problems have been minimized, but of course they tend to be more expensive (many European drivers fall into this category). Conversely, some driver design engineers, perhaps through the best intentions that ignorance so easily provides, are happy if their driver design measures reasonably flat by the conventional frequency response test, so they sincerely think their job is done, and they release their driver with the accompanying good-looking frequency response plot to an unsuspecting market (many Far East drivers fall into this category). It's worth noting that the recently reviewed Verity Tamino employs only two drivers instead of the M6's four drivers, and the Tamino is a far more expensive system, yet part of what you get for this extra money are drivers that we lauded as being audibly superb in having very low driver coloration problems.
      Which brings us to the drivers in the M6. The M6 woofer driver seems to have a spurious misbehavior problem in its plastic dust cap, which rings with a hard plastic coloration at about 1500 Hz. The similarly built M6 midrange driver seems to have the same spurious misbehavior problem in its plastic dust cap, except that its hard plastic ringing is at about 2300 Hz (due to its smaller diameter). These spuriously ringing dust caps would also impose a tonal balance emphasis, at 1500 Hz and 2300 Hz respectively.
      Fortuitously, the electrical crossover cuts the signal fed to the woofer off way down at 350 Hz, so, by the time the 1500 Hz ringing misbehavior frequency of the woofer dust cap is reached, there is virtually no electrical energy being fed into the woofer that would trigger this misbehavior. The fortunate result is that you don't have to worry about the misbehavior of the M6 woofer's dust cap. You won't hear it because it is never triggered by the input signal.

M6 Midrange Driver Problems

      The story for the M6 midrange driver does not have a similar happy ending. Quite the contrary. At the 2300 Hz frequency where the midrange driver's dust cap evinces its hard plastic ringing misbehavior and tonal balance hump, the M6 crossover is still feeding nearly full energy to the midrange driver, so this misbehavior is severely triggered by the music signal. To make matters worse, music has a lot of energy at 2300 Hz, and for most of the time, so this misbehavior is triggered with high energy by most types of music signals and for most of the time. To make matters even worse, human hearing has its maximum sensitivity, and its maximum vulnerability to being irritated and fatigued by excess energy and foreign noise, around 2300 Hz (our caveman ancestors doubtless developed this sensitivity peak, since the danger alert of a snapping wooden twig has its peak energy right around this same 2300 Hz). To make matters yet worse, if you follow the NHT setup instructions you will be sitting in the maximum lobe of the M6's radiating pattern, and this lobe is in turn most pronounced at the M6's crossover frequency, which very unfortuitously happens to also be at this same 2300 Hz. This means that you will directly hear this spurious misbehavior more than the rest of your room will, and thus all the sonic problems from this spurious misbehavior will stick out even more prominently, since you will hear them against a background of overall reverberant music energy from the room that has much less energy in precisely this same 2300 Hz problem area.
      Note that we are not saying that the M6 has a frequency response peak in the 2300 Hz region when heard from this maximum lobe. In fact, the M6 has been computer designed so that the frequency response of its directly radiated sound (directly from M6 to the microphone or you) measures reasonably flat when sampled in this maximum lobe. It's just that the M6's frequency response is very un-flat outside of this maximum lobe, for the rest of its radiation pattern as it energizes the rest of the listening room. Specifically, there's a deep notch in the general 2300 Hz crossover region for total energy radiated to the rest of the room.
      Now, human hearing judges the tonal balance of the music, heard from a loudspeaker, by the total energy received at various frequencies from all parts of the room, which is primarily reverberant energy from the loudspeaker's entire radiation pattern (with only a small fraction being energy, intercepted by the listener's ear, that came directly from the loudspeaker). Thus, the tonal balance you hear from the music coming from the M6 will have a deep valley around 2300 Hz, even if you sit in the maximum lobe.
