The Sound of Silence in the National Library

The Library of Congress occupies three massive and ornate buildings in the center of Washington, D.C. But those edifices house just part of the collection, which spans hundreds of miles of shelves across many less-interesting buildings, and extends to media beyond books.

To find the heart of the nation's audiovisual memory, I took a lovely drive in October along ever smaller highways heading southwest from Washington, D.C., to Culpeper, Virginia, where sound recordings, films, and video reside in temperature-controlled vaults beneath Mount Pony.

Passing historical sites like Manassas (where Bull Run is located) , and watching the landscape shift rapidly from government buildings and commercial high rises to strip malls to farms and antique stores, it felt as if I traveled through time as well as distance on the 75-mile trip.

But the library's Culpeper facility is firmly rooted in the 21st century, and its existence owes much to the latter half of the 20th. While the focus is on what's buried inside, it's hard to ignore the beauty of the setting, its landscaping, and the building's architecture; it's the best use of concrete that I've ever seen in interior design, and I say that completely unironically.

Bursting into flame

The Library of Congress Packard Campus for Audio-Visual Conservation has a mouthful of a name, and sits in a seemingly peculiar spot, perched on a hill above the quaint small town of Culpeper, which is chock full of shops and mildly trendy restaurants.

Gene DeAnna, the chief of recorded sound collections for the entire library, explained during a visit in October that the building's underground storage area had once belonged to the Federal Reserve, and stored enough currency to restart a cash economy east of the Mississippi River in the event of nuclear war or other massive disasters.

It became superannuated, and in 1988, the money was removed. The facility was put up for sale in 1997. David Woodley Packard, the son of HP founder Dave Packard and his wife Lucile, was instrumental in arranging for the private purchase of the site. The bunker was transformed through $155 million provided by the David and Lucile Packard Foundation and the Packard Humanities Institute, the latter of which the younger Packard heads.

Congress also allocated $82 million, but the site was under Packard's supervision, which explains why it's gorgeous. The government's brief for functional architecture doesn't include beauty (and sometimes even fails the functional part) lest officials are castigated for wasting money. Packard had a handsome building constructed for office and conservation functions, as well as upgrading and expanding storage to the precise needs of the LOC, before handing it over to the Librarian of Congress.

In a film about the facility, Packard makes this extraordinarily humble statement, that's also absolutely true: "The reason that we can do this is because of the achievement of the tens of thousands of employees of the Hewlett-Packard Company over the years, especially in the first 50 years. I think they should take the most pride and I don’t think anyone should give me credit for it."

The audiovisual conservation part of the Library of Congress has under its purview all recorded audio, film, video, and associated printed material (like movie scripts) donated to and collected by the library. (Conservation refers to keeping material stabilized, with repairs done without irreversible changes.) The Packard Campus holds much of this collection, including 124 nitrate film vaults, 3.5 million "items" of recorded sound, 1.2 million films and videos, and much more. (A smaller amount that doesn't require the same degree of care or conservation is in other LOC archives.)

If you ever wondered why we need an official repository to keep a record of our creative past that has an eternal charter, rather than allowing academic, corporate, and private libraries and archives to handle it, one only need consider the problem of nitrate film base. Kodak's founder George Eastman commercialized the use of cellulose nitrate to make rolls of film in contrast to the individual plates previously required. The ease of shooting onto continuous roll film appealed both to professional and amateur photographers, and also gave Eastman the lock on the nascent film industry. (He violated a patent and later paid out a vast sum to settle.)

Unfortunately for preservationists, nitrate film is highly flammable even when it's new, and it decomposes in storage into flammable gas. Films that haven't burst into flame — as prints did in many theatres, killing patrons, and in storage, destroying themselves and archives around them — will eventually turn into goop and dust. Kodak notes on its site, that "as the film breaks down, it gives off nitric oxide, nitrogen dioxide, and other gases that yellow the film base, yellow and soften gelatin, and oxidize the silver image. Later, the base cockles, becoming very brittle and then sticky. Finally, it disintegrates completely." Delightful.

Photographic negatives until the 1930s and motion picture negatives and prints from the start of the industry until about 1951 used nitrate film base, and those that remain are all at risk. The LOC says over 80% of movies made between 1893 and 1930 are lost for good. The vaults under Mount Pony keep remaining memories alive.

