This is a continuation of my discussion with Albert Von Schweikert of Von Schweikert Audio. Part One covered Von Schweikert’s history as a speaker designer and how he broke into the business. Part Two, below, discusses his present and future products as well as his current thinking on all things loudspeakers.
Garrett Hongo: What about your new VR-44 speaker?
Albert Von Schweikert: The VR-44 is different. It comes in either an Aktive ($20,000 USD per pair) or Passive ($17,000/pair) version. The Aktive is able to perform either close to the wall or out in the room, due to controls that adjust the bass level. The VR-44 Aktive might be seen as a full-range speaker with a built-in powered woofer system, but we also offer a Passive version that does not have the amplifier on board to drive the twin 8.8" woofers. The powered woofer system does not alter the sound of the main amplifier. The onboard amplifier takes the sound of the main amplifier and boosts it up to 300W to drive the woofers below 100Hz. It’s basically just a power booster. In effect, this design enables you to biamp without having to worry about outboard electronic crossovers, and will increase the dynamic range. If the customer owns a small tube amp but does not like the sound or large size of horn speakers, the VR-44 Aktive is their speaker. If the customer has a large amplifier with 100W or more and does not need biamp capability, the VR-44 Passive might be the better choice.
The VR-44 is a single-cabinet design with tapered sides and a slanted front baffle. It uses the same technology to damp the cabinet walls as our UniField 3. The drivers are made to our specifications and include twin 8.8" metal-alloy cones, Low Distortion motors, and a phase plug to eliminate heating effects and wind noise caused by air trapped inside the voice coil. The 6" midrange driver is made from carbon fiber, cellulose acetate pulp, and liquid PVC coating to reduce cone breakup, while the motor uses the Low Distortion patent, which consists of a copper-clad center pole-piece, shorted voice-coil windings at the front and rear of the active voice coil, and a specially designed magnetic system that greatly reduces the effects of the voice-coil motion in the magnetic gap. This is a high-cost, super-quality midrange.
On top, we use the latest-version ring tweeter with the least distortion that was developed in Denmark. To begin with, the diaphragm is pressed from a special fabric impregnated with liquid PVC, so that it has a rigid shape that will not deform under high acceleration. There is a chamber behind the ring diaphragm that is filled with a damping material to absorb the rearwave, enabling the tweeter to respond down to 500Hz, which is amazing. At the rear of the cabinet you will find our rubberized-linen mid-tweeter, used for Ambience Retrieval. It is mounted inside a waveguide to direct the off-axis response and is driven by an Ambience signal derived from the out-of-phase reverberation contained in all live recordings. This Ambience Retrieval system allows the VR-44 to envelop you with sound -- you might believe that you are listening to an omnidirectional design, although the design is closer to a cardioid microphone’s. It incorporates my earliest insights re. speaker design, so the ambience tweeter mimics the cardioid mike’s dipole pickup pattern with more radiation in front and back than side information. This permits the VR-44 to generate a spacious envelope of sound that can be compared to a "bubble" instead of a "beam" of sound, so you don’t have to sit in the direct middle to hear a 3D soundstage. All in all, the VR-44, in either Aktive or Passive form, is a very musical speaker system.
GH: How do you distinguish among various types of drivers?
AVS: Different drivers have different personalities. And it can be very difficult to decide which one of the various different types of drivers to choose, either as a listener or a designer. By far the most common are moving-coil drivers. These are the most popular because, to my mind, they have the least amount of engineering drawbacks. I myself have focused on moving-coil cone drivers in my designs. On the other hand, planar drivers, ribbons, and electrostats are drivers I admire, but that I think have engineering limitations that cannot be conquered. They have very limited dynamic ranges. They can’t move in direct correlation to the dynamic signal. They suffer compression due to lack of dynamic range, beaming effect -- just too many challenges. It’s my view you can get acceptable sound from them, but not realistic sound.
