SOUND-EQUALIZATION IN RELATION TO THE WAY WE
© 1982 The Anstendig Institute
An equalizer is too often thought to be a sophisticated accessory to a sound system. But for those who want their music to resemble the original performance, an equalizer is an essential component. In all sound-reproduction, the equalization of the original is distorted by the reproduction process, by the acoustics of the listening room, and by the peculiarities of our hearing.
Equalization has two meanings. The actual proportional balance of the frequencies in relation to each other at a given time is called “the equalization”. Changing the loudness of the frequencies in relation to each other is a process called equalizing or “equalization”. The purpose of sound-equalization is to restore the balance of frequencies of the sound reproduction to that of the original, live sounds.
I. ROOM EQUALIZATION AND PROGRAM EQUALIZATION
Room equalization is the equalization of the sound system itself. In room equalization, an audio signal, made up of frequencies of equal energy over the whole range of our hearing, is sent through the system, and the balance of those frequencies is readjusted until the sounds coming out of the speaker are also of equal loudness. In other words, room equalization adjusts the frequency balance of the system to the characteristics of the room in order to assure that the frequencies emerge from the speakers with the same balance as that of the recording.
Room equalization, though desirable, is dispensable because, when equalizing the program itself, the ear automatically compensates for the abnormalities of the room along with those of the program. In this paper, the word equalization refers to program equalization when not specifically referring to room equalization.
Program Equalization is the equalization of the actual program material itself while it is playing. In program equalization, one purposely changes the balance of the frequencies in order to make the program sound natural to one's ears. The equalizers are essentially used as elaborate tone controls.
Program equalization corrects the distortions in the equalization of the sound due the peculiarities of our hearing, the massing of overtones, and any other distortions caused by the recording and reproducing equipment. It should be emphasized that these distortions are not the slightest bit subtle. They are easily heard when correctly demonstrated and are much more disturbing, both consciously and subconsciously, than any of the more well-known distortions that are considered unacceptable by most audiophiles.
In program equalization, a typical correction can involve a reduction of some frequencies by as much as 20 to 25 decibels in the range of approximately 2,000 to 3,500 hertz (Hz). Compared to distortions that are not tolerated in other performance parameters, these figures are enormous. But most "audio purists", even those who claim to hear subtle, difficult-to-discern differences between components, will put up with these gross, major distortions in the equalization because they have not had the opportunity to compare them to correctly equalized sound.
II. SOUND-REPRODUCTION AND HEARING
In the 1930's, a famous set of tests by two Bell Telephone Laboratory scientists found that we are more sensitive to certain sound frequencies (vibrations per second) than to others. A resulting set of graphs, known as the Fletcher-Munson curves,1 plots the proportional loudness of the way we hear sound frequencies at all volume levels, up to the threshold of pain. They show
1) that our hearing distorts the relative balance of the frequencies in that we do not hear all frequencies equally loud;
2) that even small changes in overall volume change the apparent frequency balance in that we hear different frequencies proportionally louder;
3) that how much louder we hear those frequencies (i.e., their proportional loudness in relation to the other frequencies) also changes when the overall volume level is changed; and
4) that this distortion is greatest at very loud or very soft levels.
The Anstendig Institute has pointed out another distortion inherent in sound-reproduction. When sound is reproduced, all of the original overtones act as fundamentals, generating their own series of overtones. This results in a massing of overtones that must be compensated for.2
The plain fact is that without equalization recorded music will not sound natural. In relation to all recorded sound (sound-reproduction), the Fletcher-Munson curves together with the massing effect of overtones unequivocally show that equalization is necessary in all sound-reproduction because
1) at any volume level, some equalization is necessary if we are to hear the sounds in an undistorted manner that corresponds to the frequency-balance of the original;
2) this equalization has to be changed whenever the volume level is changed; and
3) increasingly larger amounts of equalization are necessary the louder or softer the volume levels become.
The same also applies to all amplified sound (sound reinforcement).
