The test tape contained the following items:

.

• 1) Tone recording on left only, right only, then left and right simultaneously. These were used for setting recording level accurately, and also for gaining an impression of distortion and wow and flutter.
• 2) Pink noise was recorded at a fairly high level to test stability (accuracy of positioning, etc.), frequency reponse and tendencies to compress the HF region.
• 3) A speech recording of the author's own voice recorded in an anechoic chamber. This is a very cruel but effective test of Dolby processing, stability, HF compression, distortion and record amplifier clipping problems. This recording also gave a good indication of record level metering characteristics.
• 4) A recording of a Steinway piano played by Tamas Vasary at a live concert recorded by the author at the Queen Elizabeth Hall, the piece being Beethoven's OplO No. 2 Piano Sonata, First Movement. This recording was used to determine transient stability, distortion, response, and the subjective effect of wow and flutter.
• 5) A pop recording, copied from a master tape "This is it" sung by Melba Moore. This recording was used to check the overall distortion performance of a loud pop track with sharp transients and strong sibilants.
• 6) After a pause without programme, the next track included a section of Tony Hatch's "Love in the Morning Sun", a light music track with very wide frequency response and encompassing light strings, bass guitar, drums and other solo instruments. This was used to give an overall impression on the subtleties of response at intermediate levels, although it also showed up low frequency response anomalies frequently.
• 7) This track incorporated a section of Stravinsky's Rite of Spring, with Pierre Boulez conducting the National Youth Orchestra which I recorded in the Royal Festival Hall. A very difficult section was chosen incorporating heavy low frequency, mid frequency and high frequency transients. The recording was made on the master tape at two different levels, very high and high, the peak levels being consistently 3.5dB apart between the two different sections of the same passage. This track enabled us to check each cassette deck at a very high input level, and it was noteworthy that the very loudest passage recorded satisfactorily on many machines, whilst on some
• the same passage sounded excruciating. The lower level recording continued on to a much quieter passage, again used for determining the recording characteristics at low and intermediate levels.
• 8) A recording was made on a stereo Nagra of underground trains entering and leaving Golders Green station to show, very clearly, transient positioning, very low frequency performance and high frequency compression, often noted when signals and points hissed as they changed. Many recorders showed bad HF compression on this track, whilst a few showed no compression at all.
• 9) The ninth and final track incorporated a copy, direct from a master, of Elton John's "Rocket Man", used frequently by the author because of its difficult high frequency sibilants and sharply percussive sounds.

.

Each subjective test was repeated in all tape positions where felt appropriate by the author (some ferrichrome tests were aborted quite early in a test because the switched position alleged to be suitable for ferrichrome was found to be inappropriate, in which case a comment is made in the review.)

Since we listened to 50 cassette decks, and some recorders were checked with up to four cassette tape types, the test programme was actually heard by the subjective panel perhaps 300 times over a period of three weeks, and so we are all heartily sick of it by now!

During each test, the reproduced quality from the cassette deck was repeatedly compared with that from the master tape played back in synchronisation, unless the deck was a 3-head type, in which case the programme was compared whilst it was being recorded.

Whenever a problem was detected an investigation was held to determine any possible causes, as an indication to the laboratory of likely problem areas for special examination.

The listening panel always included the author, and Paul Messenger, Roger Morley and Peter Willison also took part, sharing the burden generally.

Any poor points mentioned in the reviews were noted by at least two different people, and I am happy to say that there were virtually no disagreements ever about the problem areas, although the degree to which they were annoying was slightly variable at times.

I was particularly sensitive to frequency response anomalies, distortion, wow and flutter and dynamic range. Paul Messenger was particularly conscious of HF stability and positioning and transient performance, whereas Pete Willison was not quite as concerned as I was about flutter.

However, obviously, we were all very aware of any problem areas likely to be heard by the more critical listener. I mention these slight differences of priority since they are obviously important, and in the conclusions I comment on borderlines of acceptability.

During the entire subjective test programme either my hard-working secretary Barbara Meakins, or my ever-loving wife Fiona made notes on specially prepared subjective test forms concerning each recorder's behaviour - sometimes coping with an almost continuous running commentary from my colleagues and I.

At times our patience was sorely tried, especially with some of the DIN standard machines which sprouted incomprehensible DIN sockets like mushrooms. We also managed to have a few laughs here and there, although I am pleased to report that no machine emitted smoke or suffered a total breakdown (even if we got quite close sometimes!)

It was quite fun checking the ergonomics, for a few machines were strange indeed: one model actually incorporated a built-in clock which served no useful purpose whatsoever apart from telling the user the time! This model, not reviewed, might have been rather more useful if a few extra integrated circuits had been added to allow the clock to start the recorder at a predetermined time!

Another machine made some of us giddy when it was switched to rewind, since a neon light started whirling round and round in the opposite direction to the spooling! Other machines were discarded for serious design problems, or extreme difficulty in ergonomics, or because they were felt to be very poor value for money.

