When the impressive quality of 24-bit digital came along and was accompanied by ADDA converters good enough to make use of the massive increase in quality over earlier systems, I and others were foolish enough to say "Well that’s it! No more analogue for us!"

We were wrong. What we had forgotten was that, knowingly or unknowingly (and with the possible exception of classical music) we had been using the tape recorder as an effect machine. The various noise reduction (NR) systems and the reaction of the tape itself to the signals printed on it, changed the sound in many ways that we did not realise.

The biggest drawback of magnetic tape is that even with the best tape running at the highest speed and on very high quality, professional machines, unless we alter the signal in some way, there is a great deal of background noise known as tape hiss. This is comes from the fact that we are unable to make absolutely every particle on the tape do exactly what we want and so there is a small amount of random signal, which is exactly what noise is.

We concentrated all our efforts at reducing noise. The Dolby company developed compressor-expander systems that reduced the dynamic range on record (compression) and then increased the dynamic range (expansion) on playback. The most popular professional system was Dolby A that compressed and then expanded four frequency bands. This system remained the system of choice from 1965 until 1988, when it was superseded by Dolby SR (Spectral Recording) which divided the signal up into ten frequency ranges and had a much more sophisticated compression-expansion algorithms.

Dolby A was often accused of ‘muddying’ the sound and many engineers avoided using NR altogether by means of a method called pre-emphasis, which just meant that the highs were turned up on recording and turned down by the same amount on playback. As tape hiss is more or less white noise, the energy of the hiss increases with frequency. By turning down the highs, we also effectively turn down the tape hiss.

So you can see, throughout the history of analogue tape recording, we spent a great deal of energy eliminating tape hiss at the cost of dynamic range and even quality. We loaded the tape with as much signal as it could carry.

The way we worked with tape had three effects on the sound.

1. Because tape does not react in a linear manner to an increase in volume (doubling the signal does not mean that there is double the magnetic flux on the tape), we inadvertently created a gentle, multiband compressor.

2. By keeping the levels as high as possible, we introduced slight amounts of harmonic distortion on the peaks and this had the same effect as very slight amounts (or not so slight, if we went too far) of excitation , similar to slight amounts of such effects as the Aphex Aural Exciter.

3. If we used noise reduction or even just a little pre-emphasis, this helped to compress the highs - which, when combined with the other effects, is what a good de-esser does -and this gave the recordings a silky, softer sound.

Or as some of the pundits put it: ‘The warm sound of analogue.’

Even today, youngsters starting out who are lucky enough to get a job in a studio are called tape-ops (tape operators), although the original tasks that went with that title (cleaning tape heads, sorting and rewinding the tapes and labelling the boxes) have mostly disappeared. Once he had gained enough experience -and had finished sweeping the floors and making the coffee - he would be given the onerous task of aligning the heads on the 24-track tape machine.

Professional multitrack tape machines may have been completely replaced by digital harddisk recorders, but tape has been experiencing something of a comeback. The problem is that in many subtle ways, analogue tape acts as a very gentle multiband compressor and also creates very slight amounts of harmonic distortion that the listener perceives as ‘analogue warmth.’ A digital recording does not and cannot do any of that; given a good set of analogue-to-digital converters, a digital recorder records exactly what has been fed into it. There should be no noticeable difference in the control room between what comes out of the multitrack and what comes out of the live room directly.

For years we have become accustomed to the sound of analogue circuitry, magnetic tape and even valves altering the signal, each in their own special way. This type of sound has become deeply associated with rock and popular music and now we use outboard valve compressors and pugins and other effects, multiband compressors in particular, to simulate the sound of tape.

But using effects and plugins to simulate tape suffers from two massive drawbacks:

1. A digital effect can only use one calculation. That calculation may be a very complex one and even try to simulate the non-linearity of the reactions of tape, but it is one ever-present calculation. And most tape simulation plugins and effects are not very complex; all they do is add some harmonic distortion and a little compression in a totally linear fashion: quite the opposite of tape. It is the guesswork of one software designer.

2. They are all too easy to over- or under-use. Turn the knob all the way to the left and you get nothing. Turn it all the way to the right and you get almost total distortion. Somewhere in the middle is the sound of real tape, but unless you have a real tape machine to compare it with, knowing where is totally hit-and-miss.

So a growing number of studios at all levels of the business are coming to the obvious conclusion and, instead of trying to simulate tape by various means, have bought themselves good, used tape machines. Whereas before the tape machine was the workhorse of the studio and was kept running all day and tape and recording heads suffered the wear and tear of shuttling back and forth for take after take, now they are often used just once and almost as an effect. The initial recording is made digitally in the usual way and then those tracks that are to be processed are played onto the tape machine and then re-recorded into the digital system - but this time with the added flavour of having been recorded to tape.

