Understanding S/N Ratios (Sept. 1976)

Home | Audio Magazine | Stereo Review magazine | Good Sound | Troubleshooting


Departments | Features | ADs | Equipment | Music/Recordings | History





by Herman Burstein

Signal-to-noise ratios for tape decks tend to be more puzzling than ever due to the multiplicity and variation of factors involved in measuring S/N.

The maximum signal which can be put on the tape is defined in various ways.

One of these ways is the signal which causes the tape deck's meters to read 0 VU, but such a reading may represent a variety of signal levels. S/N also varies with the type and brand of tape used, and S/N varies, of course, with inclusion of a noise reduction system, such as Dolby.

Because of these and still other factors, the S/N rating for a given model tape deck may vary over a considerable range-possibly 10 dB or even more. The purpose of this article is to explain why S/N ratings for the same deck can vary so much, but first it will be useful to briefly review a bit of history concerning the subject, and to explore the meaning of S/N for tape decks.

A Bit of History

S/N of 55 dB is about the minimum for high fidelity reproduction. Ten years ago, many home tape machines achieved no better than S/N of 45 dB or so. As recently as about two years ago, a home tape deck was considered to be doing excellently if it attained 55 dB S/N. However, improvements in tape electronics, tape heads, and tape, plus the advent of noise-reducing devices such as Dolby, dbx, and JVC's ANRS, have made possible home machines with S/N ratios on the order of 65 dB and even close to 70 dB. Now that really good S/N ratios are within reach at reasonable cost, there is a good deal more candor in the tape industry about the prime importance of a high S/N ratio.

Meaning of S/N for Tape Decks

In a general manner of speaking, S/N refers to the ratio between the desired audio signal (program material) and the undesired audio signal (noise). This ratio is expressed in decibels (dB), as further discussed in the next paragraph. More specifically, the numerator of the S/N ratio refers to the maximum permissible audio signal which can be put on the tape; and the denominator refers to the noise in the tape system (produced by the record and playback electronics and the tape). Maximum permissible audio signal is generally accepted, at least for home machines, to be that which at 400 Hz results in 3% harmonic distortion on the tape. Therefore, measurement of S/N involves recording a tone of 400 Hz (or similar frequency) at a level that produces 3Y0 distortion on the tape; measuring the level of this tone in playback; re winding the tape and again putting it through the recording process but this time without an input tone of 400 Hz; measuring the playback output, now consisting entirely of noise; and expressing the ratio between the first and second playback measurements in terms of dB.

S/N of 55 dB signifies a maximum power ratio of 316,000:1. That is, assuming a fairly flat audio system, the loudspeaker delivers about 316,000 times more power for the desired program material than for the undesired noise. A S/N of 60 dB signifies a maximum power ratio of 1,000,000:1; 65 dB, 3,160,000:1; 70 dB, 10,000,000:1.

Note carefully in the preceding paragraph our reference to maximum power ratio. Much of the time the audio signal is well below its peak (maximum) level-often 20 dB, 30 dB, or still lower. Then the S/N ratio drops the same number of dB. During a very quiet passage, the audio level of pro gram material may drop as much as 45 dB (perhaps more on a disc or tape with wide dynamic range), so that the S/N ratio drops 45 dB. To illustrate, assume a tape system has a S/N ratio of 55 dB based on maximum permissible recording level. But if the audio level drops 45 dB during a quiet passage, the S/N drops 45 dB to only 10 dB; the level of the program material is now only 10 dB above the noise level.

Thus one may realize how desirable it is to attain S/N above 55 dB, particularly when dealing with program material having a wide dynamic range (ratio between the loudest and softest sounds). For example, a tape deck with 65 dB S/N would assure us that the audio signal is at least 20 dB above the noise, assuming a dynamic range of 45 dB. That is, the S/N ratio would be 20 dB on the quietest passages, more than that on the louder pas sages, and 65 dB on the loudest passages.

Reasons for Discrepancies

There are at least eight factors that explain why a given model of a given brand of tape deck can receive different S/N ratings.

1. Reference Level for Measuring S/N. As previously stated, S/N for home machines is usually based on a recording level that produces 3% harmonic distortion on the tape. This may be called the 3% reference level.

