Recording
Although digital techniques have made signal storage much more flexible and in many ways convenient, much was and is retained on tape. This medium was introduced as a vast improvement on the wire format and pioneered in Germany for communications, most notably giving the ability to record multiple independent tracks on a single spool. Wire recorders are still used in aircraft flight recorders, or 'black boxes' (which incidentally never are). The first coatings were standard gamma-phase ferric oxide, or jeweller's rouge, the initial work intended to improve ¼" tape running at 7½ips or higher.
A number of means have been used to record signals on to tape media, a permanent magnet 'erase' head has been followed by a record/play head using an equalised version of the input signal, non-return-to-zero techniques have been used to record digital data, but generally, with an analogue signal, an HF bias of 60kHz or more is used to drive the erase head, this bias then being amplitude modulated to drive the record head which is then often used to playback. Because of this economy, using mechanical switches instead of separate heads and circuitry, signal to noise ratios and performance could suffer, notably through both connections to the rec/play head having to be switched.
The commonest domestic uses for open-reel tapes and the later cassettes was for the illegal copying of vinyl records or for recording compilations and broadcasts. The fact that the medium could be reused when tastes changed, added appeal. Ideally, tapes would carry four tracks (2 x L and R) and in use the tape would have to be 'turned over' to use the other side, a bit like a vinyl LP. Semi-professional open-reel decks and the later portastudios would use the full width of the tape and in one direction only in normal use. The adoption of the ¼" width of tape was governed by the maximum speed the tape could reliably be run at without parting for hi-fi purposes, in this case 7½ips. One of the better transports used by the author for this size could be found in the Sony TCs 377, 399 and 630.
Semi- and professional machines used ½" or more, running at 7½, 15ips or higher, the higher speeds increasing the frequency response available (tables of tape lengths and play times can be found at the bottom of the page), but reducing the bass on occasion. 8-tracks used a continuous loop of ¼" tape using existing technology to provide background music for commercial premises. These ran in a continuous play mode, functions like rewind and fast forward not being possible.
The advent of the Philips compact cassette, using half the width again and running at half the speed of an average domestic tape deck, resulted in much research and development in the coatings used in order to increase the available bandwidth although originally the format had been intended for dictation only. Performance between individual cassettes could vary considerably, the bias current through the record head would then have to be adjusted by as much as +50% for each formulation apart from the use of additional equalisation. Some decks could include an oscillator (say, 400Hz) to help set the bias but ideally more than one frequency (8kHz) was required to obtain a reasonably flat response. The C-90 was the most popular size because, normally, each side could accommodate an entire vinyl LP on tracks 0.66mm wide. Very few commercial models had separate record and play heads. Portastudios using cassettes would run at 3¾ips, again to increase available bandwidth, using C-60 or C-90 types, because the thin tape used in C-120s broke easily being seven times thinner than a human hair.
Deterioration in a cassette's performance over time and with use could figure also, even with test tapes supplied by reputable manufacturers. The more frequently a test tape is used, the more the tape characteristics will change (eg; lowering of recorded level, worsening of frequency response particularly at high frequencies and an increase in wow due to tape elongation) until measured values become unreliable. Even when a tape is not used, but stored, for a long period of time, performance can drop because of self-demagnetisation. When making reference tests refer to the tape life specification and do not use for longer than it's rated life when servicing. For example, one manufacturer gave a frequence of use of not more than 20 times for each tape length and a maximum storage period of 6 months.
Cassettes used in tape streamers to backup data on PCs were formulated for high speed use (90ips). Data was transferred at some 86.3kB per second allowing 22MB to be backed up in under 4 minutes. The TEAC CT500 streamer tape cassette was expected to have a useful life of over 1,500 passes. Digital storage could give an adequate s/n ratio with tracks only 0.1mm wide.
Hi-fi 3-head Elcaset systems used ¼" tape at 3¾ips but were not a commercial success, a number of these decks not including noise reduction leaving this option to the user. Micro-cassettes would run at half the speed of a compact cassette since the intended bandwidth was that of speech only, either for dictation or surveillance, although some models had a switchable lower speed. Other packaging 'standards' were developed to meet office and communications needs.
Any tape transport mechanism, or deck, will impose itself on the signal via wow and flutter, much like a record deck will, principally because of failures of primary components or the interfaces between them. Dirt, oil or other foreign matter on belts and drive surfaces will vary speed as can failure of the pressure (pinch) roller or inadequate pressure from same. There might be drive belt deformation or excessive back tension of the supply reel or tape counter, the capstan may be bent or there can be motor failure.
