The Sound Savings conference presents an opportunity to assess the state of audio preservation programs in the United States in 2003, to examine some issues related to preservation, and to give thought to where we are headed. This is the first national conference dedicated to sound preservation. The significant number of attendees is evidence of increased interest in the challenges of audio preservation and the number of questions we have. Many archives are pursuing the transition in preservation reformatting programs from analog tape as the preservation medium to new digital formats. This transition is probably the single greatest reason that so many people are attending Sound Savings and that interest in audio preservation is burgeoning.
For years, sound archivists have been talking about the digital future. No longer do we discuss the digital future—it is the present. Preservationists and archivists are, and should be, conservative and cautious about adopting voguish trends prematurely, especially as we attempt to assure that the audio artifacts of today will be available for study and entertainment centuries from now. Yet, it is clear from my perspective that, in terms of preservation, analog is dead, or at the very least, a dead end.
I will discuss the digital present—and future as well. Yet, the digital revolution, as it is sometimes termed, is not the only subject on our minds, even if many of the other issues relate to the major transition taking place in our archives. I will touch on a variety of related audio preservation issues, but only lightly. Many of them will be covered in greater detail by our speakers this week. While I was asked to look at trends, I find myself thinking as much of the challenges.
In managing sound archives we make a distinction between conservation and preservation of audio materials. Reformatting, that is, conversion of content from one medium to another, is inevitable for most materials in sound archives. But that reformatting, to which we apply the term preservation, can be deferred for many years, if not decades, if collections are properly conserved. Unfortunately, no comprehensive, tested, and documented standards exist for the cleaning, storage, and housing of audio collections. Much work is required to develop a set of professional standards based on science for conserving our originals.
We don't know which practices constitute best care for all audio media in our collections. These practices are still being identified. For example, it is now standard practice in archives to provide the best possible care for original recordings, whether they have been re-formatted or not. This wasn't always the case. It wasn't that long ago that masters were routinely destroyed after reformatting. The procedures for conserving audio media are still evolving--more so than for other library media. What is the best packaging for original audio recordings? Do acidic record sleeves contribute to the deterioration of shellac and vinyl discs? Are compact discs potentially harmed by the paper booklets in jewel boxes? Up-to-date published standards for housing and storage of acetate tape do not exist. A number of recently formulated or revised recommendations for storage tapes circulate among archivists verbally but they are not codified within a best-practices manual.1
We await research and documentation to help archivists assess the condition of the collections within their responsibility and identify which problems demand immediate attention. All "instantaneous" audio formats (e.g., magnetic tapes and lacquer discs) are known to require eventual reformatting, yet the rates of deterioration for these formats are not known. Given content of equal cultural value and uniqueness, to which medium should an archive give priority for reformatting: hydrolysis-afflicted polyester tapes, audio cassettes comprised of cheap tape stock, or fifty-year-old lacquer discs? Definitive data on which media are most at risk are not known. Archives need documentation to help identify problems affecting media and set priorities for the limited resources available for audio reformatting programs.
In the spring of 2003, the Image Permanence Institute of the Rochester Institute of Technology issued welcome news. They announced that they will undertake a study on the Preservation of Magnetic Tape Collections. It will "focus on the deterioration of magnetic [audio and video] tape and work on creating techniques to help libraries, museums, and archives save their collections."2
Digital preservation, or reformatting audio onto a digital format, has been discussed for decades, and it has been disdained as a viable solution for nearly as long. There have been two major arguments made against digital preservation. The first is that all digital formats are susceptible to deterioration; there is no "permanent" digital format. The other objection has been that common digital formats, such as those employed for compact audio discs, employ algorithms to compress, or reduce, the data required to represent the sound.
Compression is usually inappropriate for preservation reformatting because most often the objective is to capture as full and as rich and accurate a reproduction of the original as is possible. With the cost of digital storage diminishing each year, compression is no longer considered to be a necessity.
Indeed, there is no permanent digital format. In the case of analog audio, however, there has never been a permanent, or even long-term, format. Archivists agreed many years ago that 1.5 mil quarter-inch polyester tape on an archival 10 -1/2 -inch open reel was the best medium for magnetic audio recordings.3 This format was considered then to be the best available. Yet, all reasonably priced analog audio formats are subject to deterioration and must be copied to new media eventually.4 Analog quarter-inch magnetic tape stock has become more difficult to obtain as the number of manufacturers diminishes. And, of course, each generation of analog reformatting engenders a loss of content and increase in noise.