The title "Hi-Fi Choice" does, after all, imply reviews of at least reasonable equipment by current standards, but it is only right to include a few models that are generally unsatisfactory to give a full appreciation of the differences between good and mediocre.
.

The laboratory test programme was designed to examine the mechanical, electronic and compatibility parameters of each deck and also to determine the machines' performance on the appropriate tape types.

Compatibility with external equipment was felt to be extremely important and so tests were made on the impedances, sensitivities and clipping levels of all inputs and outputs. Noise levels were measured on replay and overall, and checks were made on input noise degradation (particularly relevant to the input sockets).

The CCIR weighting was used for all measurements, but unweighted replay measurements were also taken to show up any intrusive hum or tones present; where appropriate, a spectrum analyser was used to examine noise and distortion. Any interaction between different input circuits was noted and some machines showed variable gain, for example on their DIN input when the line input controls were varied, or vice versa.
.
A special cassette was made incorporating an internal record head for testing the replay amplifier performance. A carefully compensated and equalised constant current source was fed through this head to check on replay amplifier equalisation and peaking, and distortion and clipping margins.

Dolby or other noise reduction system tracking was also checked on different levels using this 'probe' magnetic flux cassette, made in the author's laboratories. Record and replay Dolby level calibrations were checked, both on the recorder's own meters and externally, to determine compatibility and output levels. The headphone output sockets were checked into 8 ohm and 600 ohm loads to check on headphone compatibility.

The DIN input was always driven via a 470 K ohm source resistance, with the capacity between this and the recorder's input equal to that found on an average lm long DIN/DIN lead.

Nominal DIN source level was stipulated to be 470 mV from a low source impedance applied to the input of the 470 kohm DIN source resistor. Sensitivities and clippings were related to this in dBs.
.

Phono input sources varied from 160mV upwards, as required for the different tests, and the input level required for a fixed flux level on the tape was determined, of course, for sensitivity.

Input noise tests were measured using a 10kohm resistor mounted in a phono plug for the line input, or screened DIN plug incorporating a short-circuited 470kohm resistor in series with the pins (ie the resistor being between the input pin and earth). Great care wae taken to avoid creating unnecessary earth loops, in order to reduce hum problems to an absolute minimum.

The CCIR weighted noise was measured with and without noise reduction on all tape type positions as appropriate, both overall and on replay. The dBs of overall noise reduction are quoted in each review as well as the weighted signal-to-noise ratios referred to Dolby level without noise reduction.

Distortion performance was measured between inputs and the monitoring point, from replay head to the monitoring point, and also via tape at Dolby level flux and at +4dB. Throughout this book, all tape recording levels are referred to the Dolby 'B' level of 200nWb/m, measured by the McKnight Method, whether the machine incorporated Dolby B processing or ANRS or SANRS. All noise levels and tape modulation levels are thus referred to this fairly high flux level.

Frequency response charts were taken with and without Dolby noise reduction at 24dB below a level equivalent to 200nWb/m at 400Hz.

Left and right channels were charted on all appropriate tape types. Replay response was checked statically using the probe cassette head method, and dynamically with some internally calibrated test cassettes that proved to be right down the centre line in response of both the very latest BASF frequency response cassettes and those found correct made in Japan (please see section dealing with the frequency response of cassettes for further details).

Replay azimuth was checked using a laboratory standard reference tape recorded at 3kHz and monitored with a Hewlett Packard gain/phase meter, and the outputs from this meter were fed into a storage oscilloscope to check on short and long term drift. High frequency stability was also checked by recording and playing back 10kHz through the same system.

Whatever the method adopted by the manufacturer, the record level metering, was checked by introducing a tone equivalent to Dolby level and then sending bursts of this tone every few seconds for 8mS and 64mS respectively, in order to determine meter ballistics and peak reading accuracy. The response of each meter was checked to see if it was reasonably linear or if it read the equalised signal passed to the record head, the latter being generally felt very inappropriate.

Wow and flutter tests were carried out with an EMT 424 wow and flutter analyser that takes readings automatically, thus eliminating human measurement error. These readings were taken at the beginning, middle and end of a cassette respectively, and the average of the 18 readings is generally quoted.

Wind and rewind times were checked on a C90 cassette. Various other mechanical tests were introduced where necessary, particularly in response to comments made in the subjective tests. Finally, erase and crosstalk tests were introduced at three frequencies, using spectrum analysis techniques to speed up measurements.

Equipment used included two B&K 2010 BFO/Analyser systems, B&K 2307 chart recorder, B&K 1901 and 1902 control systems, Gould Advance Digital Storage Oscilloscope, Hewlett Packard and Tektronix oscilloscopes, Hewlett Packard 3580 Spectrum Analyser, Hewlett Packard gain/phase meter and other measurement equipment by EMT, Marconi, B & K, Hewlett Packard, Sound Technology, Fluke, etc. Recorders were checked at 240V in the laboratory, derived from a Variac transformer.

