- Category: General Interest & Interviews
- Created on Sunday, 01 July 2012 00:00
- Written by Peter Roth
In Part One of this feature interview with Laurence Dickie, loudspeaker designer for Vivid Audio, we explored his foundational experiences and the gestation of his philosophy of sound reproduction by high-performance audio products. In Part Two we turn to Vivid Audio specifically, and to Dickie’s challenges and achievements in the past decade. Philip Guttentag (cofounder of Vivid Audio) and Philip O’Hanlon (US distributor), also present, each occasionally joined in.
Peter Roth: Let’s turn to the present, and Vivid Audio’s Giya line of loudspeakers. I’m interested in the differences between the C125 bass driver used in the Oval series and the C125S midbass driver used in the Giyas.
Laurence Dickie: They are distinct, different drivers. Giya-series loudspeakers are true four-ways. That means that the C125S, which some would call a 6.5" driver, has to operate only from 220 to 880Hz. Accordingly, the excursion requirements are far smaller than for the standard C125, which functions deep into the bass with a diaphragm travel of +/-10mm. In the C125S application, only a couple of millimeters of driver travel are needed to give us 120dB at crossover. This means we can concentrate the magnetic flux into a smaller zone -- a smaller gap length -- and so we’ve also taken the opportunity to reduce the gap clearances. With tighter tolerances, the driver assembly is little bit fussier, and of course is more susceptible to contamination, but that is just a little reality of manufacturing -- which is fine, and in fact we have had no difficulty with it. We gain 3dB efficiency in the main passband. You wouldn’t be able to use the S driver in a full-range system because it wouldn’t plumb the depths. The excursion limit, the X-max (the length of the magnetic gap), is too small for a reasonable amount of sound output in the bass.
PR: When Doug Schneider measured the review pair of Giya G2s at Canada’s National Research Council, it was only the second speaker they’d measured to test with 0% distortion throughout the 200Hz-1kHz band, which is such an important range.
LD: Yeah, that was nice to see, actually. It was right off the top of the scale.
The C125S driver in the Giya G2
Philip Guttentag: I think the confusion is that the C125S and the standard C125 look the same from the outside. But looks are often deceiving; they are very different underneath.
PR: The Giya G3 has a new bass driver, the C135. While obviously 40mm smaller in diameter than the C175 used in the G2, what else differentiates these drivers, and what challenges -- presumably resulting from the G3’s smaller size -- was the new driver designed to meet?
LD: Because of our approach to the bass-system design, and the way we integrate the crossover and driver and enclosure together as a single system, we have more flexibility in our approach. Traditional design, particularly of bass-reflex systems, prescribes pretty tightly what the characteristics of the driver should be. But our approach is certainly different. It means that we have the advantage of using the same motor structure on the three bass units -- in the G1, the G2, and G3 -- to achieve the same filter shape, albeit with a slightly different low-frequency extension. It has allowed us to maintain the excursion capabilities of the driver. So the C135 in the G3 is able to move as far as the C225 in the G1. It has the same linear X-max as the G1. It remains capable of shifting an enormous amount of air. Frankly, the C135 gives the G3 a very satisfying low-frequency performance, which is kind of counterintuitive when you see the difference in the sizes of the two speakers. The psychology is such that you would expect the low end to be far more curtailed than what actually results. A large part of that is because of the consistent motor structure throughout the range, and the suspension design is such that, despite the reduced diameter, it is nevertheless mechanically able to travel the full distance.
PR: Your bass-system design seems quite novel. Describe, if you would, the reaction-canceling system and how it’s mechanically decoupled from the cabinet.
LD: Here’s the thing: We tie the magnets together in a reaction-canceling configuration. I have to tell you, reaction canceling totally, absolutely works. It has been used in so many different areas of engineering and industry. The Harrison Chronometer, for instance, a very early and very accurate timepiece that allowed seafarers to travel across the world and maintain some idea of longitude, depended entirely on having a pair of balanced pendulums. Move forward several hundred years and you get a Ferrari flat-12 engine or a BMW horizontally opposed engine. It simply works! When you have two equal masses traveling in opposite directions, the reaction forces are completely canceled, a phenomenon that works really well when applied in loudspeakers.
