Pearls, AKA The Fundamentals of Optimal Loudspeaker Design
The purpose of this section is to write down in one place, for the edification of the audiophile in search of truth, wisdom, and the ultimate in music fidelity, all the major “pearls of wisdom” I have learned over the last three decades of deep thought and careful study of the art and science of high-fidelity loudspeaker design. This is not a complete dissertation on the subject, but rather a brief summary which covers only the most important fundamentals of optimal loudspeaker design. If one wishes to assemble a truly satisfying music listening system, one would do well to choose a loudspeaker design which meets all of the following basic criteria.
1. The Inert Cabinet
There seems to be widespread acceptance of, and even overt desire for, loudspeaker cabinets which “sing along with the music.” This is simply anathema when attempting to achieve the highest fidelity in musical reproduction. Consider that the sound-radiating efficiency of any surface is a squared function of its surface area (or the fourth power of the diameter). Furthermore, consider that in a typical cabinet, the cabinet’s surface area can easily be a thousand times greater than the midrange cone’s surface area, making the cabinet literally a million times more efficient at radiating energy into the room. If the cabinet walls move at all in response to the midrange’s vibrations, the boxes can easily put vastly more energy into the room than the drivers themselves! And since this is a particularly pernicious form of coloration, robbing the transient attack of much of its energy, and then lingering long after the actual musical signal has ended, any loudspeaker with a less-than-herculean effort devoted to cabinet rigidity should simply be dismissed outright.
Some may remember the first production loudspeaker ever made by AudioMachina, the Ultimate Monitor. It combined a fully sculpted minimal-area front baffle machined from solid 2” thick aluminum alloy, with the world’s most advanced one-piece ellipsoidal carbon-fiber rear enclosure. No expense was spared in its construction, including the finest 100% high-modulus carbon fiber laminates, high-grade epoxy and proprietary fillers, true integral constrained-layer damping, true integral cross-bracing, etc. It probably was (and quite likely always will be) the most advanced application of high-performance composites ever produced in the field of loudspeaker design. Yet even this was not as good as the machined-from-solid SSA (Solid Slab Aluminum) cabinet construction first pioneered by AudioMachina in the stunning and revolutionary PURE System, a cabinet construction so impossibly advanced that it remains exclusive to the AudioMachina product line to this day. No other company has dared attempt it, which is why no other loudspeaker brand has achieved the vanishingly low cabinet colorations inherent in every AudioMachina loudspeaker, from the mighty-mite CRM to the just-plain-mighty Maestro S.
Furthermore, do not make the mistake of assuming that just because a cabinet is made from aluminum, that it must be just as good as an AudioMachina SSA cabinet. In fact, all the other approaches seen so far are hopelessly compromised in comparison to SSA. Consider for a moment: When the cabinet walls flex, where does the maximum flexure occur? The answer is: At the edges where two walls meet, because the angle of flexure doubles along this edge. Therefore, any attempt to bolt together flat sheets of aluminum is hopelessly lacking in critical stiffness, right at the exact location where the stiffness matters the most! On the other hand, AudioMachina’s machined-from-solid SSA cabinet achieves the optimum: The cabinet wall is not only continuous around the corners and edges, but the wall thickness is actually highest in these areas, thus achieving a level of stiffness in the overall cabinet which is orders of magnitude better than even the best bolted-sheet designs on the face of the earth.
Lastly, let’s return to the subject of cabinet surface area and radiating efficiency once again. Not only is the radiating efficiency a squared function of the surface area, but the wall stiffness is an inverse-squared function of its span, and a squared function of its thickness as well. Furthermore, fully supporting the wall edges (like in the AudioMachina SSA cabinet) gives an automatic four-fold increase in the wall stiffness over an unsupported-edge (bolted-sheet) design.
Now consider that the typical wall span of an AudioMachina cabinet is around one-third that of other competing designs, while the wall thickness is roughly twice as high. Multiply these numbers out, and you come up with the stunning realization that the AudioMachina SSA cabinet is several orders of magnitude stiffer than even the very best aluminum-sheet cabinets on earth! Add in the fact that AudioMachina speakers have vastly smaller cabinet surface areas to begin with, and suddenly the advantage owned by AudioMachina, in terms of real-world colorations due to cabinet radiation energy, rises into the thousands (or higher) compared to the very best competing designs in the world.