GH: Why do manufacturers choose paper over polypropylene, ceramic, etc.?
AVS: Designers have certain listening biases. For example, the English privilege the quality of smoothness in their speakers, so they tend to choose polypropylene over paper drivers, for the smoother sound. Paper has more rigidity and is lighter. It can trace the musical waveform more accurately than poly, but it has a harsh breakup mode. Paper goes into resonance at breakup that adds hardness to the sound. This hardness can be eliminated, though, by surface application of rubber, vinyl, or impregnated mineral compounds mixed with lossy compounds like rubber. Paper drivers are often covered in Kevlar, carbon fiber, epoxy resin, spray-coated with a clear vinyl, or a combination of some of these, to cut down on this hardness aspect.
Ceramic and metal have the same property -- a very good Young’s modulus, which is to say, a high stiffness-to-weight ratio. These are the fastest-sounding drivers, as they’re lighter and more rigid than any other driver materials -- they don’t flex and distort the waveform in response to the dynamic signal. In one way, these are the best membranes of all. But they have a huge drawback: high-Q peaks up to 15dB above an average signal level. They ring like a bell. There is a way to eliminate that, however, using a notch filter. This is a complex filter in the crossover network to pull down the Q peak. But you do end up adding more components in the signal path. Also, this notch out of the Q peak can still introduce a high-Q filter action that, in its turn, causes unwanted electrical ringing and phase shift.
I use metal-cone woofers myself -- aluminum, magnesium-aluminum alloys, and pure magnesium coated with ceramic treatments -- but I employ them several octaves below ringing. Below 200Hz the ringing isn’t an audible or measurable problem, since the ringing effect does not start until the range above 2000Hz.
GH: Why not make your own drivers?
AVS: I designed drivers at ESS and KSC Industries, for several years at each company. I realized how difficult it is to manufacture them. It takes millions of dollars of capital investment in machinery and an adequate staff of trained employees. For us at Von Schweikert Audio, it just isn’t cost-effective to DIY when large driver manufacturers can make drivers much more efficiently.
But here’s what I do. I buy off-the-shelf drivers from SEAS, Scan-Speak, and, in the past, Audax and Vifa. I put them in a prototype and test them, and critique each of the drivers for behavioral characteristics as they interact with other drivers of different membrane-material construction and characteristics. I end up with a list of preferred specifications to customize each of the three or four different types of drivers I want in a speaker so that they work optimally and coherently with each other. I send this to the manufacturer, and they then build their drivers to these specifications for me.
The big secrets are twofold: 1) the timbre of each one of the drivers has to be similar to the one next to it so they all sound the same; and 2) the transient response of each of the drivers has to be the same. You match drivers for timbre and transient response. This is why you goof around with different drivers -- ones with cone coatings, ones with alterations of structure in diaphragm, voice coil, magnet, plates, pole-piece and lug, etc. Whenever you change the shape or materials, you get a shift in transient response and distortion characteristics -- the way the voice coil operates the diaphragm, for example. You measure transient response and distortion level to get clues on how to modify them so each driver will have the same start-and-stop behavior as the one next to it. The goal is low distortion, and for all three or four drivers to have the sound of a single driver. This sounds really easy to do, but it isn’t. Typically, it takes around 12 to 16 months to build successive driver prototypes and test them, while increasing their potential by redesigning the weaknesses and enhancing the strengths. Although the largest factories in the world can afford to conduct these successive tests, it is extremely expensive, and is one of the reasons why the costs of modern speakers have risen to such heights in the last few years.
GH: Why do you think cabinet construction is important?
