Natural and Unnatural Dips.

I find the predisposition of some individuals to smooth out power-response dips and peaks in the midrange by means of an equalizer to be both interesting and problematic. Although a few enthusiasts would not be without a proper equalizer, many high-end types would not touch one with a 10-foot pole, no matter how upscale it might be. Still, the people at Lucasfilm have THX-certified a few high-quality equalizers for home-audio use, some pretty competent audio companies make them, and a lot of recording engineers work with them freely. Consequently, equalizers certainly have to be taken seriously.

I find that while equalizers are mostly used to flatten out response curves for the smoothest measured response in a listening room, under some conditions they might serve the cause of high fidelity a bit better by making a speaker's response curve less than perfectly flat. Most equalizer use involves compensating for room-induced anomalies; however, those who work carefully with equalizers (either with the help of a measuring device or simply by ear) may also be attempting to compensate for at least one genuine speaker-related artifact. This is because a number of speaker models on the market, particularly two-way designs with typical dome tweeters and 6" to 8" woofer/midrange drivers, may automatically display power-response dips at or near their crossover frequencies. This is usually in the 1.5 kHz to 3 kHz range, and the overall impact of this irregularity will also depend on the way the crossover slopes are configured and integrated. (Note that even some vertical MTM d'Appolito designs may exhibit this effect, and a fair number of three-way models may also exhibit it, because they have largish midrange drivers.)

The main cause for this behavior is that the woofer/mids tend to lose radiation efficiency as the frequency they handle climbs toward the crossover point, causing a gradual power-response slump. When the tweeter then begins to cut in, the power response flares back upward in strength, delivering a "flat/dip/flair" effect. The tweeter contributes to the flair part of the phenomenon, because the radiation efficiency of a typical tweeter is usually substantial near the bottom of its operating range, just above the crossover point. Tweeters only begin to beam and develop a power-response sag at higher frequencies. Of course, how this works will also depend on the tweeter's crossover slope and how the designer overlays it with the low-pass cut off to the woofer (in a two-way) or midrange (in a three-way). However, equalizer diddling to correct for this phenomenon may do more harm than good, and I am not talking about mysterious, non-measurable artifacts that some equalizer-haters ascribe to equalizers, but am instead referring to the impact of frequency-response corrections and differences on the direct and reverberant fields at or near the system crossover points.

I have reviewed a number of two-way speakers over the past few years (and have also reviewed a large number of reference-standard movie and music transcriptions), and the speakers that had flat-measured room response always had a different kind of hall/room "sound" from systems that were nearly as flat over the most of the same bandwidth, but which displayed that crossover-related dip in the 1.5-to-3 kHz range. This "sound" was something different from the kind of clarity and spectral-balance differences that such response irregularities also influence, and involved the sense of frontal depth, perceived size, distance, and space.

Speakers I have reviewed that exhibited such a dip to varying degrees include the Coincident Technology Triumph Signature, Tannoy Mercury M2, and the B&W DM-305. Even the Dunlavy SC-II exhibited it to a small extent, although the way the front of the cabinet of this system was designed kept the dip/flair part of the phenomenon under better control than the 6.5-inch size of the woofer/midrange drivers would have ordinarily dictated. With the SC-II, the absorptive padding around the tweeter as well as its recessed mounting position managed to minimize the flair out, and the crossover was no doubt configured to take advantage of this arrangement. (The downside to this is a loss in maximum output capability by the tweeter because of the absorption.) In addition, even the AR-310HO speaker, a decent three-way model, exhibited the effect to some extent, because the midrange-to-tweeter transition was moderately high in frequency and the dual midrange drivers were moderately large. (Incidentally, the current AR-9 model appears to be not much more than a renamed 310HO, although the tweeter may be slightly different.)

Two-way speakers that have not displayed such a power-response dip with my room-response measurement technique included the NHT VS1.2 and the Atlantic Technology 271LR and 273C, all of which have rather small woofer/ mid drivers and crossover points that do not push the mids too high. In addition, the midrange-driver response of these systems is probably adjusted to deliver flat power at the expense of a somewhat non-flat direct-field response.

