Over 400,000 Rare 78rpm Records Are Being Digitized

Prism Sound’s Titan audio interfaces are playing a key role in helping US audio and moving image preservation company George Blood L.P. to complete a unique community project, which aims to create a digitized reference collection of 78rpm records.

Entitled The Great 78 Project, this extraordinary venture has already digitized 35,000 sides and ultimately aims to digitize in excess of 400,000 sides for preservation, discovery and research.

The Titans were chosen for The Great 78 Project because these interfaces are ideally suited to its streamlined workflow, which sees audio being recorded directly onto a computer. Company founder and preservation expert George Blood also cites their reliability and sonic excellence as reasons for choosing the units – plus the high reliability of Prism Sound Hardware and the fact that he has been a Prism Sound user for more than 20 years and places great store in the relationship he has with the company.

“We have had a long, very happy history of working with Prism Sound and the technical support is fabulous,” he says. “That relationship matters a lot. We are also receiving very positive feedback from our peers in the trade for the quality of our transfers – even at the highest production scale achieved by anyone worldwide. The reliability of the Titans helps us keep the price down while delivering on the sonic fidelity that preservation requires and meets the demands of our clients and engineers.”

The Great 78 Project is currently digitizing nearly 400 sides a day at George Blood L.P.’s Philadelphia studios, which operate 16 hours a day, six days a week. Every month an HDD containing 5-7000 sides is sent to the Internet Archive (https://archive.org/details/georgeblood), which uploads them onto its servers where they can be accessed by researchers in a format that allows them to be manipulated and studied without harming the original, very delicate shellac artifacts.

To record audio from the 78rpm discs, George Blood has devised a custom turntable with four tonearms. “We capture both groove walls of four different stylus sizes, with and without EQ for each side,” he explains, “That’s 16 channels in total, which are recorded using two Prism Sound Titans. In this way the user can decide the stylus size they like best. They can even mix and match – size A for the outside of the discs, size B for the inside of the discs, size C for a damaged part of the discs, and size D for some instrument or four bar phrase.”

Apart from the Titan units, George Blood L.P. is also using six Prism Sound ADA-8XR multichannel converters, an AD-2 converter and a 2024 BitSplitter in its music studios where staff have worked on over 45 GRAMMY® nominated projects over the years.

“We have finally had to replace power supply capacitors on our 15 year old ADA-8XRs due to old age but in general Prism Sound gear just works, year in and year out,” Blood says. “I’m happy to say our Titan units are holding up as well as our ADA-8XRs have. What’s more, everyone is happy with the sound.”

The story behind The Great 78 Project goes back to 2013 when George Blood L.P was asked to digitize 10,000 sides for the US Library of Congress, taking in two collections – factory reference discs from the Thomas Edison National Historical Park and un-played or rarely played Victor discs from the Eldridge Johnson Victrola Museum pf the Delaware Division of Historical and Cultural Affairs.

“78rpm discs are not standardized for speed, stylus size, or equalization,” Blood explains. “Normally digitizing 78s entails discovering not only the correct stylus size for that label at that time, but also considers how a different stylus size can compensate for years of wear. The wear may be from normal use, playback with the incorrect stylus, abuse, or other ravages of time.”

Fortunately, the Library of Congress collections were in pristine condition and this allowed George Blood to select stylus size without consideration of wear. Further work on the Library of Congress’ OKeh label collection added to his knowledge because these discs were collected “in the wild”, so condition was a consideration in the stylus size selection.

“At about the same time we also digitized a large collection of random discs for New York University, and that served as the control for the study,” Blood explains. “All of this work led to the publication of a paper (by Indiana University) in which we analyzed our production data to determine the proper stylus size for these labels and the optimum playback speed.”

It was purely by chance that, after giving an early version of his paper at the Association for Recorded Sound Collections, Blood found himself in a hallway with Brewster Kahle, founder and Digital Librarian at the Internet Archive, having a discussion on preserving 78s ‘at scale’. This discussion evolved into The Great 78 Project, which is funded by the Kahle/Austin Foundation – a family foundation that selectively funds arts projects, especially the Internet Archive.

“From about 1898 to the 1950s, an estimated three million sides have been made on 78rpm discs,” Blood explains. “While the commercially viable recordings have been restored or remastered onto LP’s or CD, there is still research value in the artifacts and usage evidence in the often rare 78rpm discs and recordings. So far, the Internet Archive has selected over 20 collections for physical and digital preservation. By preserving the often very prominent surface noise and imperfections, and including files generated by different sizes and shapes of stylus, it is possible to facilitate many kinds of analysis.”

The Great 78 Project has plenty of work still to do and the organisers are keen to encourage members of the community to get involved. Donations are curated by B George, founder of the Archives of Contemporary Music in New York. Anyone can digitize 78s and add them to the Internet Archive and the public can also contribute metadata (whether discographic, or anecdotal), make blog postings about recordings, regroup/curate them into collections and add other materials such as biographical, bibliographic, advertisements, etc. The Project is especially eager for researchers to explore the collection and points out that, with 16 versions of every side, there is plenty of scope for a thesis. For example, can someone build a better noise reduction algorithm? Or what about exploring performance practice or social mores and norms from the era – or studying dialects and the evolution of musical styles?

If nothing else, the public are encouraged to check out this amazing resource because there is so much of interest to be discovered among these thousands of newly digitized tracks.