THE LOWDOWN ON GETTING GOOD SOUNDS FROM CASSETTE DECKS.

With the very best analog tape decks it is possible to get a recording quality that is as good as digital. Unfortunately, such decks, which run at 30 ips and use extra wide tape, are more expensive than DAT machines. Since most of us cannot even afford DATs, our problem is more prosaic: how to get acceptable sound from a cassette.

The cassette has two strikes against it. The tape is very narrow, and it moves very slowly. The effect of both factors is to reduce the amount of signal that can be recorded. Since the tape has a certain amount of noise built in, the basic quandry is to get the signal as hot as possible (to be louder than the noise) but not distorted due to tape saturation. To really understand what is going on, we need to look at three factors that are normally hidden from the user-- tape bias, equalization, and azimuth.

BIAS

All analog tape is recorded with a high frequency AC bias added to the signal to reduce distortion.

Changing the amount of bias has several effects on the result. As bias is increased, the level of the output increases up to a point, then falls off again. This effect is more pronounced at high frequency than at 1000 hz, so the overall frequency response changes. Also, as bias increases, distortion decreases, and then begins to creep back in. By the time distortion bottoms out, the signal level has usually peaked and fallen 1.5 to 2 dB. The tightness of the distortion curve depends on the brand of tape-- some require a very precise bias adjustment to give best distortion, others are more forgiving.

Bias is one of the things that is changed when you set the deck for type I or type II tape. This is often automatic, (keyed by a cutout on the cassette shell) but if it is not, you must get it right for good sound. Many cassette decks allow the user to make a small adjustment in the bias level. Generally this is used to trim up the high frequency response-- increasing bias reduces the highs.

Since bias is just a very high frequency signal, it is possible for the highest partials of the music to act as additional bias and paradoxically reduce the high frequency response. The Dolby HX PRO feature is a circuit that compensates for this, adusting the amount of bias according to the spectral content of the recorded signal.

EQUALIZATION

As a hedge against noise, the signal recorded on the tape has a high frequency emphasis. When played, the electronics reduce the high end to normal level, effectively filtering the tape noise.[1] Unfortunately, this has the side effect of making highs very likely to distort. For standard cassette tape this emphasis starts at about 1 khz[2]. With premium tape, which is quieter, the emphasis is moved up an octave [3] so that the highs will not hit saturation so soon. This is also set with the type I /type II switch, which must be checked for both recording and playback.

NOISE REDUCTION

Noise reduction systems are an elaboration on the equalization concept. All versions depend on ENCODING before recording and then DECODING upon playback, so you must be sure to do both according to the same rules. There are two basic types of noise reduction, each with a series of variations.

DBX

In the dbx system, the signal is compressed by 2:1 when recorded, and then expanded upon playback. This definitely reduces noise by 20 or 30 dB and also prevents distortion, but it can have two bad side effects. Sometimes you can hear the noise "pumping" (especially on bass tracks), and if there is any loss of high frequency signal during the recording, (common enough in cassettes) the decoding will not be accurate. DBX type II addresses the second problem by keying only on the mid band signal. This is usually all right, but material that is principally high or low may be distorted because it is read as no signal during encoding and amplified.

The best feature of DBX is that unrecorded sections of tape are silent on playback, so it is popular on multi-track decks. The down side is that a DBX encoded tape sounds horrible if played without decoding.

DOLBY

The dolby systems also compress upon recording and expand during playback, but they do this within restricted frequency bands:

Type A was the original professional Dolby system. It is still in use, but is being replaced by SR, a more sophisticated design. Both of these are too expensive to consider for cassettes.

Type B was the first cassette noise reduction system. It gives a 10 dB reduction in hiss, which effectively cuts the noise in half. Almost every cassette deck made includes Dolby B. Type C came along about 1980, and gives more extreme reduction, along with some features to make adjustment of the deck less critical. The overall noise reduction is about 20 dB.

The latest flavor of Dolby is type S. It is only available on high price decks, but does a very good job of quieting the tape.

Probably the best feature of Dolby is that B and C encoded tapes are usable without decoding. The high frequency response will not be right, but the sound is not too awful. As a matter of fact, playing an unencoded tape with type B decoding will sound quiet, and lots of decks don't have much in the way of high frequency anyway. The worst feature of B is that if the deck does loose highs in the recording process, the Dolby circuit will make things worse.

AZIMUTH

If the signal on a recorded tape were visible, it would look like a series of short bars perpendicular to the tape edge, rather like a tire track. This distance between the bars depends on the frequency and tape speed, and is in fact the wavelength of the signal on the tape. For a signal of 15khz on cassette tape, this distance is 0.000125 inches or about the width of a fine hair.

Azimuth is the angle of the tape head to the tape. If this is not the same for playback as it was for recording, the gap (which is the part of the head that actually detects magnetism) would straddle two bars, reducing the amount of signal. (The spaces between the bars represent the negative parts of the cycle.) Thus azimuth errors tend to reduce high frequency response. This is purely a mechanical matter, and is primarily affected by the cassette shell. High quality tape will be in a good shell, cheapies often are not.

SO, HOW DO YOU GET DECENT SOUND OUT OF A CASSETTE?

1. Spend some bucks and get a decent deck. The ninety dollar wonders usualy are very poor in the mechanics that assure consistant azimuth, and probably add flutter (tape speed variations caused by a not round or bent capstain) as well. Don't go overboard, of course. A deck priced in the $200 to $300 range will probably be fine. Features to look for are adjustable bias, HX Pro, and Dolby S noise reduction.

2. Use good tape, and always use the same kind. You will have to experiment with several brands to find the one that is best on your deck. Cassettes with "bridges" in them to improve tape alignment keep the most consistent azimuth. Tapes wear out faster than you think, so don't reuse old tapes for important recordings. After thirty or forty rewindings the oxide begins to flake off and the lubricant that lines the walls of the shell (there aren't any reels in there) goes away.

3. Keep the deck clean, and protect it from bumps. With wavelengths the size of a hair, it doesn't take much grunge to reduce high frequency response. Some brands of tape advertise head cleaning leader. This can't really do much, and whatever it does do will only happen the first few times the tape is played. Clean the heads with 91% isopropyl[5] alcohol. Azimuth will not change by itself, but it can when the deck is dropped, or if a cassette is forced in or out.

4. Set the deck for the proper tape type. Often this is a switch, but many decks select between standard and Type II/metal tape as indicated by a notch on the shell right next to the record protect tab. Be sure you never cover this notch. Even if the deck switches automatically, you have to use a switch to select metal or type II.

5. If the deck has adjustable bias, set it so the highs sound correct. This is easy on a three head machine (the sound quality should not change when you switch from input to tape), but on most decks you have to make a series of trial recordings.Some of the high end decks have built in tone generators that make this process easy.

6. Record at the proper level. Most people try to get too hot a signal on the tape. The meters on the deck are peak meters, so steady tones should be recorded at -4 to avoid distortion. Most decks begin to loose high frequency response above -10.

7. Label your tapes. Then you don't have to play them (and wear them out) to find out what they are. Write the type of noise reduction used on the label so it will be played properly.