by JULIAN D. HIRSCH
WHAT IS NOISE?
Current technical standards define noise as "unwanted disturbances superimposed upon a useful signal that tend to obscure its information content." A further definition, applicable to audio systems, excludes from the category of "noise" the harmonics and subharmonics of the input signals, intermodulation products, and flutter or wow.
In high-fidelity systems, noise is judged by its relative audibility. Obviously, noise that cannot be heard will not obscure "information content." The principal forms of audio noise are hiss (which is composed of a wide band of random frequencies) and hum (discrete "tones" at the a.c. power-line frequency and its harmonics). Other disturbances of a transient nature- such as crackles and pops when playing records, thumps from the action of an FM interstation noise muting circuit, or thumps and clicks caused by the operation of control switches--also fall within the definition of noise. Turntable rumble can be considered, from a subjective standpoint, as a special form of hum transposed down in frequency by one or more octaves.
Hiss is usually the most noticeable form of noise in a quality music system.
Basically, it is composed of what is called "white noise," which is a "signal" whose energy is distributed uniformly over a wide range of frequencies. White noise has equal energy in each unit of band width, no matter what the frequency.
This means that each octave of frequency has twice as much energy as the octave below it. Hence, the noise in the band from 5,000 to 10,000 Hz has twice as much power as the noise in the 2,500-to 5,000-Hz range.
This concentration of noise power at higher frequencies explains why low-frequency random noise is not a problem. Such noise is present (it sounds like a muted roar), but it is at such a low level as to be inaudible in most cases. Another factor affecting audibility is the unequal sensitivity of the human ear to different frequencies, especially at low loudness levels-the Fletcher-Munson effect. A noise measurement that gave equal weight to all frequencies could be strongly affected by relatively inaudible disturbances at both very low and very high frequencies.
For a better correlation with the way human beings hear, a "weighting" curve is usually applied to a noise measurement. The weighting curve most commonly used for this purpose is the standard "A-characteristic," which attenuates the frequencies below 500 Hz in a prescribed manner before measurement.
The highest audio frequencies are also cut back, though to a much smaller degree. Noise energy in the 1,000- to 10,000-Hz range, which is most objectionable to the listener, is given the most "weight" in the measurement.
Another important factor in noise perception is the phenomenon of masking.
Low-level sounds (most hi-fi noise falls into this category) are literally drowned out and rendered inaudible by the presence of louder sounds. In the case of a low-level single-frequency noise, such as a hum or whistle, the masking effect is greatest when the two signals are close in frequency. However, at the listening levels encountered in home music re production (greater than 60 dB SPL) the masking effect extends over a much wider range of frequencies, with the music tending to mask higher-frequency (but lower-level) noise components.
The masking effect can easily be demonstrated by playing a record with an audible noise level (it should not be too difficult to find!). Before the pickup stylus contacts the record, increase the amplifier volume-control setting until a faint hiss and/or hum can be heard. (If you are one of the fortunate few unable to achieve this condition--or can only achieve it at a very high volume setting be assured that most music systems do have some audible background noise with a high, but usable, volume-control setting.) When the pickup is lowered (gently!) to a quiet groove between bands, there will usually be a transient thump, followed by a noticeable increase in noise level. If a low-frequency, hum-like noise appears at this time, the turntable probably has some rumble (it is not unknown, of course, for records themselves to have appreciable rumble and similar noises molded in). Once the music starts, notice that the hiss and hum can no longer be heard (unless the program level is very low). Even a fairly noisy record or FM broadcast may appear to be noise-free .while the program level is high. However, during a pause or a low-level passage, the noise will become quite evident, and possibly even objectionable.
The noise level of modern phonograph records is usually low, the rumble of many modern turntables is negligible, and with reasonably good open-reel recorders and the better cassette machines, tape hiss can be held to a level no higher than that of records. Amplifier noise is usually low enough that it can be ignored. It would seem reasonable, then, to wonder why so many people are concerned about audio-system noise, and why so much effort is going into developing methods of reducing or eliminating it.
It seems to me that the explanation lies in the ever-higher standards set by dedicated audio enthusiasts and music lovers, and by the people who design the equipment and create the records they listen to. Low as the noise level of many music systems may be, it can almost always be heard under some conditions.
This is especially true when one tries to listen to music at a natural level in the confines of the home. Super-power amplifiers and rugged speakers make this possible, and under optimum conditions the results can be breathtaking. Unfortunately, when the system gain is high enough for such listening levels, even a fairly low noise level in the program material becomes very audible. Anyone who has heard music emerging at full volume from an utterly silent back ground will not readily accept a "nor mal" background noise level.
The noise problem is being attacked on several fronts. Improved record materials and cutting and pressing techniques, combined with noise-reduction systems in the transfer from tape to disc, have achieved impressive improvements in record sound quality (listen to records made ten years ago, as compared with recent releases, if you doubt this). Electronic circuit designers have devoted considerable effort to keeping the amplifier effectively out of the noise picture, and with notable success. Noise-reducing techniques for home equipment have been developed that greatly reduce the noise added to a program in the process of tape recording and playback as well as in FM broadcasting. And, in addition, some progress has been made in removing noise already present in the program-without affecting frequency response or other characteristics.
--------------------
Harman/Kardon 900+ Receiver
JBL L100 Speaker System
Pilot 211 AM/Stereo FM Tuner
Pioneer RT-1020L Tape Deck
----------
Also see:
AUDIO BASICS--Glossary of Technical Terms -10 RALPH HODGES