III. THE ANSTENDIG INSTITUTE HAS MADE NEW DISCOVERIES IN THE FIELDS OF HEARING AND SOUND-EQUALIZATION
The reasons why all sound-reproduction should be equalized are even more compelling than the Fletcher-Munson curves would suggest.
The Anstendig Institute has been able to extend research on hearing and sound-equalization into the realm of real-life listening experience. The results so far have produced the following discoveries important to understanding how we experience both recorded and live sound.
1) The first phenomenon is that frequency extremes are not heard progressively louder or softer with increased or decreased volume. The higher and lower frequencies are simply not registered until they reach a certain loudness level (threshold). In other words, when listening to a recording that contains all sound frequencies, if one starts at a low volume level and slowly turns the volume up, one will not consciously hear the frequency extremes until they reach a certain volume level. With the high frequencies, this effect begins with the important overtones around 2000 Hz. We also found that THE HIGHER THE FREQUENCIES, THE LOUDER THE VOLUME HAS TO BE FOR ONE TO REGISTER THEM. Perception of the extreme lower bass frequencies is similar but more complex, since upper bass frequencies are heard proportionally louder at lower volume levels.3
2) The Anstendig Institute has discovered that equalization plays a major role in the recognition of changes in loudness. It is well known that hearing is extremely sensitive to changes in volume levels, and it is not unusual for some people to be able to notice small changes of one decibel or less. The Anstendig Institute has made the following observations: people are less sensitive to volume changes in pure tones than they are to volume changes in program material containing large amounts of primary tones and overtones. This is a direct result of the Fletcher-Munson effect. When the sound contains a greater range of frequencies, people hear changes in the equalization with small changes in volume. Since changes in the balance of the frequencies change the characteristic sound quality of the program material, we have become conditioned to recognize those changes as changes in volume. Since pure tones do not possess complex overtone structures, their equalization does not change appreciably when their volume changes. With less change in the characteristic quality of the sound, the ear is less able to distinguish small changes in volume with pure tones.
The conclusion to be drawn is that, in real-life situations, we distinguish small differences in volume more because we notice differences in equalization than because we actually perceive the changes in loudness itself. When someone turns up the volume of a sound system and another person in the room complains that it is “too loud", what usually is meant is that it sounds distorted because of changes in perception of the equalization. The differences in perceived equalization amount to an irritation factor that is separate from, and in addition to, perception of changes in volume.
The Anstendig Institute has found that it is more difficult for people to notice changes in volume when the frequency ranges to which we are most sensitive are subdued. This corroborates our observation that we notice changes in volume due to irritation factors that are a result of changes in the equalization. We have also found that, when the sound is correctly equalized for a particular volume level, changes in volume within a range of approximately five or more decibels below the equalized level are more difficult to distinguish than with unequalized sound. But the same amount of change is readily evident when the volume exceeds the original volume level for which the sound was equalized. This is because the high frequencies, which were just below their threshold when correctly equalized, suddenly became disturbingly apparent when they reached their threshold or went beyond it.
In our tests, programs of music were first equalized at one volume level. The volume was then turned down as far as possible without the music sounding obviously unequalized. We then brought people into the room and found that, under these circumstances, as long as the volume remained between the original, equalized setting and the level to which we had reduced it, they did not exhibit the same sensitivity to volume changes. When we further reduced the high frequencies in our hearing's most sensitive range between 2000 and 5000 Hz, we were able to raise the volume substantially without anyone noticing. Our conclusions are
1) that our ability to perceive changes in volume is most sensitive in the frequency ranges that we hear loudest;
2) that most of the our much-touted sensitivity to even very small volume changes is not a recognition of degrees of loudness, but rather a recognition of changes in the quality of the sound and an awareness of degrees of irritation, i.e., whether the sound becomes more or less irritating; and
3) that the known fact that our hearing is more forgiving of distortions in equalization at low volume levels than at high volume levels is due to the absence of those high frequencies that are below our threshold of hearing.