A Mercury record.

Resuscitating the past

The LOC conducts specially arranged tours of the Packard Campus. It also has showings in its facilities of films from its collection; since the movies are shown on site, the library doesn't have to arrange permission from the copyright holders for those still under protection. The lucky folks who live within a reasonable drive of Culpeper can attend screenings for free most Thursdays, Fridays, and Saturdays of well-known movies and films that few have watched in living memory. The lovely theater even has an organ used to play along with silent films.

The library's ability to bypass copyright restrictions holds true for its sound recordings as well. In the library's Recorded Sound Reference Center in D.C., and in limited other locations, researchers and civilians may listen to parts of the collection with advance arrangement. But the limits on what and where one can listen are what brought me to Culpeper.

I'd met the recorded sound section head, DeAnna, on a previous trip to D.C. after interviewing him in June 2011 about a project he'd spearheaded called the National Jukebox. The jukebox streams audio from a digitized collection of more than 10,000 recordings from the Victor Talking Machine Company from 1901 or 1925 retrieved from wax and shellac discs.

A thoughtful man with a touch of Southern lilt in his voice and a clear love of music and his work, DeAnna showed me the labs in which technicians digitize a myriad of audio formats dating back nearly 130 years. It's a tedious process, because most media — either due to fragility, playback equipment, or other factors — must be digitized at real-time rates or even in multiple passes or fits and starts. Most labs are equipped with reel-to-reel, cassette and other formats of tape players, as well as record turntables. One even has a cylinder reader for wax and other ancient formats.

One room contains recording and playback equipment that spans the entire history of fixing sound permanently onto a medium. In a more contemporary lab, DeAnna showed me IRENE, a system developed at the Lawrence Berkeley National Laboratory to scan grooved recordings, like 78 rpm discs, and reconstruct sound without any physical contact. (You can view an annotated set of pictures from my visit at Flickr.)

Years ago, the library would play the original media on request for researchers in D.C. under carefully controlled conditions. More fragile material, such as lacquer discs and brittle or damaged tape, would be transferred to mylar tape reels for playback to keep wear and tear low on the originals. Analog audio degrades every time it's played back — some forms more than others.

Now, DeAnna said, the creation of a digital version is likely the last time the master recording will ever be played. "These copies will be the last we'll have time to make," he said. No correction is applied in the transfer, with the exception made in cases in which material is unlistenable without some editing. (One might ask why the original is maintained, since it will likely never be consulted again or will degrade beyond playing; that gets into philosophical issues.)

The library digitizes material when it's requested for research and even personal purposes as much as is feasible, as well as having a program of converting material that's falling apart or that works towards the goal of having complete collections in digital form.

To hold these digitizations, the Packard campus sports a server room that's modest in floorspace and vast in storage. It contains both spinning disks and robotically retrieved offline tape storage. The center has the capacity for 10 petabytes of tape storage currently, but that's about to quadruple with an upgrade to the tape format. The Culpeper facility is connected by fiber-optic line back to listening rooms in D.C. proper, in which researchers can hear the digitized work. Keeping up with digital storage requirements is the least of the library's problems.

Collectors of miscellany


Which brings us back to the National Jukebox. Despite the material dating back no later than 1925, launching the project required permission from Sony Music Entertainment, which is the successor through a chain of ownership to the Victor Talking Machine Company. Briefly: RCA, part of GE at the time, bought Victor in 1929. GE had to divest both RCA, which also owned NBC, in 1930. In 1986, GE bought back RCA and sold off everything but NBC, now owned by Comcast. Bertelsmann bought the music catalog, and eventually sold it to Sony in 2008. The rights followed all that. (I'm distantly related to RCA's once long-time chief, David Sarnoff, on my mother's side. He sent my parents a telegram and hefty check when they were married in 1965.)

"Wait!" I hear you cry. "Why would the Library of Congress need permission for material that old to offer it freely? Isn't it in the public domain?" That, my friends, is an involved answer, and something I talked about extensively with DeAnna during our ramble through the audio archives and labs.