However, two other reference levels--which usually tend to be about the same as each other-are also used. One of these alternative reference levels is that which produces 1% harmonic distortion. The 1%reference level tends to be a recording level about 7 or 8 dB below the 3% reference level. Therefore, when the 1% reference level is employed for measuring S/N (often the case for professional tape decks), the S/N rating drops 7 or 8 dB. For example, a 65 dB S/N rating based on 3%distortion drops to a 57 or 58 dB SIN rating based on 1%distortion.

The second alternative reference level is the recording level which causes the VU meter to read 0 VU.

Typically, VU meters in high quality machines are calibrated to read 0 VU when the recording level is that which at 400 Hz produces 1'/o harmonic distortion on the tape. In such cases, the 0 VU reference level produces the same S/N rating as does the 1 1/4 reference level. Sometimes, however, a 0 VU reading will correspond to less than or more than 1% distortion. Correspondingly, the S/N rating goes down or goes up. We should take particular note of tape decks with peak-reading VU meters, which (unlike the standard VU meter) indicate peak level rather than average level of the audio signal. Peak-reading meters tend to be calibrated so that 0 VU corresponds to the 3% reference level.

Then the S/N rating based on 0 VU tends to be the same as the S/N rating based on the 3% reference level.

2. Noise Weighting in Playback.

While the hearing range of humans extends between approximately 20 and 20,000 Hz, we do not hear all frequencies equally well, assuming that all frequencies are presented to our ears with equal acoustic power. Our hearing tends to be less sensitive at low frequencies and at high ones.

Therefore noise at low and high frequencies is less audible than noise at middle frequencies. To allow for this phenomenon, the measurement of noise in playback (as described earlier) is sometimes weighted in accord with what is believed to be the typical human change in hearing sensitivity as frequency changes. That is, the noise produced by the tape deck in playback is put through an electrical filtering device which gradually reduces the amount of noise energy at low frequencies and at high ones.

Then the noise is measured. The result of this process is to reduce the amount of measured noise.

Various "weighting curves" are employed, that is, various amounts of reduction of noise energy at low and high frequencies. An example is the ASA Standard C16.5-1961 weighting curve, adopted by NAB (National Association of Broadcasters) in its 1965 standards for tape recording. Noise reduction becomes significant (reduced 3 dB or more) below approximately 800 Hz and above approximately 7500 Hz. Noise reduction is greatest for the low frequencies, being about 32 dB at 50 Hz and still more at lower frequencies.

All in all, use of weighting in measurement of S/N results in an increase in S/N rating. It is difficult to say how much the S/N is increased. The amount of increase depends upon which weighting curve is employed. It also depends upon the particular tape deck being measured. For example, if a deck has particularly strong 60 Hz hum, it can benefit more from a weighted S/N measurement than a deck with very little hum. As a rough guess, weighting can improve the S/N rating by about 6 to 10 dB. (Thus, the NAB standard stipulates S/N ratios 10 dB higher on a weighted basis than on an unweighted basis.) A tape deck with an unimpressive 50 dB S/N rating on an unweighted basis might achieve an impressive 60 dB rating on a weighted basis.

3. Use of Dolby and Other Noise Reduction Systems. For a cassette deck to have a real claim to high fidelity, it must include a Dolby or similarly effective noise reduction system (NRS). While open-reel tape decks can achieve S/N of high fidelity caliber without NRS, there is a trend to ward inclusion of such systems; earlier discussion has pointed out the ad vantages of exceeding the minimum high fidelity requirement, namely 55 dB S/N. Dolby and similar NRS can improve S/N by amounts typically varying from 6 dB to 10 dB, particularly at lower speeds (1-7/8 and 3-3/4 ips). The improvement tends to be somewhat less at higher speeds (7 1/2 and 15 ips).

4. Tape Used. For a given amount of distortion, say 3%, some types of tape can deliver more audio signal than can others. Thus, high output tape de livers more signal than conventional tape. Further, low noise tape produces less tape noise than do some other kinds of tape, and low noise/high output tape has both ad vantages. Thus, there may be a few dB difference in S/N ratio depending on which type of tape is used. In addition, for a given type of tape (such as low noise/high output, 1 mil, ferric oxide) there may be a variation of about 1 to 3 dB among various brands of this tape.

5. Chance. Random variations in components used (transistors, resistors, capacitors, etc.) and in lead dress (wiring) may result in slight variations in S/N from one unit to another of the same model and brand of tape deck.

Such variations might be in the vicinity of 1 or 2 dB.

