Nelson-Reed 8-04/B loudspeaker Measurements

Sidebar 2: Measurements

As I was due to drive up to Colorado with my daughter for the 1991 Denver Grand Prix (footnote 1) when JGH told me about the bass problem he was having with the Nelson-Reeds, I dragged along the MLSSA measurement setup to his new home in Boulder. My room measurements confirmed the presence of the midbass suckout when the speakers were set up symmetrically, though they also showed that JGH's room has a nicely controlled and even reverberant characteristic, ranging from 0.3s at 250Hz to 0.2s at 8kHz.

Fig.1 shows the spatially averaged in-room response from 10Hz to 10kHz taken at two positions on JGH's listening couch for left and right speakers individually. (This graph includes the effect of all the room reflections as well as the direct sound from the loudspeakers.) Note how flat the upper midrange and treble are, though with some interference dips noticeable in the mid treble. This Nelson-Reed certainly gives a neutral midrange and treble balance in-room. In the bass and lower midrange, however, things are not so straightforward. The peak between 90Hz and 180Hz and the dip below between 200Hz and 400Hz are likely to be due to the interaction between the woofers and the proximity of the floor.

Fig.1 Nelson-Reed 8-04/CM, 1/3-octave-smoothed response in JGH's listening room averaged across listening seat.

The subjective importance of this kind of effect is hard to judge; the human ear is used to live sound sources being affected in this way and is therefore able to tune it out to some extent. (See "Letters" in this issue, however, for a contrary opinion on this subject.) Below 90Hz, note that though there is useful extension to below 30Hz, there is also a lack of energy between 40Hz and 80Hz—a musically most important octave—relieved only by a slight peak between 60Hz and 70Hz. This presumably correlates with JGH's feeling that the 804s had a somewhat lean midbass.

I also took some quasi-anechoic measurements with MLSSA. To the left of fig.2 can be seen the individual woofer (green and blue traces) and port (red) responses plotted up to 200Hz, taken in the nearfield. Though the port response is quite a narrow bandpass, centered on 30Hz, the woofers themselves start to roll out below 95Hz, where there is a suspicious little peak. A simplistic view of these curves would be to suggest that the woofers act in an overdamped manner so that the reinforcement in bass output supplied by the port occurs too low in frequency to be fully integrated. I suspect that things are more complex than that, however.

Fig.2 Nelson-Reed 8-04/CM, quasi-anechoic response on listening axis plotted above 1kHz at listening seat (average of both loudspeakers), corrected for microphone resposne, with nearfield woofer and port responses plotted below 200Hz,

Fig.3 shows the nearfield step responses of the two woofers and the port. It can be seen that the two woofers (blue and green traces) act in the same phase (though in JGH's system, this is with their polarity inverted). The nearfield port output (red trace) is in the opposite polarity to the woofers, as expected, and rings for most of the 100ms time window shown, suggesting a Q somewhat higher than the 0.7 specified. (This kind of behavior is an audio analog of Heisenberg's Uncertainty Principle: the more you confine a drive-unit in the frequency domain, the more indeterminate its output becomes in the time domain. The same kind of behavior can be seen with the Snell Type B's rear woofer—see RH's review in Vol.14 No.12, December 1991.)

Fig.3 Nelson-Reed 8-04/CM, step responses of upper and lower woofers (blue and green, respectively) and of port (red).

But note the negative-going spike about 2ms into the plot in fig.3. I haven't seen this kind of behavior from a loudspeaker port before; it, too, suggests that the 804's woofer and port are behaving in a complex manner. I look forward to reading designer Bill Reed's thoughts on this matter in his "Manufacturer's Comment" letter.

To the right of fig.2 is shown the average of the quasi-anechoic in-room responses of both loudspeakers at Gordon's listening position. (This is only shown above 1kHz due to the presence of early reflections of the direct sound from the speakers in Gordon's listening room rendering the data below that frequency invalid.) Not as flat as the in-room response in fig.1, an overall tilted-down balance is broken by some peakiness in the mid treble and above 19kHz. The latter is irrelevant, though the former might well be due to breakup modes in the midrange dome.—John Atkinson

Footnote 1: For those readers who feel a love of live motor-racing to be a dangerous hobby for an audio reviewer, I always wear effective ear protection. What alarms me about motor sport, however, where sound pressure levels near the track can easily reach a sustained and dangerous 120dB, is that I see so many spectators not wearing any kind of hearing protection. Couple things like that with the general predilection for listening to Walkman-type headphones too loud, and I fear that an entire generation is suffering premature hearing damage.—John Atkinson
Share | |
Site Map / Direct Links