The Masking Threshold

by

Albert Von Schweikert,

Chief Design Engineer, Von Schweikert Research

Details of the VR-6 and VR-8 Technology

It is common knowledge that conventional loudspeaker systems suffer from imperfections related to their mechanical operation which limits "realism." Lately, many designers have attacked the problems of coloration and distortion endemic to both transducers and enclosures in order to extract greater levels of clarity and transparency. It has become quite commonplace to see 200 lb enclosures advertised which utilize "exotic" drivers including Kevlar, aluminum, ceramic, and other composite diaphragms. Indeed, the level of sound quality has risen greatly from only a decade ago as cabinet resonances have been lowered while driver quality has steadily improved.

EFFICIENCY CONVERSION RATIO

However, one important system parameter has largely been neglected in this search for greater clarity, and that is the efficiency conversion ratio. A system that achieves 100% conversion of electrical watts into acoustic watts is still a dream. While very large horn systems can achieve about 40% efficiency, small acoustic suspension designs achieve less than 5%! Could this be a problem?

TACTILE PRESENCE

In the quest to develop compact, non resonant systems, engineers have sacrificed an important aspect of sound reproduction: dynamic range, or the "jump" factor. Although it is quite true that eliminating resonances from both drivers and enclosures results in greater overall clarity, the general loss of efficiency through resonance damping and cabinet downsizing has taken a toll in both dynamic range and low level clarity due to loss of efficiency. Many of us long for that missing "tactile presence."

A HIDDEN PROBLEM

The general consensus in the engineering community is that high efficiency is not really necessary with today’s high powered amplifiers. Is this really true?

Many audiophiles have noticed that comparisons between high and low efficiency speakers can not be undertaken unless the listening levels are evenly matched in sound pressure level. This is due to the fact that the ear/brain hearing mechanism will (wrongly) conclude that the louder of two sounds is the "better", even if the louder system has worse frequency response and more distortion!

Often high efficiency is achieved with the penalty of "peaky" sound quality; the extra efficiency being gained by horn loading or other manipulation of cabinet resonance leading to coloration with an attendant loss of neutrality.

Proponents of high efficiency systems often claim that large horn systems, while somewhat colored, exhibit great "presence", dynamic range, and impact. If the dispersion problems of horns and their coloration could be eliminated, the dynamic range gains would certainly be welcome.

No one who has attempted to achieve realism in his/her playback system would argue that low efficiency speakers on the order of 80-90dB have any advantages in sound quality; quite the opposite.

Low efficiency speakers sound somewhat restrained, veiled, and compressed. This is due to the fact that most high end speaker systems are quite low in conversion efficiency, usually around 5% or less! Quite simply, that lost energy has been "dissipated" due to grossly inefficient conversion of electrical power into mechanical movement. Crossover parts exact their toll on transparency, while low efficiency drivers convert very little of the signal to actual sound wave generation.

THE MASKING THRESHOLD

Several years ago I embarked on a program to increase the clarity, transparency, and "realism" of the conventional dynamic transducer-based speaker system, and made several important discoveries. The most important fact I discovered was that low level detail in the signal was being "absorbed" and dissipated by the crossover components, cabinet-wall flex, speaker diaphragm flex, and "insensitive" low efficiency transducers.

At very low playback volume levels, conventional speakers have a loss of bass response, dynamic range, depth, and image size. Most of this sonic aberration is caused by the dissipation and hence erosion of low level detail by the inefficient mechanical and electrical operation of the speaker system’s component parts.

BEYOND THE FLETCHER-MUNSON CURVE

There is a drive level threshold below which the sound reproduced by the low efficiency speaker is non linear, and to overcome this effect, more power must be dumped into the speaker system. It is thought that by simply increasing the drive level, these non-linear effects can be overcome, however it’s not as simple as that. Since the speaker can not respond to very low differences in volume level, the net result is compression of the signal, and hence, reduction of dynamic range, at any given volume level. In effect, there is always a small percentage of power needed to energize the electro- mechanical operating system, and all power input below this voltage is not only converted non-linearly, it is wasted and lost. This results in a low-level veil, or "mask", which limits overall transparency and resolution of low level details. I term this effect the masking threshold.