      Furthermore, this perceived tonal balance becomes the ear/brain's reference baseline, against which other noises are judged. Since the music as you hear it has less energy in the 2300 Hz region, a foreign noise (like that plastic dust cap breaking up at 2300 Hz) coming directly at you from the loudspeaker location will stick out more prominently above the music. And ultimately that's the only thing that counts (with all foreign noises, distortions, etc.) - do they stick out audibly above the music, or are they buried inaudibly below the music?
      The engineering decision, to deliberately design the M6 to have substantially flat on-axis response at the maximum of the 2300 Hz lobe, might seem academically to be the correct design choice, since it yields a prettier looking picture of flatter on-axis frequency response. But in this case this precise engineering decision actually worsens our ear/brain's perception of all the driver breakup misbehavior problems in the M6.
      You see, if the maximum lobe at 2300 Hz is designed to be merely flat at 2300 Hz, then the entire remainder of the M6's radiation pattern will be depressed around 2300 Hz, so the overall power response into the room will be severely depressed in the 2300 Hz region. The human ear/brain judges tonal balance largely by the overall reverberant power response, so the ear/brain will hear the M6's overall tonal balance as being severely depressed in the 2300 Hz region.
      On the other hand, the direct first arrival sound from the M6, along this maximum lobe axis, will have far more energy in the 2300 Hz region (substantially flat) than the ear/brain hears from the M6's reverberant sound field in the 2300 Hz region (severely recessed). Thus, the ear/brain will pay extra attention to this direct first arrival sound in the 2300 Hz region. It will pay extra attention because it is starved for information in the 2300 Hz region (the M6 overall power response being so depressed here), so it hungrily latches on to and pays extra attention to the direct sound information in the 2300 Hz region. The ear/brain will also pay extra attention because the direct sound is a disproportionately large percentage of the overall energy heard around 2300 Hz, so it sticks out like a sore thumb as being sonically inconsistent, for example sounding too dry and direct (for other spectral regions the direct sound is comfortably and consistently perceived as being only a small percentage of the overall energy, with the majority being comfortable reverberant sound). The sound and the music abruptly take on a different character, only around 2300 Hz, becoming primarily direct and incisive first arrival sound, rather than the indirect mixed porridge of reverberant sound that characterizes the rest of the spectrum (and is the comfortably normal way we hear most music, both live and from other loudspeakers). The ear/brain hears these strange, incongruous sonic anomalies around 2300 Hz, and so our attention is unconsciously driven to pay extra attention to these unnatural sounding phenomena sticking out from the music.
      Furthermore, the sound and the music also abruptly take on a different character around 2300 Hz because most of the energy the listener hears from the M6 in this region strangely arrives from only one direction, whereas the rest of the spectrum from the M6 arrives from all directions (virtually the entire spectrum normally arrives from all directions, with live music in a concert hall, and with other loudspeakers in listening rooms). Our inherited caveman instincts force our ear/brain to immediately pay extra attention to this strange sound (the sound from the 2300 Hz region) anomalously coming from just one direction, amidst a sea of surrounding sound coming from all directions for the rest of the spectrum (just as a caveman in the woods, surrounded by forest murmurs, would instantly focus on the sound of a snapping twig that came from just one direction).
      Now, with the ear/brain focusing extra attention on the M6's direct sound in the 2300 Hz region, and focusing extra attention on the 2300 Hz sound strangely coming from just one direction, what does it find there? It finds that a large portion of this direct sound in the 2300 Hz region consists of ugly, foreign sounding plasticky breakup misbehavior, which occurs exactly in this same spectral region. So the ear/brain winds up paying extra attention to all the sonic problems caused by the driver's breakup misbehavior, precisely (via a causal chain) because NHT intended you to listen along the maximum lobe axis and designed the M6 to have substantially flat response along this direct axis (hence depressed power response in this 2300 Hz region). Recall that a conventional frequency response measurement does not show these problems, since it can't tell the difference between good energy and breakup misbehavior garbage, so it blindly counts the foreign garbage put out by driver breakup misbehavior at 2300 Hz as part of the energy sum that it blindly reports as flat response at 2300 Hz.