Sadly, many of the open-reel preservation tapes created in the 1980s have deteriorated faster than the original media whose content they were intended to preserve. The tapes suffer from hydrolysis or "sticky-shed syndrome." The tape binder adhering the recording material to the backing absorbs moisture from the air. Upon playback the tapes squeak and break down. In cases where the original media were saved and conserved, they are often in better condition than the preservation tapes intended to save their contents.
If digital preservation is the new paradigm, what form will it take? A digital format standard comparable to that established by the Association for Recorded Sound Collections in the 1980s has not been established. Many archives are preserving content on recordable compact discs (CD-Rs). This convenient medium plays on existing compact disc players and the best available blank discs cost under $1.00 each. Yet, recordable compact audio discs hold under 700MB of data and are more prone to degradation than manufactured discs. Recordable DVDs hold much more data but at this time there are several competing recordable-DVD formats and these discs are believed to be more fragile than CD-Rs.
Given the challenges and limitations of audio preservation on analog tape, it is understandable that archivists have turned to recordable digital media. Judicious archivists see the process as an interim solution, at best, and cover their bets by making multiple copies of the recordable discs and storing the discs in separate locations.
Within these proceedings, Carl Fleischhauer describes in detail the Library of Congress's approach to digital audio preservation, the creation of digital files intended to be stored in a digital repository. Digital file repositories have been used by banks and the credit industry for decades, and are used by European broadcasting companies for storing files of audio visual content. Digital file repositories are designed to backup data systematically on the preferred storage format of the moment. The data is sustained through any number of shifts in design and configuration of the storage formats.
Digital repositories operate on the assumption that there will never be a permanent physical format. Well-designed repositories ensure the persistence of data by validating its integrity periodically when it is copied. The well-planned repository presumes media obsolescence, plans for it, and, according to its supporters, frees the archive community of the futile search for an affordable permanent medium. In the eyes of some archivists, digital repositories are liberating.
Digital repositories will be expensive to build and challenging to operate. They require a sophisticated information technology infrastructure in order to migrate files successfully, keep them accessible, and maintain their integrity. Yet, to managers at the Library of Congress and other large libraries and archives, digital preservation is dependent entirely upon the success of these repositories.
Essential in the development of repositories are safeguards to ensure their continued existence in case of a breakdown or catastrophe. Implied in a digital repository is faith in the assurances of a professional IT infrastructure. Repository systems must be networked with built-in redundancy, including mirror sites that can substitute in the advent of an adverse situation.
With wise investment of ample resources these complex systems can be built, but only by financially advantaged institutions. Smaller archives must not be left behind. For digital repository systems to be truly successful, they must accommodate collections held by institutions without the resources to build their own systems.5 With the generous support of the Packard Humanities Institute, the Library of Congress is creating a National Audio-Visual Conservation Center, now under construction in Culpeper, Virginia, 75 miles from Capitol Hill. Concurrent with the planning for this facility is the development of a digital repository for the preservation of audio, video, and other digital collections. Library officials hope that this state-of-the-art storage, processing, and preservation facility will be more than a big, new building. The intention is that it be truly national, perhaps providing storage and repository services for other institutions. The Center might also perform duplication services for other institutions if a funding mechanism and a process can be devised which do not compete unfairly with the private sector.
As Carl Fleischhauer points out, digital repositories for audio are not merely collections of sound files. The repository planned by the Library of Congress entails associating sets of files to create digital objects. Following the Open Archival Information System (OAIS) model established by NASA,6 digital objects for sound recordings in the repository will include digital images of record labels or tape boxes and other graphics or accompanying text, in addition to the audio files. The audio files themselves will be very large, recorded at a sampling rates of 96kHz or 192 kHz, with 24-bit word lengths.
The files will be described and controlled administratively by metadata (which can be partitioned into "descriptive," "structural," and "administrative" metadata) about the original recording and its digital files. Structural metadata identifies and organizes the individual files of images and sound that represent a digitized item. The metadata assist with the presentation of these related files from the digital repository. In a repository, structural metadata are called up by program scripts to reconstruct virtually the sound recording's packaging (scanned images of the covers, accompanying text, etc.) and to provide researchers with control over which audio tracks to audition.