AVS: There are two schools of thought here. One says that the speaker should create its own sound on top of the signal. The cabinets here are thin-walled, resonant, and the claim is that the cabinet itself is a musical instrument -- that it can even be "tuned." This is completely incorrect from a theoretical and even a listening standpoint. The other -- which happens to be my own school of thought -- is that the cabinet should add no sound of its own. It’s there only to hold the drivers in place. We at VSA believe cabinets should not resonate, but that there are also different ways to accomplish nonresonance. We use a triple-wall detuning of the cabinet, combining three layers of different material, each with its own separate Q, so that their resonances cancel each other when combined. To make the cabinet wall, we use a resin-impregnated medium-density fiberboard, a manufactured rocklike material made from crushed gravel and resin, and a hard felt, all glued together with a compliant rubber-based adhesive. It’s cost-effective and completely nonresonant. We call it the Triple-Wall Laminate (patent pending) and have a white paper on it.
GH: What about your crossover design, which I understand to be fourth-order? Isn’t there a problem with the complexity of a fourth-order design? For instance, isn’t the raw parts count higher than in a first-order design?
AVS: It’s theoretically true that the best crossover is no crossover. It’s not true, however, that a first-order crossover can be made simpler than a fourth-order. Our fourth-order design has a crossover circuit that divides frequencies, but it also has phase compensation, time-alignment (accomplished through the electronic delay inherent in the parts themselves), and equalization among drivers to level their outputs to a standard flat response. Technically speaking, we get a flat amplitude -- a linear amplitude at each frequency. All four major acoustic aspects of a speaker -- time alignment of the drivers, flat frequency response, coherent phase, and the realistic dimensionality in overall sound -- are blended together. We don’t create four separate circuits all in series; that causes veiling due to a slight amount of absorption of the signal by each crossover part. We use a computer to synthesize functions and reduce circuitry in order to maximize transparency. We’ve eliminated a lot of complexities. In short, we reduce the number of parts and still get same effect. Our crossover is four circuits combined into one circuit -- fewer parts to accomplish the same job. In addition, we use circuits that are not in the direct signal path, called shunts, to control most of the functions. The actual quantity of parts directly in the signal path is not higher than in the typical first-order design.
GH: What about bass-cabinet design? Each of your speakers, old or new, uses the transmission-line design rather than the ubiquitous bass reflex. First of all, what is a transmission-line speaker?
AVS: A transmission line is a box with internal compartments stuffed with damping material employed for nonresonant bass tuning. You can stuff the cabinet with Dacron, foam rubber, or fiberglass. Dacron is easiest to remove or add, so that’s what we use. Note that the terms transmission line and labyrinth are sometimes confused with each other. A transmission line is not necessarily constructed as a long tunnel (the labyrinth), and a labyrinth can sound boomy if it is not properly stuffed with an absorptive material, behaving more as a bass-reflex design. The way to determine whether a speaker is indeed a transmission line is to examine the impedance peaks of the cabinet and woofer. If you see two different peaks in the impedance measurement, you are seeing the resonance of the woofer and the resonance of the cabinet -- that is the sign of a bass-reflex cabinet design. Correctly designed transmission-line cabinets have only one impedance peak, occurring at the woofer’s free-air tuning frequency. The cabinet resonance will be damped by the stuffing material, which in turn eliminates the second of the impedance peaks. The reason we use the transmission-line design is due to the tight bass response, high efficiency, and deeper bass than can be obtained by using a sealed-box system.
GH: Why don’t you use the bass-reflex design that’s so popular even in many of today’s more sophisticated speakers?
AVS: At Caltech in 1976, we in the lab tested three designs: a sealed box, a ported box, and a transmission line. The sealed box had tight bass but it lacked dynamics, lacked efficiency. It needed a lot of power to be driven sufficiently. The bass-reflex had deep bass and dynamics but was boomy. Only the transmission line had everything.
Another difference is that a computer can model a sealed box for $300 worth of software. A bass reflex is the same. But to get a good transmission line, it’s basically trial and error, since there exists no software that can predict the outcome of the design. Designing a good transmission line takes time and a lot of listening.
