Analogue multitrack sales are in decline with many people turning towards the new generation of low-cost digital tape machines and affordable hard-disk based systems. Paul White weighs up the strengths and weaknesses of the currently available systems.

Ever since multitrack recording was conceived, the standard recording medium has been analogue tape, and the technology of analogue recording has been developed to such an incredibly high degree of refinement that it is unlikely that any further significant improvements will be made — a fact which many of the major tape machine manufacturers freely admit. Over the past couple of years, the major breakthroughs have been made in the field of tape noise-reduction systems, the most notable advances being the introduction of the Dolby SR professional process and the less-sophisticated Dolby S system, which was originally designed for consumer audio but has now found its way onto semi-pro multitrack machines.

Noise reduction has been around for a while, but these latest developments enable analogue tape to produce a dynamic range closer to that of 16-bit digital recorders with minimal side-effects. That being the case, why do people want to move up to digital machines which, on the face of it, offer similar subjective sound quality for around two or three times the price? The answer is that digital recording does have some very clear and definite advantages — and a few disadvantages. The purpose of this article is to put the advantages and disadvantages of each format into perspective to help you make the right choice for your particular application.

Analogue Multitrack

Any tape system, analogue or digital, has the inherent disadvantage that the tape is in physical contact with the heads and guides of the recorder, with the result that both the heads and the tape suffer wear with repeated playing. The other shared disadvantage is that to get from A to B on the tape, you have to wind the tape from one spool to another, which takes a finite amount of time.

The fidelity of an analogue system is dependent on both the electronic design of the recorder and the mechanics of the tape transport, not to mention the quality of the tape itself. If the tape is not drawn across the heads at a perfectly constant speed, then adverse effects such as wow and flutter result, and if these are not kept to an absolute minimum by precision mechanical design, they will show up on critical sounds. Solo piano is a good test for wow and flutter, as it has no natural vibrato whatsoever.

Assuming the electronic design is sound, and most is these days, the main limiting factors regarding fidelity are tape speed and track width. For a given tape type, the faster the tape passes over the heads and the wider the tracks, the better the sound. This is fairly logical when you think about it, because the recording is made up of a statistical summation of the magnetic charge of all the magnetic particles passing over the head at any given time. The faster the tape or the wider the tracks, the more particles are used to represent the sound — which means the result is more accurate. If you're into photography, you'll find a parallel between tape noise and the graininess of film which, interestingly, is related to film speed.

Every time you double the tape speed, you achieve a theoretical maximum improvement in noise of 6dB, and the same is true when you double the track width. Narrow format open-reel machines and cassette multitrack systems have to employ some form of noise reduction, otherwise the recording would simply be too noisy. From this basic fact, you can see why professional machines employ wide tape running at high speeds, while budget home recording equipment uses standard cassettes, often running at twice their normal speed.

All noise reduction systems have side-effects — there's no such thing as a free lunch, not even an electronic one. Even so, the latest systems, particularly Dolby SR and Dolby S, provide a very significant and worthwhile reduction in tape noise with the barest minimum of audible tradeoffs.

The older Dolby C system is slightly less effective, though still very worthwhile, with the dbx noise reduction system offering perhaps the greatest improvement in signal-to noise ratio but with, arguably, the most serious side-effects. Early machines equipped with dbx were accused of 'noise pumping' or 'breathing'. This describes the effect whereby some background tape hiss is audible during bass sounds that don't have enough high-frequency content to hide the noise. Design refinements have minimised this problem on newer machines, even on cassette multitracks, but even so, the recorders that work best with dbx are the ones that are good enough to work acceptably well without it.

Machines equipped with Dolby C typically produce a dynamic range in excess of 75dB, while Dolby S and SR can produce over 82dB, depending on the tape machine. In excess of 90dB can be achieved with dbx, but once you get to this level of performance, any further improvement is academic, because you'll be recording more noise than that from your mixer, effects units and electronic musical instruments.

That sounds like a convincing argument in favour of analogue, but the bad news is that copying a recording from one tape track to another causes a loss of quality, and when tracks are bounced, that's exactly what you are doing. On a professional machine, the quality loss is very small, whereas on a cassette-based home-recording system, even one bounce can cause a noticeable drop in quality; there's a loss of clarity and an increase in tape noise.

Switching analogue tape from one machine to another can be problematic, as small differences in the mechanical head alignment will result in a loss of sound quality. There's also the problem of whether the original tape was recorded on a machine with IEC or NAB equalisation.