      Note also that, if the ear/brain were paying less attention to direct sound at 2300 Hz and more attention to reverberant sound at 2300 Hz, then the audibility of these foreign plasticky sonic problems from driver misbehavior could be beneficially reduced and ameliorated. That's because these spurious foreign material sounds from the driver lose much of their perceived artificial quality after they've bounced off several room surfaces made of other materials, and after they've become a hidden, mixed-in part of the soup that characterizes a room's overall reverberant sound field. On the other hand, with the ear/brain paying extra attention to the direct first arrival sound around 2300 Hz from the M6, the foreign plasticky material nature of the breakup misbehavior is highlighted, as are all the further consequent sonic problems from this misbehavior. For example, the inconsistent predominance of direct first arrival energy in the 2300 Hz region summons the ear/brain's localization skills into play (skills which are especially sensitive in the twig-snapping 2300 Hz region), so the ear/brain instantly and unconsciously locks in on the M6's location as the source for this inconsistent (with the reverberant field) surfeit of 2300 Hz energy, and that's enough to destroy the spatial imaging portrayal of a sound field extending beyond the M6 locations.
      Thus, thanks to an unlucky combination of design circumstances, the spurious misbehavior of the M6's midrange driver is emphasized in several ways, especially if you obey the setup instructions, orienting the M6 with the tweeter side inward, and thereby listen within the full strength of that lobe at 2300 Hz.
      When we set up the M6 as instructed and listened along this axis, we heard in full force all the many sonic problems that arise from driver misbehavior. The misbehavior imposed an artificial, foreign material coloration on the music, which sounded like hard plastic quack or hard plastic crinkling, depending on the music (the spuriously misbehaving dust cap seems to have a hard plastic component).
      The M6 tonal balance had an artificial tonal emphasis, starting around 2300 Hz and going higher, in part because the dust cap's breakup misbehavior surely had overtones above 2300 Hz, and in part due to tweeter misbehavior [q.v. below]. This audible tonal emphasis arose not so much because the conventional sine wave frequency response measurement on this axis (an irrelevant measurement here, being time averaged and spatially averaged) had a measurable hump in this region, but rather because the temporal lingering of this plastic quack misbehavior is heard as being louder by the ear/brain, since it lingers after the triggering music transient had gone silent, and also since its foreign material noise quality (a hard plastic quack) attracts extra attention from the ear/brain.
      The added spurious misbehavior also smeared the music, due to the temporal lingering of the dust cap's spurious misbehavior. This lingering smearing also obscured subsequent genuine musical information, making the M6 less transparent. The spurious misbehavior sometimes seemed to add a kind of modulation distortion, since it was being modulated by the music signal.
      And the spatial imaging of the M6 was degraded to the point of being mediocre, with monstrous hot spots of artificial sounding and too bright energy coming directly from the loudspeaker locations. These localizable hot spots overshadowed any attempt by the M6 to portray a wide seamless curtain of sound extending between and beyond the loudspeaker locations. And of course the ear/brain's attention is further drawn to these hot spots, and distracted away from the rest of the projected spatial image, by the fact that these hot spots sounded so artificial.
      Obviously, the designer of the cone drivers selected by NHT for the M6 did not pay enough attention to controlling spurious misbehavior of all the driver's radiating parts. And, by unfortuitous coincidence, the design of the M6 further emphasizes in several ways the misbehavior of the midrange driver.

M6 Tweeter Problems

      We also heard sonic problems from the aluminum dome 1 inch tweeter selected by NHT for the M6. It sounded artificially hard and bright, and it imposed a pronounced aluminum coloration on all the music it reproduced. You know something is wrong when strings made of nylon or gut sound as though they were made of aluminum. Any time you hear a coloration from a driver, and that coloration bears the sonic signature of a physical material employed in the driver, it's a dead giveaway that the driver is not acting pistonically and is not under accurate full control of the music signal, but is instead spuriously misbehaving, vibrating on its own, and thereby emitting sounds with the characteristic sonic thumbprint of the material that is vibrating to its own tune instead of obeying the music's tune.