However, the classically built transmission line is fairly low in efficiency: 85dB. It calls for an 8’ to 12’ internal tapered tube with numerous folds that the rearwave then travels through. Von Schweikert Audio speakers are around 90dB because we use a quasi-transmission-line hybrid design that has a shorter path. We have chambers inside the cabinet formed by shelves that create a labyrinth only about 3’ to 4’ long. Then we use a certain density of stuffing material to compensate for that lack of length and tune it all by using two different methods: instrumentation and our ears. When the bass is deep and the transient response is very fast and controlled, we have the right design.
GH: How close can speakers come to the sound of a live performance? Is it possible, or are manufacturers simply building sophisticated Towers of Babel?
AVS: You can’t build a speaker to exactly reproduce a live performance, since recordings are far from perfect. This is a guesstimate, but stereo can get you only about 75% of a live performance. The microphone is actually very insensitive compared to our hearing. And, I’m afraid to say, loudspeakers are worse. Recording equipment, mikes, and loudspeakers lose about a quarter of the vibrancy of any live performance. However, this is not to say that speakers can’t make good music. I find it highly enjoyable to listen to music on a good stereo, but am not often fooled into "a willing suspension of disbelief" unless the speakers fill the room with the dimensionality of the original concert hall or recording venue. Our rear-firing Ambience Retrieval System is a method to replicate the dimensionality encoded onto the recording. This works by creating a dipolar soundfield, which behaves in a similar fashion to a cardioid recording microphone. We have a decoding circuit that subtracts the ambience of the recording. Then that ambience signal is fed to a rear-firing tweeter with waveguide. To control the size and depth of the image, we provide an Effects Level Control that enables you to dial in any amount of ambience you need, from zero to full. In the past 20 years since we have introduced this enhancement, we have been copied numerous times. What we use accomplishes a "suspension of disbelief" by having speakers fill the room with the dimensionality of the original recording, giving you the emotional impact of a live performance.
GH: What’s emotional impact?
AVS: You’re taken on a trip to the composer’s soul -- transported to his mental and emotional state of mind when he composed a particular piece of music. That is powerful. When listening to recorded music, you want to preserve this impact and not have it destroyed through coloration or distortion. That’s what you work for.
GH: What’s next for Von Schweikert Audio?
AVS: It’s a cable line we call Master-Built because I consider the Delphi Aerospace engineers who created these cables to be masters at their craft. They design wiring harnesses for satellites and space shuttles. Their cables have eight patents in the field of reactance cancellation and noise reduction. Since a $50 million satellite can’t have contaminated electrical signals that would corrupt the data, Delphi engineers created cables from materials not previously used in audio. The base metal is copper in a single-crystal form that is cast instead of extruded. But the magic of these cables is in their proprietary noise-reduction technology. All of the cables are shielded against RFI as well as EMI. The shielding resembles metallic powders coating the three different layers of insulation lying on top of the foamed Teflon. The wires are all individually insulated and are wound in a patented cross helix that almost completely eliminates reactive effects such as capacitance and inductance. These are the first cables I’ve heard that could be called game changers. I was so impressed that I bought the distribution rights for the US. We use this wire inside our expensive speaker systems, from the VR-35 Export Deluxe upward. These cables are best used as a complete wiring system and should not be used in combination with other types of cables, since their effect will be diluted by the distortion inherent in the other cables in the system. Although expensive, they are not overpriced, and cost less than many boutique brands. Delphi has also introduced an entry-level line that anyone can afford. I use it in my own system at home.
We also just released a Mk.2 version of the UniField 3, featuring parts upgrades, a change to Delphi internal wiring, and an upgraded ribbon tweeter with greater power handling and smoother sound. Owners of the original Uni 3 can upgrade theirs for $2000.
GH: Thank you, Albert. It was a pleasure.
AVS: Well, it’s a great pleasure for me to be able to speak with you at Ultra Audio. Thanks for the opportunity.
. . . Garrett Hongo