IV. HEARING CHANGES IN RELATION TO CHANGES IN THE BODY
In addition to hearing differently at different volume levels, the way we perceive both the overall volume levels and the balance of frequencies within a set volume level changes over the course of a day in relation to our changing states of physical tension and relaxation. Our bodies are not the same when we wake up as when we retire, nor are they the same when we are exercising as when we are totally relaxed. Our hearing changes with these changes in our bodies.
We hear the vibrations of our own body when vibrations from the sound source strike it, not the vibrations of the source itself. The body is like the sounding board of a piano which vibrates in sympathy with the vibrations striking it. Tensions of the external muscles act like the damper on a piano, suppressing the ability of our bodies to vibrate freely. Therefore, our hearing changes whenever our physical state changes, particularly in relation to tension and relaxation. AS one's body relaxes and calms down, there is a dual phenomenon in the way we perceive volume levels:
1) one hears louder; and
2) one is able to listen comfortably at louder volume levels.
That we hear louder can be demonstrated by disturbing one's state of relaxation after one has completely relaxed and calmed down. What was perceived as loud when completely relaxed will seem much quieter when heard again in the unrelaxed state.
This dual phenomenon is only possible to observe if the recording is equalized. Without equalization, the irritations and tensions make it impossible for the listener to relax and experience the increase in volume.
Those who practice disciplines for refining their physical state, such as autogenic training, yoga, or other relaxation techniques, can particularly notice these changes in their perception of loudness. If they have the opportunity to hear the same equalized recording before, during, and after their bodies relax, they perceive an increase in volume even though the volume level remains unchanged.
This discovery has important implications for the field of audiology and for eye, ear, nose, and throat doctors. For example, great care should be taken to give hearing tests under conditions similar to those real-life situations in which the patients would be most naturally relaxed and attentively listening. Failure to do so results in a hearing aid with a correction that does not correspond to that which the patient needs when in his own familiar environment. The fact that clinical test-situations do not correspond to the patient's normal hearing environment is probably a major reason why a great number of hearing-aid users do not like the sound-quality of their hearing-aids, find them unpleasant to use, and therefore often refuse to use them. The other senses are similarly affected by the state of the body.
V. THE DISTORTIONS OF UNEQUALIZED RECORDINGS FALSIFY THE EMOTIONAL CONTENT
Sound is the most important single influence on our lives. Hearing is the dominant of the two higher senses, i.e., when combined with sight, it determines the quality of what is experienced: the apotheosis-end of a western where the heroes ride into the sunset in a blaze of glory would be somber and funereal if accompanied by a funeral march, satirical if accompanied by a polka, sad if accompanied by a lament, etc. There is now hardly an area of life without recorded sound. The arts, even the live ones, increasingly depend on recordings or amplification; businesses use background music to aid production and morale; institutions such as hospitals use background music therapeutically; etc. But this vast amount of music is distorted not only in equalization, but also in its emotional content.
All sounds consist of the primary sound--the fundamental, the actual note being played, sung or spoken--and sympathetically vibrating overtones and undertones that, along with other qualities (intensity, focus, etc.), give the sound its peculiar character. Under normal circumstances, our hearing only consciously registers the fundamentals. If we consciously hear the overtones, they are too loud. This holds true for live as well as reproduced music. In the former, the acoustic of the hall is too bright, and in the latter both the sound-reproduction and the acoustic of the listening room have intensified those frequencies.