      The M6's on-axis measurements show that response irregularities occur at 4000 Hz and above, which suggest non-pistonic breakup misbehavior in the tweeter. They also show that the tweeter's response rises above 15 kHz, as it climbs toward a bright peak at the top of the audio range that indicates major non-pistonic breakup of the aluminum dome, which would produce excess tonal brightness plus foreign material colorations from the breakup misbehavior. We would guess that complex breakup misbehavior in the aluminum dome also might be creating difference byproducts which do fall well within the audio band, and which therefore are highly audible.
      These spurious, aluminum sounding colorations in the M6 tweeter also had the predictable further adverse sonic consequences. The M6 sounded too bright, not just from the bright rise above 15 kHz, but also because at lower treble frequencies (even below 15 kHz) the bright energy of the aluminum dome's misbehavior lingers after each strong musical transient triggering the misbehavior, and so is heard by our ear/brain as lasting longer in time than the genuine original music transient (hence having more average energy over time), and also as being much larger than the temporally succeeding genuine musical information that has neutral tonal balance, thereby skewing the apparent tonal balance of that temporally succeeding musical information to sound much too bright.
      The M6 trebles also sounded smeared, again because the spurious aluminum misbehavior lingers after each musical transient. Transparency was degraded because the lingering misbehavior obscures the subtle genuine musical information that immediately follows each strong transient that triggered the driver misbehavior.
      Spatial imaging was degraded because our ear/brain was distracted away from the genuine imaging information and instead attracted to the hot spot of extra brightness and foreign aluminum coloration of the misbehavior, which obviously was emanating only from the loudspeaker location, not from the whole stage portrayed in the image.
      Why does the M6 sound too bright in its maximum lobe? In small part, it's because the tweeter does measure too bright on or near its axis, with a generally rising trend, and then a decisive rising slope above 15 kHz. But, in larger part, the excess sonic brightness probably stems from other causes, causes which are similar to those that made the midrange driver dust cap's misbehavior more audible. Even in those spectral regions where the tweeter measures substantially flat on axis, it still sounds too bright, due to these other causes.
      First, the tweeter's spurious misbehavior and aluminum coloration is audible throughout most of the upper midrange and all treble regions, as correlates with the somewhat ragged nature of its measured on-axis response. This spurious misbehavior, being by definition out of control of the music signal, lingers after each musical transient triggering this misbehavior has already fallen silent. Thus, this misbehavior in the upper frequencies lasts longer than the triggering musical transient, and thereby has a greater power averaged over time, and furthermore is even more saliently audible because it is still playing loudly against a background of now quiet music, after the strong musical transient has fallen silent.
      Second, the foreign material nature of the tweeter's upper frequency misbehavior attracts the ear/brain's attention to it and away from the music, so this brightly biased energy is perceived as louder, relative to the spectrally balanced music, than it really measures.
      Third, the fact that this misbehavior originates in, and is clearly localizable by the ear/brain as being within, the loudspeaker location, attracts the ear/brain's attention to it and away from the music that is portrayed as a large seamless spatial image that is not within the room.
      Imagine if someone were to suddenly click some live castanets within the same room where you were listening to the large spatial image of the reproduced music. Your attention would instantly be distracted away from the spatially imaged music in the portrayed concert hall space, and toward this different noise that was clearly localizable within your room (your caveman threat response instincts would be automatically reacting).
      Thus, again the ear/brain is distracted away from the music, and focuses extra attention on this other sound that is localizable within the room, magnifying the perception of this localizable sound above the perception of the music. Since the music itself might be spectrally balanced correctly, but the intruding localizable noise is not (since it is from tweeter misbehavior and the tweeter covers only the upper frequencies), the end result is that the ear/brain winds up paying more attention to this bright localizable aspect that is mixed in with the music, so the music winds up

(Continued on page 109)