Some models used dual capstans to stabilise the tape speed as it passed the heads, the Technics RS-1500US (open-reel) and Beocord 5000 (cassette) are examples, the latter mechanism appearing in three models, both 5000s (4705 and 4715) and the Beocenter 4000. Sources of wow frequencies in the last were identified as thrust roller (1.3Hz), drive belt (1.8Hz), flywheel (5.05Hz) and capstan motor (14.5Hz). For this particular and complex chassis special gauges and tools were required which many dealers would not have held.
Splicing was used to repair snapped tape or to edit it. The tape spools were turned manually to pinpoint the splice or cutting point on the play head, which was then visually marked. This process was made easier if the recording speed was high. It was usual to slightly overlap the ends of the tape to be cut so that both would match. A cut at 45° to the sides was less audible than one at 90°, a thin adhesive tape then joined the two free ends, ordinary sellotape proving unreliable. Small 3" spools made sections easier to handle. The starting and stopping of new recordings could produce distinct clicks, although with some machines (even cassette decks) it was possible to prevent this by (relatively) slowing the contact of the heads with the tape if a mechanical pause control was available.
All magnetic medium is susceptible to damage from external fields. This can include induced noise, partial erasure and an increase in 'printing'. This last is always present as crosstalk between adjacent tracks to a greater or lesser degree but can be increased by recording at too high a level, exposure to temperatures that are too high and stray magnetic fields. To reduce the above professional tapes are usually stored, whenever off a recorder, in a metal box. If care is taken, magnetic recordings can be expected to last almost indefinitely.
Magnetisation of the heads and surrounding metalwork could occur through the earth's natural magnetism and the record bias being interrupted at peak levels. This could introduce hiss and crackling. Tape repeatedly passing magnetised metalwork can lose it's treble content and shed coating particles necessitating cleaning for optimum results. Dirty pinch rollers, drive capstans or inadequate take-up torque in cassette decks would result in the tape wrapping itself around the roller or capstan. In most domestic models there was little done to remedy this, although the Beocord 5000's bias oscillator would damp on shutdown, degaussing the record head automatically. Manual degaussing could, if care was not taken, exacerbate the problem and not reduce it. For example, the degausser used should, although being held near the heads and metalwork, never be brought into contact with them.
Although immaterial to most domestic users, motor speed variations between machines could cause problems say when two decks at the opposite limits of their tolerances were used to process the same tape. This was particularly true with longer delay effects where the same tape passed through two or more decks in which case it was wise to use the slowest machine to record and the faster (progressively, if necessary) machine(s) to playback into the mix. If this was not done, over time a tape spillage between the machines would occur. Setting up a multiple system can be very interesting, particularly if high tape speeds are involved and back-up tapes should be made beforehand to accommodate 'accidents and failure's especially if multiple operators are needed to actuate the decks simultaneously.
Younger readers might be shocked at this 'primitiveness' but probably the most famous album that used this technique is 'Dark Side of the Moon', in particular, aptly, the track 'Time'. Earlier than this, probably the best exponent of the open-reel and what it could achieve was Delia Derbyshire, who was responsible for, amongst many other things, the iconic 'Doctor Who' theme. Much of the source used in Brian Eno and David Byrne's 'My Life in the Bush of Ghosts' was on cassette, with it's attendant lack of accuracy in pinpointing edit points.
For longer delays between just two decks, a 'bin' could hold the 'overun'. This was best configured as a flat box whose depth matched the width of the tape (½" recommended). For a live show the delay would then have to be precisely timed and known by the performers if it was not possible to impart some form of 'marker' (imperceivable to the audience) on the tape or, say, use a foot-switch, or another individual who knew the piece, to operate the second deck. A clever visual device used in a show with a single musician hid the two decks processing the signal whilst a very visible third machine recorded the overall mix. Some very experienced people were left scratching their heads.
Reverse play could be achieved by either readjusting the position of the play head or mounting a new one, a similar result could be achieved by reversing the feed spool to present the 'wrong side' of the tape to the heads, but this invariably entailed a loss of treble, detail or increase in noise and problems rewinding. Varying the path of the tape as it passed the play head could produce limited phasing effects.