We also choose to "float" the drivers from the cabinet. There’s a seal -- you need to have a pneumatic seal to keep the air in -- and we’ve chosen to just leave it at that. The drivers float on that pneumatic seal. The truth is, a tight mechanical contact would still be OK, since there is theoretically no reaction force, but we rather like the fact that it floats: It provides that little extra bit of refinement. So the bass system is both reaction-canceling and floating. There is a complete absence of mechanical transmission of vibration between the motor system of the driver pair and the enclosure walls.
The Giya G3s at CES 2012 in Las Vegas
PR: In extensively listening to the G2 and G1, I’ve noticed that the bass is very impressive for a ported design -- not only the amount of bass output, but more important, the tight, integrated nature of the bass and its consistent tonality from the lowest frequencies up. Why did you choose a vented design over a sealed one?
LD: Bass reflex gives you extra output at the port tuning without cone excursion. It’s not "energy for free," as I’ve sometimes heard the expression, and you could achieve the same thing from a closed box, but it would be less efficient. Let’s turn your question around: What is the problem with it? Why wouldn’t you use bass reflex? The problems occur in application: Bass reflex is more difficult to get right, and when the approach to tuning the whole system is slightly short of the mark, the result can be overhang at the port tuning.
My lesson in this regard came at my previous outfit [B&W], where we did some work with a so-called sixth-order system. A closed box is second-order. Reflex is fourth-order. They came out with sixth order, which adds an electrical high-pass filter with a little 6dB bump at the cutoff frequency, and coincides the bump with the port tuning such that maximally flat response results. Without that electronic filter, without its boost, you would have an overdamped high-pass function. Subjectively, despite the fact that the quantity wasn’t there -- in theory, you could say it was slightly bass-light at port tuning -- what everyone appreciated was the tightness of the bottom end.
Ever since then, I’ve worked to maximize bass control. To put it into filter terminology, it is like the difference between Butterworth and Bessel filters. A Butterworth is maximally flat in the frequency domain, but a Bessel filter is maximally flat in the time domain. It’s that type of difference. We go for a high-pass filter response in the bottom that is as flat as possible in the time domain. I think that’s why ours works for many listeners. That’s why people refer to the speed of our bottom end: crisp without being light. It’s a different filter optimization.
PR: Outside the bass system, and throughout Vivid Audio’s lines of speakers, you seem to use O-rings to float the drivers.
LD: I think it’s really important to prevent the mechanical excitation of the enclosure walls by the reaction forces on the magnet at all frequencies. Inevitably, when the cone or the dome moves forward, the magnet moves back. Now with the velocities and the ratio of the masses (for instance, in a tweeter, where the mass ratio is a few hundred or something), it would seem the velocity of the magnet is insignificant. The velocity of the magnet is insignificant until you couple it to a cabinet. Then, suddenly, you have this massive radiating area. A guitar string on its own doesn’t sound like anything, but stick it on a wooden box and all of a sudden you have a guitar. It is really important that only the magnets move in reaction -- the area of the magnet is insignificant -- making sure that that vibration stays in the magnet and doesn’t get out into the cabinet. It is actually really easy to decouple everything but the bass unit. Do silicon O-rings really work? Absolutely.
PR: Andrew Jones, of TAD, tells the same tale. The acoustic logic is unassailable, yet very few speaker companies decouple their drivers from the cabinet.
LD: This is because most people still bolt on the flange, which is due to the drivers they buy. The driver manufacturers don’t stick their necks out by offering flangeless drivers with a rear magnet mounting, but that’s what you need to do. For cone drivers, you need to support the magnets at the back, with the edge just floating on a little bit of rubber, an O-ring or something. With our D26 tweeters and D50 mids, we are taking advantage of our tapered tube loading. With those tapered tubes we have an obvious point to suspend the thing at the back -- it is the logical place to stick a bolt in from the back of the cabinet. The tube, incidentally, is itself decoupled from the magnet, so the tube can be rigidly, mechanically coupled to the enclosure.