AUDIBLE RESULTS OF THRESHOLD MASKING

When the dynamic range of a signal is compressed greatly, the "jump" factor is reduced, leading to a loss of excitement and hence realism. When listening to a stereo system, one can turn the volume level down so far that the sound becomes much "smaller" than life. The effects of the Fletcher Munson curve make the problem even worse since the ear is much less sensitive to extreme bass and treble frequencies. Generally, when played at a low volume level, most speakers tend to sound like "the bottom has fallen out", since only the tweeters have received enough energy beyond the masking threshold to energize mechanical movement. Not only has the frequency response changed dramatically, but the imaging has collapsed into small "right" and "left" sound sources without any depth or dimension.

These effects have been noticed by many audiophiles who take this for granted. Indeed, it seems "normal" to increase the volume level to a point where the bass finally comes alive and the image becomes dimensional. Consequently, these effects have not been seen as problems by designers.

THE CHALLENGE

Clearly, a higher efficiency design is needed, utilizing crossover elements and drivers which have the highest conversion factor. In addition, to convert a greater percentage of diaphragm motion into sound waves, a cabinet that will not absorb any low level vibration from the drivers is mandatory.

TESTS AND MEASUREMENTS

In order to reduce the "veil" inherent to crossover parts, every type and brand of capacitor, inductor, and wire were tested and sonically evaluated over a period of several months. Many types of distortion, including E.S.R. (equivalent series resistances), non linear behavior, and saturation, were studied with the aid of a highly sensitive computerized reactance bridge and impulse response generator. Blind listening tests were also conducted to assess the sonic contribution of each part.

ELIMINATION OF CROSSOVER DISSIPATION

CAPACITOR SELECTION

It is a well known fact that capacitors have losses termed DF, or Dissipation Factor. Low level details in the signal become absorbed by the dialectric used to separate the turns of the internal element. In addition to electrical impedances, there are chemical and mechanical impedances, the sum of which are termed ESR, or Equivalent Series Resistances. These resistances corrupt the signal and affect the sound quality, adding a "veil" over the sound. All types and brands of capacitors were tested to find the best combination of both dialectric and conductive plate materials leading to the best sound quality combined with the least DF. Even vintage oil filled capacitors were tested since a highly rated $78,000 amplifier uses these capacitors.

The final capacitors chosen for the midrange circuits are solid-plate aluminum and aluminum film polypropylene high voltage units with practically unmeasurable DF and highly transparent sound quality. Indeed, on a bypass test with a wire, these caps did not add any audible or measurable distortion!

Critical treble range capacitors selected are oil filled paper and aluminum units with very large surface areas. These exotic capacitors eliminate a source of high frequency "grudge" endemic to many popular brands of vapor deposited caps used in other high end speaker systems. This grudge may be caused by unevenness in the vapor deposited conductor, imparting a "white" background which adds harshness to lower treble frequencies and reduces upper harmonics and "air".

In contrast, the oil filled units exhibit a velvety black background, rendering the transient response and harmonic integrity far more musically accurate. The lower treble harshness is eliminated, and the upper frequency harmonic structure becomes sweet and airy. The treble integrates imperceptibly with the midrange harmonic structure instead of standing out as a separate entity.

INDUCTOR SELECTION

The common use of inductors using iron cores and small gage wire (for cost reasons) can introduce distortion into the signal due to internal resistance and core saturation. Even physical location of the inductors must be carefully considered. Mutual inductance from strong magnetic fields (generated by larger coils) can radiate into neighboring inductors in close proximity, adding distortion and greatly reducing transparency, especially in the tweeter circuits.

All types of inductor coils were tested for both sound quality and static measurements; purity of the wire, effectiveness of the dialectric covering, and core type were some of the factors considered. Flat ribbon wire vs. round wire was evaluated, along with different types of dialectric coverings and air core sizes. The final inductor design for the midrange and treble sections utilize large diameter ultra high purity copper wire using large air cores to eliminate saturation and distortion. In addition, the inductors are spaced very widely so that no inductive crosstalk can occur to pollute low level signals. The bass frequencies are filtered with extremely large laminated steel plate inductors wound with 16 ga wire.

