Chapter 26. Tape Recording

26.1. Introduction

Magnetic recording underpins the business of music technology. For all its “glitz” and glamour, the music business, at its most basic, is concerned with one simple function: the recording of music signals onto tape or on disc for subsequent duplication and sale. Before the widespread advent of computer hardware, this technology was pretty well unique to the music industry. Not that this limitation did anything to thwart its proliferation—the cassette player was the second most commonplace piece of technology after the light bulb! Nowadays, with the massive expansion in data-recording products, audio—in the form of digital audio—is just another form of data to be recorded in formats and distributed via highways originally intended for other media. The long-term advantage for music recording applications is the reduction in price brought about by utilizing mass-produced products in high-performance applications that previously demanded a precision, bespoke technology.

A sound recording is made onto magnetic tape by drawing the tape past a recording head at a constant speed. The recording head (which is essentially an electromagnet) is energized by the recording amplifier of the tape recorder. The electromagnet, which forms the head itself, has a small gap so that the magnetic flux created by the action of the current in the electromagnet’s coil is concentrated at this gap. The tape is arranged so that it touches the gap in the record head and effectively “closes” the magnetic circuit, as Figure 26.1 illustrates. Because the tape moves and the energizing signal changes with time, a “record” of the flux at any given time is stored on the tape. Replaying a magnetic tape involves dragging the tape back across a similar (or sometimes identical) electromagnet called the playback head. The changing flux detected at the minute gap in the playback head causes a current to flow in the head’s coil. This is applied to an amplifier to recover the information left on the tape.

Figure 26.1. Magnetic tape and the head gap.

In an analogue tape recorder, the pattern stored on the tape is essentially a stored analogue of the original audio waveform. In a digital recorder the magnetic pattern recorded on the tape is a coded signal that must be decoded by the ensuing operation of the playback electronics. However, at a physical level, analogue and digital recordings using magnetic tape (or discs) are identical.

26.2. Magnetic Theory

Figure 26.2 illustrates the path of a magnetic tape through the head assembly of a modern analogue tape recorder. The recording tape is fed from the supply reel across an initial erase head by means of the capstan and pinch roller. The purpose of the erase head is to remove any unwanted previous magnetization on the tape. Next the tape passes the record head, where the audio signal is imprinted upon it, and the playback head, in which the magnetic patterns on the tape are converted back to an audio signal suitable for subsequent amplification and application to a loudspeaker. Finally, the tape is wound onto the take-up reel. When in playback mode, the erase head and the record head are not energized. Correspondingly, in record mode, the playback head may be used to monitor the signal off-tape to ensure that recording levels and so on are correct. Less expensive cassette tape recorders combine the record and playback heads in a composite assembly, in which case off-tape monitoring while recording is not possible.

Figure 26.2. Tape path.

26.3. The Physics of Magnetic Recording

In a tape recording, sound signals are recorded as a magnetic pattern along the length of the tape. The tape itself consists of a polyester-type plastic backing layer, on which is applied a thin coating with magnetic properties. This coating usually contains tiny particles of ferric iron oxide (so-called ferric tapes), although more expensive tapes may use chromium dioxide particles or metal alloy particles, which have superior magnetic properties (so-called chrome or metal tapes, respectively).

The properties of magnetic materials take place as a result of microscopic magnetic domains—each a tiny bar magnet—within the material. In an unmagnetized state, these domains are effectively aligned randomly so that any overall, macroscopic magnetic external field is canceled out. Only when the ferrous material is exposed to an external magnetic field do these domains start to align their axis along the axis of the applied field, the fraction of the total number of domains so aligned being dependent on the strength of the externally applied field. Most significantly, after the external field has been removed, the microscopic domains do not altogether return to their preordered state and the bulk material exhibits external magnetic poles.