PR: You essentially invented B&W’s Matrix technology in the 1980s. In the 1990s, your B&W Nautilus was unlike any conventional box speaker previously seen. Now, at Vivid Audio, the "box" is anything but. How do you approach cabinet design?
LD: There are two sides to the box: the inside and the outside. The outside needs to be smooth. That is really in accordance with the teachings of Mr. Harry Olson, from 60 or 70 years ago. Olson quite clearly showed that the worst shape for an enclosure is a square box and, at the opposite end of the spectrum, that the best shape for an enclosure is a sphere. His generalization assumes that there is a single driver. But while there are additional complications in a multi-driver system, it is clearly important that you eliminate the sharp edges, which are sources of diffraction. This dictates the external form of all our enclosures. When it comes to the inside of the enclosure, for nearly all the midrange drivers and tweeters we have rotationally symmetrical horn enclosures -- our tapered tubes. They funnel the sound away and absorb it so that it never comes back to interfere with the diaphragm. The remaining enclosure consideration is the structural performance. From that standpoint, the design goals are again fairly straightforward: all the structural, vibrational modes of the enclosure must be out of the band of the driver mounted within.
What you actually see when you look at a Giya is the bass enclosure. The bass unit crosses over at 220Hz. My goal is to keep all structural resonant modes of this bass enclosure well above the 220Hz region. That’s it, really. The external form is diffraction limiting, the internal form is to prevent standing waves and reflections, and the structure keeps resonant modes well above the highest frequency of the driver located within that enclosure.
PR: Are there a lot of new computer design tools, such as finite element analysis, available now to assist in these aspects of your designs?
LD: They are potentially there to help, yes. But, actually, it is quite remarkable: With experience in materials, processes, and structures, it’s pretty straightforward to do the job correctly. We use sandwich composites for the enclosure walls in the Giya series, which are stiff and light. Without too much effort, it is quite easy to make sure the structural modes are well outside their bass passband. Now, if this thing were going into space and weight was at a huge premium, then clearly, finite element analysis would come into its own. We could make a suitable enclosure that was a fraction of the mass, but it would clearly cost an arm and a leg. Yet our cabinets are already much lighter than much of our competition offers for the same (or less) structural performance. There is very little pressure at this stage to further optimize that aspect of the cabinet design.
We use unidirectional glass for the inner and outer skins of the sandwich. We just completed some research using carbon fiber. We could shave 40% off the weight of the cabinet, but it would double the cost -- and remember that much of the system weight derives from the magnets in the drivers rather than the enclosure itself. Nobody is asking for lighter speakers at the moment. Referring back to the subject we discussed earlier, you may well now challenge me to keep the same weight while improving the performance -- at vastly increased cost. We could actually do that, and we may choose to do that. We have reached a stage with our products with which we are very happy. But already I’m asked, "Where are you going to next?" As I feel we have tidied up the sound field pretty well, where do we go next? We will further research what I consider to be "second-order" effects. We’ve got our resonant modes so far out of band already, but we can push them even further out of band -- for a hugely increased cost. This and other second-order effects are something we can look at in the future.
PR: A few recent designs from competitors (e.g., the TAD CR-1 and the Magico Q1) are pushing what is possible from a stand-mounted speaker. Could there by a Giya G4 bookshelf speaker on the horizon? I ask because not all domestic listening rooms can accommodate a floorstanding speaker, but some enthusiasts still want the very best sound possible in such environments.
LD: You can use our C1 speakers as bookshelves or nearfield monitors on top of the desk. But . . . is there a plan for a G4 bookshelf? There hasn’t been. Yet as I was sitting there talking with one of my colleagues the other night, we thought it would be cute. It really would have the baby cuteness factor if we could make a Giya that is 20" high. But I can’t see how we would do it at this stage. It would be fun. I can imagine doing a three-way, taking that C135 bass driver and running it up to 600Hz. If that bass driver was forward facing, then you could potentially cross over to the midrange. It would be an interesting project, but hasn’t really been considered.