It is not a subtle difference, either in the numbers or in musical fidelity.
2. Time Alignment
This one is simple. In a multiway speaker, particularly in the transition between midrange and tweeter, time alignment is an absolute necessity. Consider for a moment that the human ear can easily resolve path length differences, from a sound source, on the order of 0.5″ or less. Now consider that a typical flat-baffle loudspeaker will have the tweeter’s acoustic source roughly 1″ to 2″ ahead of the midrange’s, sometimes more. This means that every transient from the tweeter will reach the ears well in advance of the midrange, lending a “harshness” to the sound which is extremely grating. This type of distortion simply does not occur anywhere in nature, so the ear finds it extremely unnatural. This time misalignment, along with the ubiquitous high-order crossover, is one of the most important contibutors to listening fatigue.
It is relatively easy to establish, by simple visual inspection, whether or not a speaker is truly time-aligned. Simply look at the junctions between the voice coils and cones (or domes). If the tweeter’s junction is not in the same plane as, or slightly set back from, the midrange’s junction, the speaker cannot possibly be time-aligned. Any speaker which is truly time-aligned will have either a radically sloped front baffle, or will have some tweeter mounting geometry which clearly places the tweeter far behind the baffle. Mildly sloping baffles or mild setbacks simply will not do the job, and are only there to fool the gullible.
3. The Crossover
There seems to be a deliberate effort, on the part of certain high-end loudspeaker companies, to confuse the terms “phase response” and “time response,” and to establish in audiophiles’ minds, consciously or subconsciously, the impression that they are one and the same. This effort is aided and abetted by the fact that magazine-based measurements in published reviews also often confuse, or even equate, the two terms. But nothing could be further from the truth, or more critical to creating loudspeakers which sound “natural” when reproducing music.
Consider the ubiquitous fourth-order crossover, widely advertised as having near-perfect “phase alignment” between the various drivers. Now, when playing steady-state sine-wave test tones, this is demonstrably true, both in terms of the physics of analog filter theory, and also in terms of real-world test results. But here is the rub: This “perfect world” is utterly and hopelessly destroyed the instant one tries to pass a transient signal through the system, i.e., the instant one tries to play real music!
Some basic analog filter physics must be covered here. Let’s consider only the crossover between the midrange and tweeter for the moment, although everything said here applies equally to woofer/mid transitions, etc. Now, in a fourth-order crossover, the reason that the drivers achieve perfect phase alignment (and remember this is only when playing steady-state sine-wave test tones) is that the tweeter’s acoustic output is always-and-forever leading the electrical input by 180 degrees of phase, and the midrange’s acoustic output is always-and-forever lagging the electrical input by 180 degrees of phase. The result is that the outputs from the two different drivers are constantly exactly 360 degrees out of phase from one another (the tweeter output leading the midrange output by exactly one full cycle), and thus (in steady-state sine-wave conditions only) they are in perfect phase alignment.
(Technical Note: This discussion is of course limited to the +/- 180 degree phase shifts which occur precisely at the crossover point. Astute readers will note that at every other frequency, the steady-state relative phase difference between drivers remains precisely 360 degrees, even though the individual driver phase shifts are no longer precisely 180 degrees each. This of course has no effect on the inherent and unfixable flaws in high-order crossovers being discussed here.)
Now, this is all fine and good, as long as one never tries to start or stop the sine-wave input signal. But let’s examine what happens when one plays a signal with any transient nature at all. It is not even necessary to play something incredibly complex and transient-dominated (like, say, real music!), only to play a simple sine-wave burst. Say we start with no input, then play 10 cycles of pure sine-wave at the crossover frequency, then stop the signal, and we look at the individual driver output responses on an oscilloscope. So, what happens?
