you are an audiophile, you probably have spent a lot of effort setting
up your home entertainment system, car system, and maybe even your
portable music player. You know all about how to select speakers,
where to place them, the kind of wire and connectors to use, treble,
midtone, and bass equalization. You know what you like, but the
way you perceive sounds is not the way many other people do. Even
if your hearing follows a 'normal' response curve per standardized
audio tests, personal preferences vary widely in part because the
way brains are wired (speed of sound processing, preferred musical
style, state of mind - maybe you're insane) and in part because
of the impact of sound waves upon your body (soft tissue and bone
pressure wave conduction, clothing worn). Pitch, timbre, tone, and
loudness has been extensively studied and quantified. While three
of the four are purely objective measurements, loudness is the one
that is subjective. Loudness, simply put, is the level of amplitude
by which frequencies other than a chosen standard (e.g., 1 kHz)
must be increased or decreased to be perceived to have an equal
level of "loudness." Research performed by Harvey Fletcher and Wilder
Munson in the 1930s used a panel of human subjects to construct
a set of curves (Fletcher-Munson curves) of equal loudness based
on frequency and the decibel level of the reference tone. Being
a lover of nice music but stuck at a relatively low income level,
I never invested in really good sound equipment; my experience with
sophisticated equalization consisted of deciding when or when not
to push the "Loudness" button on the front panel of my $50 Radio
Shack stereo unit. Now at least I know what it did. This article
goes into great detail on the concept of loudness and how stereo
systems implement compensation.
April 1957 Radio & TV News
Wax nostalgic about and learn from the history of early
electronics. See articles from
Radio & Television News, published 1919-1959. All copyrights hereby
Why Loudness Control?
By Norman H. Crowhurst
fresh approach to the need for a true Loudness control and information
on its proper adjustment.
Any subject that depends upon
appreciation by the human senses is bound to be one that comes up
for perennial discussion. This is one of them. There are two extreme
schools of thought: one of these suggests that correct reproduction
requires an accurate re-creation of the original sound wave pattern;
this means that all of the components in the original program must
be reproduced in exact proportion and at the same level; this philosophy
does not allow of anything so intangible as a loudness control.
At the other extreme we have the people who believe that
human hearing has been "scientifically" established by such data
as the Fletcher-Munson curves; therefore, to give a true subjective
impression of any reproduced program material, a loudness control
must be incorporated that introduces compensation according to the
accepted and proven data published first by Fletcher and Munson.
In between these two admittedly extremist views a number
of subsidiary questions arise and confuse the novitiate still further.
Hence the perennial topic. Let's start out from the classic point
of argument that faithful reproduction should be an exact re-creating
of the original sound wave pattern.
Pursuing this ideal
to its logical conclusion one would assume the loudspeaker should
generate a sound intensity in the listening room identical to that
received by the microphone in the recording studio. This leads us
to question number one.
When most people say this, they
subconsciously think of listening to an orchestral performance in
a typical auditorium, from a typical auditor's seat. But, if a microphone
were placed in the position of the typical auditor's seat and a
recording made, no one with any musical appreciation would want
the disc. The reproduced sound would be quite unlike the impression
conveyed when sitting in that seat personally. Why should this be?
Whether we use a stereophonic system for reproducing our
sound or not, our hearing is always binaural, whereas the microphone
is not, placed in the position of the typical auditor's seat or
wherever you will.
Some readers of this magazine have the
misfortune to be deaf in one ear. To them, the remarks which follow
will seem incomprehensible, except as pure theory. A few years ago
the author was deaf in one ear for a period, an experience which
enabled him to appreciate how different everything sounds this way.
This experience also confirmed the following explanation.
When we take up our seat in the auditorium and listen to the
program, we are primarily conscious of the orchestra; secondarily
we hear the reverberation coming from other parts of the building,
as a kind of echo, although the time interval is usually too small
for it to be noticeable as a separate entity. But our binaural capacity
for differentiating directions enables us clearly to separate, in
our subconscious perception, the original sound from the orchestra
and that due to the same program reverberated around the auditorium.
