"Fierce on the battlefield, feared by his opponents, revered by his friends, his name is legendary and synonymous with the dominant artistic and creative presence of the wider Mediterranean."


Eh, not in the case of Bernard Salabert. He is somewhat of a legend. Alas, it's of the underground, left-field kind. He perhaps could become the dominant presence in the still niche category of high-efficiency wideband ultra-performance loudspeaker drive units - if just more people knew about them. His homebrew pastis and black handrolled cigarettes are fierce. He has no enemies and doesn't believe in warfare. I'm certain he is well beloved by his friends. But as he tells the story, our modern times (the speaker establishment at large, hifi in general, the majority of press and online opinion) are
anything but kindly predisposed to the way things used to -- and should continue to -- be made. Going back 30 years to the beginnings of high fidelity in France, the days of Hiraga, Piel, Mahul and others when he ran a hifi store in Montpellier, he perceives a concerted effort to strategically eliminate the traditional 1930's craft of vintage drivers in the Western Electric/Altec vein for inefficient, multi-way speakers in small boxes which could be sold to the unsuspecting masses under the added guise of hi-tech driver materials and implied progress.

"I saw all the good drivers disappear from the catalogues at about the same time. Does that strike you as coincidence?" When Supratek's second owner reached retirement, Salabert approached him with a buyout offer but was outbid by a team which managed to crash the company within two years (it was revived later). Seeing the last two little companies making high-efficiency wideband drivers lose traction in the market; and knowing who had wound the voice coils for Supratek's founder, something the second owner no longer knew; Salabert made it his mission to step into the created void. While the last reissued Altecs disappeared and with it, certain knowledge acquired in the preceding decades, Bernard Salabert returned to making "good units". It meant articulated paper cones sourced from the old-timers who had built the original drivers for Audax, Siare, Triangle and other legendary old French transducers before the offshore misery began. It meant Talc-impregnated latex suspensions. Alnico 5 Alcomax center rather than circumferential magnets. Then he added ingredients not used before: Bronze baskets and massive dispersion rings. Vellum formers instead of Kapton. And voice coils wound on his own machine, with 0.125mm silver wire wrapped in 15 micron continuous silk. "Our unique bronze mounting rings minimize border effects with their oval profile as well as acoustic emissions from the outer part of the suspension which is in phase opposition.
More, it is a way to add mass and rigidity for the mobile equipment to get the most stable reference possible. For that you need to be very rigid and achieve the greatest possible weight ratio between moving and stationary mass. Bronze further is amagnetic and beautiful. When you listen to music at a realistic level (80 to 90dB), 92% of the movements with our efficient cones are less than 4 micrometers. When the whole cone can move 40 to 400 microns however as it does on a lot of commercial speakers, there is huge intermodulation and absorption to pay. People who judge a drive unit's action with their eye (which most of them do) are utterly fooled. The true quality of a drive unit lives in those 92% which cannot be seen without a very large microscope. It is not in the big slow movements made to try to reproduce the very low frequencies that one discovers the true quality of a driver."


"About good drive units in general, it's really very simple. Physics tell us that we must have very light moving mass since mass operates as the square of the inertial position (in fact square + 1) so you can never compensate by magnetic field strength or amplifier power. This becomes even more important when you listen to music. Music, on average, is a mix of 30% sinusoid and 70% very steep short signals (impulses, transients and such). If a sinusoidal wave is easily reproduced, transient spikes are far more difficult to render properly. For that the mobile equipment (cone + spider + leads + voice coil) should have an infinite slew rate i.e. zero mass. That's impossible of course but you must come as close as possible to 0. You need a very light cone that is rigid enough and features progressive fracturization in an hyperbolic-exponential profile. Only paper with long and short fibers can accomplish that, plush a small-diameter voice coil where one layer is inside and one outside of the former to transmit the movements of the coil with the fewest possible losses (I use impregnated vellum). It is simply in respect to basic PHYsical laws that I made these choices, not to return to the roots per se. But surely it is not a random coincidence that people from the 30s to 70s made drivers like this. They were not stupid as it seems a lot of people today like to believe. These physical laws haven't changed. Obeying them once again not only makes for better drive units, it is also a very nice way to keep alive the knowledge about these technical matters."