The Allison LC-120 speakers that I have been using as a two-way reference for some time also do not display the power-response dip/flair artifact, mainly because the system's 8-inch woofer is adjusted to give a rising response to the on-axis output as it works its way upward toward the crossover point, and the tweeter also generates enough output below the already fairly low crossover point to limit the sag. The low crossover point helps to keep the woofer from trying too hard to behave as a tweeter. All of this tends to offset power-response losses. Both the big Allison IC-20 speakers in my main system, plus the Allison AL-125 models I have used as midpriced reference systems, were also free of the dip/flair artifact, but in this case, it is because of the small size of their midrange drivers.

In reasonably large or larger rooms, with systems having average or better dispersion characteristics over most of the rest of their operating range, midrange power-response dips, among other things, are often perceived as the primary determinant of the way a system integrates into the room's space. With such systems, the result in my experience has nearly always been a sense of distance and ersatz envelopment added to the sound. There is also usually a loss in midrange bite with a lot of forward-sounding, midrange-rich recordings (including movies) that could often benefit from such a loss. Indeed, in many cases, I have found that older (and even some rather recent) classical-music recordings were substantially improved by such midrange-dip artifacts, although really high-quality pop music and movie soundtracks tended to be influenced in a somewhat less-advantageous or even negative manner by anything but ruler-flat response or maybe a slight downward overall tilt above the lower midrange.

Now, if a system does not apply such a cut automatically, as part of the dip/flair phenomenon, an equalizer can be used to deliver the same kind of effect -- almost. I say "almost," because an equalizer will cut back on the direct-field and reverberant-field response together, whereas a system with a natural power-response dip in the midrange may still have a fairly powerful direct-field signal, with the loss showing up in the reverberant field. How the difference between the two effects will influence the sense of space will involve how the chosen listening position relates to what is called the "critical distance."

Now, the "critical distance" is the point where the direct-field signals and reverberant-field signals achieve equilibrium. In typical rooms, it is where you will find the predominantly direct sound (resulting from the strong narrow-angle signals generated by speaker drivers that are large in diameter, compared to the wavelengths they are generating) passing over to the reflected sound (generated by wide-angle signals generated by drivers that are small in relation to the wavelengths they are generating). Room absorption can also play a major part in this (heavy damping will reduce the strength of the reverberant field), and if the room is highly damped, the critical distance at higher frequencies will be moved further away from the speaker system as the direct field's strength is made relatively stronger. At lower frequencies, the absorption will have less effect.

Now, you do not have to be an acoustical engineer to see that the critical distance is not going to be a fixed point at all frequencies. Its location can vary considerably over a system's bandwidth, depending on the radiation pattern of its drivers at specific frequencies and the crossover points and crossover slopes. The number and layout of the drivers on the cabinet baffle or baffles will also have an impact. At low frequencies, it will ordinarily be very close to a speaker. At high frequencies, it usually is somewhat further back; often quite some distance back, as would be the case if the system has a very large tweeter radiating surface, such as what we find with some large flat-panel systems or tall line sources.

In the midrange, however, the critical distance can be both close to the speaker at the bottom of a midrange driver's (or midrange/ woofer's driver's) operating band and rather far away from the system at the top of that driver's range. That far up, it then will abruptly move back close to the speaker again as the frequency rises still further and the typically smaller-diameter tweeter begins to deal with the upper midrange and strengthen the reverberant field. In a three-way system (at least if the woofer/ mid crossover is fairly high) it can also be rather far from the speaker at the top of a woofer's range, and then abruptly move back close to the speaker as the midrange driver cuts in at slightly higher frequencies. The peripatetic nature of the critical distance accounts for much of the sound character of speaker systems.