What definite positive attributes can we ascribe to analogue recording? For a start it's been around for a long time, so we know it works — and it's relatively cheap compared to digital machines. Analogue tape can be chopped up and spliced if need be, and we know that, under the correct conditions, recordings can be kept for many years with minimal deterioration. Also, you can overdrive analogue tape by a significant margin before audible distortion sets in, and then it comes in progressively, providing a kind of safety net or margin of error. There's no such headroom on a digital system — once all the bits are used up, the signal is clipped, and unless the periods of clipping are extremely short, the result is audible and most unpleasant.

Other, admittedly less important, aspects of analogue recording are that you can turn the tape over and play it backwards, should you want to, while varying the playback speed is simply a matter of changing the motor speed.

Should an analogue recording suffer from dropouts caused by flaws on the tape coating, the problem is instantly obvious when you play the tape back. Furthermore, any deterioration that occurs during storage is usually progressive, which means that old recordings can often be cleaned up with a little discreet signal processing.

Digital Tape

We all know what 16-bit digital audio sounds like through listening to CDs, or using DAT machines. It has a typical dynamic range of 92dB with a theoretical maximum of 96dB which, at first sight, makes it look significantly quieter than analogue tape fitted with any type of noise reduction (with the possible exception of dbx). However, to take these figures at face value would be to forget that digital systems have no headroom at above 0VU, while some analogue systems can tolerate brief excursions into the red of up to 10dB with no disastrous consequences. In order to allow a suitable safety margin when working with digital multitrack, we need to set our nominal record level at around -10VU so that when an unexpected, loud peak does come along, it won't drive the machine into sudden and horrible distortion. Because we have to allow this safety margin, the usable dynamic range is going to be 10dB or so less than the maximum quoted value, making a figure of around 82dB more realistic. This brings the practical performance of digital tape, so far as noise is concerned, directly into line with a Dolby S analogue system.

But don't digital systems produce less distortion? Digital systems work by measuring tiny slices of the incoming signal, either 44,100 or 48,000 times a second, then storing these measurements as a series of binary numbers. Each step has to be measured to the nearest whole number because there's no such thing as half a bit in digital audio — it's either a 1 or a 0. This immediately introduces a mechanism for producing distortion, because every slice, or sample, is rounded up or down to the nearest bit. Even so, a 16-bit system can sound very good because the number of measurement steps is 2 to the power of 16, which works out at 65,536.

That's fine if the signal being recorded is large because it will use most of the available bits, but a low level signal will be represented by fewer bits, making it less accurate. In other words, with digital audio, high level signals are recorded very accurately, (until the crunch point at 0VU), but lower level sounds become progressively more distorted. That's exactly the opposite of analogue tape, which is most accurate for low level signals, while higher level signals are more distorted. Ironically, because analogue tape distortion is predominantly second harmonic, small amounts tend to flatter some musical sounds, to the extent that many recording engineers bemoan the lack of it when working on digital recorders. This subtle distortion is part of the reason why analogue machines are often described as 'warm' sounding.

So far then, it seems to be a case of swings and roundabouts, but digital recording has a couple more aces up its electronic sleeve. For starters, digital recorders require no noise reduction, so there's no need to tolerate the side-effects that noise-reduction systems produce, and furthermore, because digital machines are crystal controlled, the amount of wow and flutter is effectively non-existent. The digital data is read from tape into a memory buffer from which it emerges at a precise rate, so minor fluctuations in the transport speed have no effect. There's no problem with different recording EQ curves as there is with analogue.

You can also copy data from one digital machine to another (or from one track to another on the same machine) entirely in the digital domain with no loss of sound quality. This is of particular value when mastering to a medium such as DAT, because a second DAT machine will allow you to make backup copies which are exact clones of the original. This is less of an advantage than it might appear when it comes to multitrack, though; most track bouncing will have to be accomplished in the analogue domain via a mixer, so some degree of signal degradation is inevitable.

Before drawing any firm conclusion, it would be wise to look at what happens when something goes wrong — such as when a tape becomes worn or a machine is slightly out of alignment. With an analogue machine, alignment errors usually lead to a fall-off in the high frequency response of the recording, while dropouts caused by dirt on the heads or worn tape take the form of brief drops in level. Digital machines deal with absolute numbers, and because some errors are inevitable due to dust on the tape and minor manufacturing imperfections in the tape itself, powerful error correction systems are built in.

