Krell LAT-1 loudspeaker Measurements part 2

How audible will this behavior be? The reflections themselves fall within the time window where the ear-brain integrates sounds as all being part of the first arrival, and so will not be heard as separate arrivals. In addition, the measured peaks and dips above 10kHz occur in a region where the ear's sensitivity is reduced—technically, they are all in just one "critical band"—meaning that the measured comb-filtering's audibility will be minimized. As LG was impressed by the Krell speaker's natural tonal quality in the treble, I must assume that the grille plays an essential role in the LAT-1's frequency shaping, or "voicing." And, it must be remembered, so does the speaker's off-axis behavior.

Because of the Krell's 250-lb weight, I could not maneuver it onto my speaker-measurement turntable (whose stepper motor would not have been able to handle that mass anyway). My measurements of the LAT-1's off-axis behavior therefore had to be arranged by a combination of moving the microphone and rotating the speaker on the wheeled dolly on which I'd had the speaker unpacked. There is therefore probably more error in the precise lateral off-axis angle than with the other speakers I measure, and, outside of 15 degrees off-axis, I spaced the measurements at 15 degree increments rather than the usual 5 degrees.

The result, with the grille on, is shown in fig.5, with the off-axis traces normalized to the tweeter-axis response. There is a slight off-axis energy excess in the upper midrange that will tend to fill in the depression in the same region in the on-axis response and better balance the slight excess in the presence region. But the tweeter's top octave falls rapidly to the speaker's sides, which might be expected to result in a loss of perceived "air" to the LAT-1's balance, especially in large rooms. This will, to some extent, be compensated for by the speaker's treble tilting up a little for listening axes below the 44"-high tweeter (fig.6). This graph also shows that, for axes very much above the tweeter's, a severe suckout develops in the presence region. The primary reflection of the speaker's output from the ceiling will therefore lack energy in the same region where there is an on-axis plateau, which will again tend to make the speaker sound more neutrally balanced in the low treble than its on-axis response might suggest.

Fig.5 Krell LAT-1, lateral response family at 50", normalized to response on tweeter axis, from back to front: differences in response 90 degrees-5 degrees off-axis, reference response, differences in response 5 degrees-90 degrees off-axis.

Fig.6 Krell LAT-1, vertical response family at 50", from back to front: differences in response 20 degrees-5 degrees above tweeter axis, reference response, differences in response 5 degrees-10 degrees below tweeter axis.

That the grille acts as an acoustic obstacle, particularly with respect to the tweeter output, can be seen in the LAT-1's impulse response without (fig.7) and with (fig.8) the grille. The latter features multiple high-frequency arrivals compared with the former. The step response (fig.9) indicates that the tweeter and midrange units are connected with positive acoustic polarity, the woofers with negative, but that the differentiated step of each drive-unit hands over smoothly to that of the next lower in frequency, implying good integration of frequency response.

Fig.7 Krell LAT-1, impulse response without grille on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.8 Krell LAT-1, impulse response with grille on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Fig.9 Krell LAT-1, step response with grille on tweeter axis at 50" (5ms time window, 30kHz bandwidth).

Finally, while the LAT-1's cumulative spectral-decay plot (fig.10), taken with the grille on, shows some delayed energy in the top octave due to the reflections, it is basically very clean. However, a slight ridge of delayed energy can be seen at 2.7kHz, the frequency of the on-axis peak.

Fig.10 Krell LAT-1, cumulative spectral-decay plot with grille at 50" (0.15ms risetime).

Designing a successful-sounding speaker is very much a matter of balancing areas of performance against each other, and I don't think I've encountered a design of which this has been more true than the LAT-1. There are individual aspects of the speaker's measured performance that raised my eyebrows somewhat, yet Larry Greenhill was very impressed with its sound, as was I when I auditioned it both at Home Entertainment 2001 and in one of Sound By Singer's soundrooms a few months back. It appears that, with his first speaker design, Krell's Dan D'Agostino has worked hard at achieving a successful balance.—John Atkinson

45 Connair Road
Orange, CT 06477-3650
(203) 799-9954