PHY factory in the French countryside
In the mid 90s, Salabert formed PHY-HP. That's pronounced "fee", with HP short-hand for haute parleur or loudspeakers. Today, a meager four small commercial companies use his drivers in their products: Auditorium 23 in Germany; Musical Affairs in Holland; Ocellia in France; and Tonian Labs in the US. With the volumes involved, PHY barely survives. Hence inquiries for a 15-inch unit have thus far fallen on deaf ears. Salabert doesn't have the capital to do it. His present transducer offerings include 8" and 12" widebanders, a bronze-encased high-impedance piezo tweeter and the 12-incher with the tweeter as an integral coaxial unit, all with copper or silver voice coils for eight models total. There are also PHY cables, connectors and a 15-inch tone arm.


The problem of deeper market penetration is compounded by the fact that PHY drivers cannot simply be stuck into cabinets optimized for modern drivers. Speaker manufacturers willing to step into the vintage void as it were seem few, presumably because having never heard superior examples, there's little attraction for them to pursue this route. Alas, the last two or three years have seen quite a renaissance in this sector, with solid indication that this field is gaining more widespread interest and acceptance again. Think Feastrex, Omega, Rethm, WLM and Zu for just a few examples.

Salabert's listening room and the wall

Salabert's extreme sound room affords him and visitors an unusual opportunity to test the real prowess of his drivers in a unique way. A slightly concave infinite baffle aka The Wall mounts his drivers directly on bronze plates, then sand-filled concrete. In a capacious cave built from literal tons of special concrete, deep in the country side for exceptionally low ambient noise, one can hear driver action liberated from any associated cabinet colorations. This room is a magnificent assessment tool and real-world laboratory that doesn't require modeling software to produce results.

Not your usual inwalls

One can also hear how uncompressed low frequencies ridicule even absolutely massive damping. As I heard for myself, such sound waves propagate unhindered through the walls of a large space with non-parallel walls to literally vibrate the concrete (the one-foot thick ceiling alone contains 12 tons of sand). Any illusions of successfully eliminating a driver's rear wave with a speaker cabinet as is conventionally attempted are profoundly demolished in just a few short minutes of a Salabert demonstration.

The back of the sound cave

Anyone not sufficiently impressed at that point about the seriousness of Salabert's intentions or achievements can enter the anechoic chamber next door where he conducts measurements down to 40Hz.

No reverberating sounds


But there's more to this story. Dr. Pierre Johannet's research on MDI or Micro Discharge Interface distortion conducted at the French National Electricity institute quite independent from audio applications is, according to Salabert, "the only meaningful new discovery relevant to audio over the last 40 years". Simplified, MDI deals with extremely steep high amplitude ultrasonic spikes which, in audio, become modified to resemble the sine wave component of the signal sufficiently to cause audible distortion by intermodulation. Additionally, MDI creates positive Langevin ions which change the air to less effectively support sound propagation. "Whether they knew why or not, the old Japanese who had water fountains in their listening rooms neutralized certain MDI side effects instinctively." Today, a negative ion generator could do the same of course but it won't eliminate the causes, just minimize certain symptoms. To attack the causes requires, according to Bernard and Physics, a stubborn refusal to employ any synthetic materials. "They are also bad for your health."

The first-ever commercial cabinets to employ PHY drivers, finished French Country style - the Auditorium 23 Provence

The war against MDI can't stop at the transducers. It must eventually encompass the entire audio system front to back. Accordingly, Salabert has created his own connectors -- "the world's best nobody knows about" -- which combine paper insulation (and hydrophonic cotton sleeves on the cables) with 40-micron pure silver plating to insure a silver contact with other commercial connectors where 0.2 - 0.4 microns of gold or 0.4 to 4 microns of silver are applied over a thick shiny nickel layer which, after the unavoidable scratches of insertion and removal, becomes the de facto contact material - nickel on nickel, a very bad electrical connection. PHY's RCA plug contacts for the pin and barrel sport V cutouts to make solid connections with the conductors even without solder and the pin itself is profiled with hundreds of micro pyramids to increase the contact surface on insertion.


For his own use, Salabert uses stripped-down electronics in wooden casings where capacitors are denuded and as many synthetics removed as practically feasible. While of modest origins, these modifications rather elevate the gear's performance beyond original sales costs. These components have become Bernard's listening tools for his ongoing driver design process.

The resident system

Hence, plenty of wood and bronze are in evidence chez PHY. If you want plastic, go to Las Vegas.