The difference in intensity between these two components
of sound may not be more than 10 or 20 db but, because of our subconscious
interest in the program itself, attention is focused on the orchestra
and the reverberant sound is heard merely as a background that lends
character. Because of this, our subconscious subjectively increases
the apparent difference in level to much more than the actual 10
or 20 db.
The microphone, however, being a much more non-subjective
device than a pair of human ears, would pick up the whole sound
received as a conglomerate, with only the actual 10 to 20 db differential
between the direct sound from the orchestra and the reverberant
sound from various parts of the auditorium. When this sound is reproduced
over any system whatever, even in a thoroughly acoustically damped
room, it will sound extremely reverberant and "echoy." This is because
now the direct and reverberant sound both come from the same source,
and our binaural faculty of hearing no longer can go to work on
it in the same way and direct attention to the orchestra exclusively.
Recording studio personnel, of course are thoroughly conversant
with this fact. They don't advertise it, because listeners naturally
prefer to think the recording transports them to a position in a
typical listener's seat in some auditorium. But, to get an effect
that puts the direct sound and the reverberant sound more in the
correct proportion, as the listener thinks he would hear it, the
microphone must be placed very much nearer to the orchestra, or,
what means practically the same thing (for this purpose), a highly
directional microphone must be used to favor sound from the orchestra
more than the reverberation.
Modern practice usually places
the microphone quite close to the orchestra, or uses an array of
microphones distributed among the orchestra so as first to get a
pickup that represents the original program material, exactly as
played by the orchestra, practically without any reverberation at
all. Then, if some reverberation effect is required, the studio
uses an echo chamber to add this artificially. This procedure enables
the effect to be easily controlled so the resulting sound will have
just the right amount of reverberation to give the desired effect.
this technique produces very fine sounding records, judging by recent
releases from most of the record companies. Even the sticklers for
accurate reproduction of the original sound will have to admit this
although, if they knew it was made this way, they would call it
"phony." But let's ask them, which original sound do they really
want reproduced accurately, the sound of the orchestra, as heard
by, say the conductor, or one of the microphones scattered throughout
the orchestra, or the sound as it might be heard in the echo chamber,
somewhere between the loudspeaker that produces the sound and the
microphone used to pick up the "echo" ?
It is quite obvious
that neither of these sounds will be the same as that we hope to
hear in our living room.
When you sit in the average seat
in an auditorium and listen to an orchestral concert, the sound
you hear is quite different from that heard by the conductor, or
by any of the instrumentalists in the orchestra. All of these performers
work to give the best impression to you, the paying audience. The
same is true making recordings, only here we have a few more "performers",
like the engineers who operate the various controls associated with
the microphone levels, echo chamber, etc.
thing is that the intended impression is put across, whether by
the individual performers working collectively in the orchestra
in an actual auditorium, or whether by all those who work together
to make a satisfactory record.
Picking up the program at close quarters like this gives another
advantage to the record maker. It assists in getting a better dynamic
range onto the recording. As well as giving he direct sound from
the instruments a bigger advantage over reverberant sound reflected
around the studio, it also gives the program material a better advantage
over various other stray sounds, noises that can creep into the
studio by devious means.
Fig. 1. The Fletcher-Munson curves of equal loudness at
various listening levels.
A good studio is built with all
kinds of sound insulation to keep out extraneous and undesired noises.
But all such devices only attenuate the undesired sounds, they don't
completely eliminate them. Consequently, it is still advantageous
to have a fairly high level of sound at the microphone, so as to
give the biggest possible margin over unwanted background sounds.
This also makes it easier on the recording system, because
it gets the program material that much farther above background
noise from the microphone itself and the amplifier system.