The first and still generally regarded as the most successful system was devised by Ray Dolby in the late 1960s, and was first demonstrated to the public in the UK by myself, at the Music Trades Association convention in Bournemouth, and later at the Radio Communications Exhibition in London in 1970.

The domestic 'B' system, when set up properly in an appropriate design, is basically a hiss remover. High frequencies are boosted on record and reduced on replay to varying degrees, depending upon the dynamic range; whereas at the high levels virtually no noise reduction is present even at high frequencies, as the levels decrease, noise reduction is introduced at ever decreasing frequencies.

At very low levels, such as -40dB, noise reduction operates down to below 1kHz, but the full 10dB is only present above 2.5kHz or so. Since the main background noise in a cassette system is at high frequencies, the subjective effect is to reduce overall noise by 10dB.

A manufacturer incorporating, the Dolby 'B' system has to pay Dolby laboratories a royalty on every deck sold, and so a few other companies have attempted to devise noise reduction systems of their own.

It must be appreciated, though, that Dolby laboratories spent a fortune developing and promoting the system throughout the world, and no licence is required for the use of Dolby B in pre-recorded cassette manufacture.

Philips designed their DNL system for replay noise reduction only, but this system is generally regarded as unsatisfactory because it not only reduces hiss but removes most of any magic that might be present at high frequencies as well, giving dull, lifeless reproduction with severe hiss pumping. Therefore the DNL system can only be regarded as a hiss remover in pases where the recording would otherwise be totally unacceptable.

JVC have designed their ANRS system and more recently the Super ANRS system, but early versions of these produced a brittleness and noise pumping, which I found unacceptable on models reviewed in the first Hi-Fi Choice 'Cassette Decks'. As will be seen from the patent numbers stamped on the bodies of JVC cassette decks, they are now employing elements of the Dolby B circuit in their own ANRS/SANRS systems, which are now much better and offer reasonable compatibility (see JVC reviews).

Whereas the JVC ANRS system has a similar effect to Dolby B, the SANRS system reduces HF transients on record, but expands them on replay with very good effect on some types of programme material, but with a poorer effect on others^ such as piano.

I have found, however, that if a piano recording is made with SANRS it can sometimes sound better when played back with ANRS or Dolby B, since the higher 'noise chuffs' on transients which would otherwise be present, more or less disappear, although the transients, of course, are rather duller.

The dbx domestic system has also been shown with a cassette deck by TEAC, but the machine was extremely expensive, and I found the noise pumping on some types of program most annoying, even though the noise reduction was startling.

Today's best cassette tapes on high quality decks offer a very good dynamic range with Dolby 'B', but a splendid one should be available with pure iron cassettes, which are to be introduced at around Christmas 1978.

There can be no doubt that the introduction of the Dolby 'B' noise reduction system was entirely responsible for the cassette medium being taken seriously by hi-fi manufacturers, for cassette recording quality was transformed in the first 18 months of this decade.

There is one snag with the Dolby 'B' noise reduction system, and that is the need for the sound passing through the record processor to be at the same level and to have a very similar response to that passing through the replay deprocessing system.

For this reason, many decks incorporate record Dolby 'B' calibration pre-sets, which allow a recorded tone to be adjusted to play back at a Dolby B calibration level indicated on the recorder's meters.

Without adjustment a more sensitive tape will play back at too high a level and will be audibly slightly brittle, whereas a less sensitive tape will reproduce rather dully. The Dolby B system also exaggerates any frequency response anomalies, so that a 2dB fall at 10kHz subjectively sounds rather more like a 4dB drop. It is thus more important to ensure compatibility of tape with machine to achieve high quality recordings.

As part of the Dolby licence stipulations, all decks with Dolby 'B' have to incorporate a multiplex filter which not only removes any pilot tone, but also any frequencies beyond the audio range, which might otherwise affect the record Dolby circuits by decreasing the compression, but which would not similarly affect the replay processor reciprocally, since the frequencies would not actually be recorded.

If your cassette deck contains a switchable multiplex filter rather than a permanent one, I would advise you to use it unless you find no deterioration whatsoever in overall results without it. This will preserve good tracking between record and replay, provided the cassette tape type and deck itself are aligned properly.

In the subjective tests we listened to the wow and flutter present on a recording of tone at the beginning of the test, and later checked how much subjective wow was audible on a piano recording.

It was interesting that our subjective comments did not always tie up with the laboratory measurements, and so considerable time was spent in an effort to get better correlation. The accurate measurement of wow and flutter is not simple, and most test meters require the engineer to take an average reading when the meter is bouncing around.

An EMT424 wow and flutter analyser was used to avoid human reading errors, as this meter integrates the total wow and flutter over an approximate 5 second period and gives a fixed reading, which we repeated 6 times at the beginning, middle and end of a cassette tape.
.