Bernard's CD player in a box

A driver's rear wave is 50% of its output and a vital part of the signal. In Salabert's view shared by those who embrace his drivers, it must be artfully employed rather than killed off. Eliminating it has direct implications to the remaining half. How to exactly achieve full radiation without suffering the usual open-baffle out-of-phase bass cancellation is the challenge. For a delightfully simple demonstration, Salabert placed a wind-up music box mechanism on a raw sheet of MDF, then low-grade ply. Needless to say, the MDF destroyed both tone and signal amplitude to clearly sound inferior and quieter to the ply when Bernard began turning the crank. In free air, the sound was attenuated even further and downright washed out. Within this simple experiment hides the foundation why Ocellia for example employs PHY drivers in thin-walled Spruce Ply cabinets (Spruce is commonly used in guitar bodies and regarded as a tone wood). Ocellia couples its drivers very rigidly to the tone wood carcasses for maximum energy transfer and quick release. Since this is about tone, the idea is not to deaden it with the box but help the box release unwanted resonance quickly and effectively.


"The difficulty is the reversal of phase of the rear wave, over a broad band. I didn't think it possible but when you listen to well-implemented such speakers next to my wall, they will play nearly 3dB louder because of it."

As with everything, the devil's in the details. While apparently simple and basic to the casual eye, there's always more than one reason for any given aspect of any of Salabert's designs once you ask him. And he never pretends to have invented anything. "I just didn't forget. I learned from the past. I simply prevented some of it from disappearing forever."


For example, his voice coils are wound on both the inside and outside of his Vellum formers. To track down the perfect former material meant a return to the past, investigating what was used in the best-sounding drivers then and tracking down available equivalents today. Salabert merely grins at the mention of Kapton. He leaves that hi-tech crap to those who don't know better than to design on spec rather than comparative listening.


"Specs are only based on sine waves, i.e. a mere 30% of what constitutes a complete and complex music signal. We have no simple affordable tool to look at the other 70%. That is why measurements alone give so few indications. Even if you can read and understand measurements, you'll see how incompletely they correlate to what we hear. For example, you can find two speakers with the same measurements but when you listen, they are very different. One has to be very careful of course, knowing full well how easy it is to get oneself fooled. For example, you hear bass made by a very little speaker which cannot and does not produce it yet your brain recreates the missing fundamental by using the harmonics. An easy way to recognize this is that if there is true low bass, your pants, chair and stomach will vibrate. If not, it's fake bass. Or, listen under the same conditions and to the same record before and after a meal. There will be changes. Which rendition is right? Or consider the angle made by your head with the horizontal line. Depending on whether you look straight ahead or tuck in your chin, you can lose up to 10dB in the highs (that's very easy to prove with pink noise) - 10dB is nearly 8 times louder in the treble. And a lot of people will inspect a frequency curve for a half-dB variation. Give it a try. Looking straight ahead while listening will only give you medium treble output. Tuck your chin to the chest and you'll get uncut treble."


Naturally, Salabert is opinionated. That's part of his makeup, especially when so much about audio is so ill-informed about the basics. He thinks little of the popular press which kowtows to commercial trends and understands little of that which was lost to retain a proper context for many of today's so-called advances.


Salabert is a salt of the earth type who's paid his dues in the honest currency of sacrifices. He lives exceptionally modest to pursue his craft uncontaminated by corporate compromises.


When he first appeared, the press celebrated him. But since he doesn't change his approach nor churns out new models, he's no longer news. Diamond tweeters and Beryllium mids make for far sexier copy. Meanwhile Salabert subsists on low-volume orders hoping "to be discovered" in greater measure so he can finance further R&D and perhaps launch that 15-inch dual-concentric certain folks have been itching to see him make.

Bronze tweeter housing
Salabert' hand-crafted voice coil winder accurate down into the micron range
Living quarters

"I'm very stupid. Having done this for 12 years, I should have more to show for it than I have." Salabert refers to his modest environs, his perennial evasion of impending financial doom. It's a not uncommon scenario bedeviling uncompromising characters. Do what you love, tighten the belt, KBO. Keep buggering on as Winston Churchill is reported to have put it.

PHY tone arm with, in this case, Shindo cartridge

"Here in France, we love food. We can sit for an hour discussing a recipe, whether to put the salt in before, during or after the boil." It's emblematic of all obsession. And make no mistake, Bernard Salabert is obsessed with proving that not only does the vintage high-efficiency wideband driver have ongoing relevance, it is the only way to properly hear tone and dynamics. Once you experience Samuel Furon's Ocellia Silver Grandis with the 12-inch PHY dual-concentric as I did for example, how those thin-walled cabinets seem to amplify and let go of tone, you could be humbled into recognizing this to be a straight fact rather than crooked braggadocio.