Noise reduction was used to reduce the hiss inherent with tape replay. In a good cassette deck of early '80s vintage the amplifier would give about 10dB below that of the tape. Dynamic noise reduction reduced high frequencies at low signal levels, Dolby B and C systems boosted the high frequencies during record and reduced them on replay and compression techniques would compress the dynamic range of the entire signal on record, then expand it on replay. This last was thought most successful, the noise floor on record being below the signals and then proportionately reduced further on playback. The performance of a poor cassette deck could then be dramatically improved although companding ratios of 10:1 might be required. Disadvantageously, all of these systems were incompatible.
Digital tape recorders, by their nature, could give the best noise and dynamic performances, but were invariably priced beyond most non-professional means. The Sony PCM 3324, for example, enabled spliced joins or punch-ins via computer-generated cross-fade over the join to give a continuous signal round the insert or edit point. This unit used five heads to provide 24 digital tracks, a stereo pair of analogue tracks and a time-code track. One feature allowed the rehearsal of an edit, to be able to hear what it would sound like without losing the originals. A dynamic range of >90dB was possible with all 28 tracks fitting onto ½" tape making a very handy but expensive bit of kit. This was at a time when 'Winchester' hard-drives had a capacity of 500Mb and needed a 3-phase motor to turn them. Analogue-to-digital converters were more easily applicable to video, say with an 8-bit, 5MHz 'aperture time' because this was only 20% shorter than that needed for a 16-bit 15kHz audio sampler whose accuracy needed to be better by a factor of 256. The Philips DAT system, by employing a mantissa, reduced the word length and, thus, the complexity and cost of the equipment. For the author, the point when a digital signal became indistinguishable (for all practical purposes) when switched between it's 15ips twin-track analogue equivalent (with the same signals played in synchronism at the same levels) was with the Sony PCM.F1 encoder/decoder using a Betamax video recorder in 1983.
Research was undertaken in respect of the magnetic 'grains' on tape in order to pack more information on to the media. Ordinarily, these had a long, needle-like crystal structure giving a minimum distance between which a change of magnetic flux could be recorded, offering some 15,000 bits per inch. One approach (EMI Watermark) oriented them across the magnetic track giving a greater density to the magnetic 'substrate'. Another (Vertimag), proposed standing the particles on end, perpendicular to the surface, coincidentally improving the coercivity. This gave 100k bpi and, experimentally, 400k bpi compared to 25k bpi for optically read digital data. Contemporary RAM would hold 1M per square inch compared to a possible 10G using the vertical method. Given that at 1kHz the recorded wavelength is 0.0075", 0,0036" and 0.0018" (at the tape speeds of 7.5", 3.75" and 1.875" respectively) the inferences, if taken further, for analogue signals would have been to impart higher frequency responses at much lower tape speeds.
The head gap width required to resolve the magnetic field presented to it is often less than or equal to half the required upper frequency's wavelength, or the tape speed divided by twice the frequency. The first Beocord 5000 (type 4705) mentioned above used separate heads for record and playback, the record using a 6µm gap to obtain good depth magnetisation of the tape and the play head used a 1.5µm gap in the interests of a high frequency response. These values were thought to 'secure optimum data for signal-to-noise ratio and frequency response'. The later models using this chassis used a single head to combine both functions.
Although a number of types existed, including two-layer tapes and those where the coating was plated, generally speaking a thick coating could give best results handling the low to middle frequencies, and a thin tape the higher (15-20kHz) ones. Often the frequency response could be extended at lower record levels.
Mixing
This some consider a skill, others an art. It is felt that it is true that both are necessary and this can only come with practice, and patience, because one song, piece or tape may require re-listening a 100 times before it is felt to be right. Some pieces can take years to perfect, others 15 minutes. Live performance in particular can demand a cool head to cope with aberrations, anomalies and unavoidable incursions of Murphy's Law ("Break a leg!"). After an intensive session it is wise to give the ears, and the brain, a break and, as with all things, if you start making mistakes, the best thing to do is stop. It would always be nice to have to hand the vast resources of a huge professional recording studio. Ordinarily though, circumstances will usually range from the adequate to the diabolically impossible. However, lack of resource may not necessarily hinder the proceedings. The mistakes made during the recording of the Woodstock festival, for example, are urged to be considered as the creases in a very fine leather, and quite rightly too. On this note, any form of intoxication should be reserved exclusively for listening and should be absent from any technical work otherwise hours of toil can be ruined very easily, this being the reason for a number of group's dissolutions or 'fall-outs'. If serious work, say for a commercial bid or deadline is being undertaken, those with a problem in this context should, no matter how 'amazing', be excluded from the outset, by sheer necessity. Similarly, a forceful, egotistical and/or temperamental individual's opinion may not be the right one. The adage that the needs of the many outweigh those of the few can be particularly true and an experienced professional should have no difficulty in pointing out inadequacies or difficulties to a band's management, for example, if they have one.
