.

In the tape-recording-machine field, cassette recorders may be likened to 35-mm cameras. Both require special care to approach the quality produced by larger-format equipment, while retaining the advantages of small size, convenience features, etc.

While tape overload (magnetic saturation) is the bane of tape recording in all its forms, it is nowhere so prevalent as with the cassette. All the parameters ofthe cassette medium-thin tape, narrow recorded tracks, very slow tape speed - add up to very little magnetic-oxide material per inch of tape, and very little material per unit of time. To skimp on oxide is to invite magnetic saturation, and this is likely to occur most often at high frequencies.

This is due to the treble boost applied to offset inevitable high-frequency losses and to fight the hiss that is rampant with cassettes. The combination record-playback head of most cassette decks is an additional complication. A head that is ideal for playback is itself prone to overload when recording high-level signals.
.

Many recordists - particularly those who tape live music - are finding that working with cassettes can be a game of blindman's buff. The game is blind because the recording-level meters, your hedge against distortion-producing levels, are not always to be trusted; dB numbers that are "safe" in one situation can lead you astray in another for a variety of reasons.

For example, too many of the meters have response characteristics approximating those of VU devices, and while this may sound admirably rigorous and "professional," this type of meter is not really appropriate for cassette-type recording.

True VU meters conform to a well-defined standard of ballistic damping. As a result, they respond to the recording signal in a kind of lagging, lingering manner. This is good, because it makes the meters readable. An undamped meter, responding to every brief spike of the musical waveform, would wiggle, shudder, and overshoot its way across the dial in a blur of movement. However, it's also bad because a VU meter is incapable of registering momentary peaks that might be just long and strong enough to cause audible trouble. This is very much a matter of degree.

At one time the VU meter was admired for its tendency to stolidly ignore a little bit of "blasting," since the ear is tolerant of very short bursts of distortion, and to distract the recording engineer with split-second overloads that no one would ever hear seemed self-defeating. But mind you, these rules applied only to the finest recording equipment of that time. And even so, a good deal of practical experience was necessary to interpret a VU's readings accurately and safely in every situation.

With cassettes, and particularly in live recording situations, a VU meter is definitely not the level indicator of preference. The cassette has a hair-trigger readiness to overload not shared by studio recording equipment. Signal peaks of very short duration must be taken into account; they can produce very audible distortion. And large peaks come along frequently in live recording.

When you dub a disc, FM broadcast, or any kind of second-hand program source onto a cassette, you are dealing with material that has been "pre-limited" - all the severe peaks have been chopped off (not audibly, one hopes). This stands to reason when you consider that a full-scale live performance may have a dynamic range well exceeding 80 dB, whereas a disc or tape recording with a 60-dB dynamic range is really pushing the state of the art.

Most cassette decks have been designed with copying chores in mind; and in this application, brief meter excursions up to 0 dB and a little above usually create no problem with standard tape. But the live cassette recordist quickly finds that, to keep the meter needle from going berserk on fortis-simos, he must choose levels that have it resting on its bottom peg, barely moving, for distressingly long perids. And even if he doesn't succumb to the temptation to inch the level controls up gradually (sheer disaster when a climax takes him by surprise), he may still not be entirely safe if the meters aren't fast enough.

The new breed of super tapes, however, demand more drive. If your recorder doesn't have a high-output tape bias switch that recalibrates the level meter, as well as a standard tape bias position, then musical climaxes should be recorded with the meter needle moving deep into the red (say, +2 dB). Another bias switch, often available, is for use with chromium-dioxide tape.
.

Of late the recording industry has been taking peak levels very seriously, and many of the newer consoles and mixing desks are festooned with light displays that register and hold peaks, and even "remember," for later reference, the maximum levels achieved during a recording session. Some of these devices have already begun filtering down to the consumer sphere, where they exist - in a simpler, modified form - in some cassette decks. Among these devices are:
.

