Tim de Paravicini: Dialing-in the Original Experience

Meeting Englishman Tim de Paravicini for the first time, you start to wonder if your mind has slipped a gear, whether premature brain fade has cut in. The conversation seems not only to be racing by unexpectedly quickly, but also subjects you hadn't even realized were subjects are being examined in knowledgeable depth. It was at the end of the 1970s that I bumped into Tim at a trade show in the UK; having wanted to ask his opinion of tube-amp design, knowing that the gangling, wispy-bearded, Nigeria-born, one-time resident of South Africa and Japan, ex-Lux engineer (footnote 1) had cast a magic wand over the Michaelson & Austin product line, I found myself instead being treated to an exposition of color phosphor problems in TV monitors. For Tim is a true polymath, his mind seemingly capable of running at high speed along several sets of tracks simultaneously.

But it is in the areas of tube-amp design and transformer winding that he has netted a worldwide reputation. Dick Olsher and Jerry Novetsky interviewed Tim on these subjects for Stereophile back in 1984 (footnote 2) his Esoteric Audio Research (EAR) amplifiers are sought after, particularly in Japan, for their solid-state-esque bass performance; we bought a pair of his natural-sounding tube microphones and an EAR tube mike preamp for use on Stereophile's recording projects; and his HEAD step-up transformer is still remembered by many audiophiles as one of the most musical. These days Tim has been turning his attention to professional audio, in particular the restoration of classic tube tape recorders and analog disc cutting. It was Tim, in fact, who cut the lacquers for Stereophile's Poem album at John Dent's Exchange Studio in London, and he has done excellent work for both Water Lily Acoustics and Chesky Records in the US, and Island Records in the UK. I had lunch with Tim at the 1990 Consumer Electronics Show and started by asking how he'd gotten involved in the high-end industry's sharp end:

Tim de Paravicini: It goes back 20 years, when I was involved with rock and roll bands. Having worked on all aspects of the studio industry, when I ran into John Dent, who was then the cutting engineer for Island, five years ago or so, we got talking and I said, having thought about it over the years, I didn't like the sonic signatures that were being cut onto records. Knowing what master tapes sounded like, I didn't like what the finished products sounded like. John Dent gave me the opportunity to prove that I could design electronics for a cutting system that got rid of the sonic signatures of the cutter head, that I could make a transcription system that had the minimum audio degradation between master tape and playback.

The biggest impact on the sound is the way the motional feedback on the cutter head works. A typical cutter head is a mechanical system with mechanical colorations, the end result being that there is a mechanical "clunkiness," for want of a better description, that I could hear on records. And because it is a mechanical system, the ability to capture detail on the record is also somewhat missing.

Once he'd got agreement from Island (footnote 3), John Dent gave me the funding to go ahead with the project. I spent some three months designing a system that would deal with these problems. The first cutting head I used was a Western Electric, but since, whether they're large or small, they're all basically similar—they all have a pair of drive coils coupled to an amplifier to move the chisel and a pair of feedback coils—the exercise is roughly the same in all cases. Each one, however, has a different sonic signature because of the size and the way it's built.

John Atkinson: And, of course, they're not intrinsically flat in frequency response.

TdP: No, traditionally, stereo cutter heads are designed with a resonance at 1kHz which is used to take care of part of the RIAA net curve; ie, the fact that the cutter head rolls off at 6dB/octave below 1kHz takes care of the 318-microsecond time constant of the RIAA curve. (Ortofon take a slightly different approach to the RIAA correction; the Ortofon cutter head designs are around 2kHz.) So you only need a bit of bass correction for the 50-cycle turnover point. But on the opposite side of 1kHz, you then have to boost massively to put in the RIAA pre-emphasis. Now the cutterhead falls off by some 30 or 40dB by 20kHz, so you've got to actually boost the drive to the cutter head by about 50dB at 20k...

The very old electrical mono cutter heads, such as the Grampian, for 78rpm cutting in the electrical domain rather than the acoustic, were designed in a different way. The original approach was to design a cutter head which had an electrical resonance of somewhere between 10 and 20kHz...78s were therefore cut with what was effectively a flat, constant-velocity characteristic. So driving these old primitive mono cutter heads with an ordinary 10W tube amp, with no correction whatsoever, gave them an approximately ideal characteristic for replay...While they had high-frequency coloration, they actually did a very good job in the midband. But when Neumann came in with motional feedback, they had to design the head a different way...the resonance was brought right downband and then you had what would be called the second-order resonances, which were typically around 10kHz.

JA: The combination would give you this "clunky" sound...

TdP: Yes. What I would call a "quacky" sound. What I specifically did different was how I read the motional feedback signal, how I read the sonic signature of the head, and how I negated it. All my research was done by the simplistic method of bonding a cartridge to a cutter head rather than actually cutting records, in order to cut down time and to avoid wasting lacquers. And immediately I could recognize all the colorations because I was able to treat this as a real-time, line-in/line-out system. The end result was a system that degraded as minimally as possible: ie, the level of comparison, one could say, is that with all my electronics, the replay deck and cartridge, and the replay phono equalization, the degradation is about equivalent to that of putting audio through one 741 op-amp chip. That's my point of reference.

JA: But that sounds terrible! One antique op-amp chip?

TdP: But no other retrieval system can even approach that...no other recording system of any description can come even close to that minimum standard. I'm just trying to put it into context; that to get that far is no mean achievement for a complete system, including a needle, a lump of rock, roughing around in a groove.

JA: Were the cutter heads modified mechanically? Or is it all done electronically?

