The Timbral Character of the Loudspeaker

Given the above, the following informal observations can be made about the timbral behavior of traditional loudspeakers:

Divided Spectrum

No single transducer can reliably reproduce a bandwidth of greater than 3 octaves (8:1) without electronic intervention. Therefore, common practice is to distribute the ten-octave audible spectrum among multiple transducers. Timbre is audibly affected by the choice of transducers, their assigned bandwidth, the crossover circuitry used to distribute spectra to the transducers, and their relative positions in an assembled loudspeaker system.

Formants

All acoustical instruments have distinctive sets of formants, or resonant structures that are a consequence of the size, shape and materials of the instrument. Such formant sets are, in general, independent of the pitch or frequency content of a given note. The spectral character of the formant set is probably more important to the perceived timbre of a given instrument than is the specific spectrum of any given note played on that instrument. In fact, the individual note spectra (and waveform) change dramatically from note to note.

The loudspeaker, on the other hand, seeks to have no formants or resonant structures, except as induced by ported low frequency enclosure topologies and as a natural, if unintended, consequence of crossover behaviors (i.e. the effects related to two drivers at different points in space generating the same frequency, a behavior that is unique to loudspeakers among musical instruments). This speaks to the Universal Sound Generator quality of loudspeakers. Unlike all other instruments, the acoustically observed spectrum and shape of any given waveform played by the loudspeaker should remain constant, regardless of frequency.

Directivity at low, mid and high frequencies

The output of loudspeakers tends to be omnidirectional at low frequencies and unidirectional at high frequencies, yielding a power response that declines with increasing frequency. Significant interference effects exist at crossover frequencies. These directional behaviors may be the single most characteristic timbral element of loudspeakers, vis-à-vis other instruments.

Crossover Design

Crossover design and implementation has a pervasive effect on timbre. Numerous schools of thought and practice exist. These all introduce variability into the timbral behavior of loudspeakers.

Low Frequency Performance

Low frequency performance of loudspeakers is generally compromised by the desire to reduce the cost and size of loudspeakers. Numerous schools of thought and practice exist regarding low-frequency transducer/enclosure topology, again having a significant effect on timbre.

Interaction with rooms at low frequencies.

A loudspeaker’s performance varies as a function of the environment it is operating within, its position within that environment, and the individual frequency content at any given moment. There is no perfect solution to the problem of this variability.

A Brief comparison with conventional acoustic instruments.

Below is a table showing informal estimations of the frequency ranges, approximate directivity and sound pressure level of a variety of acoustical instruments vis-à-vis both a typical good loudspeaker and an ideal one.

Instrument	Low freq. fundamental	High freq. fundamental	Overtones to:	Low freq. radiation directivity	High freq. radiation directivity	Min dB SPL (3m)	Max dB SPL (3m)
Flute	250	2K	10k	omni, transverse	forward horizontal	50	90
Clarinet	150	1.2K	5K	omni, downward	lateral horizontal	50	90
Oboe	250	1.2K	15K	weak omni	omni	60	85
Bassoon	75	400	15K	weak omni	omni	50	80
Saxophone, tenor	125	600	15K	omni, forward	lateral horizontal	55	90
Trumpet	180	1.2K	20K+	forward	lateral, vertical	60	100
French Horn	100	800	5K	rearward	rear, lateral	55	95
Trombone	40	500	20K+	forward	lateral, vertical	60	105
Tuba	30	250	8K	forward	omni	60	90
Violin	200	2K	20K+	weak omni	Vertical	45	85
Viola	150	1.5K	20K+	weak omni	Vertical	45	80
Cello	60	1K	20K+	forward omni	Forward	45	85
Doublebass	40	500	20K+	forward omni	Forward	45	80
Snare Drum	200	4K	20K+	vertical, horizontal dipole	omni	55	105
Cymbal	500	10K	20K+	vertical, horizontal dipole	omni	50	95
Tympani	40	250	1K	omni	weak omni	40	95
Glockenspiel	250	1K	20K+	omni	omni	60	95
Marimba	250	1K	2K	omni	weak omni	50	80
Xylophone	250	1K	5K	omni	omni	50	85
Triangle	1K	2K	20K+	none	omni	50	90
Bells	200	800	15K	omni, downward	horizontal plane	50	80
Bass Drum	20	100	2K	horizontal, vertical dipole	weak omni	40	105
Piano	30	2K	15K	omni, vertical, horizontal dipole	omni, forward	40	100
Pipe Organ	20	2K	12K	omni	omni	50	105
Loudspeaker (typical traditional)	50	20K	-	omni	narrow forward	0 dB	99
Loudspeaker (ideal)	20	20K	-	omni	horizontal	0 dB	120

