue tape, and in professional systems the setting of bias is part of the day-to-day alignment of the tape machine (see ‘Tape machine alignment’, below).

Bias requirements vary from tape to tape as has been said, and cassette tapes are grouped into four ‘slots’: Type I (ferric), Type II (‘chrome’ or CrO₂), Type III (ferri-chrome), and Type IV (metal). Cassette formulations are expected to conform closely to one of the above groups with respect to required bias level so that good results are achieved when various different brands are used with a given machine. This is important because variations in bias level cause significant variations in high-frequency performance. The domestic user cannot be expected to optimise his or her cassette machine for a particular type of tape as is routine in professional open-reel usage, although some cassette machines do offer fine bias adjustment, either manual or automatic. It is usually found that a given cassette machine will perform better with one brand of tape than with another depending on exactly how the manufacturer or distributor has aligned the machine, and it is worth experimenting with different types in order to find the one which gives the best performance on that machine.

Print-through

Print-through is caused by a modulated area of tape inducing its magnetism into the adjacent layer of tape on the spool during storage, rather in the manner that a pin, when left stuck to a magnet, will itself become magnetised. This manifests itself as pre-echo or post-echo, depending on which way the tape is wound off the machine, meaning that one can hear the beginning of a movement or programme after a pause at a very low level a second or so before it actually starts. At the end of the programme, faint traces of the last second or so can be heard repeated, although this is usually effectively masked by decaying reverberation or, if no significant reverberation is present, some leader tape will be spliced on immediately after the programme material has ceased.

The storing of master tapes ‘tail out’ – standard industry practice, indicated by red leader tape which should always be present on the end of every master tape – helps to avoid pre-echo because now a silent section will lie adjacent to the previous programme material rather than adjacent to the beginning of the following section. Tail-out storage of master tapes is also desirable because tape machines invariably give neater spooling on replay than on fast wind or rewind, and so a tape which has just been played and is now tail out will be stored in this neat state which helps to maintain the tape in good condition. Post-echo is preferable because it is at least more natural to hear a faint echo of what one has just heard than to hear a preview of a following section. Dying reverberation also helps to mask it as has been said.

Noise reduction (see Chapter 7) helps to reduce the consequences of print-through because on replay the decoding process pushes low-level signals further down in level as part of the expansion process. Since print-through signals are introduced in between noise reduction encoding and decoding they are therefore reduced in level.

The tape recorder

Studio recorder

Professional open-reel recorders fall into two categories: console mounted and portable. The stereo console recorder, intended for permanent or semi-permanent installation in a recording studio, outside broadcast truck or whatever generally sports rather few facilities, but has balanced inputs and outputs at line level (no microphone inputs), transport controls, editing modes, possibly a headphone socket, a tape counter (often in real time rather than in arbitrary numbers or revs), tape speed selector, reel size selector, and probably (though not always) a pair of level meters. It is deliberately simple because its job is to accept a signal, store it as faithfully as possible, and then reproduce it on call. It is also robustly built, stays aligned for long periods without the need for frequent adjustment, and will be expected to perform reliably for long periods. A typical example is pictured in Figure 6.7.

The inputs of such a machine will be capable of accepting high electrical levels – up to at least +20 dBu or around 8 volts – so that there is virtually no possibility of electrical input overload. The input impedance will be at least 10 kΩ. The outputs will be capable of driving impedances down to 600 ohms, and will have a source impedance of below 100 ohms. A facility will be provided for connecting a remote control unit so that the transport can be controlled from the mixing console, for instance. Often the real-time tape counter can also be remotely displayed so that the machine itself can virtually be ignored during a recording session. A noise reduction system can often be located within the housing of the machine, and the record button can send a DC control voltage to this which will automatically switch the noise reduction to encode during recording. Input and output level, bias and EQ controls will be provided, but tucked away so that they cannot be accidently misaligned. Often, small screwdriver holes are employed for these.

Figure 6.7   A typical professional open-reel two-track analogue tape recorder: the Studer A807-TC. (Courtesy of FWO Bauch Ltd)

Depending on the machine, record ‘safe’ and ‘ready’ switches are normally provided for each track, as well as the ability to switch between sync replay and normal replay if it is a multitrack machine (see below). Varispeed controls often exist too, being a means of finely adjusting the speed of the machine either side of a standard play speed. There may also be the option to lock the machine to an external speed reference for synchronisation purposes.