But picking up sound for the record at this high level and recording
it through a system with standard equalization characteristics,
means that the original sound pattern can only hope to be truly
re-created when played back through the correct compensating playback
equalizer characteristic and reproduced over a loudspeaker system
at the original sound intensity. This will be too loud for comfort
- much louder than one would receive sound in the average seat of
an auditorium. The intensity of sound in the living room, reproduced
this way, would be comparable to the level of sound in the area
occupied by the orchestra itself.
To get the true perspective
of the music. one needs to be a little farther away or to hear the
sound at somewhat lower level. Consequently it is usually desirable
to reproduce the sound at a somewhat lower intensity than it is
picked up at the microphone for recording.
Let us now consider
another angle. Live musicians can, and do, provide program music
for any of the purposes for which we use reproduced music. As well
as playing to an audience in an auditorium, musicians, upon occasion,
play to entertain people in a living room, or to provide background
music in a restaurant or club, while the occupants talk. A group
of musicians playing under any of these conditions will naturally
adjust their performance to suit the purpose in hand. They will
play louder in an auditorium and much quieter in a restaurant or
club. How do musicians themselves make this loudness adjustment?
This does not mean that each instrumentalist will reduce
his volume by a precise number of decibels. Rather, each reduces
his own loudness so that the same sense of balance is achieved at
the lower loudness level. This, naturally enough, is done according
to the judgment of the musicians' ears.
hearing certainly is differently conditioned from that of most of
us (which is what we infer by crediting a person with a "musical
ear"), there is no intrinsic difference between the ears of musicians
and those of the rest of the population. The famous Fletcher-Munson
loudness contours were based on the hearing of musicians as well
as of other groups of people. So one would expect an average musician
to have a sense of comparative loudness similar to an average person
of any other group.
As this is the case, each instrumentalist
will automatically adjust the loudness of his playing so as to obtain
a perspective of over-all balance that agrees with the general pattern
of human loudness sensation.
If we base this on the experience represented in the Fletcher-Munson
contours, shown in Fig. 1, and assume that the difference in loudness
from the concert hall level to the background music level is from
the 70 curve to the 40 curve, this will be a difference of 30 decibels
at 1000 cycles and also at most frequencies above this. So musicians
playing instruments with frequencies from 1000 cycles and up - or
even 500 cycles and up, will reduce the intensity of tone from their
instruments by about 30 decibels, or a power ratio of 1000 to 1.
But, going down to instruments like the string bass, which may be
playing tones in the region from 40 to 80 cycles (taking 60 cycles
as an average), the difference between the curve marked 70 and the
curve marked 40 is only about 12 decibels.
Fig. 2. Comparison between range of control afforded by
the average tone control and the kind of change in response
required for satisfactory loudness control.
While the bass
player will un-doubtedly reduce the apparent loudness of his instrument
by the same amount as other players in the orchestra, the actual
intensity difference is much less. The higher instruments reduce
by 30 decibels, while the bass player only reduces by 12.
As we have seen in the preceding discussion, music is never
recorded at such a low level. If you want to use recorded music
for this purpose, you will have to turn the gain or volume control
down by 30 decibels or so, and if you want to play at a level suitable
for the average living room you will still need to turn it down,
by possibly 12 or 15 decibels, from the intensity at which the music
was originally recorded.
But to retain the balance at which
the group of musicians would naturally play, the bass frequencies
must be turned down to a lesser extent than the middle and higher
frequencies. This is our argument for using a loudness control.
If an ordinary volume control is used (which should more accurately
be termed a gain control), to produce this effect, the balance is
The effect produced is not that of the orchestra
playing more quietly but of its being farther away.
is because all frequencies are attenuated in proportion, which is
exactly what happens when you listen to a musical program from a
distance, unless of course there is exceptional transmission through
floors and ceiling, such as may occur if the "distance" is from
one floor to another of an apartment house!