Multitrack recording can become complex and trying to remember every setting may not be helpful. Notebooks and track sheets will help enormously. Apart from identifying the recording medium, title, date and performers, columns can cover the initial take and overdubs, transfers or track combinations and the setup for remix.
Although many recordings are effected with only the level controls on the recorder, digital or otherwise, a mixer will enable enormous flexibility. Although at first glance, a mixing desk, particularly a professional one, can appear to be intimidatingly complex, they all employ the same principles which are then repeated to give the required number of channels. Invariably, the object is to mix a large number of inputs to a smaller number of outputs or busses, say L and R for example.
Input channels
These consist of a vertical row of knobs with a slider volume control, or fader, at the bottom and layouts can vary although a general pattern is often seen. At the top, ideally, should be a meter which gives an indication of what that particular channel 'is doing', then there may be an input selector for microphone/line and tape. Next, a trim or sensitivity control sets the gain of the channel between that required for a dynamic microphone, line (about 50dB with respect to each other) or other input levels between these like condenser microphones and hot guitar pickups. Below this may be an auxillary or effects send control to vary an unequalised output of the channel to, say, an effects unit via the auxillary buss. Next will be the equalisation or tone controls which are used to iron out peaks or troughs. Unsophisticated models will have simple bass and treble controls (about ±10dB), like those found on a hi-fi amplifier. More sophisticated parametric controls (about ±15dB) allow the centre frequency of the filter to be varied and a Q control allows adjustment of the width of the band either side of this frequency. With these care should be taken if the frequencies, on one channel, are set to an overlapping frequency. For example, with two controls set so that the high end of the low frequency band coincides with the low end of the high frequency, double the gain or cut and boost will be possible (say ±30dB) which, normally, will be excessive. If adjusting by ear, first set the gain to maximum boost or cut, then sweep the frequency to identify the one required, then adjust the gain to suit. Next, may be a pan control. This is like a balance control and moves the mono signal of the channel between the two output (L and R) busses adjusting the signal's position in the soundfield. For example, a lead vocal will, generally, be set for the centre. An additional switch may allow the signal to be moved between busses. A cue control can allow a signal to be passed to a cue output (usually mono) which then feeds a satellite speaker/s off-stage where other performers or speakers may be waiting. This can also be used to supply stage directions, via earpiece to the 'visible performer', that are inaudible to the audience. A pre-fade listen, or PFL, button selects that particular channel for listening using the PFL switch on the monitor. This is usually done through headphones and, during live performance, is useful for fine-tuning an individual input or finding one that is lost. An output trim may then be available to allow the output of the channel to be set to whatever levels are required by external equipment. This may allow switching between the input and the recorder. Lastly, comes the fader which varies the output of that channel from zero to maximum.
Output and monitor channels
Meters will be included for the outputs even if there are none for the inputs. A buss monitor enables selection and volume control of the busses. Tape cue controls will allow the output from a tape deck (in this case) to be varied and the record channels to be selected. Line level inputs should have sensitivity and pan controls. Output equalisation is useful on a mixer intended for live performance with the same performers. Individual instrument's settings can then be retained whilst different venue's inequalities can be 'straightened' without affecting these. An auxillary or effects receive control allows the mixing in of a signal from external equipment fed by the send controls on the input channels. A headphone output should include sockets for two sets, a variable output and a means of switching between the busses and the PFL, if included. Lastly, the 'master' fader sets the overall output level of the master busses (L and R say) and in performance is used to fade the mixer in or out to the PA, or when recording, to the recorder.
Setting up
It is impossible to account for all situations but very basic advice is given here. If starting from scratch set all the controls to their minimum or neutral position. Firstly, set up the output/monitor levels using a known line-level source, set the output faders to 0dB and adjust the busses levels and equalisation for best results from the sound system used. At this point, it should be realised that replaying any recorded material may require adjustment of these on any subsequent system because of the variations in speaker types, listening room, etc. Once these are established and any recording media similarly accommodated, individual inputs can be addressed. Plug in an input and set the fader to it's 0dB position, then adjust that channel's sensitivity so that 0db is shown by the meter on peaks. Input level matching, in all audio equipment, is important. Too low and background noise will become obvious. Too high and level adjustments become awkwardly critical. Mismatches between input and output impedances can cause problems. Usually, the input impedance of a device will be 50 times the output impedance of that which feeds it. Never plug the output of a power amplifier directly into an input without a suitable attenuator between them as smoke will be the likeliest result.
