midrange and tweeter drivers, originating in the drivers themselves and thereby attracting the listener's attention to the loudspeaker locations, were now no longer audible. Thus, with the distracting hot spots gone, our ear/brain could now pay attention to the M6's portrayal of just the spatial imaging information from the recording itself. There was now an even curtain of sound in the portrayed image between loudspeaker locations, with no hot spots at the loudspeaker locations.
      Flipping the M6 upside down also dramatically expanded the width of the portrayed spatial image, so that it now extended far beyond the loudspeaker locations. This improvement was largely caused by the fact that the main energy lobe of the M6's radiation pattern was now directed at the room's side walls, especially in the 2300 Hz region, with much less energy directed at our listening position. This naturally changed the balance of direct to reverberant sound we heard, which in itself can be importantly beneficial to spatial imaging (especially to our perception of hall ambience, as encoded in a recording). But it also beneficially provided a needed increase in the amount of reverberant energy coming at us from our sides.
      The enormous width and spaciousness of the spatial image you hear live in a concert hall is largely due to the fact that reverberant energy bounces off the side walls of the hall and then comes at you from the left and right sides of your head. To test this, the next time you attend a live concert simply hold your hands, flat and palm side inward, about 6 inches away from each ear, to block the incoming reverberant energy coming at you from the sides, but still allow in the direct energy coming at you from the orchestra and the reverberant energy coming at you from the front and rear of the hall. You'll hear that the width of the spatial image completely collapses (and also that hall ambience dramatically decreases, so that the giant hall sounds like a small room). This proves how important it is to have significant reverberant energy coming at you from your left and right sides. And flipping the M6 upside down accomplishes exactly this.
      Flipping the M6 upside down wreaks such a dramatic change, and such dramatic sonic improvements in so many ways, that we strongly recommend you do likewise. As we said before, you get very different tonal balances (bright vs. mellow) when you make the choice to listen to the M6 tweeter side inboard as the manufacturer instructs, or upside down with the tweeter side outboard as we tried. And, if it were only a matter of altered tonal balance, we'd happily recommend that you pick the tonal balance personality you prefer. But, as we've seen, there is far more at stake in this setup choice than merely tonal balance. Your choice here is really between a lemon and lemonade.

