412 North Chester Road, Swarthmore, PA 19081 - 610.690.1703 [voice] - 610.328.6355 [fax]

Free shipping to the United States, Canada, and Mexico!



"Wizards of sound: through their art, transfer engineers bring us closer and closer to performances of the past"

by James Camner, Opera News, October 1996

The emergence of digital recording technology and the compact disc has held an unexpected bonus for historical vocal enthusiasts. The increased fidelity resulting from advanced recording techniques would have appeared to doom older recordings, particularly those of the pre-LP era. But contrary to all such expectations, even the earliest recordings sound better than ever when reengineered for compact disc. With the push-button ease of choosing individual tracks, this has made the digital medium ideal for conserving and exploring a wealth of historical performances.

From Caruso to Callas, there has been an explosion in historical vocal releases in the past decade. While many come from major labels -- EMI, BMG, Sony, Philips, DG -- eager to capitalize on the contents of their vaults, many more have appeared on enterprising small labels like Romophone, Pearl and VAI, which rely on the generous resources of enlightened collectors. In both instances, a handful of dedicated American sound engineers have produced some of the most significant results. "None of us are in this for the money," observes Ward Marston, a pioneer in transferring old 78s. "We're all in it for the love of singing, for the love of old records. Most of us are collectors. Danny [Hladik] is a collector, Mark [Obert-Thorn] is a collector. Seth [Winner] is a big tape collector."

A legend among sound engineers, Marston performs his technical wizardry in a well-appointed studio in his Philadelphia home. With skill, technology and an innate genius that is part absolute pitch and part intuition, he has blazed new territory in raising the transfer of 78s to an art form. By working to correct and eliminate the surface noises, clicking, clanking, blasting, distortion and pitch variations that shroud old 78s, to allow the original performance to come through, Marston approaches his goal. "I always try to make my 78 transfers sound like real music, rather than records. My secret is the combination of good sources and a good pair of ears.... CEDAR and NoNoise [sophisticated computer-driven technologies] are not the end of the world. When used with no ears, they can be a dangerous thing."

Marston, who is blind, showed an early gift for music, playing the piano at age four; by the time he was eight, he was fooling around with tape transfers of old records. While he disparages his earliest work, in fact his efforts for the International Piano Library and CBS in 1976 were superior to anything else being done at the time. For Pearl, Marston has engineered a complete Caruso set and a remarkable-sounding Louise Homer disc. For VAI, he achieved miracles with the historic 1912 Pathe Romeo et Juliette, a recording originally made with arcane Rube Goldberg-like technology that he likens to "two tin cans and a tight string." His deluxe series of complete collections for Romophone has only added to his reputation. Included in this group are Rosa Ponselle, Claudia Muzio, Pol Plan,con, Amelita Galli-Curci, Emmy Destinn, Emma Eames, Mary Garden and, most recently and spectacularly, Lucrezia Bori.

Though Bori was one of the Metropolitan Opera's most beloved sopranos, many of her records, played at 78 rpm, sound shrill and unpleasant, and they have played no small part in harming her reputation among collectors. In Marston's two compilations for Romophone, the soprano's famous warmth and communication have been restored. Marston explains the difficult, almost mystical art of pitching vocal records. "A collector I once knew said he could always tell when a singer was properly pitched, because when you find the right speed, there gets to be a halo around the voice. Something happens to the sound of the voice when you get it into the groove-- you sigh with relief, That's where it belongs.' The lowest speeds for Bori are in the session from 1921, I believe, of two recordings, 'When Love is Kind' and the Cosi aria. They play about 71.5. It's a mystery why they recorded these so slow, but few records actually play at 78. Most play between 75 and 77, and some change speed from the beginning to the end of the record."

Marston is also a master of the tricky business of splicing side joins, crucial for recordings that are longer than one side of a 78, including, of course, complete operas and symphonies. To explain the process, Marston cites Stokowski's recording of the Beethoven Seventh Symphony, in which each record begins at 78.6 and slows to 76 by the end, as typical of the problem. To correct this and make seamless joins, the pitch must be adjusted during playback. The engineer must then reequalize the sound and pick up the music from the next disc without missing a note.

Future projects include complete collections for Romophone of Ernestine Schumann-Heink, Emma Calve, Marcella Sembrich, Frances Alda, John McCormack, Giovanni Martinelli and Luisa Tetrazzini. Marston also talks of starting his own label, hoping to issue mid-priced editions of complete operas.

For Dan Hladik, who has done transfers of exceptional clarity for the Metropolitan Opera Guild's recording projects, the judicious use of filtering is an important characteristic of his work. "The real tragedy of a lot of transfers is that they actually remove parts of the alphabet, so that all the sibilants are missing. It's a bad mistake to remove brightness. You're going to affect the balance of the words."