Small errors can be repaired completely, as the format uses a system of redundant data to determine the missing numbers and replace them, while more serious errors require the machine to make a guess at the missing data by examining the good data on either side of the error. If longer errors occur, then the machine doesn't know what to make of the situation so it mutes the audio output until the problem is over. This is a major disadvantage, because the first you know of a problem is when the audio starts dropping out — there's no gradual loss of sound quality to give the game away. One day a tape might appear to play perfectly, while at a later date it might exhibit dropouts. Furthermore, head-wear on the recorder could have the same effect; one day the error correction system will be able to cope — the next day it won't.

The real problem is that these are early days and low-cost digital multitrack hasn't been around long enough for us to know how well it will hold up. On the positive side, the head life should be significantly longer than for analogue machines and there's every reason to believe that head replacement will be less costly.

Some lesser disadvantages you may want to consider are: digital tape can't be edited using razor blades (at least not on the machines we can afford) though in a multiple digital machine environment, editing can be achieved by copy and paste from one machine to another with no loss in quality; you can't play the tape backwards to create reversed sounds or backwards reverb; and you can't play the tape at double or half speed to produce special effects — though there is some degree of variation via the varispeed controls.

Hard Realities

The last medium I'd like to look at in this article is hard-disk recording, which is finally coming down in price sufficiently to attract the home user. This is a spin-off from the computer industry, which has been using hard disks for many years to store program data, and the operating principle isn't dissimilar to that of the popular floppy disk. The main difference is that the hard disk holds a lot more data, but it is also significant that the head is spaced just above the surface of the disk rather than in physical contact with it, which means that there should be no wear. However, as hard disks spin continuously from the moment you switch on to the moment you switch off, the bearings do wear out eventually.

"Hard disk offers the most benefits in that you get a combination of digital sound quality and a supremely flexible system for editing."

Standard hard drives are known as 'fixed media' , which means you can't take the disk out. You can buy removable disk systems where the disk itself comes out in the form of a cartridge, but these are, naturally, more expensive. Magnetic drives of this kind are available in 44Mbyte and 88Mbyte versions, while magneto-optical drives can be many times larger.

Recording in this way demands a lot of disk capacity, and as a rule of thumb, one track-minute of audio eats up around 5Mbytes; that's 10Mbytes for a minute of stereo, 20Mbytes for a minute of four-track and so on. The recording format is exactly the same as that used on digital tape, so given that hard-disks cost vastly more than a reel of tape, why use them? The short answer is random access — a process admirably explained in the Digital Basics II feature in the December 1992 issue of RM. With just a little electronic trickery, it is possible to jump, quite seamlessly, from one part of a hard disk to another in real time, allowing the recorded material to be played back in a different order to that in which it was recorded — which is why hard-disk editors are so powerful. You simply mark your music up into sections such as chorus, verse, intro, bridge and so on, then reassemble the parts to produce a brand new arrangement, quickly and seamlessly.


This feature also allows digital audio to be integrated into MIDI sequencing packages, because anything you have recorded can be accessed in an instant and arranged in the same way as you arrange your sequenced MIDI patterns. If you produce your music mainly using MIDI instruments, this is a convenient way to add high quality audio such as vocals. Even a system offering only a few minutes of recording can be deceptively powerful, because much of what you record can be used several times. For example, if you record a song on tape, you have to sing it all the way through, whereas on hard disk, you need only sing the chorus once because you can use the same one as many times as you like!
It seems that the practical advantages of hard disk outweigh tape, purely because any material can be accessed in an instant with no need to wind back and forth. Furthermore, in a tape system, the tracks are recorded side by side on the same piece of tape, which means you have no flexibility to move things around; with a properly designed hard-disk system, you can run any part of any track alongside any part of any other. For these reasons, I feel certain that the future of multitrack must rest in some form of hard-disk system, but at the moment, there are several significant drawbacks, mainly to do with cost.

Tape recorders have the massive advantage that tape is cheap, and once a reel or cassette is full, you just put it away and start on another. Hard disk systems using fixed drives can hold up to an hour or more of stereo, but once the drive is full, you can't do any new work unless you erase the disk or back up the data in some way. DAT backup systems are available — indeed I use one with my Sound Tools system — but an hour of stereo audio takes an hour to back up and another hour to put back onto the hard drive when you want to resume work. And that's only stereo — a four-track recording would take twice as long to back up.

Removable drives look far more attractive from a purely practical viewpoint, but the basic drive units cost significantly more than fixed drives, while the cartridges themselves cost far too much to be used as permanent storage; a thousand pound's worth of cartridges might still hold less data than a digital tape cassette costing just a few pounds! Apart from their obvious editing advantages, current systems are only really suitable for short MIDI/digital audio projects or post production work for video, such as soundtracks for TV commercials.