Semi-professional recorder

Its semi-professional counterpart will be capable at its best of a performance that is little inferior, and in addition to being smaller and lighter will sport rather more facilities such as microphone inputs and various alternative input and output options. Headphone outlets will be provided along with record-level meters, source/tape monitor switching, variable output level, and perhaps ‘sound on sound’-type facilities for simple overdub work. A typical example is shown in Figure 6.8. The semi-professional machine will not usually be as robustly constructed, this being of particular concern for machines which are to be transported since rough treatment can easily send a chassis askew, causing misalignment of the tape transport system which will be virtually impossible to correct. Some chassis are constructed of pressed steel which is not very rigid. A casting is much better.

Figure 6.8   A typical semi-professional two-track recorder: the Revox PR99. (Courtesy of Revox UK Ltd)

Input and output connectors on a semi-pro machine will generally be of the domestic type – phono and unbalanced jack – and voltage levels may well be referenced around −10 dBV instead of +4 dBu. Tape guides may well not include the rollers and low-friction ball races of the professional machine. Facilities for interfacing with noise reduction systems will probably be absent. Such a machine is considerably cheaper than its fully professional counterpart, but the best examples of such machines do, however, return excellent performance consistently and reliably.

The portable machine

The professional portable tape machine, unlike its console equivalent, needs to offer a wide range of facilities since it will be required to provide such things as balanced outputs and inputs, both at line and microphone level, phantom and A−B mic powering, metering, battery operation which allows usefully long recording times, the facility to record timecode and pilot tone for use in TV and film work, illumination of the important controls and meters, and possibly even basic mixing facilities. It must be robust to stand up to professional field use, and small enough to be carried easily. Nevertheless, it should also be capable of accepting professional 10 inch (25 cm) reels, and adaptors are usually available to facilitate this. A lot has to be provided in a small package, and the miniaturisation necessary does not come cheap. The audio performance of such machines is at least as good as that of a studio recorder. A typical commercial example is pictured in Figure 6.9.

Figure 6.9   A typical professional portable two-track recorder: the Nagra IV-S. (Courtesy of Nagra Kudelski (GB) Ltd)

The multitrack machine

Multitrack machines come in a variety of track configurations and quality levels. The professional multitrack machine tends to be quite massively engineered and is designed to give consistent, reliable performance on a par with the stereo mastering machine. The transport needs to be particularly fine so that consistent performance across the tracks is achieved. A full reel of 2 inch tape is quite heavy, and powerful spooling motors and brakes are required to keep it under control. Apart from the increased number of tracks, multitrack machines are basically the same as their stereo counterparts and manufacturers tend to offer a range of track configurations within a given model type. Alignment of course takes a lot longer, and computer control of this is most welcome when one considers that 24 tracks implies 168 separate adjustments! A typical 24 track tape machine is pictured in Figure 6.10.

A useful feature to have on a multitrack recorder is an automatic repeat function or autolocate. The real-time counter can be programmed so that the machine will repeat a section of the tape over and over again within the specified start and end points to facilitate mixdown rehearsals. Multitrack recorders will be equipped with a number of unique features which are vital during recording sessions. For example, sync replay (see Fact File 6.3), gapless, noiseless punch-in (allowing any track to be dropped into record at any point without introducing a gap or a click) and spot erasure (allowing a track to be erased manually over a very small portion of tape).

Figure 6.10   A typical professional multitrack recorder: the Saturn 824. (Courtesy of Saturn Research)

Various semi-professional multitrack machines have appeared over the years, making use of decreased track widths in the interests of cheaper transports and lower tape running costs at the expense of quality. Quarter-inch tape has been used for four, eight and even sixteen tracks; 1 inch tape has also been used for eight and sixteen tracks. Some models incorporate domestic-type noise reduction systems such as Dolby C, and a recent machine makes use of the Dolby S system, derived from professional SR (see ‘Dolby SR’ Chapter 7).

Cheap multitrack machines will have domestically orientated input and output sockets, and will be unbalanced. Comprehensive alignment facilities may not be provided. Phase errors between the tracks can be high, and the transport will not be first class. Crosstalk between tracks can be poor. The machine may not last as long as its professional counterpart, nor may it stay in alignment for as long.