"Naturally, a driver like our 12-incher which is linear to 8kHz must be properly articulated to where its center acts independent from the edge for high frequencies. This can't be a single or sharp demarcation but rather, must be a gradual action. This is achieved with various cone thicknesses, how the paper pulp is successfully layered onto the molds, from what direction. It's the artisanal approach you can't learn from books or online."


"To convey the tiniest of signal nuances, the voice coil windings must be supremely accurate. When the present owners of Supravox stumbled over a forgotten pair of original drivers in some dusty corner of their plant, they compared them to current production. Next thing I know, I get a call. Would I wind their voice coils for them? I'd do them in copper for them but not silver. Silver remains exclusive to PHY and I'm the only one in the world doing it."


Nothing like turning down a paying gig if it involves having to compete with Chinese labor rates? Quality costs money and quality backed by exclusive know-how even more. PHY drivers aren't cheap. But then, you only need one per side.

Silver and copper VCs for the 8" and 12" PHY drivers
Motor assembly, Alnico 5 core to extreme right


"The stereo illusion relies on
extreme accuracy. We match our drivers to within 1Hz of resonant frequency, 0.1g of moving mass as well as 0.1 ohm of DC resistance. Call me with a serial number and I can read you back the original measurements of that driver. It takes more time but is important to doing things the proper way. That's also why we give a life-time warranty to the first owner. Did I mention that we perform very serious burn-in of all units prior to packing?"

Bernard's assembly accomplice to left, Samuel Furon of Occelia to right

"Done properly, high efficiency, wide bandwidth and good performance are interrelated and synonymous." Don't get Salabert started on the status quo. He has little patience for it. He watched the emergence of the French underground audiophile press, the subsequent evolution of the establishment and the onset of disdain for past accomplishments. He is a living link to the past many no longer know about. Once retro pioneers like Salabert die out or close shop, current generations will lose context for the larger picture. Based on what I heard at PHY and Ocellia during my recent visit to France, that would be beaucoup merde indeed.

Hillside view from PHY

For now, PHY is alive and kicking. If more commercial makers visit Bédarieux to hear what a single driver with an auxiliary tweeter brought in at 10kHz and 40dB/octave can do in a simple infinite baffle, there's even the off chance that we'll see Salabert develop that dual-concentric monster he's been dreaming about.

PHY office

There's junk food, instant gratification, the wholesale eradication of species. In the commercial sector, conglomerates displace the single proprietor outfits. It's all in the name of progress. Once our outer environment is as toxic as our own bodies are becoming, with cancers and heart failures on the rise, our plant and animal kingdoms depleted of variety... then we'll perhaps begin to appreciate what we sold out for. In the meantime, audio has its own parallels. How we treat our resources determines our future. "There was a reason Western Electric enjoyed the renown and success they did in their heyday. They had 1200 of the brightest brains working for them."


History has much to teach us. New isn't always better. PHY-HP reminds us of such matters and offers alternatives as long as there's enough people caring enough to support this tiny outfit and its colorful crusty core character called Bernard Salabert.

"Much about modern hifi is bullshit. The basics are firmly rooted in physical laws. Check out an old military radio box, the kind soldiers used to wear on their backs. You'll learn a lot. Here, have another glass of pastis."
It's hard to imagine today that Bernard ever sang solo soprano in a chorus. Well, that was before his voice dropped. He did continue as solo bass afterwards. He also played piano and pipe organ. But that's the past. The present -- say sometime in the first quarter of 2008 -- will see something else than his voice drop: price. Salabert is preparing to launch the present 8-incher in a ferrite motor version with stamped steel basket to attenuate the retail price below 200 euros per unit. Everything else -- cone, suspension, voice coil, efficiency -- will remain identical. "There's a simple trick not practiced elsewhere that produces astonishingly good results from ferrite magnets. This saves a lot of money over my bronze/Alnico drivers and will broaden the appeal of our PHY products."


That's terrific news both for DIYers and manufacturers who already knew about PHY but had written their products off as too expensive for their applications. It's the same PHY for a lower fee. That even rhymes. How about zat?
PHY-HP's website