Recorded sound has always been about creating a moment, and the tools to accomplish this only get more advanced as time goes on. Our ability to record and endlessly tweak audio in so many different ways these days is a perk of living in our era. (Who would have thought slogging through 2020 had its perks?) The tools at our disposal allow us to artificially create moments that seem as natural and realistic as possible— the proverbial “lie that tells the truth.”
Recording in a practical sense started in 1877 when Thomas Edison created the wax cylinder recorder, but ensuing decades and today’s audio tools pull us further and further away from what it originally started out as—a simple audio snapshot of a moment, replete with all the plusses and minuses that real life serves up, whether it’s that the trumpet player was on fire, the singer missed the high note, or that the drummer sped up and slowed down like, well, 2020.
Even in Edison’s time, people wanted to hear immersive audio, or at least audio from more than one source. Inventors explored what would become known as stereophonic sound as early as 1881 with the creation of the théâtrophone, a French subscription service that delivered live performances in simulated stereo over phone lines, though the performances weren’t recorded; the technology wasn’t there yet. For years, the earliest known stereo recording was created in the 1920s, but now an even earlier moment captured in stereo has been brought to light and preserved for the ages.
Indiana University’s Archives of Traditional Music in Bloomington, IN, recently digitized a number of wax cylinders from 1901 featuring Shanghai musicians performing traditional Chinese folk and opera music; when combined, they create the oldest stereo recordings known to exist. Ironically, they weren’t recorded with the intention of creating stereo sound. In fact, they weren’t meant to be heard by the public at all.
In 1901, the American Museum of Natural History (AMNH) in New York sent German anthropologist Berthold Laufer to China to research and obtain items for the museum. There, he recorded the musicians—not with an eye toward preserving their music for the ages, but rather as an archival ethnographic research exercise. Laufer used two cylinder recorders operating at the same time in order to capture the singers and the instruments separately, with each group placed slightly apart as they performed together. While there was inevitable audio bleed, with the vocal-oriented cylinder capturing some of the instruments’ ambient sound and vice versa, recording the vocals and instruments separately, albeit at the same time, allowed Laufer to more clearly transcribe each later on.
The AMNH gave thousands of wax cylinders to the IU archives in the 1940s; in return, the archive preserved and made them accessible as needed, according to Dr. Alan Burdette, director of the Archives of Traditional Music. “We’ve always known the cylinders were there [but] the stereo aspect of it has been a new discovery,” he told me. “We copied our cylinders to tape in the mid-1980s to provide access to patrons by way of copies. More recently, we got a grant from the National Endowment for the Humanities to digitize the cylinders themselves—and that opened up some new realizations of possibilities.”
Over a period of two years starting in 2017, the archive’s collection of 7,000 wax cylinders was digitized, capturing each two- to three-minute recording. Laufer’s cylinders stood out among the selections as unusually high-quality recordings considering the format and the era. While researching those recordings at the AMNH, Burdette discovered a letter dated Sept. 19, 1901, from the anthropologist to his superiors, explaining how he was using two cylinder machines simultaneously. Realizing they essentially had 119-year-old wax multi-tracks, the archive suddenly saw its digital recordings in a whole new light.
“This digitization process made certain things possible that would have been nearly impossible to do in the past, and that is getting cylinder recordings that have variable, hard-to-define speeds to match up,” said Burdette. “In some cases, getting those matched would have been really difficult in the open tape era. In the digital era, it’s still a real challenge, but it’s conceivable that you might actually be able to do it.”
IU media preservation specialist Patrick Feaster took on that challenge, and it was hardly a case of dropping files in Pro Tools and calling it a day. Finding which cylinders matched up required further research, with Feaster poring through both Laufer’s notes and the cryptic, often multi-language notation on the cylinder boxes, and then ultimately writing some code himself to digitally stabilize the recordings’ pitch. The result is a handful of proof-of-concept stereo recordings—which you can hear for yourself—along with an in-depth technical account of how Feaster pulled it off, all on his academic blog at www.bit.ly/317MKQN.
In the meantime, the archive is working to make all 400 of Laufer’s cylinders publicly available. Enticingly, Laufer appears to have recorded an entire Beijing opera in his accidental stereo format, stopping and starting musicians repeatedly so he could swap out cylinders, resulting in a whopping 72-cylinder series. “We made a trip to China last spring,” said Burdette, “and there’s a lot of excitement there among scholars about these recordings. They’re from the time [from which] a lot of the art forms are no longer practiced or have been lost—there has been quite a bit of cultural and political upheaval in China since 1901—so for a lot of scholars, it was truly remarkable to be able to hear things that were this old.”