These oversized units have the ability to handle 1500 watt transients before core saturation vs. the industry standard 40 watt powdered iron core. The audible result of this enormous headroom is a bass clarity which is uncompressed, tight, and transparent, even at high volume levels.

INTERNAL WIRE SELECTION

Not surprisingly, internal wire can greatly affect the signal transparency. Coloration can be caused by impurities in the metal as well as the dialectric jacket. Stranded vs. solid core wires were studied; arcing effects in stranded wire were found to be undesirable as arcing "unfocused" the image specificity, not to mention altering the harmonic structure. The more commonly noted effects of capacitance, resistance, and inductance alter the tonal balance, as would be expected, so these reactances were eliminated by the use of large gage solid copper wire randomly spaced.

A premium quality 99.999% pure copper solid core wire covered with a foamed Teflon dialectric jacket was manufactured exclusively for the VR-6 and VR-8 models. This wire was designed to be not only neutral and transparent, but non reactive to other brands of cables used to connect and interface the speaker and amplifier. Sixteen gage is used on the crossover, with multiple runs equaling 12ga being used for the woofer connections and 16ga for midrange and treble connections. Lead-free silver/ tin alloy solder is used exclusively, as is point to point wiring. We have found that the common ground plane and thin traces normally used on circuit boards leads to an increase in resistance, so we have employed a star grounding system and hard wired crossover boards to increase transparency.

VR-6/VR-8 CIRCUIT ARCHITECTURE

An entirely new crossover design was deemed necessary to achieve the goals of both high efficiency and driver control. Since simple first order circuits allow driver overlap leading to distortion by modulation and beaming problems in the soundwave radiation pattern, steeper crossover slopes are utilized in the new design. Although complex, our proprietary circuit design does not come at the penalty of excessive phase shift or ringing. The Global Axis Integration Network tm was designed to allow steep 24dB/octave acoustic slopes to be formed with cascades of 1st, 2nd, and 3rd order circuits, depending on the frequency vs. the radiation angle over a 360 degree global axis. Equalization and phase compensation is accomplished with shunts to ground, which are parallel circuits not affecting the signal path. In addition, the GAIN circuit enables wide dispersion soundwave patterns to be radiated by the drivers, since beaming problems are eliminated.

Our contemporary minimum baffle design, in conjunction with the spherical wave pattern enhanced by the GAIN circuit, reduces diffraction and allows the recreation of the original soundstage. We have used our principle of Acoustic Inverse Replication tm, or the speaker’s ability to recreate the microphone’s pickup pattern in reverse, to ensure accurate spatial imaging. A rear firing ambience driver adds depth perception and high frequency "air" normally lost in point source designs since they do not normally achieve a 360 degree soundfield. This additional ambience retrieval combined with the wide dispersion allows faithful spatial recreation of the signal, especially effective when combined with a high efficiency/high resolution design. A Spatial Dimension control is provided to adjust the depth effect for any type of room acoustics or listening taste.

The VR-6/VR-8 circuits are extremely transparent and "direct" sounding due to the parts quality and labor intensive, hard wired, hand made circuit boards. Inductors are spaced widely apart to eliminate magnetic crosstalk and the boards are mechanically isolated from the enclosure by elastic damping materials. There are three separate boards employed for bass, midrange, and treble, and each of these boards are housed in different parts of the enclosures for complete isolation and signal purity. In addition, the bass section is electrically isolated from the midrange and treble sections by separate inputs which also allow biwiring and biamping.

MECHANICAL CROSSTALK ELIMINATION

It is well known that cabinet panels will vibrate when activated by the driver cone motion. The panel vibration not only adds unwanted resonances which color the sound, but also interferes with the driver’s ability to start and stop quickly with the electrical signal. For this reason, we not only employ massive cabinet panels (up to 3" thick) which are heavily cross braced with an internal honeycomb, but are isolating the driver frames from the cabinet itself. We have engineered gaskets from a plasticene mastic that is 99% effective in eliminating transmission of the frame vibration to the enclosure panels, allowing almost 2dB of quieting at certain frequencies.