Sound recordings typically have two sets of rights: a copyright in the underlying material, such as a composition or an oral work (even if the oral work is a public extemporaneous performance), and a sound copyright, sometimes called a phonogram right, which covers the audio part. (The phonogram right has a neat Ⓟ symbol that has a distinct meaning apart from copyright's ©.)

Until February 15, 1972, the sound copyright fell under the common law of whatever state in which a recording was made. Common law had no expiration date on rights, and also made it murky as to what party possessed the right to reproduce audio. The singer? Musicians? The studio? The producer? Some experts believe that whoever owns the master recordings (in whatever form: wax, lacquer, tape, metal, and so forth) also owns the audio rights. Studios are reluctant to test this in court.

Recordings created after February 1972 have a federally regulated expiration, and phonogram rights can be vested and assigned just like the copyright in a book. (For the uglier details, read "Copyright Term and the Public Domain in the United States," a one-page summary by Peter B. Hirtle at Cornell of all of America's copyright vagaries.)

Federal rules leave work under state protection until February 15, 2067. On that date, a peculiar legal event will occur. All material under protection is federalized for just a moment, and then instantly put into the public domain. It's a form of copyright magic. You read those dates correctly, too. Edison's first utterances in 1877 onto tinfoil (lost now) may bear 190 years of protection. (I dive into this particular aspect in greater depth in my Economist blog entry about the rights issues associated with the Jukebox.)

There's an effort underway by the Copyright Office to fix part of this by retroactively federalizing the phonogram right for all works before February 15, 1972, and assigning sensible expiration dates of 95 years from the date of publication or, for unpublished works, 120 years from when it was fixed in an audio medium.

That would expire the phonogram right for all published audio from 1917 and earlier (unpublished audio, 1892 and earlier), representing a wonderfully large hunk of the LOC's collection, along with material in other governmental hands and in academic and private collections. It's also intentionally close to the 1923 cutoff before which published printed works vest in the public domain with no ambiguity. (Books, films, compositions, and other material published in 1923 through 1972 had retrospective extensions of 75 years, later extended to 95 years, that mean that anything published in 1923 doesn't enter the public domain until January 1, 2019.)

Artifacts of a past age.

Listen carefully

The current state of things clearly pains DeAnna, who is rightly proud of the remarkable range of material in the library's collection. He said the LOC was once culturally snooty and rejected popular music in favor of classical, which led to lacunae, some of which were later backfilled by donations and collection efforts.

Universal Music Group donated 200,000 master recordings (nearly a mile of storage) in 2011 covering released and unreleased material from the late 1920s to the late 1940s, but not the rights to that material. Universal managed to shift the cost of maintaining these archives to the LOC without providing any cash or revenue stream, although it has given permission for parts of the archive to be streamed in the future.

Much of the library's audio holdings, even most of that Universal gift, have little current commercial value to the parties that retain the phonogram right and often the associated copyright. In one lab, I looked into a bin and spotted a box containing the October 20, 1942 NBC radio broadcast captured on a lacquer disc of several programs, including an episode of "Fibber McGee and Molly."

NBC meticulously recorded its radio broadcasts, and donated the archives in 1978 (150,000 lacquer discs from the 1930s to 1970s). When NBC gets an inquiry about using this material, DeAnna said the interested party is referred to the library, which digitizes it and charges its cost in doing so. NBC separately negotiates its own licensing fees.

Thousands upon thousands of hours of broadcasts, including an enormous amount of news recorded during World World II, are locked up. DeAnna would like to set them free, but he lacks the authority, and copyright and phonogram rightsholders have little motivation to let their rights go as they cling to the notion that some portion of donated or library-collected material could potentially generate revenue. DeAnna would make NBC's material broadly available, if only someone with authority at the network had the interest in signing over the necessary permissions.

The nation's audio history remained tantalizingly at my fingertips as I toured around the main building and, later, through one of the 17 vaults dedicated to recorded sound. In the future, the library hopes to create dozens of outposts under its control across the country to allow more people access to digitized material of all kinds.

DeAnna doesn't want to deny rightsholders ownership nor commercial exploitation of valuable recordings, but that remains a tiny part of what is in the library's hands. The rights situation means, for now, the sound of silence reigns over our past — unless you can make a trip to a listening room in D.C. or cadge a visit to Culpeper.