Well, the tweeter cannot possibly start moving before it receives an electrical input signal, but it immediately tries its very best to achieve its 180-degree phase lead, starting the instant the signal arrives. Thus, for the first few cycles of the sine-wave burst, it rapidly advances its output phase (and in the process, actually goes temporarily to a higher pitch!) until its output phase finally reaches the steady-state 180-degrees-ahead-of-input-phase condition. By the time the 10 cycles are over, it is practically there, with its output phase now leading its input phase by 180 degrees.
At precisely the same time, the midrange is doing its very best to achieve its steady-state 180-degree phase lag relative to the input signal. So, while the midrange must perforce begin moving the instant the input signal is applied, it doesn’t really want to move just yet; it really wants to be 180 degrees behind the input. So, over the first few cycles of the sine-wave burst, it deliberately slows down (and in the process, actually goes temporarily to a lower pitch!) in an effort to achieve its steady-state 180-degree phase lag relative to the input. Thus, by the end of the 10 cycles of sine-wave, its output phase is nearly perfectly 180 degrees behind the input signal phase.
And lo and behold, now the tweeter and midrange are nearly perfectly 360 degrees out of phase with one another, and the outputs sum perfectly in phase and amplitude, and everything is right with the world. Or is it?
Finally, we are on the threshold of exposing the “dirty little secret” of high-order crossovers, whose proponents so desperately want you never to discover. To expose the flaw, we need only think about what happens to the summed output of the tweeter and midrange, during those moments in the transient when each is still only partway to achieving their respective +/- 180-degree phase shifts.
Just choose something simple, like halfway there, and the tweeter is now leading the input phase by 90 degrees, and the midrange is simultaneously lagging the input phase by 90 degrees. Sum their acoustic outputs at this precise moment in time, and what do you get? Simple: You get the two output components exactly 180 degrees out of phase from one another, at the same amplitude, which is (are you ready for this?) a perfect null cancellation!
So, any time the high-order system sees a transient, not only are the two drivers both producing pitch errors, but their summed amplitude is completely wrong as well! And at every other point in time, from the beginning of the transient to when the two outputs finally reach 360 degrees phase shift, the summed output is similarly mangled compared to the input, that is to say, the output vectors cannot possibly sum to the original input.
So, what happened to the integrity and fidelity of the original transient input signal? It’s gone, destroyed irretrievably by the mere existence of a high-order crossover in the signal path. In other words, any time one tries to pass anything resembling a transient (like, say, real music!) through such a system, the combined output of the drivers bears at best only a passing resemblance to the original input signal! And this occurs even though the steady-state frequency response and phase response is demonstrably perfect!
Now, although I won’t go into the physics here, it can be shown conclusively that every other form of analog high-order crossover is similarly hopelessly flawed under transient conditions. On the other hand, there is one, and only one, type of analog filter in which the output of the two drivers is summed perfectly, in both steady-state and transient conditions: The first-order filter. Care to guess what is used exclusively in all AudioMachina designs?
The real problem with high-order filters is that the ear/brain system, being the highly sophisticated processor that it is, instantly recognizes the flaws of high-order filters on a subconscious level while listening to music. It doesn’t know precisely what is wrong with what it’s hearing, as this type of transient distortion has no analogue in all of evolution or natural history; it only knows that it simply doesn’t sound “natural,” and consequently it is impossible to relax and simply enjoy the music. This is one of the most important contributors to “listening fatigue.”
Inevitably, when presented with these facts, the proponents of high-order filters will trot out all kinds of supposed “advantages” of high-order filters, including higher power handling, reduced driver spectral overlap, more uniform polar response, blah blah blah. But it begs the question: If it doesn’t (and can’t) sound natural when playing real music to begin with, isn’t everything else kind of meaningless?
4. On Ported Boxes
The ubiquitous ported (bass reflex) box has long been employed for additional power handling and bass extension over a sealed box of equal size. Unfortunately, precisely as in the “Golden Rule of Economics,” there is no such thing as a free lunch. For in addition, you get not only the infamous “one-note boom,” but also lousy transient response, lousy power handling below resonance, and last but not least, a deep-bass response that falls off a cliff. All of these problems are simply inherent in the physics of ported boxes, and very little can be done to fix them, except to sell the speakers and buy something better.