6. Manufacturer's Conservatism.

The conscientious manufacturer, wishing to attain and live up to a good reputation, will be conservative in stating specifications for his tape deck, including its S/N rating. He will make allowance for chance variation from one deck to another, as just de scribed, for different tapes that may be used with his unit, and possibly for other factors. His may be a "worst condition" specification for S/N.

Therefore, the typical purchaser of his tape deck may find that actual S/N performance exceeds rated S/N by several dB.

The reverse can also be true, particularly for tape decks of generally lower price and quality. Rated S/N may be a "best condition" specification; it may assume that a selected tape deck from a given model line is employed, and that it is used with the best of tapes. Thus, the typical purchaser may find actual S/N below specification.

And of course we can have the intermediate situation, where many purchasers find actual S/N above specification, and about an equal number find actual S/N below specification.

7. Quality Control. Control over the quality of tape decks reaching the purchaser may range from very rigorous to quite loose. Quality control is one of the unseen things (such as extensive research) that go into a costly machine. In the case of S/N, tight quality control helps insure that there is little variation from one tape deck to another of the same model and that none fall below specification. Loose quality control makes possible fairly extensive variation.

8. Design Improvements. During the lifetime of a given model of tape deck, which may be several years, the manufacturer typically makes changes, some of which may result in improved S/N. These changes are based on continued research by the manufacturer, new developments by others (such as the supplier of tape heads), and feedback from dealers, purchasers, equipment reviewers, and others. Therefore, a later version of a given model (higher serial number) may have a few dB better S/N ratio than an earlier version.

Conclusions

While the foregoing discussion hopefully is enlightening, it can hardly be satisfying. It cannot be satisfying to know that S/N ratings for tape decks vary not only because of inherent differences in quality of the decks but also because of differences in methods of measurement and differences in tapes used for measurement.

How is the audiophile to compare the S/N of one tape deck with that of another? How is he to know the S/N of a given tape deck-simply just to know it or to compare it with the S/N of other audio equipment? It seems this situation can be re solved by the audiophile asking two basic questions and by the industry--manufacturers, dealers, and others concerned--being prepared with the answers for each brand and model of tape deck. The two questions are as follows.

Based on (a) the tape recommended by the manufacturer for his tape deck, (1) (b) the 3% reference level, and (c) measurement on an unweighted basis:

1. What is the S/N ratio without use of NRS (noise reduction system, such as Dolby)?

2. If the tape deck includes NRS what is the S/N with use of NRS? We have to be able to compare tape decks' S/N ratios on a common basis. Until the day when all tape decks include built-in NRS, the common basis for measuring S/N is necessarily without use of built-in NRS.

Therefore, when a tape deck has built-in NRS, we need the answer to Question 2 as well as the answer to Question 1. (Of course, when a tape deck lacks built-in NRS, we can only raise Question 1.) NRS are available today either built in or as an external device made by several companies. Therefore, one may rightfully want to know the S/N of one's tape system including NRS. If NRS is built-in, this information is pro vided by the answer to Question 2. If not, we can get a reasonable approximation by adding the answer to Question 2 to the S/N improvement claimed by the maker of the external NRS. To illustrate, assume a tape deck without NRS has a rated S/N of 55 dB, based on the 3% reference level and unweighted measurement of noise.

And assume that 8 dB S/N improvement is claimed for the NRS device.

Then the S/N of the tape system, including the NRS device, amounts to about 63 dB.

-----------

(1) It is unlikely that the manufacturer would recommend a tape chosen solely because it serves to maximize S/N. Other factors in his choice of tape will include its frequency response, availability, cost, physical characteristics (such as oxide shedding, tendency to cup or curl, accurate dimension, etc.)

Probably a better alternative to (a) would be an industry-accepted standard tape, issued under the auspices of NAB, which all tape deck manufacturers would agree to use for the purpose of specifying S/N. (A manufacturer could still recommend commercially available tape or tapes to be used with his deck.)

(Audio magazine; Sept. 1976)

Also see:

Harmonic Distortion by Richard C. Heyser (Feb. 1976)

New Tests and Standards for Tuners and Receivers (Jan. 1976)

Reading VU Meters (Sept. 1976)

How We Hear (May 1977)

= = = =

Prev. | Next

Top of Page    Home

Updated: Tuesday, 2018-10-23 5:36 PST