Hladik points to his La Gioconda compilation in the Guild's survey, Great Operas at the Met. "It's only by accident that sibilants like sh' or th' can be recorded acoustically, but you can almost hear the s' in Emmy Destinn's Suicidio.' Even on electrics, if you don't judge your filtering carefully in the 6,000-9,000 Hz range, you're going to give singers a lisp. My transfers may be noisier, but I'm not going to give them speech impediments. Ten Cents,' not Ten Thenth'!" --referring to Ruth Etting's crystal-clear diction in "Ten Cents a Dance" from the MOO's Original Cast! 100 Years of American Musical Theater.

Based in Philadelphia, Mark Obert-Thorn is another notable engineer with several fine transfers to his credit, including a "Potted Ring" (Pearl) and a complete Beniamino Gigli Victor collection (Romophone). In his search for prime source material, he finds "Americans have a great natural advantage, in that some of the quietest shellac pressings of 78s ever made -- Victor Z' pressings from the mid-thirties Columbia Viva-Tonals and laminated Brunswick re-pressings of Polydor originals -- were manufactured here. By contrast, British HMV and English Decca are notoriously noisy. If there was no U.S. release of a particular recording, the experienced engineer will know what the next-best alternative is -- a German Electrola, French Disque Gramophone or Australian HMV pressing."

But finding the right source doesn't end there. There is the problem of dubbings. "Dubbings are re-recordings onto a new master to form a new stamper. They were made because the originals played too loud and wore prematurely on home systems. Every dubbing degrades the original sound, and the trick is to find an early issue that wasn't dubbed. Once the best record is found, the choice of a correct stylus, to hug the groove walls and minimize rumble, distortion and swish, is essential." Obert-Thorn's future projects include Acts I and II of Die Walbure with Lauritz Melchior and Lotte Lehmann, Kirsten Flagstad's Philadelphia recordings, the Aureliano Pertile Trovatore and the Piero Coppola Carmen.

Seth Winner enjoys some unique advantages in his work, since he is a sound preservations engineer at the New York Public Library, in the Rodgers and Hammerstein Archives of Recorded Sound, and oversees the preservation of sound recordings in the Toscanini Legacy collection there. At his own state-of-the-art Seth B. Winner Sound Studios, he's done several projects for Pearl, including a lovely Grace Moore disc, the landmark twelve-CD survey Music From the New York Stage, 1890-1920, and a fine Giuseppe De Luca set.

Winner's work with the Toscanini Legacy has included transferring, for BMG, the Maestro's Fidelio, Otello, Traviata and Boheme recordings. With John Taddei, a retired CBS engineer, Winner has made, for the library, revelatory tape transfers of recordings made by the Austrian Broadcasting Corporation on a machine called the Selenophone, which cut a variable-density soundtrack on 8mm nitrate-based film. Anyone who has heard the private-label recordings of Toscanini's Salzburg Die Zauberflote, Falstaff and Die Meistersinger or Bruno Walter's Don Giovanni and Nozze di Figaro that have circulated for years in dreadful sound quality (originating from lacquer dubs made in the 1940s) would be astonished by the full, rich sound and definition of these performances on Winner's tapes. It is fervently to be hoped that these priceless recordings will be made available to the public in state-of-the-art transfers.

One of the library's greatest treasures, a unique mint set of the legendary Columbia Grand Opera Series from 1902-03, the first important commercial opera recordings made in America, will be available in the new Sony "Masterworks Heritage" series, along with Bidu Sayao and Alexander Kipnis collections, all in superb transfers by Winner.

A frequent collaborator of both Winner and Marston, Jon Samuels is employed as a full-time sound engineer at BMG, transferring recordings by virtually every vocal star who ever worked at RCA, beginning with Enrico Caruso. While he has done his share of work with 78s, Samuels' BMG projects involve to a great extent the conservation and transfer of analog tape to compact disc, a process that has its own set of requirements. In this work, the transfer engineer has to find the proper playback curve -- the level at which the original recording was made -- either through written records or by trial and error. The latter is often the case, as, according to Samuels, "Until 1958, there is no source-by-source blueprint." Although tape speed is always standardized, locating an original, even at a record company, is not always possible. After establishing the playback curve, it becomes possible to manipulate the results in all sorts of ways, including "correcting" equalization and even adding reverb (artificial hall atmosphere), which Samuels does "reluctantly, if at all."

These American transfer engineers are well aware that technology, continuing to improve, will change the nature of their work in the near future. Winner, speculating on imminent breakthroughs, anticipates a 24-bit transfer rate that will improve the capacity to record low-level passages while correcting certain anomalies in the CEDAR and NoNoise systems. Ward Marston looks to the Internet as a possible alternative to CDs and expects advances inplaying 78rpm discs with laser technology, thereby eliminating stylus friction. What seems certain is that thanks to them and their colleagues, listeners can continue getting closer and closer to the great performances of the past. We are left with the happy thought that our beloved Caruso and Callas records can only get better and better.


next article >