This has about
sated the esthetic factors that form the basis for discussion as
to whether the loudness control is needed at all and, if so, what
kind of control should be provided. Having settled that we do need
control facilities to make this adjustment, the next point is if
a special loudness control is needed, or whether an ordinary volume
(or gain) control can be used satisfactorily in conjunction with
the regular tone controls?
The alternatives in this stage
of the argument are: whether we have a loudness control that automatically
increases the relative amount of bass, measured by intensity, as
the volume goes down, or whether we use a "straight" gain or volume
control and then use the bass knob of the tone controls to get more
bass when working at a lower loudness level, if necessary also giving
a little boost to the treble end for the same reason. Don't both
of these methods really achieve the same thing? And, using just
a gain control with tone controls, surely, is simpler than having
to "bother with" an additional loudness control?
when it was argued that equalization facilities were not required
for all of the different recording characteristics - the tone control
would serve for this purpose as well. Going back still further,
there was a time when tone control consisted of just a variable
resistance in series with a capacitor, conveniently placed somewhere
in the circuit, which produced only a variable treble cut. There
was no true bass control. All too often, this control was operated
with the treble fully cut, because this removed a maximum of unwanted
background noise, scratches, and plops, either due to scratch on
the record or due to static coming in on the radio. The average
listener of those times couldn't tell the difference between the
maximum and minimum position of the control, as far as the quality
of the program itself was concerned. The only effect noticed was
the apparent reduction of background noise.
But, since then,
listeners have become educated to high fidelity. And the fidelity,
both of equipment and of recordings, has improved tremendously.
Nowadays, the comparative novice in high fidelity can tell the difference
between bass boost and treble cut, or vice versa. It requires a
little more education to discern the difference between records
played with the correct equalization characteristic and with the
wrong one, but it certainly is possible to tell the difference.
But, a decade or so ago, strong arguments were put forward
that the tone control should serve this function in addition to
any other tone compensation that may be desired. Nowadays, any self-respecting
preamplifier provides equalization characteristics for RIAA, and
usually one or two other recording characteristics that may be encountered;
and the average high-fidelity listener appreciates this position
and reserves the use of tone controls for their proper function
- adjusting for differences in balance that may become apparent,
due to differences in the acoustic characteristics of the studio
and the listening room, or due to any other variations in the balance
of musical composition of the program.
Now the extremist
comes along with the loudness control and says this also is a must,
in addition to tone controls and equalization. Are we going to accept
this further stage as a necessity, along with equalization, or is
this really a luxury that is completely unnecessary?
would seem to depend upon the accuracy of your perceptive powers.
This can be illustrated at Fig. 2, which shows a comparison. between
the range of control given by the average tone control and the kind
of change in response needed in a loudness control, to follow the
variation predicted by the Fletcher-Munson curves. For the tone
control to get enough boost in level at the low-frequency end, gives
too much boost at frequencies higher up. The boost has to start
in the region of 500 cycles. whereas the loudness contour shows
not very much difference at 500 cycles based on true loudness contours.
As the region in question is from 100 to 500 cycles, which corresponds
to the middle register of music, use of the low-frequency boost
in conjunction with the volume control will over-emphasize the middle
register which is not at all what is required for true loudness
impression. So the average tone control puts the low-frequency accentuation
in the wrong place, and by the wrong degree.
Fig. 3. Tapped pot used as loudness control.
If you can
hear this difference. which an educated high-fidelity listener certainly
can, then a loudness control is definitely an advantage, in comparison
with the use of a volume control with separate tone controls.
The next question is, what arrangement of loudness control
and gain or volume control should be used? Do we need both controls,
or is one control that can serve both purposes enough?
preamplifiers have appeared with a single control, and a loudness
volume switch, which alters the function of the control from one
to the other. A little thought will show that this achieves practically
no advantage over an amplifier provided with only one or the other.
To be effective, a loudness control must produce the right
contour appropriate to the level at which the program is played.