The DIN peak weighting curve peaks up at between 4 and 10Hz, and falls off either side of this pass band. It is my opinion that this curve does not correlate sufficiently well with subjective wow and flutter of the type generally heard in cassette decks.

For example, any little tape judders are very noticeable, but do not contribute significantly to the reading; similarly a very slow wow may cause some listeners to feel slightly giddy, but insufficient account of this is taken in the measurement.

However, we found that moving around the room whilst listening varied the annoyance of the wow quite considerably, so we also tried listening to the wow and flutter on headphones, finding, generally, that it was much less annoying.

Somewhat surprisingly, there was better correlation with the measurements when listening on headphones. Thus, whilst measurements will show how good any machine is basically, please note any subjective comments, as these are also inportant.

Some types of cassette tape produced more audible wow than other types on average, and it was interesting that wow and flutter, and especially any form of scrape flutter, was more annoying when the dynamic range was wider.

Machines employing a combined record/replay head sometimes produce subjective dropouts or azimuth wandering, and this was occasionally more annoying, subjectively, than some of the measurements indicated. There is still much to be learnt about cassette tape guidance over combined heads, and tensioning problems sometimes caused exaggeration of various mechanical effects.
.

Some machines wound tapes very fast, making it difficult to back step a short way, whilst others spooled very slowly. Winding speed is rather a subjective matter, but spooling could be rather untidy and damage might be caused to some types of cassette tape if very fast. On the other hand, very slow spooling can, of course, be irritating.

Memory tape counters and types of tape position indicator are considered useful by some, but I have not placed too much priority on their functions as so many users are not too bothered with them. Occasionally, we were all very impressed (or unimpressed) with such a device and comments are made where appropriate.
.

There was considerable variation in the ease with which cassettes can be inserted and withdrawn, and in one or two cases the cassette itself became rather too warm inside the machine, and thus any print-through tendency of the tape could be exacerbated.

It is only fair to comment, though, that once one is accustomed to working a particular deck, cassette loading and unloading usually becomes relatively simple, even if your friends might get a bit confused!

It is sometimes useful to be able to transfer directly from play to wind, and later, back again, and this was possible on some machines (see text). A few allowed cueing on rewind, which can be very helpful if trying to find the beginning of a particular part of the programme.

Some machines have remote control facilities, but no one supplied us with a remote clock switching device. One model submitted incorporated a clock which was interconnected with the recorder for automatic starting, etc, in models supplied outside the UK, but because of the rather annoying BEAB regulations (British Electrical Approvals Board) all interconnections between the mains operated clock and the recorder had to be removed.

I sometimes begin to wonder if some of the BEAB regulations are getting much too finicky, and more or less tend to assume that every user is an idiot.

As an aside, I would point out that if somebody wants to kill himself he is not likely to make a point of pushing his finger round and round in circles inside a piece of electrical equipment in order to find the mains. (mains sind die 240 Volt Anschlüsse an der Wand)
.

It is important for the heads of all machines to be aligned with respect to azimuth so that they will record and replay tapes in a compatible way with other machines. A machine which has a head which is slightly out of vertical alignment will replay a standard test tape or a pre-recorded cassette with high frequency loss.

The azimuth of each machine was checked with a special test tape, and was adjusted if necessary, so that our frequency response cassettes were in alignment with the recorder. All further tests were made with the azimuth corrected.

Unfortunately, some prerecorded cassettes are themselves recorded slightly out of azimuth and so some differences between tapes may be detected.

Some machines having 3 heads have a user azimuth control on the record head, in order to give optimum azimuth between record and replay on any required blank cassette. Some machines required continual adjustment, which was annoying, whereas others required hardly any adjustment of this control, even when changed from one make of tape to another.

We checked the type of azimuth indication to see if it was effective and easy to operate. Since with the cassette tape medium one is dealing with recorded wavelengths of as short as 3 microns (1 micron is one millionth of a metre) it is obvious that a very small misalignment in the vertical angle of the record or replay head gap can have a very marked effect on the reproduction.
.

I must admit that I still marvel at the quality available on cassette tapes today, and the development of the system over the years has been a magnificent achievement of the industry, especially with reference to frequency response, dynamic range and general tape stability.

Notwithstanding this, it is important to be rational in making criticisms, since one manufacturer may overcome a particular problem area so much better than another.

The ear is not equally sensitive to noise at all frequencies, and so in the laboratory we used what is known as a CCIR weighting filter, which exaggerates noise present in the frequency region that is most subjectively annoying, while reducing the output level measurement in parts of the audio range where the ear is not so sensitive.

Unity gain at 1kHz was employed for all the filters used and RMS reading meters have been used throughout, since this is the standard we have established for some years in our laboratory.

Some cassette decks produce more inherent noise in their replay amplifiers than others, and this can have a significant effect in adding to the noise present on a recorded cassette.

Ideally, the replay amplifier should be 10dB quieter than the noise generated by the tape and record electronics, but few (wenige) machines were anywhere near as good as this.