TdP: They were all electronically fixed. If you do it right, you can make it work. Because the whole point had to be that it had to work with a stock system. It's no good having a system that is not compatible because day-to-day commercial running of the studio means you have to have serviceability and reliability. I have done the same thing with both Neumann and Ortofon heads, and the end product, the record, nobody knows which records were cut with the old Western Electric or which were cut with an Ortofon, or the Neumann. Nobody knows. And I'm not going to tell them which ones are which.

JA: When the studio tried your system out, what was the reaction of the artists?

TdP: Some of John's better clients immediately heard this quality. People like Jim Capaldi's manager, they were aware of it. For one Jim Capaldi record, they were that impressed that, rather than sending copy masters around the world, John had to originate the lacquers for the world production of the record. They wanted all the records pressed to have the same quality. With others, there was a degree of, "It doesn't sound like a lump of vinyl." Their expectations were screwed up, because they'd produce a master tape in the studio, and because they know that they're not going to get the same thing on vinyl, they tend to be artistically pumped up, making something up for the end sound.

JA: My own experience when I was involved in making records was that when the final two-track master was mixed, it never sounded like the way we wanted the record to sound. Only when that same tape was transferred to disc, did it sound right.

TdP: That is a problem with some people. But Dire Straits and one or two other bands do try to produce what I call very pure, very clean sounds.

JA: And then you cut the groove with as clean a signal path as you can.

TdP: Yes, I want to retain the master tape in toto. I want the person out there to experience the same thing that I'm experiencing.

We talked for a while about the problems in obtaining a consistent supply of lacquers, the acetate-covered aluminum blanks on which the cutting engineer enscribes the grooves, and about how lacquers from different companies produce different sounds when the LP is finally pressed from the resultant metalwork. And the metalwork itself presents significant problems. Tim now uses a company in Leicester, England, who've recently installed a new plant:

TdP: Getting good metalwork is a problem. Unfortunately one of the best metal-plating plants in America, Allied, decided to close. EMI in England lost their quality many, many years ago. You see, the problem is with all these processes, because it's a black art it's really the skill of the people involved. Now, the skilled guys at EMI were already pensioned off and buried; the new guys who replaced them hadn't acquired those subtle skills...so those skills have died. There are very, very few places in the world where these traditional black-art skills are maintained.

JA: But surely the basis of economic development in the modern world is that men with skills be replaced by machines that can be operated by relatively unskilled labor? A record company can't stay financially viable if it has to depend on a small number of skilled people in order for its products to be produced. That's one of the reasons why machine-produced CD has been so attractive to these companies—it's a real mass-production medium that doesn't need arcane and archaic skills for it to be produced at an acceptable and consistent level of quality once the capital outlay on the plant has been made.

TdP: That is the whole point. And the point is that there are intangible qualities. If you take a parallel, Rolls-Royce use craftsmen for their cars; no amount of computers and robots can build a car with that indefinable property that a Rolls-Royce, or a Jaguar, or a Mercedes has. The Japanese, for example, are trying to build fine motorcars that will do everything correctly, but they just don't have that indefinable property that is an art form. You can't become a Rolls-Royce overnight. You can't make a machine do a Picasso. So it depends what you want. Are you after something that is a good art form—you want to do nice music?—or are we going to have it computerized? Instead of buying records, you could just buy a program that plays your MIDI [synthesizer] at home, and you would have electronically recreated music in your home. But it would be sterile; it would be the same as having a speech recognition system that then tells the other end to regenerate a facsimile of your speech, instead of transmitting your voice down the telephone. We are trying to dehumanize this real-time life experience...

JA: And then you're no longer involved with quality.

TdP: Exactly. And this comes down to a commercial decision rather than an artistic decision. The same way that television or FM radio [operates] on a level that satisfies the bulk of people rather than all the people. I personally prefer to have the best. You can either buy cheap pianos or you can buy Steinways. You can either buy a Toyota or an Aston-Martin.

JA: To return to records...

TdP: There was a recent letter from Bob Ludwig in The Absolute Sound. [Issue 61, p.24, about the fact that the one-step process of which TdP is an advocate and which was used for the remastered version of Stereophile's Poem LP is less accurate than the traditional three-step, due to the fact that the freshly cut lacquer has to be sprayed with silver before it can be electroplated to make a stamper, and that that silver layer remains in position on the stamper]. Well, the silvering left on the nickel stamper is literally only of the order of a molecule thick. And that is worn off in the first four stampings or so. Because it is only a molecule thick, it doesn't actually have any tangible degradation because it will still be pulled off, when they plate the interstage, second, and third plate to the final stamper. And the second problem is because of thermal cycling, the fact that the nickel plate, when they grow the second and third generations in the tank, changes temperature. You have a bimetallic action going on. And so you do get problems with the three-step that the one-step doesn't have. Another problem is that no metal deposits absolutely uniformly. So you get grain boundaries that produce a higher noise on the third generation that the first generation doesn't have.

JA: So you feel the one-step process, even with the silver temporarily left on the stamper, is still the more true representation of what was cut into the lacquer.

TdP: It's very good, yes, because this new company [in England], their nickel plating is so good that they're actually doing laser hologram credit-card stampers by exactly the same process. That is molecular-level information. The same as the record.

JA: I think that a lot of people don't realize that the information on an LP can be an order of magnitude smaller than the pits on a CD.

Footnote 1: Now married to a Japanese wife, Oliva, Tim is one of the few Western engineers to have been employed in a creative role by a Japanese company. The classic Lux 3045 tube power amplifier is a TdP design, as were the monstrous M6000 and M4000 power amplifiers, two of the first Japanese transistor muscle amps.

Footnote 2: In the "Almost All-Tube Issue," Vol.7 No.3, May 1984.

Footnote 3: About three years ago, John Dent decided to take the gamble and go it alone. He and his partner, Island's second cutting engineer, hocked their houses, rented a premise in Camden, and called their cutting studio The Exchange.