New machines, capable of analyzing the balance of frequencies in “realtime” (as they are happening), have shown that the overtones of most sounds peak (are strongest) at approximately the same frequencies to which our hearing is most sensitive. As already mentioned, the sounds produced by a loudspeaker, which already include the overtones of the original sound source, produce their own additional set of overtones. Thus, there is a "balling up", a massing of overtones at those very frequencies that we hear loudest. These frequencies, especially those between 2000 and 4000 Hz, cause the strongest physical reactions in us, generally in the nature of physical tensions. (This explains why women, who are more sensitive to high frequencies than men, often complain that loud recorded music makes them nervous and irritable.) With unequalized recordings, the unnaturally exaggerated frequencies between 2000 and 4000 Hz stimulate physical tensions and other bodily and mental reactions in the listener which degrade the way the listener actually experiences the emotional content of the program material. One hears the emotional content of the original combined with and falsified by one's own bodily reactions to the irritation caused by those frequency peaks. One literally hears and experiences a combination of the expressive content of the original and one's own unrelated and quite different bodily reactions to a physical irritant. If one is listening to music that has a particularly fine expressive-emotional content, either one will not experience the emotion at all or one will experience a quite different, coarser (less fine) emotional content that bears similarities to the original, but is, in reality, quite different. Equalization eliminates this problem because restoring the original frequency balance eliminates the irritating peaks.
VI. PRACTICAL ASPECTS OF THE NEED FOR EQUALIZATION
In the 1930's, the primitive recording equipment added its own imbalances to the frequency spectrum, and the recording industry had no practical means of changing the equalization to offer to the public. Today, equalizers capable of adequately adjusting the frequency balances for all volume levels exist, but the need for more sophisticated equalization than simple high and low tone controls is not made known. While the field of sound-reproduction pays some attention to room-equalization, which compensates for the acoustics of the listening room, the need for equalization of the actual program material has been ignored.
An important reason why equalization is shunned is that, among audiophiles, it is quite logically believed that the more electronic devices through which the recorded signal travels, the worse the sound must be. To them, it does not matter whether these devices are equalizers, tone-controls, noise-suppressors, or anything else, since each part of a sound system adds its own small or large degree of distortion. Thus most audiophiles have avoided equalization of any kind. But the distortions in equalization are much greater than the sum of any distortions that well-designed components would introduce. The ideal would be to build better equalizers that essentially do not degrade the signal. Quality components that do not degrade the sound have been achieved in other parts of the system, including equalizers in the phono-amplification stage (RIAA equalization). The same quality should be possible in equalizers (a number of much-improved equalizers have in fact recently been introduced).
Another objection is that equalizers are not easy to use and are very easy to misuse, thus worsening the sound quality. That argument holds for almost anything beneficial, from medicines to automobiles. The answer is to educate the users and make available clear, easily understandable, detailed explanations and instructions.
Hearing is the most important of our senses. Recordings permeate our whole culture. Everyone listens to them. Yet the public has already become conditioned to accept flawed, shrill, distorted, unequalized recorded sound as accurate and to demand similar sound of live events. Concert halls have been designed and built to sound the same as the flawed recorded sound that the architects are used to hearing, and musicians are changing their technique to simulate the distorted, wrong sounds they hear on recordings and in these halls.
The raison d'etre of music is its expressive content. A much-lamented, universal lack of real expressivity in today's music-making can be traced to the fact that performers listen to, and even study their scores with, unequalized recordings. Of course, they are not hearing the fine interpretive details of expression. With the recent availability of equalizers in all price ranges, the importance of the need for equalization should finally be made known and the public should be instructed in how to use them.
An important aspect of equalization is the ability to restore the sound of old recordings to a comfortably listenable quality. Interestingly enough, the dynamic qualities that make up the expressive content of the performances on these older recordings have been faithfully captured. The main causes of the distorted sound were the inability of the old machines to record all the frequencies equally loud and certain so-called "resonances" that caused some small parts of the frequency range to stand out much more audibly than others. These problems can, to a very large degree, be corrected with equalizers, opening up a broad range of historic listening experiences that is one of our legacies and treasures. The now famous Sound-Stream restorations of old classics, such as the Caruso recordings, are essentially a highly sophisticated application of equalization.
Proper equalization can play an important role in reducing the dangers of listening to loud disco and rock music. Most of the negative effects of loud music that have been noticed in fans of rock and disco are due to a lack of equalization: the distortions of balance, particularly the frequency peaks, are enormously exaggerated at the customary super-loud volume levels. Since equalization eliminates these peaks, music can be safely listened to at substantially louder volume levels, within reason. The same applies to other instances of dangerously loud sound-reproduction.