However, most machines were adequate. I am concerned that some machines were not correctly equalised on playback to a replay equalisation curve now more or less agreed around the world (please see section on frequency response standards).

Machines incorporating more HF lift on replay, such as the Nakamichi 1000 II, will naturally be more hissy than those that are flat at 10kHz, and other things being equal, the additional hiss is about proportional to the amount of lift at HF.

When Dolby 'B' deprocessing is switched in, the replay amplifier hiss should reduce by around 10dB. Switching from ferric to chrome or ferrichrome equalisation on replay should reduce the hiss even more, by about an additional 4dB.

As well as checking replay noise in various equalisation positions overall noise was also measured, and whilst sometimes the noise levels were poor because of noisy replay and record amplifiers, a few cassette tape types were found to be significantly noisier than others, and this should be borne in mind when consulting the cassette tape section.

Unfortunately, some machines presented noise problems on the record input circuits, and in particular, almost all DIN input circuits produced more noise than the cassette tape produced on replay with noise reduction switched in.

In general, the newer decks reviewed in this survey had relatively good hum levels throughout. However, hum loops can be encountered when interconnecting the cassette deck with receivers, etc, and to experiment with connection leads and mains earthing to get the best overall performance is the best way to tackle any problems.

Sometimes, a hum loop can be created if the cassette deck is earthed to the mains as well as being connected to external equipment which is also earthed. Theoretically, earth loops should not present a problem, but in practice they can be a pest, but care must be exercised because if an equipment fault develops, it is possible to get a nasty electric shock.

Decks using just a 2-wire mains lead with a double insulated mains transformer that meets BEAB approval can often cause less aggravation than ones incorporating a mains earth wire.

Whilst the basic distortion caused by the tape medium is odd harmonics and odd-order inter-modulation, sometimes even-order distortions (ie. 2nd harmonics) can be present in the electronics.

The basic harmonic distortion of both record and replay circuitry have been checked and comments are made in the reviews if problems have been noted. 2nd harmonic distortion is not quite as annoying as 3rd harmonic, and it is, frankly, quite remarkable how much distortion the average person can tolerate before throwing his hands in the air!

Although 5% 3rd harmonic distortion at middle frequencies is easily noticeable, it need not be excruciating on program, and I have slightly changed my mind about the tolerable amounts of distortion at middle frequencies, bearing in mind the biasing conditions of the tape and its high frequency performance.

If a recorder is biased to give very low distortion at low and middle frequencies, it may well show marked HF compression, and we all tended to prefer an intermediate bias setting which gave approximately 2% distortion or so at +4dB, rather than a setting which gave significantly lower figures than this.

Some machines were clearly overbiased, producing amazingly low distortion figures on appropriate tape types at 333Hz, for example, but HF compression was almost always very poor in such cases.

However, normal chrome tapes gave such high values of distortion at reasonable programme levels that machines set for such tapes did not do very well subjectively, with relatively few exceptions.

We have measured distortion via tape at Dolby level and at +4dB, but comments are also made on the subjective distortion performance of each machine. Since tapes can compress quite badly at high frequencies, and in some cases the cassette decks could not even cope with high frequency transients, particular attention should be paid to comments on high frequency compression in the reviews.

Quite frankly, a substitution of a better cassette tape can make a world of difference to sound quality, and a number of cassette deck manufacturers were recommending what to me seemed inappropriate tape types for their recorders.

Some did not even want to recommend any tape at all, and this was most tiresome since we then had to spend considerable time choosing a reasonably compatible one ourselves.

If you use the cassette tape section guide, you should be able to find various types of tape that are similar in performance. But so many technical considerations in the cassette deck affect tape performance that listening tests on your own machine on different tape types must be advised, especially as no deck will be identically set up to another sample of the same model.

Since pure iron pre-recorded cassettes may be forthcoming one day, we have checked each recorder's capability of playing them back satisfactorily. However, hardly any decks currently available will be satisfactory for recording on the new tapes when they are available (NB: Tandberg and Philips reviews).

Bad distortion can be introduced if signal levels
are put into the recorder's input circuits which are above the maximum designed levels. An effect called 'clipping' is produced, and this is particularly marked if inappropriate use is made of a DIN input socket.

If the sound is completely clean on the monitor circuit whilst recording, then any distortion present on replay is likely to be produced in the tape itself, or perhaps in the record electronics.

If any distortion is heard whilst recording and monitoring the input, the deck's input circuitry is almost certainly overloading, providing the program source is clean. This may be caused by using the wrong interconnections or leads.

If the record level controls have a very low setting but the meters are indicating a high record level, there is probably an excessive input level. However, if it is necessary to have the record level controls at a very high setting the source levels are too low and hiss may be introduced.

We also checked to ensure that the noise reduction circuits were not adding distortion at lower levels, and most Dolby B circuits now incorporate distortion compensation to improve this.