The relation between the magnetizing field (H) and the resultant induction (B) in an iron sample (assumed, initially, to be in a completely demagnetized condition) may be plotted as shown in Figure 26.3. Tracing the path from the origin, note that the first section of the looped curve rises slowly at first (between O and B1), then more rapidly (between B1 and B2), and finally more and more gradually as it approaches a point where only a very few magnetic domains remain left to be aligned. At this point (B3) the ferrous material is said to be saturated. Significantly, when the magnetizing force (H) is reduced, the magnetic induction (B) does not retrace its path along the curve B3–B2–B1–O, instead it falls along a different path, B3–B4, at which point the magnetizing force is zero again, but the ferrous material remains magnetized with the residual induction B4. This remaining force is referred to as remnance. For this remnance to be neutralized, an opposite magnetic force must be applied, which accounts for the rest of the looped curve in Figure 26.3. The magnitude of the applied magnetic force required to reduce the remnance to zero is termed coercivity (the ideal magnetic tape exhibiting both high remnance and high coercivity).

Figure 26.3. BH curve.

26.4. Bias

If a sound recording and reproduction system is to perform without adding discernible distortion, a high degree of linearity is required. In the case of tape recording this implies the necessity for a direct relationship between the applied magnetic force and the resultant induction on the tape. Looking again at Figure 26.3, it is apparent that the only linear region over which this relationship holds is between B1 and B2, with the relationship being particularly nonlinear about the origin. The situation may be compared to a transistor amplifier, which exhibits a high degree of nonlinearity in the saturation and cut-off region and a linear portion in between. The essence of good design, in the case of the transistor stage, is appropriately to bias the amplifier in its linear region by means of a steady DC potential. And so it is with magnetic recording. In principle, a steady magnetic force may be applied, in conjunction with the varying force dependent on the audio signal, thereby biasing the audio signal portion of the overall magnetic effect into the initial linear region of the BH loop. In practice, such a scheme has a number of practical disadvantages. Instead a system of ultrasonic AC bias is employed, which mixes the audio signal with a high-frequency signal current. This bias signal, as it is known, does not get recorded because the wavelength of the signal is so small that the magnetic domains resulting from it neutralize themselves naturally. It acts solely to ensure that the audio modulation component of the overall magnetic force influences the tape in its linear region. Figure 26.4 illustrates the mechanism.

Figure 26.4. Linearizing effect of AC bias.

It is hardly surprising that the amplitude of the superimposed high-frequency bias signal is important in obtaining the best performance from an analogue tape machine and a given tape. Too high an amplitude and high-frequency response suffers; too low a value and distortion rises dramatically. Different tape formulations differ in their ideal biasing requirements, although international standardization work [by the International Electrotechnical Commission (IEC)] has provided recommendations for the formulation of “standard” tape types.

26.5. Equalization

For a number of reasons, both the signal that is imprinted upon the tape by the action of the record current in the record head and the signal arising as a consequence of the induced current in the playback head are heavily distorted with respect to frequency and must both be equalized. This is an area where standardization between different manufacturers is particularly important because, without it, tapes recorded on one machine would not be reproducible on another.

In itself, this would not be such a problem were it not for the fact that, due to differences in head geometry and construction, the electrical equalization differs markedly from manufacturer to manufacturer. The IEC provided an ingenious solution to widespread standardization by providing a series of standard prerecorded tapes on which are recorded frequency sweeps and spot levels. The intention was that these must be reproduced (played back) with a level flat-frequency response characteristic, with the individual manufacturer responsible for choosing the appropriate electrical equalization to affect this situation. This appears to leave the situation concerning record equalization undefined, but this is not the case because it is intended that the manufacturer chooses record equalization curves so that tapes recorded on any particular machine must result in a flat-frequency response when replayed using the manufacturer’s own IEC standard replay equalization characteristic.

The issue of “portability” should not be overlooked, and any serious studio that still relies on analogue recording must ensure that its analogue tape equipment is aligned (and continues to remain aligned—usually the duty of the maintenance engineer) to the relevant IEC standards. This, unfortunately, necessarily involves the purchase of the relevant standard alignment tapes.