The importance of the placement of microphones (see sound pickups) cannot be underestimated. Close-miking a musical instrument may give a higher signal and reduce extraneous noise (or feedback at a gig, say) but because of the lack of natural resonance may impart a dead quality that lacks life. Even small changes in a mic's position, or the choice of mic, can affect sound quality dramatically. Using one's own ears to determine the optimum position for a mic is acceptable but should not be attempted with a drum-kit since the transients produced can cause irreparable ear damage. However, nothing beats the empirical approach which adds to experience, even if it means that the best result is given by putting a mic into a bucket, dust-bin or shower cubicle! If it works, then it's right.
Some mixers have switchable padding which can reduce noise on high impedance inputs. Hum can have a number of origins - inadequate screening, earthing or supply smoothing. All connecting media must be sound and clean. If a connection causes an earth loop (disconcerting hum), disconnect one of the concerned grounds. A colossal hum, accompanied by a chorus of radio stations, usually indicates a disconnected signal earth. Some equipment is fitted with an earth-lift switch which isolates the signal ground from the (safety) chassis ground. Some 'separates' designs will have, say PCB selectable, ground options. If a ground is lifted, an HF path to chassis (capacitor) usually improves and is best made permanent. Mains transients can be removed with filters.
After these points have been satisfied attention should be given to the equalisation to give the most acceptable sound, trimming the sensitivity as required to keep the signal within reasonable limits. When recording, even experienced engineers will forget that cutting the low frequencies will have a similar subjective effect to boosting the higher ones. Cutting will give more headroom and less likelihood of clipping. So instead of raising the treble, say, cut the bass and boost the overall gain and see if this sounds better than simply raising the treble. Again, good mic placement will reduce the need to apply equalisation in an attempt to remedy the defficiencies given by a poor position. Any instrument that employs tone and volume controls of it's own should have these set flat, or agreement made beforehand over their manipulation in order to avoid arguments ("But I always play it like this.", etc) and wasting time. This is simply because a signal can then overdrive into clipping or disappear in a mix quite easily and the cause may not be immediately apparent and an unnecessary distraction is caused. Expensive studio time will not be wasted if everybody knows what they are doing. This is what practice sessions are for and what makes a truly skilful performer.
Vocalists should be aware, or made aware, of how the output from a microphone falls off dramatically as the distance between it and the singer's mouth increases. This can play havoc particularly in a live show. For example, a singer may have acquired the habit of screaming into a mic whilst nearly swallowing it, and then moving a fair way away from the mic for a gentler chorus. 'Hunting' in live performance for a subdued vocal by increasing the gain of that particular channel will invariably lead to unwelcome, unimpressive and unprofessional feedback, or 'howlaround' which can damage speakers. The same can happen with instruments. To gauge an adequate distance some will need a microphone stand or something to hang on to (even a partial one, like Freddy Mercury did), others will not. Dancers, for example, might use radio mikes and 'covert' earpiece/boom microphones offering a reasonably constant mouth-to-mic distance relationship. Sometimes a combination of the two can work, the 'obvious' mic and stand being mere props or back-ups.
Where it is not possible or desirable to close-mike vocalists or performers pressure-zone mics, like the Crown PZM, using the pressure zone principle and well known for it's excellent characteristics, can be used. This however is simply an omni-directional capsule mounted close to a hard, flat surface which shields in one direction and reflects in another. This gives, approximately, a -3dB rear lobe and +3dB forward lobe, giving an apparent 6dB 'gain'. Improved bass response will be given if the mic is suspended above the performers mounted on a metre square sheet, perspex improving 'invisibility'. This solution is more suitable than, say rifle or parabolic mics, whose more directional characteristics can pick up the movements of scene shifters and other stage crew.