The Disadvantages of Tape

If you read some of the stuff I have written here (and some of the wild praise heaped upon analogue tape by those who are either too young or too old to remember what a pain in the butt it could be) you might be tempted to think that we were fools to ever opt for digital.

Well, before you decide to flog-off your state of the art DAW and scour the small ads for a mankey old reel-to-reel, let me remind you of a few facts.

Do you have one of those cassette recorders with two drives so that you can make copies? You do, good! Make a copy of a short section of a cassette. Now make a copy of that copy. Now make a copy of that. That is three generations of copies. What does it sound like? Compare it to the original - don’t tell me, it sounds dreadful! The background noise will be almost as loud as the music and the music will be totally distorted.

OK, that was a somewhat unfair test as cassettes run at 1 7/8 inches per second (ips) which is one eighth of the slowest professional speed of 15ips. The usual speeds are: 15/16ips (the old dicaphones using micro cassettes); 1 7/8ips (cassettes); 3 3/4ips (domestic reel-to-reel and the old 8-track cartridges); 7 1/2ips (better domestic and semi-pro); 15ips (professional) and 30ips (high quality professional).

But you get the picture: multiple copying and bouncing tracks down into convenient sub-mixes (e.g. mixing the drums together with the bass into a stereo pair to make the final mix quicker and easier) are to be avoided if possible. A good clean mix and then straight to the master machine gives the best quality.

Then there is the question of maintenance. Those heads only last so long before they need re-lapping (polishing down to give a good, clean surface once more) and after a couple of re-laps, they will need replacing. And not just the heads, all those drives, belts, clutches and pulleys will only last so long.

But to understand why tape alters the sound in the way that it does, we must understand how it works in the first place.

The tape recorder started out as a wartime German invention called the Magnetofon and the first machines recorded sound onto metal wire and later onto paper tape coated with iron-oxide. One of the uses was to messages via radio in such a way that Allied forces could not decipher them and this was done by playing the messages back at four times real speed. Obviously this method was not very secure (if caught in time, the messages could be recorded onto disc at 78rpm and this could be played back four times slower) so the German military opted for the Enigma Machine - which proves something about reliance on technology - but let’s get back to the tape recorder.

After the War, Telefunken’s Magnetofon gained the interest from the broadcasters and in Germany in particular. In Britain, the only broadcaster was the BBC and they opted for the far higher quality (at the time) of direct to disc recording.

The tape recorder (now known in Germany as the Tonbandgerät) works by writing little copies (analogies) of the sound waves onto the passing tape. There are three heads, erase, record and playback and they are arranged in that order as the tape passes from left to right. In the record head there is a circle of iron with a coil wound around it that has the signal pumped through it, rather like the signal is pumped through the coil of a loudspeaker. This causes a magnetic flux to be created in the iron circle. At the tape side of the circle, there is a very small gap (filled with a nonferrous hard metal) and a magnetic field is created at that gap. The molecules of the ferrous compound on the tape copy those magnetic fields as they pass and keep them.

The first attempts showed that those molecules were more compliant when there were high-pitched sounds in the signal and soon a very high pitched sine wave was incorporated into the signal to ‘jiggle’ those molecules so that they would be more ready to adopt the patterns of the magnetic flux. It’s a bit like tapping a magnetic compass that is sticking to make it swing round and show where North is.

This high-pitched sine wave is called the bias signal and somewhere between 120kHz and 180kHz. On a professional machine the volume and even the frequency can be altered according to the type of tape used.

The replay head sits close behind the record head and is optimised to give the best performance in replay mode only. The erase head gets a signal similar to the bias signal that wipes out any existing recordings on that track.

The tape itself comes in four widths, 1/4" 1/2" 1" and 2" and in the early days all kinds of professional and semi professional formats were tried. There were so many that it would difficult to list them all, but in the end, the professional standards were 2-track 1/4", 2-track 1/2" and 24-track 2", all running at 15 or 30ips. The 2-track formats are for mastering (storing the final mix) and the 24-track is for tracking (jargon for recording the original tracks).

The 2-track formats are sometimes called half-tracks and come with different gap sizes between the top and bottom tracks. The National Association of Broadcasters (NAB) set a larger gap than the Deutsche Industrie Normgesellschaft (DIN). These two august bodies also set different plug norms, reel sizes, output voltages, eq settings for tapes and anything else that needed a standard to be set. The DIN (as its name, which translates to the German Industrial Norm Society, suggests) set all industrial norms such as screw-thread depths, electrical supply voltages, paper sizes (A3, A4, etc.), oil viscosity standards and even the sizes of egg cartons.