Attention was also paid to distortion in the headphone circuits, for some machines gave problems with some types of headphone.
.

Various types of indicator can be provided to show the user the recording level being presented to the tape. The "VU"-meter was originally established just before World War II as a broadcast standard instrument, and all too many cassette decks incorporating so-called 'VU' meters in no way come up to the correct published standard for such meters.

They are intended to show the average level during any passage of music, but in no way will they indicate the level of short transient sounds accurately. Speech, for example, may under-read as much as 10dB, whereas a long continuous low frequency note may well read fairly accurately.

In order to give better meter accuracy peak-programme meters or indicators are used on some decks. These should show the highest level of transients, thus enabling the recording level to be set quite accurately, and, avoiding tape compression and overloading.

In my opinion peak-reading type meters should show the peak-level of the program being recorded before Dolby processing or equalisation, but some manufacturers prefer to indicate the peak-levels present on the feed to the record head.

In practice, this may tend to cause the user to record at a somewhat lower level than he might otherwise do, this was found particularly on the Eumig machine, whose meter was hitting the end stop on a tape that was not audibly distorting to any significant degree.

This meter is a typical example of one reading a massive treble boost, thus grossly exaggerating the program levels at high frequencies.

Peak-level indicators of one form or another are on most of the decks, and these light up when a particular level has been exceeded.

The Sony liquid crystal display was much liked by all of us, and was particularly interesting (model TCK 8B), although the price difference between this model and the almost identical model TCK 7B II is rather large.

In many cases, the peak reading indicators were set at inappropriate levels, and so comments are made on this. The tone burst test was introduced to ascertain how appropriately any particular meter read a typical programme peak or whether a tendency to severe under-reading was present.

Ordinary VU-meters usually presented Dolby calibration level at +3dB, whereas peak reading types had this level somewhat lower, or even did not indicate Dolby level at all.

An average reading meter, as will be found on most decks, will be indicating correct recording levels if the average programme is not allowed to reach more than the zero dB mark. However, many types of program may be over- or under-reading at this setting, and so on a particular machine I suggest that one should experiment with recording levels before attempting any serious permanent recordings.

The Dolby calibration marks were checked by replaying a standard Dolby level test tape made in my own laboratory, and in general most meters were acceptably calibrated.
.

Cassette decks usually have 3 separate output connections: line out (phono) sockets, the output pins of the 5-pole DIN socket, and a 3-pole stereo headphone jack socket.

The line output sockets usually present typical maximum output levels between 750 mV and 2V on an average programme.

Sometimes a gain control operates before the final output amplifier, but as often as not this control works on the actual audio output; some machines employing an output control after the final transistor stages run into clipping problems on program peaks, especially if very high recording levels are present.

It is far better to have the volume control immediately prior to the output stage, so that a greater overload margin is available. It is possible that in the next few years pure iron pre-recorded cassettes will become available, and if so, they are potentially likely to reproduce with considerably better quality than normal ones.

However, they will have up to 6dB more level at all frequencies on them, on average, and it is thus important that a modern cassette deck should be able to accommodate such tapes if they become available. Comments are made in the reviews on this, where appropriate.

The 5-pole DIN socket outputs, on pins 3/5, are sometimes at the same level as the line output sockets, but are often at a somewhat lower level and from a rather higher source impedance for better compatibility with DIN standardised receivers.

In general, unless you have good reason to use the DIN socket, always use the line-output phono ones.

The headphone sockets should be capable of driving all normal types of headphone from 8 ohm impedance to as high as 2 kohm impedance, as high quality models are available over this somewhat large impedance range.

Many cassette decks could drive low impedance phones satisfactorily, but were incapable of driving high impedance ones at a sufficiently high level. Sometimes clipping was audible on some types of headphone before the normal line outputs were distorting, and this is due to inappropriate headphone amplifier design.

Again, relevant comments are made in the reviews. Although the majority of machines employed 3-pole stereo jack sockets, one or two used DIN headphone sockets, which I found rather annoying.

I would earnestly suggest that manufacturers should standardise on the normal jack socket, which would make it less annoying for the average user, who will almost certainly be far more easily able to purchase headphones fitted with a jack plug than with a special DIN plug.
.

Usually, the output sockets present the input program whilst recording is taking place, although the DIN socket should be muted. Some machines, when the Dolby circuits are operating, present the Multiplex Filtered signal at the output, whereas others take the monitor circuit from before the Dolby filter circuit.

It thus becomes possible to use headphones, etc, whilst recording, and this can be most useful. Earlier JVC models employing the ANRS system used to present the process signal to the monitoring circuits whilst recording and thus no real idea of the quality of the input programme could be gained; fortunately, this has now been rectified in JVC's more recent designs.
.

Three types of input are normally available on a cassette deck:

1. microphone,
2. line input with phono sockets, and
3. DIN inputs.

Ideally, the line inputs should feed directly through to the record gain control but the microphone and DIN inputs require considerable amplification.