Although perhaps there was once a time when the “balance of virtues” could justify the use of ported boxes, if only barely and in spite of their severe audible problems, that time went the way of the dinosaur with the introduction of the AudioMachina MAPS system. Suddenly, with MAPS (Monitor and Powered Subwoofer) technology, it is possible to have truly world-class bass extension, power, and control, along with superb transient response and musical fidelity, and to top it off, to have all of this in an impossibly small cabinet volume.
With AudioMachina’s fully optimized active analog filter system and matching high-powered ICE amplifier technology, all packaged neatly within the solid aluminum cabinet, the state of the art in bass performance has now reached a level that no ported box (and, for that matter, no passive sealed box) will ever dream of touching. Which, of course, raises the obvious question: Are you still stuck in the age of the dinosaur?
5. The MAPS (Monitor and Powered Subwoofer) System
Only a few short years ago, the MAPS system would not have been a part of the “Pearls” discussion, as I would not have imagined that it could be so important as to warrant a place here. When AudioMachina first conceived and developed the MAPS concept for use in the revolutionary PURE System, I felt that it would offer significantly higher bass performance than traditional speaker designs could achieve. And of course, that proved to be true: It did indeed give far deeper, more powerful, and higher-fidelity bass than had ever been achieved in the entire history of small loudspeaker cabinets, in what can only be described in hindsight as a quantum leap. But it turned out that its real breakthrough was even more momentous, and in an entirely unexpected area.
In time, I began to understand that the real-world advantage of MAPS is vastly more important than I had thought at first, and it is not exclusive to its obvious technical and musical advantages in the bass. After several years of living with the PURE System and its progeny, it became apparent that the real advantage of MAPS extended far beyond the bass, and opened up a whole new world of performance.
The problem with traditional passive loudspeakers, particularly those which aspire to true full-range performance (that is, true bass extension to at least 20Hz), is that it is almost impossible to find an amplifier, any amplifier, which can properly drive such a speaker and simultaneously achieve true high fidelity. There are plenty of amps which have the brute force necessary to drive massive woofers, and there are also plenty of amps with state-of-the-art transparency, delicacy, and musicality, but try finding one which combines these two requirements. Not only are they exceedingly rare, but the few that can be found are, without exception, monstrously, hideously expensive.
However, with the MAPS system, the main system amplifier is suddenly relieved of the duty of driving the massive woofers, which opens up a world of possibilities in amplifier choice, which in turn allows the use of a very wide selection of truly world-class moderately-powered amplifiers. Most tube amps, for example, will struggle badly when trying to properly control massive woofers, but will suddenly sound utterly spectacular and effortless when driving an AudioMachina MAPS system. The difference in the musical performance capability in a given amplifier, between driving a large full-range passive speaker and an AudioMachina MAPS system, is quite simply mind-boggling.
Thus, in the real world, the superiority of MAPS in actual system performance is vastly higher than would be expected from talking about “speakers” or “amplifiers” in isolation, and vastly higher than would be expected by focusing merely on bass performance. And this superiority holds true regardless of the cost or performance level of the speaker, from AudioMachina’s least expensive MAPS system to the most advanced. By making the job of the main system amplifier so much easier, a whole new level of performance suddenly becomes possible, even at relatively modest price points. Furthermore, as one rises into the upper echelons of amplifier and loudspeaker performance, the advantage of MAPS in overall system performance over passive speakers still remains. All else being equal, the AudioMachina MAPS system simply achieves vastly higher total-system performance than is possible with a “comparable” passive loudspeaker.
This, in turn, dramatically shifts the relative system budget priorities to a different optimum balance, justifying placing a far larger percentage of the total system investment into an AudioMachina loudspeaker. It is not a stretch to recommend placing fully two-thirds of the total system investment into a pair of AudioMachina loudspeakers, knowing that with a simple high-quality source and a modest integrated amplifier, it is possible to achieve total system performance of a level normally attainable only with vastly more expensive electronics.
And then, of course, one can rest comfortably in the knowledge that not only does the present system give spectacularly musical performance, but that further upgrades to the front end of the system will be rewarded handsomely as well, simply by making a wise upfront investment in the world’s finest loudspeakers, AudioMachina.
Dr. Karl Schuemann