Basso Profundo:
Page 4 In a room 13' wide, 17' long, and 9 ½' high with an average low-frequency loss, this means the speakers will sound tolerably well-balanced if placed with their radiating centers about 2' from the floor and 3 ½' from the side walls, plus 2 ¾' from the back. Note that these three dimensions are unequal, with the deliberate intention of avoiding a coincidence in spacing to these three local boundaries. Interferences from the energy reflected by the local boundaries is thus spread over a range of frequencies, reducing their effects. While interference produces sharp notches in the frequency response, the general action of the local boundaries is to act as reflectors, progressively boosting the bass level. This boost is real enough, typically adding 6dB in the 30-40Hz range. This is highly convenient: If our compact speaker is well-tuned in the bass and critically damped for optimum transient performance, it can be designed to be typically -6dB in absolute response at the low-frequency limit. This is why good small speakers, when properly aligned and placed, can produce reasonable bass down to 35Hz in many rooms.

Room boundaries
In order to provide an agreed basis for the mathematical analysis of the low-frequency response of loudspeakers, academic acousticians made the assumption that loudspeaker theory is based on "2Pi" environments. The "Pi" nomenclature refers to the solid angle, in radians, into which the speaker fires: 4Pi is thus free space, conferring truly spherical radiation at low frequencies, while 2Pi represents a hemisphere where the speakers are effectively flush-mounted in an infinitely large wall. However, a room is rather more contained than this, and at frequencies below 500Hz the radiated sound wavelengths are sufficiently large to suffer both cancellation and boosting, or reflection, from the local boundaries.

For a "free-space," stand-mounted speaker system, the first cancellation is at approximately 160Hz and is due to the floor. The back wall comes into play next, followed by the side walls. In-phase reflection results in a theoretical boost of 3dB per boundary; with idealized non-coincident spacings and good room proportions, the maximum lift could be 2.5dB at 80Hz, 6dB by 50Hz, and 8dB by 35Hz. This underlies the uneven but progressive room gain present at lower and lower frequencies. In normal rooms, it explains why a small speaker with a free-space response 6dB down at 50Hz under test chamber conditions can still produce some audible 40Hz in the listening room. It also explains the oft-noted difference in bass quality between traditional acoustic-suspension and bass-reflex systems. The former have been described as providing more even, more extended, and less colored bass than the latter.

Traditionally, reflex systems have been described as "boomy," yet are often found to provide less power in the low bass than the specifications suggest. Those specifications were based on test-chamber measurements and a rated -3dB rolloff in the bass. However, the important factor turns out to be what happens below rolloff. The aural sensitivity to changes at low frequencies is great, and we have room gain adding to the available response, helping to counteract the rolloff. A well-damped sealed-box system can have a desirably slow rolloff rate significantly complemented by room gain, thus extending the overall useful response. Below box resonance, the output from a bass-reflex system usually falls rapidly, too quickly for the room boundaries to help out; no low bass is heard. In addition, the reflex system is likely to have a sharper, squarer response "cover" at rolloff; mild room gain at this frequency can easily turn the corner into an audible lump---the notorious "boom."

Variation due to speaker type
There is insufficient space here to cover various speaker types, though the open-panel dipole is worth some attention. A box speaker's lower range increases in level as it nears the floor, wall, or corner. Perversely, for a panel speaker---Magneplanar, Apogee, and particularly the super-light diaphragmed electrostatic type such as the Quad---the effective low-range output is decreased. When very close to the wall, the output from a Quad approaches zero as the listening back wave is in opposition, out of phase with the moving element and almost canceling out the audio output. Moreover, the cancellation is anything but uniform, as the cancellation zeros appear at multiples of the audio wavelength. This "comb" of notches presents quite serious coloration and explains why the positioning of panel speakers in rooms is highly critical.

Generalizations are not possible here. Full-range electrostatics perform best well away from the back wall. For example, the Quad ESL-63 can produce serious bass in large spaces, but becomes lightweight when backed up in a small room. Conversely, a full-range Apogee requires either a smaller room where a tendency to bass excess can be judiciously tamed by exploiting some back-wall cancellation, or a very big space where little or no room lift is present in the main low-frequency range.

Active control
Devices are available which can be placed in regions of greatest low-frequency standing-wave energy, the Phantom Acoustics Shadow reviewed by RH in Vol.12 No.12 being the only example commercially available in the US. Set correctly, these monitor the incoming waves and generate anti-phase energy, sapping the power of the standing wave. If used with care, this expensive technology can be quite effective.

Summary
Opinions vary greatly concerning reproduced bass sound. Many factors play their parts, from such acoustic fundamentals as the standing wave or resonant behavior of these long audio wavelengths in small listening rooms, to the influence of loudspeaker type (open-panel or box designs). Our hearing characteristics also play a part, with a combination of poor absolute sensitivity in the bass and greater sensitivity to changes in level in the audible range.

Subjective characterization of bass quality is important, determined both by the final frequency response perceived by the listener and its qualities of time coherence with the remainder of the frequency range. A rhythmically involving if restricted bass may prove of higher quality in terms of enjoyment than a more extended bass register lacking pace and speed. Sheer bass quality, in level or extension, is no yardstick by which to measure musical quality. The negative influence of the normal technical presentations intended to depict bass sound quality needs to be appreciated. In the usual frequency-response graph the bass region is not accorded its rightful visual weight. In the light of the psychoacoustic responses there is a good case for weighting the bass region by an expansion of, say, 1.5x in amplitude and 2x in frequency on the graph, to allow the observer to make more accurate assessments of bass performance.

At very low frequencies, every Hz counts!