Philip O’Hanlon: I don’t know, off the top of my head, of any other speaker manufacturer that makes a range of speakers appropriate to the room size without sacrificing any quality. For example, take Wilson Audio. The aural differences between the Sasha, MAXX 3, and Alexandria are huge. With the Giyas, there isn’t really a huge difference in performance as long as people apply the right speaker to the right room.
PG: It is not as if any Giya is designed for a specific room volume, although the G3 will obviously fit in more spaces, into smaller rooms. There are some trade-offs, but when you listen to the G3, you hear a full-range speaker cut from the same sonic cloth as the G1 and G2.
PR: I had seen pictures of the Giya G3 and read the brochure. It’s 45" tall and contains just one-fourth the volume of the G1, yet only when I sat beside the speaker "in the flesh" did its physical and visual compactness come across. Sculpturally, the architectural structure is now of a scale that it will fit into almost any domestic environment. It’s not as extreme or demanding physically or visually, and doesn’t dominate like the G1 (or, to a lesser degree, the G2), but rather complements its environment. It’s really beautiful.
LD: We’ve softened people up with the most extreme in the Giya G1. To those who find the shape a challenge, the G1 was a real challenge. People are now getting used to the shape. In fact, it’s remarkable how people warm up to the look, even those who quite seriously oppose it initially. With those two things -- growing acceptance of the form and the reduction in size -- working together, I think it’s going to be very well received. In fact, the reaction we’ve had . . . people are reacting very positively.
The Giya G3s at High End 2012 in Munich
PR: You have a personal and professional history with active loudspeaker systems. Active speakers are commonplace in professional audio, but rare among consumer products. Setting aside the question of the commercial viability of a fully active consumer speaker model, how does the active approach solve problems and expand performance potential?
LD: When I first joined forces with Philip [Guttentag], I was firmly dedicated to active-only solutions. I thought that the only way of getting a decent, matched, prescribed filter function was to use an active filter. But one of the essential parts of why our partnership works is that Philip has a realistic perspective on things. He was quite frank: "To start with an active solution, we might as well just chuck the money into the sea and go fishing. We would absolutely be shooting ourselves in the foot to start with active loudspeakers. We are making a product that is difficult enough as it is. From a design perspective it is an expensive loudspeaker from an unknown brand. All those things. We do not need to add ‘active’ to the list of reasons why people won’t buy it."
So OK. I was forced to get back into the world of passive crossovers. But things had moved on since I first took a go at passive crossovers. There now exist some excellent computer-aided design tools. Network modeling and optimization have radically changed things. I was so forced into it -- but once I got there, I was amazed at the possibilities in passive networks. The important thing is to do your driver-design homework in the first place -- drivers which behave in a linear fashion, which have constant properties so that, irrespective of where the voice-coil is, the driver impedance remains the same. With such linear, constant drivers, you don’t modulate the performance of the passive crossover. Also, frankly, taking care with acoustics so that actual driver performance within the cabinet is a fairly smooth shape. It doesn’t matter too much what the shape is, as long as it is free of nasty wiggles. The crossover-design software works incredibly well, and we achieve matches between measured performance and the target function within 0.5dB. It is amazing! I would never have thought. Ultimately, I think it’s probably true that the active solution has to have the slight edge, but only a marginal one compared to what I used to think.
So I have been converted somewhat, assuming we’re talking about minimum-phase crossover function. But if you move the debate into the digital active realm, then a whole new world opens up. We could do linear-phase filtering with a digital active crossover. This means we could actually end up with a system capable of reproducing squarewaves properly, with flat group delay. How sensitive our ears are to that phase distortion is a matter of debate, but certainly we can tell the difference. If you switch between a system with a flat group delay and one with a phased response, you can hear a difference. But actually, it is rather hard to tell which is the correct reproduction. Nevertheless, we know the linear-phase alternative is correct by definition. At some point, I’m sure we’re going to go down that avenue. Whether we supply an active filter module or work with an electronics manufacturer to supply a complete package with amplifiers, there exists that possibility.