​This article appeared previously in ​The Magazine, an iOS periodical.


  1. We have family just down Route 3 and I have also been there.  We got a peek into the hole when they were doing some initial excavation, and that place is seriously huge, and deep.  I swear it went down 20 stories…  I always wondered what was going to become of it, since it was just around the corner from that ‘listening post’ facility.  I understand that area is filled with underground facilities. There is a decent Thai restaurant in the old hotel downtown.  One time we were lucky enough to be visiting on the anniversary of the Wilderness Battle, and they had all the houses open to the public, including the property where Stonewall Jackson’s arm is buried.  The inside of the house, which switched hands many times during the Civil War, and was used as headquarters and hospital, had graffiti on the walls from those who stayed there during the war.  And let me tell you, they had great penmanship back then.

    1. Not quite 20 stories, haha. Our vaults are located on only 3 levels. Our office building is 3 main levels with a 4th level expansion area.

  2. What an excellent article!  Thanks for posting this. 

    You mention why maintain the original, when we have the digital recording. I’m not sure it’s as much a philosophical as a practical question, at least for me. When you are looking at these things as a source, it’s more than just the recording – you want to look at the physical record itself, as that is an important source as well. While having the music itself matters, how these items looked, feel, what they say on the album itself matters, too.

    Also, as this is funded in part by HP, I wonder if this is going to be the wave of the future – museums and archives in a private-public partnership. I’m also thinking of (is it?) Paul Allen’s museum projects (Experience Music and didn’t he found a sci-fi or a computer museum?). What does this mean for access for academics, journalists, ordinary joes, etc, if museums/archives are partially in private hands? I know that corporate archives are at times hard to get into and then there tends to be expectations that you won’t write anything negative about the company. Coupled with that is the tough time that public archives are under. Here in GA, we have an ongoing issue with the state archives being possibly closed, with accessing only being by appointment. There was an outcry and a petition drive, which the governor responded to, but there is still some uncertainty about the whole thing.

    Also, has anyone read David Suisman’s Selling Sounds. It’s an excellent study of the early recording industry.

    1. Not funded by HP directly — by money that came from HP. That’s why David Packard is so humble. He wasn’t stating he had any special grace to have the money to pay for a chunk of the facility. Rather, the work of HP employees made the stock have value that led to the foundation’s having capital to contribute to the public good.

        1. There are many other reasons for maintaining the original after digital copying, not the least being the fragility of many digital formats.  I’m surprised at the return to tape-based systems.  The Library itself will admit that most digitization being done is at a “reference” level which is very basic with little or no processing on the majority of items.  IF a user needs a better quality transfer, then it can be done on an individual basis later on IF the original has been maintained.  So many of the recordings in the National Jukebox are at the wrong speed, and the streaming does not allow for speed correction.  And there are so many other problems with the recordings that U of Cal-Santa Barbara outsourced — and if you knew how those transfers were being done you would know why.  Finally, we researchers often need the original disc for the real information.  The catalog info is often useless for anything but title and artist.  Broadcast lacquers are never identified by which studio recorded it, what city — important for network broadcasts for quality — manufacturer of the lacquer blank, disc size, speed, if it is an uncoated aluminum disc or lacquer, if it is glass base,  and other things that strongly affected my research.  The people at LC know that I can often tell more about a recording by looking at it than listening to it.  I gathered a lot of info going thru the collection back in the 70s when it was still in the Madison building. A record is more than the sound on it.  Film researchers often have to look at the film itself on rewinds to see how it was printed, what defects are printed in or on this specific print, are the splices printed in or on this print, what sound recording system was used, and other things never noted in the catalogs.  The originals MUST be maintained.. 

  3. Culpepper has a lot of other interesting data centers there too.  The DNS Root Zone key is stored in an ICANN space there (along with one in LA) and a lot of other US Federal data centers.  Plus there is a historic distillery not far from the historic downtown. 

    The rumor was it was deemed “far enough” from a nuke blast in DC to survive so the DoD had backups stored there. Probably not true, but makes a good Cold War story.  