Fine Tuning M6 Orientation

      The sonic improvements from flipping the M6 upside down were so many and so dramatic, that for a while we simply enjoyed listening to the Evolution system in this configuration, on a variety of music. But, when listening critically, we heard a couple of minor points that could still use a little further improving. First, the tonal balance, though acceptably mellow on the whole and very relaxing, was just a bit too polite and recessed in the midrange to upper midrange. Since this corresponded approximately to the 2300 Hz region where the null in the radiation pattern was deepest, this told us that listening in the deepest part of the null took away a little too much spectral energy in this 2300 Hz region (and at higher nearby frequencies). Secondly, although the spatial imaging was now wonderfully wide and seamless, the center image between loudspeakers had a slightly ghostly or phantom quality, without enough tactile solidity and presence. This too might relate to there being not enough spectral energy in the 2300 Hz region when the M6 was heard in the deepest part of its radiation pattern null.
      The above clues suggested that it would be sonically profitable to try fine tuning the M6's orientation relative to the optimum listening seat. We could try slightly changing the height of the stand supporting the M6, so we would be listening along a slightly different vertical axis. The manufacturer's stands for the M6, the B6 subwoofer or the P6 pedestal, put the M6 at a height where the seated listener's ears are exactly at the height of the midrange and tweeter drivers, so the listener is hearing the M6 on a vertical axis that is exactly in line with the height of these side-by-side drivers. If we raised or lowered the height of the M6, changing our vertical listening axis, we would be slightly changing the difference in path lengths to our ears from the midrange vs. the tweeter drivers, hence slightly changing the time delay, hence slightly changing the frequency of the radiation pattern minimum as we heard it. We might also be changing the depth of the null.
      We could also try slightly rotating the M6 horizontally to the left or right, or try slightly moving our listening chair forward or rearward (on its convenient casters). This would slightly change the horizontal axis along which we heard the M6. And this would reduce the depth of the null, since we would now be situated to the left or right of the deepest part of the null, thereby giving us more spectral energy in that 2300 Hz region where the lobing radiation pattern has its strongest peaks and nulls. This change in horizontal listening axis might also slightly change the frequency of the radiation pattern minimum as we heard it, since we'd be slightly changing the path length difference to our ears from the midrange vs. tweeter drivers.
      We tried fine tuning both vertical and horizontal listening angles, and both turned out to yield wonderful sonic improvements, in precisely the aspects we were hoping for. First we tried lowering the height of the M6, and we found a magic spot when our vertical listening axis was changed so that we were exactly aligned with the height of the center of the top M6 woofer (as opposed to being aligned with the height of the midrange and tweeter drivers). At this magic vertical listening axis, there was more spectral energy in the 3000 Hz region, just enough to make the tonal balance sound flawlessly consistent, and restoring a needed bit of upper midrange sparkle - but not so much as to start making the gremlins of the midrange driver's spurious misbehavior audible again. And there was a further sonic benefit. The center portion of the spatial image between loudspeakers no longer sounded ghostly in quality, but instead became much more solid, tactile, and believably real. So both of the slight remaining sonic problems we had heard were now cured by simply listening to the M6 along the vertical axis of the woofer.
      We listened along the axis of the center of the top woofer, achieved by lowering the height of the M6, which was very convenient (as will be discussed below), and which also afforded a solid, non-rocking mechanical foundation for the M6. You could also try raising the height of the M6, until you are listening along the vertical axis aligned with the center of the bottom woofer. Raising the M6 to this higher height would produce a floor bounce cancellation at a different frequency in the warmth region than lowering the M6 to a lesser height did, so you can try both, listening for differences in the warmth region, to decide which you prefer. We can't give you a simple prediction and recommendation here, since the exact frequency of this floor bounce cancellation effect depends not only on the height of the loudspeaker above the floor but also on your listening distance from the loudspeaker, and the latter of course will vary significantly among you.
      Next we tried experimenting with fine tuning the horizontal listening axis. Our discoveries here put the icing on the cake. The basic idea was to change the horizontal listening axis slightly, to get out of the deepest part of the radiation pattern null, whose greatest effect was in the 2300 Hz crossover region. Now, we could move horizontally, out of the deepest part of the radiation pattern null, either to the left or right of the deepest part. Which should we try first? Which might theoretically provide better sound? And how far out of the null should we try listening?
      We were mentally debating this, when we suddenly realized something serendipitously special. Rather than moving to the left side or the right side of the deepest part of the null for both loudspeakers, we could instead move to the inboard or outboard side of the deepest part of the null for both loudspeakers. And, if we moved slightly to the outboard side for both loudspeakers, something very special should theoretically happen. We could place our nearer ear (the ear nearer a given loudspeaker) on the outboard side of a given loudspeaker's deepest null, and just far enough outside the null so that our farther ear (the ear farther from that same given loudspeaker) was precisely in the deepest part of the null. This should theoretically enhance the M6's spatial imaging even more, by giving us a pure form of sonic holography (purer than the Carver electronic version, since no electronic signal scrambling and degradation would be involved). Each of our ears would hear more of the nearer loudspeaker and less of the farther loudspeaker, thereby theoretically giving us purer, more accurate spatial imaging.
      So we tried this. Would the predictions of theory work in practice? Cowabunga! The center image became even more solid and believable, and the portrayed width now extended even beyond the room walls, making our listening room totally disappear for the first time with the Evolution. Getting the boundaries of your room to aurally disappear is of course crucial to achieving the ultimate goal of a high end stereo or surround sound system, suspending your disbelief and aurally transporting you to the alternative venue of the recording.
      We tested and confirmed this discovery, by simply rolling our listening chair slightly forward and backward. This took our farther ear slightly in and out of the precise deepest part of the M6's radiation pattern null. Sure enough, it was easy to repeatedly lock in on the same precise location, i.e. the same precise horizontal listening angle, to achieve this spatial imaging magic. With this tactic, the M6's spatial imaging rating went up yet another notch, from very good to excellent.
      How can you duplicate this magic with the M6? We measured the horizontal angle to our farther ear, and it was 22 degrees off axis. You can keep the M6s pointing straight ahead into the room, and move your seat forward or backward until you match this angle. An alternative, and probably better, setup plan is as follows.
      First, set up the two front main M6s to meet as well as possible all of the usual common sense guidelines: (1) so they are the usual reasonably large distance from the walls to their side and rear (4 feet or more is best), (2) so there is the usual 45 to 90 degrees (depending on your preferences and imaging goals) of included angle between their locations as perceived from the listening location, and (3) so they are at the usual locations about 1/4 to 1/3 of a room width away from the walls to their sides. Install them at a height so that your ears are vertically in line with the exact center of either the top or bottom woofer. Set up your prime listening seat in the best location relative to your room acoustics. Then position the surround M6s around a clock at the general angles (relative to your prime listening seat) that you favor (some people favor the side/rear surrounds exactly at their sides, other people prefer them diagonally to their rear, etc.). And finally, rotate each M6 so that you are listening off to the midrange driver side, at an angle of 22 degrees to your farther ear (for simplicity, you can use an angle of 20 degrees to the center of your head).
      To measure this angle, put the baseline of a protractor so it lines up with the M6's front panel top edge, and sight the protractor's center pivot point so it is over the midpoint between the midrange and tweeter drivers (which by the way is not at the center of the M6's front face). If you already happen to have about 44 degrees of included angle between any pair of M6s (not counting the center channel one), then of course the amount of further rotation you do will be minimal.
      These precise angles obviously pertain to your single best listening seat, and the sonic results will be different (and presumably not as wonderful) for other listening seats in your room. But that's also true for virtually all other loudspeakers. Furthermore, there can only be one best listening seat, regardless of loudspeaker radiating patterns. That's because you can achieve ideal arrival time correction and simultaneity from plural loudspeaker locations (even for a simple two loudspeaker stereo setup) only for one best listening seat. So, from the outset, we all simply have to accept the fact that the best sound is reserved for just one seat in the house.
      Our unorthodox setup, quite different from the manufacturer's instructions, transforms the M6 into a true high class loudspeaker, worthy of partnering with the best subwoofers (see below). The thusly transformed M6 turns the whole NHT Evolution system from being merely a good value into being an amazing bargain, giving you champagne sound for a beer budget price. The only remaining sonic limitation that keeps the M6 from being a truly great loudspeaker is the slight sonic veiling attributable to its budget crossover parts, which a hobbyist could cure by upgrading the crossover parts (or perhaps NHT could follow the lead of some other manufacturers who offer Signature versions of their loudspeakers with upgraded parts). With our setup, the tonal balance personality of the M6 is mellow rather than aggressively bright, but within this context it does an excellent job. And you can still play with choice number 2 discussed above, to see if you prefer the more direct, involving, exciting upper frequency sound you get by inverting the polarity of your connection to the M6, or whether you prefer the more relaxing sound of the straight polarity connection.
      Similar logic pertains to the best tactic for setting up the M6 you use as a center channel loudspeaker. You could install this center M6 vertically, and simply rotate it 22 degrees to one side (assuming your display or screen is mounted high, or you have an acoustically transparent screen). If you install the M6 horizontally, simply install it low enough or high enough so that your ears' listening height is 20 degrees off (above or below) the midrange driver side (obviously you won't get the holography effect to your farther ear from horizontal M6 installation, but that's OK from the center channel).
      If you opt for the less expensive M5 instead of the M6, all the above comments still generally apply, since the M5 has the same design and employs the same type of drivers as the M6, with the twin woofers and midrange driver merely being slightly smaller than they are on the M6. We did not test an M5, but we can predict that its performance and its ideal setup parameters would be very similar to the M6.
      The small differences should be as follows. The M5 has slightly less extension down into the upper bass, slightly less bass power capability, and slightly less overall loudness capability, due to its smaller drivers. The spurious misbehavior of the midrange driver's dust cap should be at a slightly higher frequency, because the driver is smaller, so presumably the dust cap is also smaller in diameter. The M5 crossover is still at 2300 Hz, so the interaction between the lobing radiation pattern and the midrange driver's spurious misbehavior should be slightly different. The optimum listening angles might be slightly different, both vertically and horizontally. The overall cabinet dimensions are slightly smaller, so the best vertical listening angle, away from the midrange/tweeter height, might be slightly different than at exactly woofer height.
      Some of these small differences might actually offset each other, such that the best listening angles for the M5 might wind up being very close to those for the M6. We suggest that you start with the same alignment for the M5 that we recommend for the M6. Then, using our text as a guide, listen for the sonic changes we described, to zero in on exactly the best alignment for the M5 that gives you the same sonic benefits we described.