Table 1: Various acoustical instruments vis-à-vis a typical “good” loudspeaker and an “ideal” loudspeaker, showing the range limits of their various fundamental frequencies and informal estimations of the upper limits of their significant harmonics, and their patterns of polar radiation at low and high frequencies, plus estimations of the minimum and maximum Sound Pressure Levels such instruments can obtain as observed at 3 meters in a free field.

This table illuminates several interesting things.

First, all traditional instruments generate complex waveforms with extensive overtone structures. As a general rule, the frequency range for the fundamental frequencies of musical notes spans four octaves, from 65 to 1040 Hz. (low C to high C), while overtones extend up to five octaves above that range. Loudspeakers, on the other hand, do not generate overtones as such, except when driven into distortion.

Second, the directivity of acoustical instruments varies widely at both high and low frequencies. There is no uniformity, and no particular trend. In a collective sense, we can generalize that high frequencies are radiated laterally and upward, occasionally forward, while low frequency radiation tends to be either bidirectional or somewhat directional as a function of horn coupling with free space. It needs also to be noted that the timbral results of such directivity is also affected by the comparatively large size of performance venues – the strident harshness of a violin in a small practice room, with extremely strong high frequency early reflections off a low ceiling is in marked contrast to the warmth of the same instrument in a concert hall, with the attenuated vertical high frequency radiation lending a silky reverberant patina to the sound.

In contrast, traditional loudspeakers have a comparatively distinctive pattern of directivity, as noted above, a pattern generally unlike almost all acoustical instruments. While low and mid frequencies are radiated approximately omnidirectionally, high frequencies are beamed narrowly, so narrowly that for the most part a listener cannot sit more than 15° off-axis and still perceive the spectrum the designer intended. The result of this is a particular interaction with rooms that sounds quite unlike other instruments.

It is therefore quite important to note that no other instrument requires such a precise orientation by the listener as does the loudspeaker, simply in order to take in the spectral range of the loudspeaker. Also, as noted earlier, the volley of early reflections are spectrally deficient vis-à-vis the on-axis radiation. This has led to compensations, in many instances, during the act of recording and production. By tradition, many recordings are overly bright to fill in for this spectral deficiency and to compensate for off-axis listening.

Finally, when we compare the maximum sound pressure levels that can be obtained by loudspeakers vs. acoustical instruments, we see that loudspeakers have a much larger dynamic range and appear to be equivalent or greater in range in almost all regards. However, comparisons here are extremely tricky when we start accounting for variables such as concert hall vs. playback room, listeners’ distance from the source, etc. Further, instruments aggregate into ensembles that may include as many as 250 members (orchestra and chorus). Finally, the subjective perceived loudness from many sources is not easily matched by the phantom images generated by a loudspeaker pair.

At best, loudspeakers can generate sound pressure levels in a small listening room that effectively mimic the subjective quality of loudness obtained by a full orchestra, but it is safe to say that they cannot replicate the amplitude at bandwidth that an orchestra can achieve. Further, the mimicry of a live rock or heavy metal band is always going to be reduced in subjective loudness (and this is probably as it should be).

We can further generalize that the amplifier power available, when coupled with the given sensitivity of most domestic loudspeakers, is usually insufficient to achieve such levels, introducing a further constraint on loudspeaker playback.