As different programs are often recorded at quite widely differing
levels (as regards the input delivered to the preamplifier), there
is no guarantee that setting the volume/loudness dial to a specific
position (say 5) will always reproduce a program at the same loudness,
regardless of what disc is being played. But this is how a loudness
control should operate, otherwise the loudness contours will not
come in the right place.
What we need then is two separate
controls: a gain or volume control, to adjust the input received
from the recording, so that setting the loudness control to a particular
position will give the right loudness in the room for that setting.
For example, if you normally listen at a level corresponding to
50 in the loudness contour family, the loudness control should be
set to a position which makes this correction from the average recording
level of 70. Then the gain control should be adjusted on each particular
recording, so the loudness in the room corresponds with the average
to which you have become accustomed.
Then, once you have
set the gain control for this particular degree of loudness, you
can alter the loudness control to play the music softer or louder.
Of course, you do not always have to go through this routine just
to get it right for a particular recording. It is much simpler to
set the loudness control for the desired loudness. This you will
become used to with a little practice. Then adjust the gain control
so the actual loudness corresponds with the setting on the loudness
Some loudness controls, instead of providing a
continuously adjustable range, use a multi-position switch, that
may have three positions for example, marked low, normal and high,
referring to the degree of loudness. This control can then be set
according to how you want the music to play in the room and the
normal volume control can then be used to get the loudness right.
This will generally give a pretty close approximation to the right
loudness contour and from this point the tone control can be used
to adjust for slight deviations in balance, if it does not sound
to the best advantage.
The switched arrangement has the
advantage that the correct contour can be tailored into the circuit.
Some loudness controls use just a simple tapped potentiometer, the
schematic of which is shown in Fig. 3. This does not give sufficient
steepness to the loudness compensation, particularly for playing
at low levels. In fact, it is little better than the use of the
tone control. Maybe, if you have a separate tone control and a loudness
control of this type, the two can be used together to get results
that come nearer to being right than just the tone control by itself.
But the best arrangement, if you have a loudness control
that is continuously variable (like most volume controls), is to
use one which comes as a ganged potentiometer, two or more decks
with associated components wired on to the potentiometers to produce
the required variation of loudness contour as the control is turned.
But discussion of the precise way of getting the right characteristics
into a loudness control is another article altogether. It must be
said that, after becoming accustomed to the correct use of a good
loudness control, going back to a system that only has the regular
volume and tone controls seems to lack the degree of smoothness
in reproduction obtainable with the loudness control. Especially
is it difficult to get satisfaction for low level listening, which
is where the good loudness control really does its job well.
This can easily be demonstrated on a preamplifier that has both
forms of control, playing a good wide-range program that has plenty
of bass and treble. Turn down, first of all, the loudness control
and the effect is that the orchestra or program is being played
more quietly, but still in the room with you. Now restore the loudness
control to the maximum comfortable level and turn down the volume
control. This time the effect is not of the program being played
more quietly in the same room, but it seems to go farther away as
if it is no longer in the room with you but playing from somewhere
down the street.
Sometimes it is nice to have both of these
facilities available, but if you principally want to present music
that sounds good, you certainly want to have it sound as if it is
in the room with you and not being played from a distance away.
So it looks as if we really need loudness controls.
a final word, however. This decision is not going overboard on the
second extreme mentioned at the beginning of this article. The Fletcher-Munson
curves may be (and are) scientific, but they do not represent the
exact loudness contours of your ears. In their investigation, it
appeared that individual hearing is so divergent that it is extremely
unlikely that anyone has average hearing. But use of these
average curves assures a control that should satisfy.
each of us has hearing that diverges one way or another from these
curves, they represent the shape of almost everyone's hearing characteristic.
This means that our individual loudness difference contours will
be very close, although the actual contours deviate widely. So a
well-designed loudness control, based on loudness differences,
will sound right almost anyone.
Posted November 23, 2013