Unfortunately, microphones are so insensitive that their amplifiers require around 30dB more gain than the optimum DIN input socket needs, but all too many decks employ the microphone input amplifier for the DIN input as well.

In order to reduce the level of the DIN input sufficiently to avoid clipping the microphone amplifier's input circuit, the level has to be attenuated to such a degree before amplification, that hiss usually develops.
.

I have been somewhat hard on recorders having an inappropriately designed DIN input circuit which is more noisy (ie. adds more hiss) than the line input in almost every case.

The ideal situation would be for a manufacturer to incorporate a variable gain switch with a pre-amp operating at around 15kohms input impedance with a consequent level of around 15mV for DIN, increasing in gain by 26dB or so when the microphone jacks are inserted, and also disconnecting the DIN input.

Only European designed machines have, in general, optimised their DIN inputs properly, and some Japanese models add so much noise as to render the Dolby B circuits rather inappropriate!

Some decks have added too much gain after the recording level control in order to attempt to optimise the microphone/DIN input, even if they have incorporated a line/microphone switch.

The Sansui model SC1110, for example, attenuates the line input level down to just a few mVs on the record level slider, and this has then to be amplified up again with hiss (unless the input signal is at a high level itself), which allows the record gain control to be used at a very low setting and improves the hiss level by presenting a much lower source impedance to the succeeding stage.

Most recorders have inadequate sensitivity on their microphone inputs because of the attempted compatibility with the DIN input.

However, I must state that I abhore (verabscheuen) DIN 5-pole input standard anyway, which was originally designed at least 25 years ago for interconnections between valve receivers and valve recorders!

If I had my way, all DIN inputs would be withdrawn from cassette decks, thus properly optimising the microphone input and easing the line input compatibility by allowing less gain to be used after the record gain control. After measuring around 150 receivers in the last few years, I can categorically state that the majority of receivers are not fully compatible with the majority of decks, and results are almost always better when the phono sockets on both equipment are interconnected rather than DIN ones.

Worse still is the habit of using leads with phono plugs one end and a DIN plug on the other, for normally either high frequencies will be lost and levels will be severely attenuated, or severe clipping can result. If you do wish to use such a lead though, you can purchase DIN socket adaptors with built in resistors to attenuate signals, but surely this is rather ridiculous in this present age of high technology.

The DIN 5-pole socket uses pins 1/4 for record and 3/5 for replay, but note that on a properly designed DIN compatible recorder, pins 3/5 should be muted inside the deck whilst recording is in progress to reduce crosstalk at high frequencies between the output and input circuits.

Many decks don't do this, but some mute the line out phono sockets as well. Some recorders are festooned with DIN sockets which are totally incomprehensible to the average person unless a lengthy study is made of what I term the "destruction" book. Even after this, other members of the family are likely to be confused!
.

I know that this is one area in which I am prejudiced, and in reviewing machines having only DIN sockets I have attempted to remove my prejudices. But I am delighted to see at least two manufacturers, Eumig and Philips, get away from exclusive DIN standardisation by introducing phono sockets to meet world-wide demand outside Germany (and perhaps in Germany too?).

A recorder should have a microphone sensitivity of, ideally, around 150 uV to meet all normal live recording requirements, provided reasonably sensitive microphones are used. However, sometimes a user will want to record very loud sounds, so clipping levels as high as 30mV are desirable. A DIN input should be provided, theoretically, for 1uA current, which is
equivalent in voltage terms to lmV per kohm of the recorder's input impedance.

If the latter is below 10kohm or so, and the DIN source is at its usual very high impedance, hiss may be apparent. Although the DIN standard specifies a maximum sensitivity of 0.2mV per kohm, I would prefer to see this amended, since an input sensitivity greater than 0.5mV per kohm introduces so much hiss as to render the system rather ridiculous.

If we really must keep the DIN system then I would prefer to see levels of 5mV per kohm which would make life for the sensible designers very much easier; I cannot remember measuring any model which actually clips at anywhere near as low a level as this.
Line-in or phono inputs are basically flat, high impedance inputs intended for direct connections to low impedance outputs from tuners, amplifiers, receivers and other signal sources.

I do not like to see a maximum sensitivity greater than lOOmV, and most input levels presented to cassette decks average between 250mV and IV. These can easily be accommodated on all the decks reviewed, although not when using the DIN in/out 5-pole sockets.

All cassette decks incorporate a high frequency RF oscillator running at around 100-150 kHz, which is used to develop an alternating field in the erase head. This is required to erase any trace of a previous recording whilst a new one is being made.

A very small amount of this erase frequency is fed through to the record head via potentiometers of one form or another, and this current is called RF bias, or more simply - bias. Bias is required to enable the recording tape to accept audio magnetisation optimally, but its very presence has some undesirable effects on the overall quality.