Track formats

Mono, two-track and stereo formats

The professional stereo format is known as half track, because each track occupies approximately half of the tape width. Full-track mono recorders record the signal across the whole of the tape width (see Figure 6.11). Domestic open-reel two-track machines conformed to the quarter-track format, in which left and right channels were recorded in the first (upper) and third quarters of the tape’s width. The tape could then be turned over so that fresh material could be recorded on the other two quarters of the tape. Twice the recording time could therefore be accommodated. The ‘other side’ of the tape, as it is conveniently called, is actually the same side but uses a different part of the tape’s area. The disadvantages are that reduced track widths mean higher distortion and poorer signal-to-noise ratios; greater possibility of ‘drop-out’ (momentary loss of signal) due to head-to-tape contact being more critical; and the fact that the editing of a programme will also chop up any material recorded on the other side. Some semi-professional machines are quarter track, and one occasionally finds such a machine in a studio to cater for quarter-track tapes which may sometimes be brought in. Alternatively, some half-track machines have a second replay head in quarter-track form.

Figure 6.11   Track patterns for quarter-inch recording. (a) Mono, full-track. (b) Stereo, half-track. (c) Four track, or stereo quarter-track (tracks 1 and 3 in one direction and tracks 2 and 4 in the other)

Two different, professional two-track formats exist, NAB and DIN, as described in Fact File 6.4.

Multitrack formats

The professional standard for tape width adheres as far as possible to the scaling derived from the quarter-inch two-track mastering machine. A four-track machine therefore uses half-inch-width tape. Next comes the eight-track which makes use of 1 inch tape. Sixteen track utilises 2 inch tape, and 24 tracks are also accommodated across 2 inches. Comparable quality levels of all the tracks across all the formats are therefore achieved, with no audio degradation as one moves, say, from four track to eight track.

Tracks are normally numbered from 1 at the top of the tape to the highest number at the bottom.

Magnetic recording levels

It has already been said that the equivalent of electrical current in magnetic terms is magnetic flux, and it is necessary to understand the relationship between electrical levels and magnetic recording levels on tape (a little was said about this in Fact File 5.9). The performance of an analogue tape recorder depends very much on the magnetic level recorded on the tape, since at high levels one encounters distortion and saturation, whilst at low levels there is noise (see Figure 6.12). A window exists, between the noise and the distortion, in which the audio signal must be recorded, and the recording level must be controlled to lie optimally within this region. For this reason the relationship between the electrical input level to the tape machine and the flux level on tape must be established so that the engineer knows what meter indication on a mixer corresponds to what magnetic flux level. Once a relationship has been set up it is possible largely to forget about magnetic flux levels and concentrate on the meters. Fact File 6.5 discusses magnetic flux reference levels.

Figure 6.12   The available dynamic range on an analogue tape lies between the noise floor and the MOL. Precise figures depend on tape and tape machine

What are test tapes for?

A test tape is a reference standard recording containing pre-recorded tones at a guaranteed magnetic flux level. A test tape is the only starting point for aligning a tape machine, since otherwise there is no way of knowing what magnetic level will end up on the tape during recording. During alignment, the test tape is replayed, and a 1 kHz tone at the specified magnetic flux level (say 320 nWbm⁻¹) produces a certain electrical level at the machine’s output. The output level would then be adjusted for the desired electrical level, according to the studio’s standard (say 0 dBu), to read at a standard meter indication (say PPM 4). It is then absolutely clear that if the output level of the tape machine is 0 dBu then the magnetic level on tape is 320 nWbm⁻¹. After this relationship has been set up it is then possible to record a signal on tape at a known magnetic level – for example, a 1 kHz tone at 0 dBu could be fed to the input of the tape machine, and the input level adjusted until the output read 0 dBu also. The 1 kHz tone would then be recording at a flux level of 320 nWbm⁻¹.

Test tapes also contain tones at other frequencies for such purposes as azimuth alignment of heads and for frequency response calibration of replay EQ (see below). A test tape with the required magnetic reference level should be used, and it should also conform to the correct EQ standard (NAB or CCIR, see ‘Equalisation’, above). Tapes are available at all speeds, standards and widths, with most being recorded across the full width of the tape.

