A primary reason you almost always see modern speakers using a vertical array of drive-units is that this places any crossover interference effects in the vertical plane, where they will be both predictable and have minimal effect on the speaker's perceived balance. When you place drive-units side by side, however, there will be suckouts to the speaker's sides, due to destructive interference when the path difference to the listening or measuring position combines with the phase shift due to the crossover to give cancellation of the two drive-units' outputs.

This can be seen in fig.4, which shows the S38's lateral dispersion. (This graph shows only the differences between the off- and on-axis responses.) The speaker is very directional in the midrange, though its dispersion is otherwise well-managed. The off-axis behavior is smoother in the treble on the tweeter side of the baffle—shown to this 3-D graph's rear—which presumably is why JBL recommends the tweeters be placed on the speakers' outside edges.

Fig.4 JBL S38, lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90 degrees-5 degrees off-axis on tweeter side of baffle, reference response, differences in response 5 degrees-90 degrees off-axis on woofer side of baffle.

In the vertical plane (fig.5), a big suckout appears at the upper crossover frequency for listeners sitting above the tweeter axis. The speakers should be used on fairly high stands, or at least aimed up at the listener if they have to be used on low stands.

Fig.5 JBL S38, vertical response family at 50", from back to front: differences in response 45 degrees-5 degrees above tweeter axis, reference response, differences in response 5 degrees-45 degrees below tweeter axis.

Fig.6 shows the JBL's step response, taken on the tweeter axis. The first up/down spike is the tweeter's output, followed first by a down/up step from the midrange unit and a slower, up/down step from the woofer. The tweeter and woofer are thus both connected with positive acoustic polarity, the midrange unit with negative polarity. This combines with the phase shift due to the crossover filters to give a summed flat response in the crossover regions.

Fig.6 JBL S38, step response on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Finally, the S38's cumulative spectral-decay or "waterfall" plot (fig.7) is relatively clean for what is quite an inexpensive speaker. Nevertheless, it can be seen that the on-axis suckouts mentioned earlier are each associated with a ridge of delayed energy. While the lower-frequency suckout probably arises from some sort of cancellation, the ridge of energy at 7kHz indicates a resonance of some kind, probably in the midrange cone.

Fig.7 JBL S38, cumulative spectral-decay plot at 50" (0.15ms risetime).

All in all, this isn't a bad set of measurements for a budget-priced speaker. In keeping with JBL's tradition, its high sensitivity and generous bass make the S38 a speaker that will suit rock music. But I would use it on stands away from room boundaries rather than on bookshelves close to room boundaries.—John Atkinson