Floor Loudspeaker Reviews

Although we have long preferred the Janszen Z-600 to any other speaker system in or about its price range, is nonetheless an outstanding example of the old saw that nothing is perfect and loudspeakers are less perfect than most. For while has excellent overall balance, unsurpassed high-end range and transient response, superb detail, a deep, rich low end and a remarkanle degree of musical naturalness, it also has rather poor stereo imaging, a slight tendency toward a rising high end in most listening environments (because of its treble-beaming proclivities, a little less low-end range and definition than it might have, and a very slightly depressed middle range that takes some of the fatness and warmth from cellos and trombones and tends to make things sound a bit farther away than they are supposed to.

According to Neshaminy, the Z-960 is exactly the same as the Z-600, except that the woofer enclosure is larger (for better low end), the woofer cone is lighter (for better transient response and more midrange output), and there are three electrostatic elements instead of the two in the Z-600 (to give broader, more-uniform dispersion, for better imaging and less beaming).

Z-960 Sound Quality
We were not surprised to find that the Z-960 did not sound at all like the Z-600.

Our first reaction to the Z-960 at a hi-fi show demonstration was that it was superb. It had that detailed, liquid transparency that we have always liked in the Z-600, and it obviously had less of the Z-600's imperfections. And when we first tried them in our Stereophile listening rooms, we had the same reaction all over again.

After a while, though, we began to notice something amiss in the middle range. The sound was certainly fatter and more forward-sounding than that of the Z-600s, but there was a certain deadness about it that did not seem in any way related to high-end response. There were plenty of highs, with the usual superb detail and even greater smoothness than from the Z-600s. But trumpets sounded muted, woodwinds sounded subtly nasal, voices sounded a bit as though their owners had head colds, and the lower parts of the overtone range from many instruments sounded rather more prominent than they should.

Deactivating the tweeters showed that the coloration we observed was in the woofer section, but audio oscillator checks did not reveal anything that might pinpoint the cause of what we were hearing. So, we braved the North-Eastern winter weather and took one Z-960 and a good capacitor microphone outdoors (to avoid problems with room acoustics) for some response measurements

Since the microphone has never been calibrated, and the measurements were made under rather haphazard conditions, we can make no claims for the accuracy of the. results in terms of decibel values, but the final averaged response curve certainly showed a tendency toward the kind of deviations that would explain what we had heard.

Above about 1500Hz, where the tweeters start to take over, we could almost have drawn the measured response with a ruler, and the crossover region seemed free from the kind of sharp deviations that might suggest woofer/tweeter misphasing.

We did however find a fairly large, fairly broad response dip centered at about 600Hz, followed by a rise to narrow but moderate humps at around 150 and 300Hz, followed by a general downward trend below that.

The low-end rolloff, by the way, was not worthy of note because one speaker out of doors will never develop nearly as much low-frequency air pressure as two speakers within the confines of a room. Had the Z-960 measured flat in our tests, the two of them would have sounded much too bottom-heavy under normal-use conditions.

As a matter of fact, the low-end performance of both speakers indoors proved to be excellent, with usable response extending down to 35Hz, and with very good definition as well, despite an observed tendency for the rear panel of the enclosure to flap like a sheet when the woofer was pumping out lows.

There are many nice things that can be said about the Z-960, such as the fact that its high end is beautifully smooth, detailed and lucid, its stereo imaging is very good although not outstanding, and it is somewhat more efficient than most "low-efficiency" speaker systems. In fact, if we were to assign equal values to every aspect of loudspeaker performance, we would have to conclude that the Z-960 is at least as good as any other $300 system we have heard and is better than most.

But we don't really feel that one can assign equal values here, for if a speaker imparts significant coloration to the timbres of musical sounds, no amount of low-frequency range or treble smoothness or dispersion can make it a more accurate reproducer of music.

Just because we happen to feel this way does not imply or suggest that every other listener will agree with us, though. People who feel more strongly about performance aspects other than naturalness alone may well opt for the Z-960, simply because in most respects it is such an excellent system. But when we apply our particular set of sonic biases to a comparison between the Z-960 and the Z-600, and then note that the latter costs about $50 less, there is no doubt in our minds as to what our choice would be.

Z-960 Manufacturer's Comment
The report stating that the Z-960 did not sound "at all like" the Z-600, and the conclusions oased on the reviewer's response curves are totally inconsistent with the measurements, subjective and objective, that we have ever made at Neshaminy.

In our opinion, meaningful frequency response measurements can only be made under free-field conditions, which means in an anechoic chamber or in an open space outdoors. Laying the speaker on its back and placing the microphone above it reduces misleading measured response irregularities caused by phasing cancellations due to the surrounding planar surface (the lawn). To obtain even further accuracy, the speaker should be buried with its front surface flush with the ground in order to eliminate all reflective interference.