Tape machine alignment

Head inspection and demagnetisation

Heads and tape guides must be periodically inspected for wear. Flats on guides and head surfaces should be looked for; sometimes it is possible to rotate a guide so that a fresh portion contacts the tape. Badly worn guides and heads cause sharp angles to contact the tape which can damage the oxide layer. Heads have been made of several materials. Mu-metal heads have good electromagnetic properties, but are not particularly hard wearing. Ferrite heads wear extremely slowly and their gaps can be machined to tight tolerances. The gap edges can, however, be rather brittle and require careful handling. Permalloy heads last a long time and give a good overall performance, and are often chosen. Head wear is revealed by the presence of a flat area on the surface which contacts the tape. Slight wear does not necessarily indicate that head replacement is required, and if performance is found to be satisfactory during alignment with a test tape then no action need be taken.

Replay-head wear is often signified by exceptionally good high-frequency response, requiring replay EQ to be reduced to the lower limit of its range. This seems odd but is because the replay gap on many designs gets slightly narrower as the head wears down, and is at its narrowest just before it collapses!

Heads should be cleaned regularly using either isopropyl alcohol and a cotton bud, or a freon spray. They should also be demagnetised fairly regularly, since heads can gradually become slightly permanently magnetised, especially on older machines, resulting in increased noise and a type of ‘bubbling’ modulation noise in the background on recordings. A demagnetiser is a strong AC electromagnet which should be switched on well away from the tape machine, keeping it clear of anything else magnetic or metal. This device will erase a tape if placed near one! Once turned on the demagger should be drawn smoothly and slowly along the tape path (without a tape present), across the guides and heads, and drawn away gently on the far side. Only then should it be turned off.

Replay alignment

Replay alignment should be carried out before record alignment, as explained above. The method for setting replay and record levels has already been covered in the previous section. HF tones for azimuth adjustment normally follow (see Fact File 6.6). The test tape will contain a sequence of tones for replay frequency response alignment, often at 10 or 20 dB below reference level so that tape saturation is avoided at frequency extremes, starting with a 1 kHz reference followed by, say, 31.5 Hz, 63 Hz, 125 Hz, 250 Hz, 500 Hz, 2 kHz, 4 kHz, 8 kHz and 16 kHz. Spoken identification of each section is provided. As the tape runs, the replay equalisation is adjusted so as to achieve the flattest frequency response. Often both LF and HF replay adjustment is provided, sometimes just HF, but normally one should only adjust HF response on replay, since LF can suffer from the head bumps described in Fact File 6.2 and a peak or dip of response may coincide with a frequency on the test tape, leading to potential misalignment. Also full-track test tapes can cause ‘fringing’ at LF, whereby flux from the guard band leaks on to adjacent tracks. (Although it seems strange, replay LF EQ is normally adjusted during recording, to obtain the flattest record–replay response.)

Record alignment

The frequency response of the machine during recording is considerably affected by bias adjustment, and therefore bias is aligned first before record equalisation. The effects and alignment of bias are described in Fact File 6.7. It is wise to set a roughly correct input level before adjusting bias, by sending a 1 kHz tone at reference level to the tape machine and adjusting the input gain until it replays at the same level.

After bias levels have been set, record azimuth can be adjusted if necessary (see Fact File 6.6) by recording an HF tone and monitoring the now correctly aligned replay output. It may also be necessary to go back and check the 1 kHz record level if large changes have been made to bias.

Record equalisation can now be aligned. Normally only HF EQ is available on record. A 1 kHz tone is recorded at between 10 and 20 dB below reference level and the meter gain adjusted so that this can be seen easily on replay. Spot frequencies are then recorded to check the machine’s frequency response, normally only at the extremes of the range. A 5 kHz tone, followed by tones at 10 kHz and 15 kHz can be recorded and monitored off tape. The HF EQ is adjusted for the flattest possible response. The LF replay EQ (see above) can similarly be adjusted, sweeping the oscillator over a range of frequencies from, say, 40 Hz to 150 Hz, and adjusting for the best compromise between the upper and lower limits of the ‘head bumps’.

Some machines have a built-in computer which will automatically align it to any tape. The tape is loaded and the command given, and the machine itself runs the tape adjusting bias, level and EQ as it goes. This takes literally seconds. Several settings can be stored in its memory so that a change of tape type can be accompanied simply by telling the machine which type is to be used, and it will automatically set its bias and EQ to the previously stored values. This is of particular value when aligning multitrack machines!