    1. Actually, the bunker we’re located in could withstand a nuclear attack on DC and the surrounding area. It could also keep dozens of people alive for approx. 2 years without opening the door. Before the renovation, the doors here weighed several hundred pounds and could crush you if you stood in the way while opening. Also, there were arrowslits in the walls outside of the main bunker for gunners as well as a shooting range for survivors to practice.

  4. What is the 190 year-old audio format (mentioned in passing)?

    1823… all I can imagine might be some sort of mechanical construction with pegs like a music box or holes like a player piano.

  5. Just FYI, Glenn. The discs we digitized for the Jukebox project were primarily shellac 78’s, not wax discs.

      1. Not a problem and I didn’t really mean to nitpick, haha. I think the confusion there is that the original recording–when the musicians/singers were there in the studio–is made of wax because you needed a material pliable enough to etch the sound into the grooves. Those discs were then used to make the molds and masters to pump out the mass produced shellacs for sale. These are the discs we used in transfers. While we have collections of wax materials (mostly cylinders), they are usually not for playback since it is likely that if we were to play them back in any analog manner, it would effectively destroy the item or render it useless. While IRENE is very promising and gives hope to the idea that we may be able to capture audio from materials not fit for playback, it is still very infantile and may be several years for any real breakthrough. Sadly, there isn’t much monetary interest in investing in this kind of technology.

        Also, I forgot to say this earlier: wonderful story! It’s always a treat to see press on our mission here. Thank you very much for the interest.

        p.s. I am in no way authorized to speak on behalf of the Library, its policies, or mission. I am simply writing as a fellow audiovisual enthusiast/copyright philosopher and loyal BB reader. (Had to cover my hindquarters lest HR or OIG come down on me!)

    1. Sort of related, I found the switch over to vinyl from shellac rather interesting. I was reading old Variety and Billboard issues from the end of the second WW, and this was a big deal. The industry had been collecting shellac records, and melting them down to press new records (since shellac supplies came from SE Asia, which was of course occupied by Japan during the war; also collectors were kind of upset about so many records being melted down, but it probably increased the value of the ones that were left). It was largely due to government internvention that vinyl became preferred in the postwar period, at least that’s how I interpreted the sources I found. The records sent overseas for Victory Records were vinyl (much less fragile). And I read about how the government began to help subsidize vinyl for record labels around this time, too. Prior to that, vinyl was generally only used in records for radio stations – at least according to the sources I read on the matter. They kept saying it had been too expensive. But by early 1950, it was clear that shellac was on the way out and vinyl was to become the mainstay of pressing records. I’ll have to look into the role the petroleum industry played in this…

      1. V-Discs were never called Victory Discs.  Not ever.  In fact, the head of the producing unit, G. Robert Vincent, always joked that he named them after himself!  The V-Discs pressed by Victor were also pressed of a similar plastic called (I think) Formvar.  The ones Columbia pressed were usually laminated shellac. There were some collectors during the war who had made deals with stores taking in the old records to go thru the piles and take out the valuable ones.  So some valuable records were saved because of the shellac drives!

  6. “Universal Music Group donated 200,000 master recordings (nearly a mile of storage) in 2011 covering released and unreleased material from the late 1920s to the late 1940s, but not the rights to that material. Universal managed to shift the cost of maintaining these archives to the LOC without providing any cash or revenue stream, although it has given permission for parts of the archive to be streamed in the future.”

    Will music companies never cease to be scumbags?  

        1. It should be noted that the Universal donation came after they had lost an even greater number of masters in the disastrous Universal Studios warehouse fire a few years ago. Since they know that they can no longer be trusted with this stuff, they are now letting the experts watch over what is left.  We probably lost most of the wide-range vertical lacquer masters of full recording sessions from the 40s, and metal parts that had not been considered important enough to transfer or reissue. 

  7. I spent a year volunteering with the LOC Motion Picture Conservation folks when they were in an old WWII-era facility on Wright Patterson AFB in Dayton. It was dirty, smelly, cold, and wonderful. Best year of my life. Those people are unsung heroes. There was a great documentary made a few years ago about the National Film Registry with lots of footage and interviews with the people there in Culpepper. It’s called THESE AMAZING SHADOWS. Good stuff!

Comments are closed.