If the bias is set too low for the tape being used, then low frequencies will be very distorted at high levels, whilst high frequencies may well be too shrill. Furthermore, the audio magnetisation will not go deeply enough into the oxide, and so surface variations will cause more obvious output variations, described aptly as "dropouts".

However, as the bias level is increased LF and MF distortion is reduced, but high frequency response gradually deteriorates. Above optimum bias the response falls very rapidly indeed as bias is increased, and in addition HF compression becomes noticeable.

Unfortunately, an RF bias setting for one tape may well be anything but optimum for another brand, and the cassette tape section refers to this in greater detail.

Very approximately, regarding the average budget ferric tape as zero dB bias, hi-fi cassette tapes require between 1 and 2dB more bias, whilst one or two other ferric tapes require slightly more still.

Ferrichrome types require at least 2.5dB more bias than budget ferries, about 1.5dB more than average ferries, while chrome and pseudo-chromes ideally require about 4.5dB more than average ferries.

The bias switch on the deck normally alters the bias appropriately for the different tape types, whilst the equalisation switch selects the appropriate curve.

Some recorders have their bias variable by the user, and if this control is moved in a negative direction, bias is decreased and high notes will be boosted, whereas when the control is moved in a positive direction, high notes will become more muffled whilst low ones become less distorted. Unfortunately, some modern types of record head become saturated at very high bias level, so when the audio current is passed through as well, distortion may result.

For this reason, all too many cassette decks cannot provide sufficient bias for ideal results in the chromium position, so sometimes bad distortion figures will be seen here usually due to this 'saturation' problem.

I have only rarely met with this problem in 3-head decks, where the record gap is somewhat wider.

When cassette decks and tapes were first introduced over twelve years ago in 1963, Philips worked in co-operation with German tape manufacturers BASF and AGFA to establish response test tapes which should have indicated the correct replay equalisation (originally at 1590/120usec.)

After a few years, it was realised that the originally designed 7dB bass cut at 50Hz on replay was ridiculous, and so by international agreement the time constant became 3180/120usec, which gives only 3dB cut at 50Hz.

The Japanese studied the original Philips specifications very carefully, and many manufacturers came to the conclusion that the BASF response test tapes were in error at high frequencies.

My own research led me to the opinion that the BASF test tapes had approximately 3dB too much level at 10kHz, and Japanese Teac and other test tapes seemed to replay more in accordance with, what seemed to me, a correct 120usec curve.

In the early summer of 1977 I published details of this controversy, and was backed by many manufacturers throughout the world.

At the time, BASF took up the cudgel by stating that their tapes were the original standard, but I disputed this, pointing out that the Philips written specification was the standard that most people accepted.

We have had, therefore, a situation where almost all European manufacturers have been adjusting their replay equalisation to the BASF test tapes, but virtually all the Japanese decks that I have reviewed in the last few years have been far more compatible with Japanese test tapes.

What is perhaps more serious is that prerecorded cassette manufacturers in the UK have been observing the BASF replay standard. So many pre-recorded cassettes have sounded rather brittle at lower and intermediate levels but compressed at high frequencies at high levels, since, if there is more treble cut on replay for the BASF curve, it is necessary to attempt to put more HF on the tape.

It is for this reason that many pre-recorded cassettes have such poor high frequency compression. The situation now would seem to be changing, in that the latest very expensive BASF frequency response test tapes, having frequencies up to 18kHz on them, fall virtually perfectly along a straight line equalisation up to at least 10kHz, with what I have always claimed as the correct equalisation. This seems ironic since they are surely admitting that I have been right in claiming that their earlier tapes were in error.

All the cassette decks reviewed in this book have been tested on replay with tapes conforming to the latest BASF standard with which I totally agree, and which incidentally, seems to be gradually being accepted by all.

The 3180/70usec replay curve required for ferrichrome and all chrome and pseudo-chrome types, and which is almost certainly to be used for pure iron replay, requires just over 4dB cut at 10kHz compared with ferric replay of 120usec, and thus the replay noise using 70usec should be up to 4dB better, thus giving a greater dynamic range potentiality provided the tape itself is sufficiently improved over normal ferric types at high frequencies.

Dolby level is specified as 200nWb/m using the American McKnight Method. Dolby level test tapes are available from Metrosound on the UK market, and are also exported throughout the world.

Such tapes should replay on the Dolby mark indicated on almost all meters. There is no recording standard equalisation for it is stipulated that the equipment should be equalised on record, in order to give a flat overall response at low and intermediate volume levels.

The amount of record equalisation necessary will, of course, vary from head type to head type, as well as from tape to tape. However, all recorders should now incorporate a 3dB bass lift at 50Hz in the record amplifier, to offset the standardised equivalent cut on replay.

All the measurements concerned with response and level in this tape survey, are related to the latest BASF test tapes, and my own international Dolby level calibration tapes that I supply to both Dolby laboratories and Metrosound, which thus should set the international standard originally devised by Ray Dolby himself.

.