Once the tape machine has been correctly aligned for record and replay, a series of tones should be recorded at the beginning of every tape made on the machine. This allows the replay response of any machine which might subsequently be used for replaying the tape to be adjusted so as to replay the tape with a flat frequency response. The minimum requirement should be a tone at 1 kHz at reference level, followed by tones at HF and LF (say 10 kHz and 63 Hz) at either reference level (if the tape can cope) or at −10 dB. The levels and frequencies of these tones must be marked on the tape box (e.g.: ‘Tones @ 1 kHz, 320 nWbm⁻¹ (= 0 dB); 10 kHz and 63 Hz @ −10 dB). Designations on the tape box such as ‘1 kHz @ 0 VU’ mean almost nothing, since 0 VU is not a magnetic level. What the engineer means in this case is that he/she sent a tone from his/her desk to the tape machine, measuring 0 VU on the meters, but this gives no indication of the magnetic level that resulted on the tape. Noted on the box should also be an indication of where peak recording level lies in relation to the 1 kHz reference level (e.g.: ‘peak recording level @ 8 dB above 320 nWbm⁻¹), in order that the replay chain can be set up to accommodate the likely signal peaks. In broadcasting, for example, it is most important to know where the peak signal level will be, since this must be set to peak at PPM 6 on a program meter, corresponding to maximum transmitter modulation.

When this tape comes to be replayed, the engineer will adjust the replay level and EQ controls of the relevant machine, along with replay azimuth, to ensure that the recorded magnetic reference level replays at his or her studio’s electrical reference level, and to ensure a flat response. This is the only way of ensuring that a tape made on one machine replays correctly on another day or on another machine.

Mechanical transport functions

Properly, mechanical alignment of the tape transport should be looked at before electrical alignment, because the electromagnetic performance is affected by it, but the converse is not the case. Mechanical alignment should be required far less frequently than electrical adjustments, and sometimes it also requires rather specialised tools. Because most mechanical alignments are fairly specialised, and because they differ with each tape machine, detailed techniques will not be covered further here. The manual for a machine normally details the necessary procedures. Looking at the diagram in Figure 6.13, it can be seen that the tape unwinds from the reel on the left, passes through various guides on its way to the head block, and then through various further guides and on to the take-up reel on the right. Some tape guides may be loaded with floppy springs which give on the instant of start-up, then slowly swing back in order to control the tension of the tape as the machine starts. The capstan is the shaft of a motor which pokes up through the deck of the machine by a couple of centimetres or so (more of course for multitrack machines with their increased tape widths) and lies fairly close to the tape when the tape is at rest, on the right-hand side of the head block. A large rubber wheel will be located close to the capstan but on the opposite side of the tape. This is called the pinch roller or pinch wheel. The capstan motor rotates at a constant and carefully controlled speed, and its speed of rotation defines the speed at which the tape runs. When record or play is selected the pinch roller rapidly moves towards the capstan, firmly sandwiching the tape in between the two. The rotation of the capstan now controls the speed of tape travel across the heads.

Figure 6.13   Typical layout of mechanical components on the deckplate of an analogue open-reel recorder

The take-up reel is controlled by a motor which applies a low anticlockwise torque so that the tape is wound on to it. The supply reel on the left is also controlled by a motor, which now applies a low clockwise torque, attempting to drag the tape back in the opposite direction, and this ‘back tension’ keeps the tape in firm contact with the heads. Different reel sizes require different degrees of back tension for optimum spooling, and a reel size switch will usually be provided although this is sometimes automatic. One or two transports have been designed without pinch rollers, an enlarged diameter capstan on its own providing speed control. The reel motors need to be rather more finely controlled during record and replay so as to avoid tape slippage across the capstan. Even capstanless transports have appeared, the tape speed being governed entirely by the reel motors.

When fast wind or rewind is selected the tape is lifted away from the heads by tape lifters, whilst spooling motors apply an appropriately high torque to the reel which is to take up the tape and a low reverse torque to the supply reel to control back tension. The tape is kept away from the heads so that its rapid movement does not cause excessive heating and wear of the tape heads. Also, very high-level, high-frequency energy is induced into the playback head if the tape is in contact with it which can easily damage speakers, particularly tweeters and HF horns. Nevertheless, a facility for moving the tape into contact with the heads during fast spooling is provided so that a particular point in the tape can be listened for.