26.6. Tape Speed

Clearly another (in fact, the earliest) candidate for standardization was the choice of linear speed of the tape through the tape path. Without this the signals recorded on one machine replay at a different pitch when replayed on another. While this effect offers important artistic possibilities (see later in this chapter), it is clearly undesirable in most operational circumstances. Table 26.1 lists the standard tape speeds in metric (centimeters per second, cm/s) and imperial measures (inches per second, ips) and their likely applications.

Table 26.1 Standard tape speeds in metric and imperial measures.

Tape speed (ips)	cm/s	Application
30	76	Top professional quality
15	38	Top professional quality
7.5	19	Professional quality (with noise reduction)
3.75	9.5	Semiprofessional quality (with noise reduction)
1.875	4.75	Domestic quality (with noise reduction)

26.7. Speed Stability

Once standardized, the speed of the tape must remain consistent over both long and short terms. Failure to establish this results in audible effects known, respectively, as wow and flutter. However, these onomatopoeic terms relate to comparatively coarse effects. What is often less appreciated is the action of speed stability upon the purity of audio signals—a fact that is easier to appreciate if speed instability is regarded as a frequency modulation effect. We know that FM modulation results in an infinite set of sidebands around the frequency-modulated carrier. The effect of speed instability in an analogue tape recorder may be appreciated in these terms by looking at the output of a pure sine tone recorded and played back analyzed on a spectrum analyzer, as shown in Figure 26.5. Note that the tone is surrounded by a “shoulder” of sidebands around the original tone.

Figure 26.5. FM sidebands as a result of speed instability.

Happily, the widespread adoption of digital recording has rendered much of the aforementioned obsolete, especially in relation to two-track masters. Where analogue tape machines are still ubiquitous (e.g., in the case of multitrack recorders), engineering excellence is a necessary byword, as is the inevitable high cost that this implies. In addition, alignment and calibration to recognized standards must be performed regularly (as well as regular cleaning) in order to ensure that multitrack tapes can be recorded and mixed in different studios.

26.8. Recording Formats—Analogue Machines

Early tape recorders recorded a single channel of audio across the whole tape width. Pressure to decrease expensive tape usage led to the development of the concept of using “both sides” of a tape by recording the signal across half the tape width and subsequently flipping over the tape to record the remaining unrecorded half in the opposite direction. The advent of stereo increased the total number of audio tracks to four: two in one direction, two in the other. This format is standard in the familiar analogue cassette. From stereo it is a small conceptual step (albeit a very large practical one) to 4, 8, 16, or more tracks being recorded across the width of a single tape. Such a development demanded various technological innovations, the first was the development of composite multiple head assemblies. Figure 26.6 illustrates the general principle. Given the dimensions, the construction of high-quality head assemblies was no mean achievement. The second was the combination of record and replay heads. Without this development, the signal “coming off” tape would be later than the signal recorded onto the tape, a limitation that would make multitrack recording impossible. In early machines, the record head was often made to do temporary duty as playback head during the recording stages of a multitrack session, with its less than perfect response characteristic being adequate as a cue track. The optimized playback head was reserved for mix down only.

Figure 26.6. Multiple tape tracks across width of the tape.

Despite this, the number of tracks that it is practical to record across a given width of tape is not governed by head construction limitations only, but by considerations of the signal-to-noise ratio. As shown earlier, the signal recorded onto tape is left as a physical arrangement of magnetic domains. Without an audio signal, these domains remain unmagnetized and persist in a random state. These cause noise when the tape is replayed. Similarly, at some point, when a strong signal is recorded, all the domains are “used up” and the tape saturates. A simple rule applies in audio applications: the more domains onto which the signal is imprinted, the better, up to the point just below saturation. This may be achieved in various ways: by running the tape faster and by using a greater tape width for a given number of tracks. Figure 26.7 illustrates this by depicting the saturation levels of a commercial tape at various speeds. This simple principle accounts for the many different tape formats that exist. Each is an attempt to redefine the balance among complexity, sound quality, and tape cost appropriate to a certain market sector. Table 26.2 lists some of the major analogue recording formats. Note that the format of a tape relates to its width, specified in inches.