This, to me, is one reason that speakers will often seem to change in terms of their soundstaging, spaciousness, and dimensionally as one moves the listening position closer to or further from them, or moves the speakers closer to or further from the listening position.

Getting back to dips and equalizers, while this kind of response diddling with such a device can be interesting in larger rooms, it is in smaller rooms that the manipulation of midrange frequency response, either by an equalizer or because of the way the speaker systems themselves are designed in terms of their midrange response, can sometimes do wondrous things.

In smaller rooms, speakers with really flat on-axis response often have a degree of forwardness that works against happy listening. Consequently, using an equalizer to install a modest sag centered at about 2 kHz might work even better in such a room than what we find when we apply such manipulations in a larger room. (Note that a properly sloped sag really cannot be delivered with a typical one-octave equalizer; a good 1/3-octave or parametric model is a much better bet.) As I have noted, I have reviewed a number of speakers that display a power-response sag in this range, while having a flat on-axis response. If the small room has a proper blend of absorbing and reflecting materials in its makeup, this kind of phenomenon, combined with the crispness that results from a flat direct-field response, a good tweeter, and a proper toe-in with both speakers (and a good, sweet-spot listening position) should result in a rather crisp midrange attack and somewhat attenuated midrange reverberation. The upshot is a subjective soundstage that mysteriously appears to move the speakers away from the listener and makes the room actually seem larger and more concert-hall like.

Using an equalizer to pull down part of the midrange power response of a system that does not have an inherent midrange power sag will of course have an impact on both the direct and reverberant fields. Consequently, some of that first-arrival crispness noted above is lost. However, a substantial amount of that crispness is actually delivered by the tweeter as it deals with still higher "attack" frequencies that are not influenced by the equalizer cut, and so the effect can still sometimes be beneficial.

In other words, multiple speakers having a natural or equalizer-generated dip in the range centered at about 2-to-2.5 kHz can sometimes sound subjectively better in smaller rooms than speakers that have flatter overall response, at least with some program sources. This is particularly the case with recordings that might have a tendency to be a bit bright, forward, and brittle sounding in the midrange, but which are still weak enough at higher frequencies to benefit from the flat, extended output of a tweeter above the sag. I believe that this is why many people who have speaker arrays that employ two-way speakers (and subwoofers) set up in rather small rooms are so often able to get a surprisingly large-scale, concert-hall effect. Yes, if you have it, surround-sound processing will deliver an even greater improvement, but the midrange power-response dip of certain speakers will have a great deal of impact, too.

Incidentally, part of the reason that some recordings dovetail so nicely with speakers that we have to admit are not all that frequency-response accurate is that some control-room monitors are similar to the above designs when it comes to their direct-field and reverberant-field balance, and the engineers adjust the mix accordingly. The control rooms themselves may also simulate small-room acoustics, and the speakers are often listened to from fairly close up. With home-based speakers that are actually more accurate in terms of power-response uniformity, direct-field and reverberant-field blend, and critical-distance stability than those control-room monitor combinations, the resulting transcriptions may come across as a bit forward-sounding and not as pleasing or as subjectively realistic as one might like.

Hence, equalizer users may be justified in applying compensation to offset the deficiencies in those recordings, and those who are enamored of speakers with the midrange-dip deficiency built into them may also be somewhat justified in their preference for models that are actually less than perfectly accurate. At least this is the case when listening from close up in small rooms.

Room size notwithstanding, just how much this midrange-sag phenomenon (be it one that mainly effects the reverberant field because of speaker design, or instead effects both the reverberant and direct fields, by means of outboard equalization) will modify our subjective impressions may ultimately depend on the distance between the listener and the speakers. This listening distance will play a part, because where one sits in relation to the critical distance between the reverberant and direct fields appears to have a lot to do with the perceived sense of frontal depth and spatiality -- in other words, the perceived sense of playback realism. The sense of spaciousness and depth of a recorded program, particularly in smaller rooms, is going to be strongly influenced by the direct-field vs. reverberant-field ratio throughout the midrange. -HF