Motion sensing and logic control is an important feature of a modern open-reel machine. Because the transport controls are electronically governed on modern machines, one can go straight from, say, rewind to play, leaving the machine itself to store the command and bring the tape safely to a halt before allowing the pinch wheel to approach the capstan. Motion sensing can be implemented by a number of means, often either by sensing the speed of the reel motors using tachometers, or by counting pulses from a roller guide.

The tape counter is usually driven by a rotating roller between the head block and that reel. Slight slippage can be expected, this being cumulative over a complete reel of tape, but remarkably accurate real-time counters are nevertheless to be found.

The Compact Cassette

Background

The Compact Cassette was invented by Philips, and was launched in 1963. It was originally intended as a convenient low-quality format suitable for office dictation machines and the like. It was envisaged that domestic tape recording would be open-reel, and a boom in this area was predicted. Pre-recorded open-reel tapes were launched. The expected boom never really materialised, however, and the sheer convenience of the cassette medium meant that it began to make inroads into the domestic environment. The format consists of tape one-eighth of an inch wide (3 mm), quarter track, running at a speed of 1.875 ips. Such drastically reduced dimensions and running speed compared with open-reel returned a poor level of audio performance, and if it was to be used for reasonable-quality music reproduction considerable development was needed.

Leaving the standards which had been set for the Compact Cassette unaltered, tape and machine manufacturers worked hard to develop the format, and the level of performance now available from this medium is quite impressive given its humble beginnings. It is worth mentioning that in the mid 1970s Sony introduced a rival called the Elcaset. The cassette housing was larger to accommodate 0.25 inch (6 mm) wide tape and the tape speed was 3.75 ips, promising rather better quality. But the format came too late, the Compact Cassette already being well established particularly since Dolby B noise reduction had been widely exploited since the early 1970s.

Cassette housing and transport

The cassette housing incorporates a felt pressure pad on the opposite side of the tape’s recording surface which maintains the tape in intimate contact with the head during record and replay, this being particularly important when dealing with such small dimensions. During record and replay the heads are moved towards the cassette housing, and through appropriate access holes to contact the tape. Originally specified as a two-head format, an erase and a single record/replay head, no provision was made for a third head which would allow record and replay to be carried out by separate dedicated heads optimised for each purpose. In the continuing process of development manufacturers wished to add a third head and some chose to incorporate the record and playback heads in one housing, to be positioned in the normal record/replay-head position. Others chose to use a separate housing locating a third head elsewhere, using a separate access hole in the cassette. No pressure pad is provided for this, and the machine’s transport needs to be up to the task of providing optimum back tension of the tape to give good head-to-tape contact. Three heads of course means that off-tape monitoring is possible and ‘Double Dolby’, ‘Double dbx’, etc. noise reduction circuits are employed which encode the incoming signal and simultaneously decode the off-tape signal for monitoring.

Dual-capstan drives have been offered with a capstan and pinch roller placed at each end of the cassette. The capstan which is placed at the end which supplies the tape to the heads is engineered to run at a marginally slower speed than the other one, the latter defining the actual tape speed. Such an arrangement ensures very consistent tensioning of the tape across the heads and the mechanical performance of the cassette reels no longer affects the results. Some machines push the pressure pad away from the head so that it does not influence the performance. Auto-reverse has been seen, as has automatic ‘preview’ whereby the machine senses the silences between recorded items on the tape and then plays the first few seconds of each item before fast winding on to the next.

Tape selection

The different tape formulations were discussed in ‘Cassette tape’, above. A number of machines automatically select EQ and bias for the tape type which is loaded, sensing special holes in the cassette housing which are provided for this purpose (see Figure 6.14). Internal bias adjustment is usually provided in a cassette machine, but record and replay EQ are usually fixed. If it is desired to bias a machine for a particular tape, a good way to do this in the absence of EQ adjustments is to record pink noise at a level 20 dB below zero on the meters from a test record, or ‘white’ noise provided by FM tuner interstation noise, increasing the bias from a low level until the sound coming off tape is as near as possible indistinguishable from the sound going on. Too much bias produces a dull sound; too little gives an overbright sound. With a three-head machine this is very easy since the source/tape switch can be flicked to and fro for instant comparison, but two-head machines require regular rewinding and comparison.