The need for sound equalization in the playback of recordings was proved in the 1930's, but since adequate means of accomplishing it were not available until recently, this need has not been pointed out to the public. It remains a basic prerequisite to natural-sounding sound reproduction.
VII. METHODS OF USING EQUALIZATION IN RESEARCH OF HEARING AND SOUND-REPRODUCTION
For research to reflect the way we normally hear, real-life program material must be used. In real life, our hearing and experiencing of sound differs in many ways from that of listening to test tones and other test objects that are not a familiar part of our lives. In order to analyze scientifically the idiosyncrasies of how we hear and experience sound in real life situations, it is necessary to create a test object that is easily recognizable, repeatable, and familiar enough for us to detect whether it changes under various listening conditions. Sound conveys expression as well as information, so it must also be a meaningful hearing experience, not simply meaningless, clinical test tones. Obviously, recordings offer the only possibility of this, but only if they are equalized.
Since our perception of sound does not register all frequencies with equal loudness and because the process of sound-reproduction also distorts the balance of frequencies in its own, different way, the sound we ultimately hear when playing an unequalized recording has a completely distorted frequency balance in relation to how we would have heard the original live. But the listener does not know what the original sounded like, and the recorded sound does not duplicate anything familiar which can be used as a frame of reference. Therefore, a repeatable test of hearing in real listening situations is impossible unless the sound is equalized to sound the way we are used to hearing it live.
It should be pointed out that the only capacity of any of our senses that can be described as exact is the direct comparison of shades of color tones that are immediately adjacent to each other. It is the only capacity that does not rely on memory, which is the most undependable aspect of our sensory perceptions. Hearing does not enjoy this possibility of direct, simultaneous comparison of sound impressions and has to rely on memory. This weakness makes it impossible for us to observe, recognize, and compare distortions of musical sounds, unless we know what they sound like undistorted. A meaningful comparison and evaluation of distorted versions of anything such as music that is meant to achieve a certain result is not possible unless one is familiar with the undistorted version. If we listen to music that is already distorted and repeat it in a different, but also distorted manner, there is no way for our ears to accurately differentiate between the two or the way each distorts the original. But if we familiarize ourselves with something wherein the balance of frequencies has been corrected so that it sounds natural and "right" for our ears and then repeat it in a distorted way, we immediately hear that it is different and can re-equalize it to sound right again. The human voice is an excellent object because we are familiar with what a voice sounds like and can recognize when it sounds unnatural. Without equalization, sound research amounts to comparing one distortion to another without being able to directly compare them to each other (as one can with sight impressions).
When investigating how we hear at two different volume levels, the difference in the amount of equalization necessary to correct the sound at each volume level is the difference in how we hear at those two volume levels. By using sound analyzers that can measure and plot this graphically, we can compare the two graphs, subtract the values of one from the other, and arrive at measured values of the difference.
New possibilities of equalization that have been introduced in the last few years have given us a means of implementing the above scientific procedures. At the same time, the recent availability to the public of serviceable, moderately priced equalizers of good quality has made that research practical and necessary.
1 Copies are available from The Anstendig Institute.
2 Refer to our paper "The Massing of Overtones".
3 This is due, in part, to the threshold effect of the upper frequencies: with higher volume levels, the high frequencies are loud enough to be consciously heard and add to the overall impression of the volume which has the effect of balancing and even masking the lower frequencies. With lower volume levels, this is not the case, making the mid bass seem proportionately louder. When the sound is correctly equalized at each volume level, this effect is less apparent, because, when equalized, the higher frequencies that make up the overtone structures are balanced so that they add their coloring to the sound, but are not consciously heard.
The Anstendig Institute is a non-profit, tax-exempt, research institute that was founded to investigate the vibrational influences in our lives and to pursue research in the fields of sight and sound; to provide material designed to help the public become aware of and understand stressful vibrational influences; to instruct the public in how to improve the quality of those influences in their lives; and to provide the research and explanations that are necessary to understand the psychology of how we see and hear.