Catch a .wav

Information obtained from http://www.gcn.com/print/25_10/40581-1.html

Inside a digitization project at the Library of Congress

If you’re on your third copy of the White Album or you can’t find a phonograph needle to play your Elmore James 78s, you can understand the plight of the Library of Congress: A collection of nearly 3 million recordings in every format from wax cylinders, to disks and tapes, to CDs, most of them playable only on obsolete equipment.

Eugene DeAnna, head of the library’s Recorded Sound Section, calls format obsolescence “the plague of audio recording from the beginning.”

Just as troublesome is the deteriorating condition of physical media, which are becoming more difficult to play back safely over time. So archivists want a standard for preserving recorded audio that will ensure recordings remain available for future generations.

“It is alarming to realize that nearly all recorded sound is in peril of disappearing or becoming inaccessible within a few generations,” the LOC’s National Recording Preservation Board warned in a recent report on best practices for capturing analog sound.

The task faced by the library continues to grow. Michael Taft, archivist for the National Folklife Center, announced recently the acquisition of a collection of 186 one-sided 78-rpm test pressings from artists including Sonny Boy Williams, Lil Johnson and two Blind Willies—Johnson and McTell. Perhaps most significant are 25 sides by seminal blues singer Robert Johnson. Johnson recorded only 29 songs in his brief career but was a major influence in blues and on British and American rock music. “You can’t hear him any clearer than you can on these recordings,” Taft said.

DeAnna also announced the discovery in the archives of unsuspected recordings of a jam session with jazz great Lester Young.

“No one was looking for it,” DeAnna said. “No one knew it existed.” (For the latest additions to the National Recording Registry, go to GCN.com, Quickfind 561.)

Now, these treasures have to be preserved.

Seeking standards

The board, which was created by the National Recording Preservation Act of 2000, convened a panel of experts in March to establish digital file standards and metadata schema for preserving old recordings.

Although digital formats seems like a logical choice for preservation, archivists are a cautious lot and did not adopt digital technology until it was forced upon them.

“We were fairly late in coming to digital,” DeAnna said. “As late as 1999 the standard archival approach was that it is better to stick to tape. Two years later, nobody was making tape.”

Now the writing is on the wall, and the Library of Congress is moving to digitize its valuable archives.

The move to digital preservation comes as the library’s Motion Picture, Broadcasting and Recorded Sound Division is moving its collections from the LOC’s Madison Building in Washington to new storage facilities near Culpeper, Va. “We ran out of storage space 15 years ago,” DeAnna said.

They found a new home in a Cold War bunker complex under Virginia’s Mount Pony, where bank vaults were built to hold enough cash to jump-start the U.S. economy after a nuclear war. The vaults have been gutted and climate-controlled to provide subfreezing temperatures needed to store color motion picture film. The facility has 57 miles of shelving on three floors with 20 years of growth space. About 1 million items have been moved to their new home.

“There’s a lot more to go,” DeAnna said, including some of the most fragile items, a collection of 16-inch lacquer-coated glass disks that contain NBC Radio broadcasts from World War II.

“The coating peels and the glass is incredibly thin,” DeAnna said. “Those we’re moving last and we’re trying to come up with a process to do it safely. They have been preserved to tape, but with the state of digital processing we can do much better.”

How much better? That depends on the standards used. The accepted format for audio archiving today is an uncompressed WAVE File. “There is not much debate about that, since it is an uncompressed format,” DeAnna said.

But there is no standard for the sampling rate used in recording analog sound digitally. The International Association for Sound and Video Archives, a European organization, recommends capturing a 96-KHz frequency range with a 24-bit rate.

This means the digital recording has 24 bits to describe each sound, which provides 16.8 million possible levels of audio. A standard CD today has a rate of 44 KHz with a 16-bit rate. Some engineers believe the sampling rate for archiving should be even higher than the current 96-24 recommendation.

Why the high sampling rate? “You already are exceeding what the human ear can hear,” DeAnna said. “But by capturing at a wider bandwidth you are making more data available for re-storation.”

The library does not do restoration when it makes archive copies of recordings. It copies with all of the clicks, pops, cracks and other artifacts included. But if the recording is restored later, the additional data makes it possible to clean it up without making it sterile.

“The danger in digitization is that you can clean it up too much,” DeAnna said, losing some frequencies.

Audio experts convened in March to come up with a recommendation for digital audio archiving standards. With recent advances in digital storage, there is no practical reason not to increase sampling rates to include more data, DeAnna said.

“The argument against it used to be the files were too big,” he said. “That’s a joke, now. Nobody discusses it.”

Whatever the sampling rate used for archiving, another format will have to be used for listening.

“A WAVE file at that resolution is not a consumer format,” DeAnna said. Digital files probably will be converted to MP3 format for listening. MP3 has become the commercial standard for digital music even though “it’s not nearly as good as a CD.”

Currently, the library plays original recordings for listeners. Once the archive standard has moved to digital files, MP3 copies could be created from archival files and made available from the Culpeper facility to the reading room in Washington.

“Our plan to is to have a derivative server to deliver MP3 files to the local server on the Hill,” DeAnna said.

Security of the network will be a primary concern because copyright laws prohibit any copying of the files. In fact, “just by preserving sound digitally, we’re on the edge of the law,” DeAnna said.

Copyright laws permit playing an original recording for listeners, and making a limited number of copies for archival purposes. But copying a file for listeners, as the library envisions, requires the copyright holder’s consent. The Library of Congress is not exempt from these rules (see sidebar).

“Who knows what we’ll find out in the future?” he said. “We’re only a few years away from ‘born digital.’ ”

Music releases now submitted to the LOC must be the “best copy,” which today means a CD, even if it also is available as an MP3 download. But DeAnna sees the imminent demise of the CD as a commercial medium.

“Clearly, that’s what is coming; iTunes has made that perfectly clear. Nobody is making money on a CD anymore.”

Guess we’ll have to buy the White Album again.