These are warning flashers that supplement the meters (usually more-or-less conventional VU-type level indicators) by flickering on in the presence of sharp, potentially troublesome transients that the meters are too slow to register. Logically enough, LED's (light-emitting diodes), with their virtually instantaneous response, are becoming rather popular in this application. The levels that trigger these indicators* and the time constants they follow, are discretionary with the manufacturer, who has presumably taken the trouble to find out what types of signal are likeliest to cause audible distortion with his machine.
.

As noted earlier, an unaided meter fast enough to respond accurately to brief musical transients would have to be so under-damped that it would whip all over the dial in an undecipherable fandango.

But if a conventional meter is driven by its own little amplifier, the time constants of the meter's response to the recording signal can be controlled electrically for whatever result is desired. The combination of a meter and a driving amplifier with a fast attack time and slow release time (the electronics hold onto the transient somewhat longer than it actually lasts) produces a reasonably valid peak-reading indicator. (Fig 1.)

One manufacturer's device of this type gets to within 2 dB of actual value for a signal lasting only 50 milliseconds. (Compare this with a VU meter's attack time, which is somewhere around 2 to 4 tenths of a second.) Another manufacturer advertises an attack time of 20 microseconds (and release time of 70 milliseconds) for his meters' amplifiers.

Because of their slow release times, peak-reading meters tend to display a somewhat compressed picture of the actual signal dynamics. This tends to make them fairly readable. At the same time, since they ride along the peaks of the musical waveform, ample indication of excessive levels is given.
.

A conventional recording-level meter reads the "flat" signal as it comes from the inputs of the tape recorder. The signal applied to the tape, however, has received a strong high-frequency boost to offset hiss and treble losses. This means that the tape can easily get into trouble with excessive high-frequency levels that the meters know nothing about. The obvious remedy is to have the meters monitor levels after the recording-equalization stage - in other words, equalized meters. As a rule, this requires separate driving amplifiers for the meters, which may or may not have peak-reading time constants. Also, the meter movements should be able to tolerate substantial amounts of high-frequency energy without burning out.
.

A limiting circuit controls the gain of the recording amplifiers, typically acting only when a preset threshold level (usually a level that produces 1 to 3 percent distortion) is reached.

A good limiter just won't let the signal going onto the tape get any stronger than that, no matter what the level at the inputs. (Fig 2.) All the worrisome transient peaks are thereby simply eliminated.

An audibly unobtrusive limiting circuit has to be designed with a rather fast attack time and a very slow release time - several seconds or so. Otherwise the decaying tails of cymbal crashes and similar abrupt, loud sounds would disconcertingly "bob up" in level.

Therefore, any limiter which is working (in many live-recording situations, it will work quite a lot) is going to introduce considerable compression in the recorded signal. And since no limiter found in consumer equipment is likely to be sophisticated enough to be inaudible under every circumstance, there should be a switch to take it out of the circuit when it's not wanted.

But how does the ear react to being deprived of the high-level musical transients we tend to associate with excitement and "realism" in reproduced sound?

Up to a point, the ear is forgiving. The action of a limiter will be heard if it is allowed to affect the average levels of the program. But if the recordist chooses levels that result in only the briefest clipping of peaks, the limiter may very well go unnoticed in the final recording.
.

All over the recording world, the perils of high-level transients have been a little slow in receiving acknowledgment. In fact, none of the developments that made the cassette a high-fidelity medium has any direct bearing on the tape-overload situation.

The "B"-Type Dolby noise reduction system is a help only because it permits lower recording levels overall while still retaining a good signal-to-noise ratio.

Chromium-dioxide tape is somewhat more resistant to overload distortion than iron oxide, but it's clear from the proposed standards for CK)2 playback equalization that its potential is generally being exploited in another way - to increase the recorded levels of high frequencies and thereby further improve the S/N. This is probably the correct priority to follow for most cassette users. However, it eases the plight of the overload-prone cassette medium hardly at all.

Complete mastery of reading level meters and other devices in every situation probably won't come until you've had a chance to assess what went wrong with a number of previous recordings. Which is another way of saying that there's no substitute for experience in tape recording, as seasoned recordists are already aware.
□