PR: For a long time I’ve thought that a converter plus amplifier per driver in an active setup could produce the best of all possible worlds -- theoretically, at least, and assuming high enough quality in the D-to-A conversion. Perhaps we could have our cake and eat it too. But all attempts I’ve heard to date have left me wondering.
LD: You mustn’t forget that an active system, even with very accurate correction, is never going to make up for a poor driver. First and foremost, it is most important to start with a bunch of good drivers in a properly designed cabinet. Maybe some of the products you heard involved less-than-ideal driver combinations, where the designers had hoped that the mere application of correction would overcome those limitations.
PR: What do you wish consumers better understood about speaker design? How could enthusiasts better approach the selection of a speaker system for their listening spaces and musical preferences?
LD: This revisits our discussion about speakers with character. I understand why people like speakers with character, those that are non-neutral. It’s like constantly putting MSG on your food or something. It might seem nice in the moment, but it is not right, and it may not be a recipe for long-term satisfaction. I believe that a wider range of music will be more enjoyable if you have a more transparent, neutral system. It is tempting to go for something with a bit of bite, with a bit of shout, with a little something that slams you in your chest, but while that might work well with certain program material, it isn’t accurate, even if sometimes it’s quite good fun.
Some people get what we do straight away. Others seem to think that there’s something missing. There probably is something missing -- a coloration that they’ve gotten used to -- but I’m always very careful with these things since, at the end of the day, it’s a matter of taste. I guess there are as many ideal loudspeakers as there are pairs of ears in the world. Far be it from me to say we’re right, you’re wrong. What we have noticed is that, as new loudspeakers are developed to be progressively less colored, we see a convergence toward a consensus that it’s a good-sounding product.
Perhaps this is the point where I should air one of my recurring themes. I call it consistency of philosophy. Basically, I believe that you are far more likely to get a realistic, homogeneous sound if each frequency band uses more or less similar transducers loaded the same way. So I’d be very suspicious of a system with, say, a thin 15" paper cone in a box for the bass, an electrostatic dipole midrange, and a horn-loaded bullet tweeter on top. While it might be perfectly possible to tweak the crossover to give a decent on-axis frequency response, the chances are that when you listen it’ll be disjointed. On the contrary, a system with, say, horn-loaded everything -- where the horn shapes are different only in scale, and the polar responses match from one band to the next -- that will probably jell. Our speakers are like that, with our metal drivers on the ends of tubes for each band and broad dispersion at all frequencies.
The other thing consumers should remember, of course, is that you have to be careful about the rest of your system. That, actually, was something we found with the original work on the Nautilus. The more imperfections you remove, the more noticeable those that remain become. It’s an obvious conclusion, if you think about it. But it is a really important thing to understand. If you put a pair of G3s on an otherwise compromised system, and suddenly you hear those system-based compromises, you may incorrectly ascribe the problem to the speakers.
PR: What drives you crazy in competitors’ designs, or in the industry at large?
LD: Midrange cones! Really, though, the single biggest thing that bugs me in the world of audio is bad science. I don’t mind designs based on gut feel, even if I might disagree with the conclusion, but the thing that really winds me up is when I see scientific untruths used to bamboozle the naïve customer.
PR: I keep hearing, from sophisticated industry sources who should know, that you have technical expertise beyond all but a very select few, and that your designs make 100% perfect sense from an acoustical point of view. How do you respond to such praise?
LD: That’s very flattering to hear, of course! I know some of my elders were surprised by my career choice, and thought maybe I should have gone to work for NASA or Airbus. What can I say? I’m just passionate about speaker design and sound reproduction. I’m really into it, and very lucky to be doing the thing I enjoy. I still find it so very fascinating in all its challenges.
At the end of the day, the purpose of all this is the music. Music is such an important thing to humanity. Maybe I’m a frustrated artist, and should have actually been more serious about playing the guitar or something -- but I do the second best, which is to help people enjoy the performances of other artists. It provides a real service to make life more enjoyable. I’m still having fun and I’m still really interested. It connects to the primeval feel-good part of my brain!
. . . Peter Roth
For more information about Vivid Audio, please visit www.vividaudio.com.