Floor Loudspeaker Reviews

Sidebar 3: Measurements

Starting with the Studio 20's impedance magnitude and phase (fig.1)—all the measurements were performed by John Atkinson, using DRA Labs' MLSSA system—we can see the double impedance peak in the bass, indicative of a reflex port tuning at 44Hz with the woofer tuning at about 70Hz. The minimum impedance is 5.5 ohms at 235Hz, and stays above 8 ohms beyond 800Hz. This moderate to high impedance, coupled with a benign phase angle (dotted line in fig.1), suggest the Studio 20s will present an easy load to an amplifier.

The 20's impulse response, shown in fig.2, is dominated by the tweeter's large ultrasonic ringing. An examination of the individual drive-units' step responses (fig.3) revealed that both drivers are connected with the same inverted polarity, and that the woofer exhibits some ringing at or above the top of its passband, probably from a high-Q cone resonance (more on this later).

The curve most revealing of the Studio 20 is fig.4, the response averaged over a 30° lateral window with the nearfield woofer and nearfield port curves attached (with approximate level matching). The most prominent feature is the peak at 5.7kHz and the "hole" in the response between 6kHz and 10kHz. This aberration is apparently due to a severe mismatch between the woofer and tweeter at the crossover, the 5.7kHz peak being the woofer's primary resonant mode (explored in more detail later). As bad as this looks on paper, the subjective perception was far less than one would imagine (footnote 1). My criticism of cymbals sounding "whitish" and lacking body is perhaps explained by the reduction in energy between 6kHz and 10kHz; cymbals' lower harmonics—which provide a sense of fullness—are attenuated in relation to the upper harmonics that contribute to their shimmer. Cymbals' "shhh" component was lower in level than their "ssss" component.

Except for this problem, the response is fairly flat, with a top-octave output well matched to the rest of the spectrum. The tweeter's very high "oil can" resonance can be seen at about 25kHz. Another feature is a gentle rise in output between 500Hz and 1.5kHz. Although this peak is low in amplitude, it covers a fairly wide bandwidth and is in an area where the ear is very sensitive. This perhaps explains my comments about certain midrange instruments lacking warmth and body; the region below 500Hz is shelved down in relation to the band between 500Hz and 5.7kHz.

The woofer's rolloff below 100Hz is augmented by the port's output (curve with peak at the far left-hand side), effectively increasing the system's extension. Some spurious output from the port can be seen between 250Hz and 350Hz, but this is well down in level.

Looking next at the family of normalized curves revealing the speaker's horizontal dispersion (fig.5), the lowermost curve, taken 30° off-axis (as when the loudspeakers are pointed straight ahead with the listener sitting in the nearfield), shows a shelved-down treble and a large dip in the crossover region. These curves show the respective differences from the on-axis response, so that the large peak at about 6kHz will fill in the on-axis mismatch between the two drive-units. Experimenting with the exact degree of toe-in will therefore give the optimum balance through the treble.

The vertical family of responses is shown in fig.6. The curves are, from top to bottom, taken at 7.5° above cabinet top, level with cabinet top, on the tweeter axis, midway between the woofer and tweeter, and on the woofer axis. The response is smoothest below the tweeter axis, but the 20's short stature means most listening positions will be higher than this. As the listening axis becomes lower, however, the woofer's resonant peak at 5.7kHz is emphasized. The rise in output between 1kHz and 3kHz is greater above the tweeter axis, and the treble suckout is also more severe. A low listening chair is therefore recommended.

The Studio 20's cumulative spectral decay, or "waterfall," plot is shown in fig.7. The peak at 5.7kHz is very apparent, accompanied by a ridge indicating the release of energy at this frequency after the input is removed—some sort of resonance. This is quite a severe problem and is explored in more detail later. The very high amplitude tweeter resonance is visible at 25kHz. Overall, however, the decay is quick and clean. (The magnitude of the ultrasonic tweeter peak uses up much of this graph's 30dB dynamic range, however, meaning that it looks cleaner than it really is.)

Looking at the waterfall plot of the woofer only (fig.8) confirms the hypothesis that the woofer's resonance was responsible for the peak at 5.7kHz seen in fig.4 and the ridge at that frequency in fig.7. In addition to the primary mode at 5.7kHz, there are a series of subsidiary modes, with a strong one appearing at about 17kHz. These modes are seen as peaks along the right-hand edge of the rolloff and the peak by itself in the top octave. Note also the broad and gentle rise between 500Hz and 1.5kHz seen in fig.6.

Further confirming the woofer's misbehavior is fig.9, which shows the FFT-derived response of the woofer (blue trace) and tweeter (red) separately. The woofer's output at its resonant frequency is equal to that of the tweeter, even though the electrical crossover point is 3.2kHz, a full 2.5kHz below the point where the drivers have equal output. Note also the woofer's other resonant modes, seen as spikes in its rolloff.

Finally, driving the Studio 20 with a variable-frequency sinewave oscillator produced fairly loud spurious noise from the port between 45Hz and 60Hz. The enclosure was quite inert through the bass, but was lively between 350Hz and 550Hz with a strong mode at 450Hz. These cabinet resonances were felt toward the enclosure top; the sand-filled bottom was better damped. The port noise may have been the source of the "wooden" character on some bass notes, and the cabinet resonance centered at 450Hz may have been responsible for the forwardness of some notes, especially noticeable on piano.

Had I seen the Studio 20's measurements without auditioning them, I would have been inclined to dismiss the 20s based on the severe amplitude errors between 5kHz and 10kHz and the woofer's resonance problems. Indeed, comparing the Studio 20's measurements to those of the Snell Type B also reviewed in this issue would indicate that the Type B is a far "better" loudspeaker. During the auditioning, however, I preferred the Studio 20s—by a wide margin. Despite the response irregularities, the 20s had a better ability to disappear into the music and make the listener forget he was hearing loudspeakers. Apart from my minor criticisms related to these aberrations (lack of warmth in mids, "whitish" upper treble), their subjective effects were far more benign than the measurements would indicate.—Robert Harley

---

Footnote 1: Bob Harley did all his auditioning before looking at the measurements that I had carried out earlier. When I listened to the Studio 20s in his room, therefore, I already knew about the woofer's resonance problem above crossover and fully expected to hear the drive-unit singing away. I was astonished, however, to find that I was only occasionally made aware of a problem in this frequency region, and then only on high-level choral recordings, such as the Hyperion Parry collection, where boy sopranos both acquired a little more flutiness and lost their exact tonal center on some high notes compared with others. But oh, did I find that open-sounding, hugely transparent soundstage thrown by the Studio 20s appealing!—John Atkinson