Figure 26.7. The effects of tape speed on saturation and distortion.

Table 26.2 Major analogue recording formats.

Tracks	Format	Medium/speed	Application
2	½″ stereo	½″ 7.5–30 ips	High-quality mastering
2	¼″ stereo	¼″ 7.5–30 ips	High-quality mastering
2	Cassette	¹/8″ 15/8″ ips	Medium quality replay
4	½″ four track	½″ 7.5–30 ips	High-quality mastering
4	Cassette	¹/8″ 3.75 ips	Personal multitrack
8	¼″ multitrack	7.5–15 ips	Semipro multitrack
16	½″ multitrack	½″ 15–30 ips	Professional multitrack
16	1″ multitrack	1″ 30 ips	High-quality multitrack
16–24	1″ multitrack	1″ 30 ips	High-quality multitrack
24	2″ multitrack	2″ 30 ips	High-quality multitrack

26.8.1. Analogue Mastering

Analogue mastering is now very rare, this office having been made the exclusive domain of digital audio tape. A typical high-quality two-track mastering recorder is illustrated in Figure 26.8.

Figure 26.8. Analogue mastering recorder.

26.8.2. Analogue Multitrack Tape Machines

As mentioned earlier, analogue multitrack machines betray their quality roughly in proportion to the width of the tape utilized for a given number of tracks. A 2-inch tape, 24 track, which utilizes a 2-inch width tape drawn across 24 parallel heads, is therefore better than a 1-inch, 24 track, but not necessarily better than a 1/2-inch, two track! Not only does a greater head-to-tape contact guarantee a higher tape signal-to-noise ratio (i.e., more domains are usefully magnetized) but it also secures less tape dropout. Dropout is an effect where the contact between tape and head is broken microscopically for a small period during which the signal level falls drastically. Sometimes dropout is due to irregularities in the tape or to the ingress of a tiny particle of dust; whichever, the more tape passing by an individual recording or replay head, the better chance there is of dropouts occurring infrequently. Analogue tape machines are gradually becoming obsolete in multitrack sound recording; however, the huge installed base of these machines means they will be a part of sound recording for many years to come.

26.8.3. Cassette-Based Multitracks

Figure 26.9 illustrates a typical analogue cassette-based portable multitrack recorder and mixer combined. This type of low-end “recording studio in a box” is often termed a Portastudio and these units are widespread as personal recording “notebooks.” Typically four tracks are available and are recorded across the entire width of the cassette tape (which is intended to be recorded in one direction only). The cassette tape usually runs at twice normal speed, 3.75 ips. Individual products vary but the mixer of the unit illustrated in Figure 26.9 allows for two (unbalanced) microphone inputs and a further four inputs at line level, of which only two are routed to the tape tracks. Each of the first four inputs may be switched between INPUT, OFF, and TAPE (return). Selecting INPUT will (when the record button is engaged on the tape transport buttons) switch the track to record. The mixer also incorporates two send–return loops and the extra line level inputs mentioned earlier. In addition, an extra monitor mixer is provided, the output of this being selectable via the monitor output. It is thus a tiny split multitrack console. Despite the inevitable compromises inherent in such a piece of equipment, many portable multitrack units are capable of remarkably high quality and many have been used to record and mix release-quality material. Indeed, so popular has this format proved to be that digital versions have begun to appear, products that offer musical notebook convenience with exemplary sound quality. One such is illustrated in Figure 26.10.

Figure 26.9. Cassette-based “notebook” multitrack.

Figure 26.10. Roland digital multitrack.