Before we begin our consideration of loudspeakers per se, we need to consider how they are regarded by humans. Here we have a very special case: humans tend not to perceive loudspeakers as sound sources. Instead, we are strongly inclined to perceive the virtual source signals in the program material presented by the loudspeaker. We assign cognitive identification to those virtual sources, including their imagined position(s) and environs. This tendency is pervasive, and it tends to make loudspeakers sonically imperceptible as such. We call this substitution of illusion for reality a “willing suspension of disbelief.”

The cognitive act of substitution of imagined signal sources for the physical reality of the loudspeaker generally blinds us to the real qualities and characteristics of the loudspeaker and its behaviors. Fortunately, such an act of cognitive substitution also greatly eases the task faced by recording engineers and producers in terms of creating an illusion of a “previous sonic event.” Listeners willingly accept, even seek out, such an illusion and consciously regard the virtual sources as perceived reality in place of the actual physical source.[1] The action is voluntary, but generally not acknowledged by the listener.

Functional Perspectives from which to regard Loudspeakers

There are numerous ways to think of a loudspeaker, based on specific applications and functions. The following perspectives are not mutually exclusive, but rather overlap significantly. However, they emphasize different characteristics of the loudspeaker and illuminate qualities that may not be recognized from other perspectives. These perspectives include: Reinforcer, Mimic, Reproducer, Unique Source, and Musical Instrument.

The Loudspeaker As a Reinforcer.

One view of the loudspeaker is as a “reinforcer.” In such a worldview, the loudspeaker “makes other instruments sound louder”. Such an approach implies that loudspeakers “change” other instruments, instead of replacing them (the change being to make them louder). While the loudspeaker may actually in fact acoustically replace the source instrument(s) by masking it with a louder iteration of the source sound, it MUST NOT APPEAR to replace it. Such an illusion is central to the use of loudspeakers to support live instruments and voices in sound reinforcement for theater, concert halls and any other live performance.

The Loudspeaker As a Mimic.

The loudspeaker can also be thought of as a “mimic,” which is generally the viewpoint held by synthesists and music producers. In such a world-view, the loudspeaker “mimics” the sounds of other sound sources: a flute, a snare drum, a rock band, an orchestra, etc. Questions of accuracy and realism exist. Does the loudspeaker sound “exactly” like a guitar? In this perspective, we are concerned with the loudspeaker’s sonic resemblance to other sound sources. We also encounter distinctions between “natural” sound sources (i.e. a flute) and “artificial” sound sources (i.e. the loudspeaker).

The Loudspeaker As a Reproducer.

The most traditional way to think of a loudspeaker is as a “reproducer.” As a reproducer, the loudspeaker and its drive mechanism “reproduce” a previous sonic event, as observed by a microphone(s) and stored as an electronic signal trace(s). Instead of being concerned with the loudspeaker’s sonic resemblance to other sources, we are concerned with it’s accuracy of reproduction of a prior event.

When thinking of a loudspeaker in these terms, questions of “accuracy” and “realism” arise, and the metaphor of loudspeaker as lab instrument is applicable. For instance, how “closely” does the reproduction resemble the original? Is the reproduction indistinguishable from the original? If not, how different is it, and what are the causes of those differences? Such questions, of course, imply the existence of a previous “sonic event” and the answers are also dependent on the recording practices used. In current commercial recording practice, there is often no such unique coherent event.

The Loudspeaker As a Unique Sound Source.

From a fourth viewpoint, we can think of a loudspeaker as a Unique “Sound Source With(out) Its Own Sound Quality.” This is an unconventional view that arises from the willing suspension of disbelief, but nonetheless a quite useful perspective for studying the behavior of loudspeakers.

It is important to note that we can simply and unequivocally imagine or remember the sound of a tuba, without making mnemonic reference to any other sound. This is generally impossible to do with loudspeakers. We have no way to “imagine” what a loudspeaker sounds like, except in terms of how it “sounds” relative to its abilities to reproduce, mimic and reinforce other sounds, or in terms of its imperfections.

What is the characteristic sound, or timbre, of a loudspeaker? It has none, or rather, it has an essentially infinite range of such timbres. Therefore, the ideal loudspeaker apparently has NO particular distinctive timbral characteristics. We can distinguish a bassoon from a loudspeaker, for instance, simply because the bassoon has a constrained set of “characteristic” sounds and a specific timbral character and range, while the loudspeaker does not.

This leads us to a description of an idealized loudspeaker, as a device that can create:

any reasonable sonic spectrum, with
any reasonable temporal envelope, at
any reasonable loudness, in
a reproducible manner.

As such, the loudspeaker can be regarded as a “Universal Sound Generator.”

To date, there have been two significant constraints for such a loudspeaker, constraints that both limit its performance and tend to cause it to have both a distinct timbral behavior in spite of the above criteria, as well as significant variability in reproduction.

