Hey there, human — the robots need you! Vote for IEEE’s Robots Guide in the Webby Awards.

Close bar

Engineer Remasters Thomas Edison’s Musical Recordings

Marc Hildebrant finds the missing notes in mechanically recorded performances

3 min read

sketch of early sound recording with pianist
Hit it, Maestro: Early sound recording relied on the direct mechanical force of acoustic vibrations to etch patterns in cylinders and discs.
Photo: AP Photo

One of the best parts of working for IEEE Spectrum is the opportunity to discover intriguing tech projects that are pure labors of love, whether it’s building a 16-bit CPU out of discrete transistors or keeping a collection of vintage personal computers alive. So when I heard about Marc Hildebrant’s work on using digital techniques to restore mechanically recorded music to its full glory, I had to know more.

The earliest sound-recording and playback technology was pioneered by Thomas Edison in 1877. The system was purely mechanical: To record a performance, musicians arranged themselves around a recording horn, at the other end of which a diaphragm vibrated in accordance with the incoming sounds. A needle attached to the diaphragm etched a track in the recording medium—first cylinders, and later discs. Playback was simply a reversal of this process, with a needle vibrating a diaphragm at the base of a horn. In the 1920s, the recording switched to an electrical system using microphones. By the 1930s, electrical playback was also available.

Before and After

Original excerpt from Red Hot Mama by the Georgia Melodians (1924).

Processed excerpt from Red Hot Mama by the Georgia Melodians (1924).

Edison’s early mass-market catalog represents a cultural trove: Many famous musicians of the day made recordings. But these recordings have significant sound-quality problems. Most sound-restoration projects eliminate the obvious issues that plague decades-old analog storage media, such as scratches or hissing. But Hildebrant is going deeper and fixing some of the distortions introduced by the original mechanical recording process.

“That area has always been kind of abandoned by the restoration people.... Pretty much what’s available today are recordings where they just clean up the noise. They don’t really change any of the sound of the music,” says Hildebrant.

The distortion comes from the limited frequency response of the mechanical recording system—essentially many high and low frequencies simply couldn’t be captured, so even after being cleaned up, the recordings sound tinny. Hildebrant focused on restoring the missing frequencies below 300 hertz to fill out the bass notes. Trying to simply boost low frequencies below 300 Hz would serve only to amplify noise. Instead, Hildebrant realized that voices or instruments present in the medium frequencies would likely have had harmonics in the missing frequencies originally and so could be used to reconstruct the missing sound signals.

Using commercially available software from Diamond Cut Productions, which can generate subharmonics as part of a digital signal processing chain, Hildebrant was able to generate missing bass frequencies in old Edison recordings, as well as filter out high-frequency noise and make other adjustments.

Because the subharmonic filter will create a low-frequency sound for any strong signal in the midranges, it can produce unwanted frequencies—for example, adding too much bass to a singer’s voice—so the process depends to a large extent on Hildebrant tweaking the processing chain for each track, guided in part by his extensive experience with audio recording. The result is a digital file that sounds much more natural than the original recording. (A before-and-after sample of Hildebrant’s work is available below.)

Currently semiretired, Hildebrant says he has been working on this audio-rejuvenation project for about five years, after working on restorations of early electrically recorded music. “Like everything in engineering, you’re always working on improving,” he says. “An engineer’s life is never done.”

His interest in the recordings goes beyond the merely technical, however, and into the music itself. “It’s hard to find decent Edison music because he himself decided what should be played and what not. He didn’t like jazz music, so he hardly recorded any,” says Hildebrant. But “there’s a small window of time where he perfected the [mechanical] process, and this is the area I concentrate on. He didn’t really care anymore about who recorded, so the jazz people started recording. You’re hearing jazz played by young musicians with Benny Goodman and Tommy Dorsey, and all those people right at their prime when they just started. So I always thought if I could fix that music up, then I’d really have something unique.”

This article appears in the February 2017 print issue as “The Music Man.”

This article is for IEEE members only. Join IEEE to access our full archive.

Join the world’s largest professional organization devoted to engineering and applied sciences and get access to all of Spectrum’s articles, podcasts, and special reports. Learn more →

If you're already an IEEE member, please sign in to continue reading.

Membership includes:

  • Get unlimited access to IEEE Spectrum content
  • Follow your favorite topics to create a personalized feed of IEEE Spectrum content
  • Save Spectrum articles to read later
  • Network with other technology professionals
  • Establish a professional profile
  • Create a group to share and collaborate on projects
  • Discover IEEE events and activities
  • Join and participate in discussions