Subwoofers

      To complement its two satellite monitors, the Evolution system offers you not just one subwoofer, but a choice of four subwoofer models, as described above. The Evolution W1 subwoofer is the star performer of this group. It is a very good subwoofer on all counts, and at $550 it is such an extraordinary value that you can (and should) use at least two of them. With at least two (widely spaced) subwoofers, you'll experience immersion in the space of the alternative recorded venue much better, since you'll hear the original bass waves wafting with velocity across a giant space, whereas with a single subwoofer at a single room location you totally miss out on this experience, and the only thing you experience instead is the one-dimensional pressurization of your puny little room.
      The Evolution subwoofers are not actively powered subwoofers. Some other active subwoofers we have evaluated seem limited by the performance of their built-in power amplifiers, and of course in those cases you are stuck with that power amplifier. So it is a potential advantage that the Evolution offers you the opportunity to drive the subwoofers with a power amplifier whose power, bass performance, and price match your taste and your budget. We evaluated the Evolution W1 with a variety of power amplifiers, and the W1 clearly revealed sonic differences in the bass performance of various power amplifiers, thereby proving NHT's wisdom in leaving this choice up to you.
      The dual driver Evolution subwoofer models (the W1 and B6) offer only a single terminal input, so there is no facility for connecting the two drivers to separate signals from separate power amplifiers (nor should you want to, since non-identical signals fed to two woofers sharing the same acoustic suspension cabinet would compromise the linearity of the acoustic suspension function, and in the case of the W1 would also compromise the sonic benefits of having the two woofers acting identically but in opposite directions). If, heaven forbid, you should want to combine plural bass signals into a single subwoofer location, the Evolution's X1 electronic crossover offers you the facility for electrically combining up to three channels of bass (two regular channels plus the LFE channel) into one.
      Before we plunge into explanations of why the W1 sounds the way it does, let's briefly summarize the various aspects of its sonic performance. Its bass extension is very good, reaching flat down to 27 Hz. Only the bottom half octave of low bass (20-27 Hz) is weak, and you'll miss that only on certain select sound effects and music recordings (for example, the famous organ pedal notes on Reference Recordings' track called The Vikings). The W1's bass impact is likewise very good, giving you a solid kick in the stomach (this is obviously somewhat dependent on the power amplifier you choose). The W1 is very good in bass definition, with just a slight degree of slowness and lingering overhang keeping it from earning an excellent rating (see below for explanation). The

(Continued on page 111)