These constraints are that a typical loudspeaker can only obtain its optimum performance in:

one direction, and
in a free field (or very large volume space, where the Schroeder frequency is near or below 20 Hz.).

These constraints are significant because essentially all applications of loudspeakers (for listening to music) occur in small reverberant spaces. Timbre is rendered variable as a function of degraded off-axis spectrum and energy folded back into the room via early and reverberant reflections, at both high and low frequencies (short and long wavelengths, relative to room size).

There is an additional feature of loudspeakers that needs to be noted in this regard. Loudspeakers, when combined in phase-locked arrays, can:

reproduce or mimic any reasonable acoustic space within another smaller space,
reasonably represent the phantom position of any reproduced sound source within that virtual space, and
reasonably envelop the listener within that virtual space.

Therefore, a phase-locked array of loudspeakers can be thought of as a “Universal Sound Environment Generator.” The explanations regarding why this is so are beyond the scope of this paper.

It is these qualities of “Universal Sound Generator” and “Universal Sound Environment Generator” that make the loudspeaker so hard to sonically characterize, as well as extremely useful to us. If the loudspeaker can make “any” sound, all sounds would appear to be equally “characteristic” of a loudspeaker. Another way to describe this is to say that the most distinguishing sonic feature of the ideal loudspeaker is its sonic featurelessness!

A requirement for such idealized loudspeakers is, of course, that the differences between them should be inaudible. This condition is necessary if reproducibility of common signals is to be consistent and reliable.

Meanwhile, there is a commonly held view that “accuracy” (which is actually defined as “the absence of error”) of reproduction is a desirable characteristic of loudspeaker performance, and that a more accurate reproduction is a better one. From this viewpoint, it follows that there must be a strong positive correlation between “accurate” and “beautiful.” If this is true, then we should find that as loudspeaker playback of music becomes more accurate, it should also become more beautiful.

However, current recording practice does not maintain rigorous standards of accuracy. There are no standards of absolute accuracy at all (absolute accuracy referring to conformance to absolute values, such as dB SPL, time-of-day, geophysical location or direction, etc.), and only a few standards of relative accuracy. No records are kept, for instance, of absolute magnitude, direction, or time. Further, little attention is paid to the relative accuracy of these attributes, nor is there any effort expended in maintaining them. They are routinely set aside, in practice, for the purpose of improving the “quality of the production,” at, of course, the expense of such accuracy. Similar changes are routinely made to spectra as well.

It has been my experience that while there is a significant overlap between the qualities of “accuracy” and “beauty,” particularly at the “poor quality” ends of their particular ranges, the qualities of accuracy and beauty tend to diverge as resolution increases, or at least as resolution reaches a point of diminishing enhanced audibility. In our recording practice AND in our loudspeaker design practice, when faced with this divergence, we generally choose to pursue beauty at the expense of accuracy, for very good artistic, aesthetic and professional reasons.

This is in conflict, of course, with the reproducer modality. Precise, accurate reproduction can be thought of as an activity suitable in the laboratory – the loudspeaker as a laboratory instrument. Such a function is not necessarily musical and can be inherently unmusical. That our current recording practice subverts the reproducer modality through the routine introduction of massive errors as a matter of course indicates how important this conflict is.

This leads us to a fifth viewpoint, the point of view that is the basis of this paper: the loudspeaker as musical instrument.

When we think of the loudspeaker from this perspective, we can think about it “playing music,” in fulfillment of all of the artistic premises and aesthetic principles that are present in musical performance in general. This turns out to be extremely useful and also quite revealing.

We are able to consider the timbral behavior of the loudspeaker, vis-à-vis other instruments and vis-à-vis the environment. We are further able to examine the aesthetic qualities of “loudspeaker music,” and to examine the differences between that particular instrumental genre and other instrumental genres.

In the light of conventional definitions, a loudspeaker certainly qualifies as a musical instrument. A loudspeaker “plays music” subject to the same acoustical laws and principles as any other musical instrument, and with the added characteristic that unlike any other musical instrument, the loudspeaker also undertakes to create “acoustic musical environments.” There are, of course, some interesting and unique features of the loudspeaker in this regard, the most important of which is that the loudspeaker is operated by remote mechanical control, in parallel with hundreds or thousands of other loudspeakers in other environments at other times, by an ensemble of performers (the artist, producer and engineer) that does not get to hear each individual loudspeaker or its environment, but can only estimate “what will probably sound good” in the broad range of cases.

At the same time, the loudspeaker also continues to function as Reinforcer, Mimic, Reproducer and/or Universal Sound Generator, depending on the specific application at any given moment.

1. Moulton, David - Numerous informal classroom experiments with students in Critical Listening and other audio classes. When asked to identify sound sources, no student has ever identified a loudspeaker, even though all examples have been via loudspeakers.

NEXT> Music performance

1 2 3 4 5 >

Login | Register

More Info

Mailing List

The loudspeaker as musical instrument

Humans’ willing suspension of disbelief in regard to loudspeakers