Cambridge SoundWorks Ambiance loudspeaker Measurements

Sidebar 2: Measurements

Regarding measurements, I use a mixture of nearfield, in-room, and quasi-anechoic FFT techniques (using the MLSSA system from DRA Labs) to investigate objective factors that might explain the sound heard. (Stereophile's measuring microphone is a calibrated B&K 4006.) The speakers' impedance phase and magnitude were measured using Stereophile's Audio Precision System One.

Looking first at the speaker's impedance (fig.1, note that the phase angle is incorrectly inverted), the sealed-box woofer tuning is revealed by the peak around 75Hz. As the impedance only drops below 8 ohms in the upper bass and ranges between 12 and 16 ohms in the treble, the Ambiance will be an easy load for even inexpensive amplification to drive, particularly as the phase angle is also minimal throughout the midrange and treble. Its low sensitivity, however, will mean that realistic sound levels will not be within its reach, even with 100W amplifiers. (Assessed with a 1kHz-centered, 1/3-octave warble tone, the Ambiance is 3dB more sensitive than the LS3/5a, but 5–7dB less sensitive than the Celestion 3 and Monitor Audio 7.)

Fig.1 Cambridge SoundWorks Ambiance, electrical impedance (solid) and phase (dashed). (2 ohms/vertical div.)

Turning to the time domain, fig.2 shows the anechoic portion of the Ambiance's impulse response, 48" away on the tweeter axis, as calculated by the MLSSA software. (The measurement bandwidth was 30kHz, equivalent to a sampling rate of just under 91kHz.) The shape is typical of a design with a second-order crossover, while the HF ripples are presumably due to a tweeter resonance. What appears to be a reflection around the 7.5ms point is not from anything in the room, the speakers being sited some 4' above the floor and 5' from the walls for this measurement. (Simple trigonometry indicates that the first reflection in these circumstances, that from the floor, will arrive about 8ms after the first arrival.) I suspect that this reflection is related to the stand used. The step response is shown in fig.3; it indicates that both drive-units are connected in positive acoutsic polarity.

Fig.2 Cambridge SoundWorks Ambiance, impulse response on tweeter axis at 48" (5ms time window, 30kHz bandwidth).

Fig.3 Cambridge SoundWorks Ambiance, step response on tweeter axis at 48" (5ms time window, 30kHz bandwidth).

Performing a Fast Fourier Transform on the anechoic section of the impulse response gives the speaker's anechoic frequency response above 200Hz or so on this axis, which is shown to the right of fig.4. A strong peak in the upper midrange can be seen around 1400Hz, above which the highs gently roll off, relieved by what is presumably a resonance peak in the top octave. The lefthand side of fig.4 shows the woofer's nearfield response measured with the microphone against the grille (which I was unable to remove). The response starts to slowly roll off below 110Hz, with a half-power point at 63Hz. For comparison, the larger Celestion 3, also a sealed box, has a nearfield half-power point of 53Hz. Despite Cambridge SoundWorks' claim that it "has more output in the 40Hz region than any speaker its size we've encountered," the Ambiance is definitely a mini-monitor in this respect.

Fig.4 Cambridge SoundWorks Ambiance, anechoic response with grille on tweeter axis at 48", corrected for microphone response, with the nearfield woofer response plotted below 300Hz.

Table 1: Distortion Levels


This is also revealed by looking at the distortion levels in the bass, shown in Table One. (The spls indicated are those measured at an 18" distance at 50Hz and 100Hz, equivalent, due to the speaker's early LF roll-off, to midband levels of around 86dB and 96dB.) At any reasonable spl, the 4.5" woofer is working very hard at low frequencies, the third harmonic dominating at 50Hz and the second at 100Hz. (For those not familiar with my preferred way of expressing distortion as a dB level below the level of the fundamental: –10dB is equivalent to 31.6% distortion; –20dB, 10%; –30dB, 3.2%; –40dB, 1%; –50dB, 0.32%; and –60dB, 0.1%.) The Ambiance is not a speaker suitable for the Annual Headbangers' Ball.

Fig.5 shows how the speaker's anechoic response translates to its subjective balance in a real room. The curve is derived by measuring the spectral balance of left and right speakers individually, using pink noise and an Audio Control Industrial 1/3-octave analyzer at 10 different positions in a window around the listening position. Averaging the 20 different spectra eliminates the effect of LF room resonances on the measurement to a large extent, while it also integrates the direct sound of the speaker with that of the delayed room reflections to an extent that mimics the behavior of the ear. The mid-treble can be seen to have a rather peaky nature, which correlates with the somewhat lispy nature of the speaker noted during the auditioning. The smooth response trend through the midrange and low treble, however, is impressive for such an inexpensive model, and the individual spectral measurements taken to derive fig.5 indicate that the Ambiance has a smooth and wide dispersion above 1kHz, implying a lack of criticality concerning optimum positioning. The speakers were well away from the room boundaries for this measurement, which explains why the bass rolls off below 160Hz. Note, however, the gentle rate of the LF roll-off, the 30–50Hz band being on average 10dB down from the midband level.

Fig.5 Cambridge SoundWorks Ambiance, spatially averaged, 1/3-octave response in JA's Santa Fe listening room.

Finally, I looked at the Ambiance's resonant behavior using MLSSA. The classic "waterfall" or cumulative spectral decay plot, showing how the speaker's response changes as the sound decays, can be seen in fig.6. The plot suggests that the peak in the high treble, centered on 14kHz, is mainly due to one or more resonances, though with the exception of the 14kHz one, these decay quickly. Lower down in the treble, the two small peaks at 3450Hz and 3750Hz, as well as that an octave higher, are also resonant effects, as is evidenced by the prominent and long-lasting ridges that develop at these frequencies. Though these look like two very small peaks in the frequency domain, their persistence and the fact that they are positioned right in the ear's region of maximum sensitivity will undoubtedly contribute to the subjective lispiness and feeling of emphasis in this frequency region. A strong resonance can also be seen at 1420Hz, the cursor position, which might correlate with both the coloration and the lack of clarity noted in the upper midband.

Fig.6 Cambridge SoundWorks Ambiance, cumulative spectral-decay plot at 50" (0.15ms risetime).

The overall impression I get from this set of measurements is of a competent design executed from what are actually not very promising ingredients: I suspect that this is a recipe for commercial success with an inexpensive loudspeaker.—John Atkinson

Cambridge SoundWorks
120 Water Street
North Andover, MA 01845
(800) 367-4434