Sometimes a performer will have forgotten to attach their microphone, or switch it on. One pair, specialising in 'lewd' humour, did not understand until it was too late that their excellent and expensive Shure radio-microphones were not indestructible or impervious to liberal quantities of vaseline. A used condenser microphone's or effect pedal's battery will always fail mid-gig. For these and many other reasons and examples, a back-up must be instantly available, especially on-stage. The same can be safely said about signal leads (if not some performers!). As an 'engineer' one can be seen by some as obliged to effect instant repairs to faulty equipment at no cost. Some 'arty' types can become quite upset if one does not oblige and it is pointed out that the tools of their performance are their responsibility. True professionals just get up and play, don't make a fuss, are modest and are a joy to work with. Some advice to those with immense opinions of themselves - be nice to others and they, in turn, will invariably be nice and helpful in return. Live work is particularly, in the author's view, the most demanding and the least and most rewardable.
The '75 Neu! 'black' album, the author considers interesting. Memorable pieces can be produced using the simplest of techniques. The track 'E-music' shows what can be done simply with the effects send controls on a mixer, and live even.
Often one will see all of the faders on a mixing desk in their maximum position, bar those actually used for incoming signals. This is intended to sum the self-generated noise from the unused channels thus cancelling that in the others and giving a lower noise floor and greater dynamic range to the overall mix. The sensitivity, gain or trim controls of these channels will be set to a minimum, the input sockets often automatically shorting to ground with nothing plugged in.
Initial recording
Before pressing record, reset the counter or timer on the recorder so that it is then easy to locate passages when required. Generally, a rhythm line will be the first track/s laid. Assume a drum synth is used to create a stereo recording on two channels. The pan controls for each input will be turned to their furthest extremes (L and R) creating a stereo soundfield and levels set so that the playback is the same as the inputs.
Overdubbing
Overdubbing, or sync recording, is where multitrack recording comes into it's own. A single musician can play all the instruments building up a track, channel by channel. The previously recorded backing or rhythm track is replayed and, using this as the reference, new channels are added or recorded. Backing vocals and fills are usually next because these can be combined and bounced, leaving room for more parts. Leads and solos are usually last so that they are first generation recordings on the master tape. To obtain the necessary 'texture' one vocalist overlaid her vocals in excess of 60 times. With this kind of attention to detail or if much work is being undertaken it makes sense if the artist has means to record at home. Naturally, the results must be easily compatible with any mastering equipment used elsewhere. The internet has opened very many exciting opportunities in this context.
Ping-pong
Ping-pong recording, or bouncing, is used to combine tracks in order to make room for others, especially in systems that have a limited number of channels. Before combining tracks it must be ascertained that these are of the standard or quality required.
Mixdown
When all the tracks are completed it is then time to combine or mix them to give the greatest or desired effect and establish the stereo perspective (L to R placement) of each element or track. With small portastudios this is usually recorded on another machine. It is usually during mixdown that a new idea may prompt the re-recording of a part making the overall piece stronger. In this context, the author has found that where improvisation is involved, the first take can often be the best. Originality can tire easily.
Stereo and multi-mono
In a true stereo recording two mics will be carefully placed to capture the image and two tracks recorded. Alternatively, more mics can be added and be recorded on their own channels, either as stereo pairs or independent tracks. In multi-mono a number of single, or mono, tracks are recorded and then panned to simulate a stereo soundfield. Most recordings are made in this way. For example, the drums may be set so that the hi-hat or crash/ride cymbal are at extreme right, the lowest tom, extreme left and the bass in the centre. Another stereo source, like piano, will be mixed between the stereo spread and the final sources, like leads and solos, are centred. There are no hard and fast rules, only guidelines, and ultimately, if you like the sound, use it.
Punch-in
This is used for correcting minor mistakes without the need for recording an entire track again. Timing is critical since any insert will erase previously recorded material. Re-doing a coherent section will be more successful than trying to replace just a couple of notes. If the insert coincides with major downbeats, small timing variations can be masked. Often, on a portastudio, say, a foot-operated punch-in/out switch is used.
Back-up
With the advent of digital techniques, PCs, etc, this is now much easier and physically separable and portable back-ups should be made whenever a critical point has been achieved. Always understand that, if it can go wrong, it invariably will and any consequent loss of work and time will not be appreciated. One studio in the Netherlands was considered very cool because it was on a houseboat, until it sank.
Given the breadth of this field, a few books are recommended, in no particular order;
Home Recording for Musicians by Craig Anderton,
Building a Recording Studio by Jeff Cooper and
Basic Disc Mastering by Larry Boden.
especially if you want additional content to this page
or if you find any links that don't work. Don't forget
to add the page title or URL. Take care!