As can be seen from our frequency response curves (fig.1), taken outdoors ("free field") with the loudspeaker facing upwards, the response of the two units is almost identical except for the slight additional bass response of the Z-960. This is consistent with the original design parameters of the larger cabinet.

The second set of curves (fig.2), which were taken in a room too small to support deep bass, also shows the responses of the two speakers to be almost identical. In fact, through the range in question, the Z-960 has a slightly smoother response than the Z-600.

From a subjective viewpoint, we can only conclude that if the trumpets sounded "muted" and the woodwinds sounded "subtly nasal" and the voices sounded a bit as though the owners had "head colds," then the trumpets were in fact muted, the woodwinds were in fact "subtly nasal" and the singers, in fact, had "head colds" or maybe even the flu.—Frank Wetherill, Neshaminy Corp.

Reviewer's Addendum
Despite the imprecision of our measurements, there is a fairly close resemblance between our curve (fig.3) and those that Neshaminy obtained with both loudspeakers standing up. Since we did not, however, run any similar curves for a Z-600, we haven't any comparable objective measurements to present. All we can say is that, in direct comparisons under the same circumstances—ie, with the tweeters turned off and both speaker systems auditioned under the same conditions—the Z-960s sounded considerably more colored than the Z-600s.

As for the measurements involved, we do not feel nearly as confident about these as Neshaminy apparently does. A high-fidelity loudspeaker is after all, intended for use in a typically echoic room rather than in an anechoic chamber or out on the lawn. Since every room adds its own irregularities to the frequency response, particularly at middle and upper-bass frequencies, free-field tests are the only way of determining the potential smootness of a speaker through these ranges. And the tester must bear in mind that the deep-bass output of any loudspeaker is going to be increased by the proximity of a room's walls and ceiling, as well as by augmentation from the second speaker of a stereo pair.

In addition, it is an established fact that the response smoothness of a speaker system is significantly affected by the size and shape of its enclosure. The right-angle corners of the typical box-shaped loudspeaker enclosure produce diffraction effects that add humps and dips to the response, and it has been demonstrated that even the rear edges of a cabinet contribute to these irregularities. And since the cabinet is decidedly an integral part of any loudspeaker, regardless of what room it is used in, response measurements must include the effects of cabinet shape in order to be a valid reflection of the speaker's potential performance.

Standing a loudspeaker on its back changes its entire diffractive characteristics, and we suspect that it was this, rather than the elimination of cancellations due to the ground, which accounted for the difference that Neshaminy had observed. But while the speaker's response was measurably smoother with the cabinet lying on its back, it is very doubtful that the resulting curve would reflect the potential performance of ions of home use.

Burying the speaker flush with the ground just carries this self deception one step further, by eliminating virtually all traces of diffraction interference, which may explain why loudspeakers that have been design-tested in this way often have pronounced midrange coloration in actual use, unless custom-installed in a similar manner—that is, flush with the room walls.

Since Neshaminy did measure both the Z-960 and the Z-600 in their standing-up position also, we can only I guess that something audible (to us, anyway) was not being measured. Neshaminy informs us that their measurements were made with the test microphone on the woofer's axis. perhaps some off-axis tests might have been more revealing.

PostScript After the above was written, we noted with interest a similar discussion in the April 1969 issue of Stereo Review between Acoustic Research's Roy Allison and the magazine's tester-in-chief, Julian Hirsch.

Mr. Hirsch, in testing a new AR speaker system in "a normal 'live' room," had found a response discontinuity in the middle range. Mr. Allison denied that any such discontinuity existed, citing AR's own response tests in an anechoic chamber and dismissing Mr. Hirsch's findings as "probably due to conditions in the listening environment where the tests were made."

Like Mr. Allison, we are not convinced that measurements conducted in a live room can yield valid data even for comparative purposes, because the effects of standing waves and interference patterns on measured frequency response depend as much on the speaker's dispersion characteristics as on room acoustics. But anechoic chamber measurements too can yield misleading data if they do not include the effects of enclosure shape.

Some designers, for instance, make anechoic chamber tests with the speaker mounted flush in the wall of the chamber, leaving the corners of the enclosure outside where they cannot influence what goes on inside the chamber. This, pretty much duplicates the free-field test with the buried speaker, and the results are not likely to be any more valid in terms of actual performance. We do not know how AR conducts their anechoic chamber tests, but if these do not include the effects of the enclosure, they would not necessarily be much more meaningful than measurements obtained in a typical echoic room.
J. Gordon Holt