Having seemingly put the financial dampers on hard disk recording for general music use, the future is always just around the corner. Sony are currently launching their new MiniDisc player/recorder which uses a data compression system to record over an hour of reasonably high quality audio (better than analogue cassette, but still not quite as good as CD or DAT) onto a tiny removable disk smaller than your average floppy. What's more, the format looks very promising for semi-pro multitrack, with somewhere between four and six tracks per disk being theoretically possible. There have been no formally announced plans for MiniDisc Portastudios, but if the format becomes established, then I feel they will eventually materialise. And as the recording medium is a consumer format, the blank disks cost under £10. A four-track based on this technology would give something like half an hour's recording for £10 — a far cry from a couple of thousand for a large-capacity hard drive. Though the data is compressed and is admittedly inferior to CD in terms of sound quality, the subjective difference is surprisingly small. As a substitute for tape in home recording systems, this format or one very like it has a lot going for it, but the compromises made necessary by the data compression mean that the audio quality is not up to professional standards. At the time of writing, MiniDisc (and the competing Digital Compact Cassette) systems cost rather more than budget DAT machines, which means that DAT is still the preferred mastering format.

Summary

For the majority of applications, analogue tape running with Dolby S will give the same subjective quality as a digital machine for a lot less outlay. In fact it can be argued that the vast majority of noise that ends up on a recording is nothing to do with the tape, in which case a lesser noise reduction system such as Dolby C is also quite adequate. If you don't believe me, check out one of your own recordings and listen to the change in hiss level as you move from blank, unrecorded tape to the supposedly silent section just before or just after one of your songs! The noise really comes from your mixer and your outboard gear, not to mention noisy guitar amps, hissy synths and rustling lyric sheets.

Digital recording (and Dolby S analogue, for that matter) can only deliver its true potential when every other item in the system is chosen for its low noise and high fidelity. What you actually gain from digital tape, in practical terms, is the absence of wow and flutter and less quality loss when you bounce tracks. Multiple digital machines under central control can also provide some of the copy and paste editing facilities of a hard disk system, albeit at a slower rate of working.

But when you consider how much cheaper analogue machines are, you could buy more tracks for less money, enabling you to do away with the need to bounce altogether. This is particularly relevant to those people using multitrack tape alongside MIDI sequencing, as the majority of the MIDI stuff never has to go to tape anyway unless you run out of polyphony on your MIDI instrument.

There is little to choose between the operational ease of analogue and digital tape machines, in that you still have to wind through the tape to get from one part of the recording to another. However, you can't edit digital tape with a razor blade, as previously mentioned.

Hard disk offers the most benefits in that you get a combination of digital sound quality and a supremely flexible system for editing and rearranging your material without having to make any permanent changes to the original recording. Instant access means that sections can be played in any order, and the inclusion of DSP (Digital Signal Processing) in systems such as Digidesign's Sound Tools enables the user to call upon such facilities as digital equalisation, sample-rate conversion, compression, gating and time manipulation. Current systems are expensive, especially those using removable media, and backing up to DAT has to be done in real-time. Consequently, the most practical uses of current disk-based system are editing (including album compilation), short soundtrack jobs and providing one or two audio tracks to run with a MIDI sequencer.

The whole point of this article is not to promote one system above any other, but rather to make you think about what you really need before you go out and spend a lot of money on something that might not answer your problems. I know that a lot of newcomers to recording go into their dealers and enquire about hard disk systems because that appears to be the state of the art. It may well be, but when you consider the price of the recording medium, it only makes sense for very specialised applications. Likewise, a digital tape recorder seems like a very nice idea, but my feeling is that a lot of inexperienced purchasers may be disappointed when they find that their recordings are just as noisy as they ever were for all the reasons I've mentioned.

Current open-reel analogue recorders are capable of excellent results and many a hit record has been produced on so-called semi-professional gear. Just because analogue is old technology, don't write it off. The fact that in an ever-changing world it represents at least one known quantity could be a big point in its favour.

For more background on current recording formats see also:
Tascam 238S Dolby S cassette multitrack recorder review, August 1992.
Fostex R8 and Tascam TSR8 analogue 8-track reel to reel recorder reviews, August 1992.
Dolby S Noise Reduction Explained article, August 1992.
Alesis ADAT digital 8-track tape recorder review, September 1992.
Philips Digital Compact Cassette recorder review, October 1992.
Digital Audio Basics articles, November and December 1992.
Tascam DA88 digital 8-track tape recorder review, March 1993.

Back issues are available at a cost of £2 each, including postage and packing, from RM Back Issues, Recording Musician, £5">(Contact Details).