The combination of dead wall panels and damped driver mounting results in virtually no crosstalk at all. The cabinet’s mass and rigidity allows every minute driver vibration to be transferred into sound waves with maximum efficiency. In contrast, conventional cabinet designs with low mass and flexing walls actually absorb low level signal detail, contributing to unintended low level masking.

The VR-6 system utilizes twin 9" Eton woofers from Germany with cast frames, massive magnets, and a proprietary triple Kevlar sandwich cone with internal honeycomb construction. This cone exhibits the highest Young’s Modulus (the ratio of stiffness to weight) we have ever measured, 7 times better than polypropylene! Their transient response speed is more accurate than conventional woofers due to the lighter moving mass (26 gms) and very high driving force (BL product of 8.5 Tesla/Meters). When loaded into the modified aperiodic vented enclosure, the bass is extremely powerful and sensitive (96dB) with clarity and transient response matching the midrange and treble drivers perfectly.

APERIODICALLY TUNED ENCLOSURE

The medium sized enclosure (12"W x 26"H x 22"D) is massively over-built with four tone chambers, each damped with Gradient Density Dacron stuffing to eliminate cavity resonance. Each of the tone chambers is tuned with staggered internal dimensions and 5" flared vents, ranging from 20Hz to 40Hz in bandwidth to prevent "one note bass" endemic to conventional bass reflex enclosures. Two of these flared vents operate into the room, acting as both pressure release valves and tuning ducts. The cabinet wall thickness is staggered and ranges between 1" and 3", with cross braced honeycomb internal walls. In addition, the exterior side walls are further damped with "cheeks" filled with damping materials which prevent the transmission of panel resonance into the air.

Using impulse response tests with 1/1,000 second tone bursts, the woofer system has been developed to be the state-of-the-art in transient response speed. Kettle drum rolls that sound like a roar on conventional systems can be heard to be individual drum strokes, with each mallet strike being quite apparent. The pitch differentiation and timbre of each bass instrument is now easily discerned; in fact, the VR-6 woofer system can easily let you know what brand of strings the bass guitarist is using!

VR-8 WOOFER SYSTEM

When the ultimate bass system is needed due to a larger room size, the VR-8 is the answer. Although large (18"W x 48"H x 28"D), the twin stacked enclosure is elegant in appearance. A SuperDuty 13" professional subwoofer with 22 lb magnet assembly is loaded into a five cubic foot aperiodically tuned enclosure, using quad tuning chambers as described above. A SuperDuty 10" professional woofer with 10 lb magnetic assembly is loaded into a two cubic foot quad tuned enclosure. Both enclosures feature 90% Dacron Gradient Density stuffing and massive cross bracing with honeycomb internal shelving. As in the VR-6 enclosure described above, the tuning of the enclosures is staggered between 20Hz and 40Hz to eliminate the dreaded "one note bass" common to ported enclosures. The Theile/Small alignment is an Extended Bass Shelf (EBS) design with a highly damped Qts of approximately 0.5. (Commonly used Butterworth alignments with a Qts of "1" sound thick and slow, but are routinely used by other manufacturers.)

The VR-8 woofer system has extremely high speed transient response due to the lightweight pulp cones employed, very large voice coils (3" diameter edge wound ribbons), and oversized magnetic assemblies. Although capable of moving tremendous quantities of air at very deep bass frequencies, the response is quite accurate and revealing of delicate sonic nuances at the deepest bass frequencies, comparable to the VR-6 woofer system described above.

VR-6/VR-8 MIDRANGE/TREBLE MODULE

To match the high efficiency and sensitivity of the woofer enclosures, a state-of-the-art midrange/treble array was developed using the highest resolution drivers available.

MIDRANGE DRIVERS:

A 5.5" full range driver with a response of 57Hz to 11kHz was developed to replicate midrange frequencies with the least possible signal degradation. Our research into diaphragm materials several years ago for our VR-4 speaker system proved that woven carbon fiber fabric is the finest available material for midrange reproduction. Not only is the Young’s Modulus (ratio of weight to stiffness) very high, but the tiny traps formed by the fabric weave greatly reduce energy storage on the surface of the diaphragm. This high tech cone material allows very fast, detailed, and neutral replication of sonic information. Since the diaphragm is extremely light and stiff, it can be accelerated and decelerated very quickly (3kHz = 3,000 vibrations per second!), with a minimum of cone flexing. This high degree of accuracy translates into very high transparency and clarity not heard with ordinary drivers. In contrast, soft materials such as doped paper and polypropylene allow flexing which in turn dissipates mechanical energy which should instead be transferred into the air as sound.