Figure 6.14   Holes in the top edge of a compact cassette may be uncovered to signify different tape types and to prevent recording

It is very important to set the same replay level for the noise source and the off-tape noise, otherwise the ear hears frequency response differences which may simply be a subjective result caused by the different levels. The process should be carried out with noise reduction switched off, as this will exaggerate response errors. After alignment the noise reduction should be switched on and another section of noise recorded. The noise reduction should not introduce significant degradation of the signal off tape. In many cheaper recorders the bias control either affects all tape types equally, or simply works for ferric tapes which have the widest range of requirements.

Other alignments

Cassette test tapes are available, enabling frequency response and azimuth checks. Thorough cleaning and demagnetisation of the machine should be carried out before one is used. Small azimuth adjustments can bring particularly worthwhile improvements in cassette performance, especially when replaying tapes recorded on another machine. Azimuth can simply be adjusted for maximum subjective HF response using the little spring-loaded screw on one side of the record/replay head. Some machines incorporate computer systems similar to those found in certain professional open-reel models which automatically align the machine for a particular type of tape. Settings for several types can be stored in the computer’s memory.

Automatic replay azimuth adjustment is also possible. The two channels of the stereo output are filtered, converted into square waves and then fed to a comparator. Phase differences produce an output control voltage and this drives a small low-speed motor which adjusts the azimuth setting of the replay head. When azimuth is correct no control voltage is produced and the azimuth is left alone. The system is continuously active throughout the replay process and it is designed to extract the best performance from pre-recorded musicassettes and recordings made on other machines.

Multitrack cassette recorders

In the late 1970s the Japanese TEAC company introduced a machine called the Portastudio. It was a four-channel multitrack cassette recorder with mixing facilities and multiple inputs built in. The tape ran at twice normal speed, 3.75 ips, and the four tracks were recorded across the full width of the tape. Each track could be recorded on separately, sync facilities were provided, and ‘bounce down’ could be achieved in the manner of a professional multitrack machine whereby signals recorded on, say, tracks 1, 2 and 3 could be mixed and recorded on to the fourth track, freeing the other three tracks for further use. The final four-track tape could then be mixed down into stereo, these stereo outputs being fed to a conventional cassette recorder (or even open-reel).

One mixer company even offered an eight-track cassette-based system which incorporated a mixing section that offered facilities such as multiband EQ and auxiliary sends.

Cassette duplication

Cassette duplication is an important area of activity, pre-recorded musicassettes being very popular. The vast majority of such cassettes are produced on special machines which run the tape at 16, 32 or 64 times normal speed so that a length of tape for a 20 minute cassette can be duplicated in a few seconds. In one system, used mainly for short runs, banks of cassettes are duplicated in one go by each particular duplicator, and the labelling and packaging is carried out automatically. Master tapes to be duplicated are copied on to a half-inch ‘loop-bin master’ tape which is fed from a vacuum bin where the tape is stored in loose loops to allow for high-speed repeated reproduction. Tones are recorded on to the tape during duplication which tell the cassette loading machine where the beginning and end of each section is.

A speed 32 times normal for duplication requires record bias frequencies in the megahertz range and record-head gaps need to be rather less than a micron in width for high frequencies to be adequately recorded. Level and frequency response errors will be magnified by Dolby encode/decode processes. Consistent head-to-tape contact is difficult to maintain at such high speeds, high frequency loss normally being the result. Such duplicating equipment therefore needs to be maintained in first-class working order if the final cassette is to stand comparison with an equivalent home recorded example. Many musicassettes are found to be recorded at too low a level, failing to exploit the maximum dynamic range of the medium. Many sound dull, and several actually sound rather better with Dolby switched out during replay. Some manufacturers use higher-quality tape than others, and several chromium dioxide examples are available, but recorded using 120 µs equalisation.

An alternative to high-speed duplication is real-time copying. Instead of the highspeed duplicators a bank of carefully maintained cassette recorders are used which all chug away in real time, programme signal being provided by an open-reel machine or even a digital recorder, linked to a distribution amplifier to feed the cassette machines. Such a setup is ideal for the production of relatively low numbers of cassettes when it may not be economical to prepare the special production tapes necessary for the high-speed duplication process. Sound quality can of course be as good as the medium is capable of.

Recommended further reading

Jorgensen, F. (1995) The Complete Handbook of Magnetic Recording. 4th Edition. McGraw-Hill

See also General further reading at the end of this book.