In addition, the carbon fiber cone has been flared in an exponential shape to enable a very wide dispersion pattern. In contrast, the normally used conical shape does not allow wide dispersion of higher frequencies. The edge surround of this full range driver is formed of a patented dampening agent called Norsorex. Instead of allowing energy to be stored by a "live" rubber surround, this proprietary edge absorbs vibration. Hence, the next incoming signal does not have to "fight" a signal being stored as vibration in the mechanical suspension. In addition, another form of distortion was eliminated by venting the rear suspension "spider" which normally traps air within the voice coil gap. Instead of the commonly used multiple layer voice coil which can reduce high frequency range, a single layer of ribbon wire, wound on it’s edge, is employed. This results in lighter weight, faster transient response, and better high frequency range. These innovations all contribute to this driver’s inherent superiority to even electrostatic or ribbon drivers, and can be considered as a reference.

In order to move the quantity of air necessary to match the sensitivity of the woofer system, two of these drivers are utilized, wired in parallel. The resulting sensitivity is 96dB, with far better power handling and lower distortion than a single driver since two voice coils share the load. These midrange drivers are mounted in a narrow profile sealed air suspension enclosure, damped with 100% fill of Gradient Density Dacron to eliminate cavity resonance and backwave interference.

TWEETER:

For extreme accuracy and realism, a 1" inverted titanium dome damped with a layer of titanium dioxide to prevent ringing, has been chosen. To enhance transient response speed and signal accuracy, the light weight diaphragm is driven by an enormous 1200 gram magnet. The rear wave is vented into a transmission line cavity damped with absorbent stuffing to reduce coloration and prevent phase shift in the audible band. In addition, Ferrofluid is used to both cool the voice coil and damp any mechanical vibration, adding to it’s sweetness at high SPL’s. The pulse replication of this device is incredible due to the amount of flux density driving an extremely low mass diaphragm, resulting in the most accurate yet smooth replication of high frequencies possible.

AMBIENCE DRIVER:

An identical tweeter is used on the rear of the enclosure to match the frequency response and soundwave pattern. However, a special circuit is used to drive the ambience tweeter. Since the rear wave should ideally be the inverse of what was picked up at the rear of the recording mic, a specially equalized and processed signal is fed to the driver. This signal is derived from the ambience in the program material and has a tailored frequency curve to enable the system a quasi-omnidirectional response. As it’s signal is out of phase with the front tweeter, the response in effect is dipolar in nature and enables correct depth replication similar to planar designs.

TIME ALIGNMENT:

To replicate correct depth and focus, the acoustic centers of the midrange drivers and tweeter have been phase aligned. Rather than centering the tweeter between two midranges as is the current "rage", we have stacked the drivers to allow correct high frequency dispersion without the tunneling effect that a center mounted tweeter exhibits. In addition, the close spacing of the twin midrange drivers allows them to radiate a coherent, spherical wavefront instead of the directional beam set up by spaced drivers. This opens up the sound considerably and results in an acoustic image that is more stable to listeners either standing or sitting, even far off axis.

NEW SYSTEM PARAMETERS:

The sum total of the increased efficiency, reduced dissipation, and heightened transparency results in new system parameters. The VR-6 and VR-8 models exhibit the highest degree of accuracy to the electrical signal, but more importantly, speak with the voice of truth. It is immediately evident to the seasoned listener that a new level of realism has been attained: dramatic shadings of expression are now revealed in the most mundane of recordings, not just sonic spectaculars. The wide dynamic range allows explosive crescendos with frightening power to be combined with microscopic enlargement of harmonic fractiles. This combination brings forth sonic details previously unheard, heightening the illusion of "live sound".

The sum of these engineering benefits result in what may be considered the most emotionally involving speaker system ever created. Indeed, every listening session becomes an event not soon forgotten....