The Library of Congress is working with other institutions to develop the Metadata Encoding and Transmission Standard (METS) metadata format to facilitate the documentation, maintenance, and presentation of these files.7 METS is not a universally-accepted standard, and we can't guarantee that it will be the standard, but I believe that it is the right direction and has the granularity that would facilitate the migration to another standard if necessary.
METS is complicated. Because it requires populating a very large number of fields, at the present, it is time-consuming to create a full METS record. Officials at the Library of Congress hope and expect to "develop tools for automatically creating metadata," as recommended in a study of challenges related to the preservation of digital content.8
A preservation manager is quoted by Richard Griscom as stating that, "to date no one can prove that any digital version will survive and be accessible beyond a few decades, despite much talk of migration and emulation. [Libraries should] exploit access capabilities of digital technology and combine them with the longevity of proven preservation methods."9 This prescription has not yet been disproved, but in the case of audio preservation, unlike print microfilming practices, there is no proven analog preservation practice. Digital preservation in some form is here to stay and many capable people are working to ensure that its products remain permanent as well.
The creation of repositories and files to store in them is but one challenge sound preservationists face. Much work remains to ensure that it is possible to re-format our vast collections. There is an enormous variety of media in need of reformatting. The media fall into broad categories (magnetic tape, discs, etc.), but each medium presents its own challenges for the best possible reformatting. I am not convinced that we are fully prepared to meet these challenges. The development of audio preservation standards and professional practices is in its infancy. Tests must be conducted, best practices proved and documented, and training provided.
Systematic development of a body of professional knowledge about audio preservation will take place on many fronts. The most common preservation practice for tapes afflicted by hydrolysis is to bake them at a low heat in a scientific oven and then re-format them. However, some engineers disdain this approach and argue for alternative methods to "dry" the tapes. I am unaware of documented scientific tests proving any approach and look forward to such research. A catalog of common problems encountered in audio media and their recommended solutions would be of value to all archivists.
Most recognized experts in signal capture from legacy analog media are over sixty years of age, and their methods and tricks may retire with them. There is no systematic program to document senior engineers' most successful procedures and ensure that their wisdom is passed on to future generations. Competence with digital recording tools is not always synonymous with expertise in capturing sound from antique media. Documentation and training for safe and effective capture of sound are as necessary as proven guidelines for housing and storage.
A recent study commissioned by the Council on Library and Information Resources concluded that, "Many libraries, and especially smaller ones, need outside help for their preservation programs in the form of advice, instruction, opportunity for learning, contact with those active in the field, involvement in collaborative efforts, and funding."10
The Library of Congress National Audio-Visual Conservation Center is intended to support audio-visual preservation education. A training program, administered cooperatively with a local community college, is being considered. Training specialists will have to be funded and hired, and a curriculum written. This program offers great potential but developing it presents significant challenges. In order for such a program to be successful, it will require the support and collaboration of the preservation community.
Other significant work is underway. The conferences of the Association for Recorded Sound Collections (ARSC) have provided important venues for communication among engineers experienced in the playback of legacy media. The ARSC Journal has published some of the valuable results of their work together. The association also has taken the lead in developing guidelines for the design of archival cylinder players and stylus design for playback of cylinders.
Playback, or signal capture, methods where nothing physical comes in contact with the recording are being explored by a number of scientists. Vitaliy Fadeyev and Carl Haber at the Lawrence Berkeley National Laboratory are experimenting with using high-energy physics instruments to map the grooves of 78-rpm discs via precision metrology and digital image processing.
Their experiments thus far entail both two-dimensional and three-dimensional mapping of the record grooves. With these detailed representations of the forms of the disc grooves, the scientists hope to be able interpolate corrections to rectify scratches and other groove wear. The approach shows promise but is still in the developmental phase. Currently, it requires over 100 hours to map one side of a ten-inch 78-rpm disc in three dimensions.11 Independent of this effort, a group from the University of Applied Sciences at Fribourg, Switzerland, and the Fonoteca Nazionale Svizzera are experimenting with taking high resolution photographs of disc recordings, digitizing the photographs, and applying algorithms to extract sound from the images.12 In a third project, engineers at Syracuse University are developing a laser player for cylinders.
Fadeyev and Haber have expressed hope that their work will lead to large-scale digitization preservation projects. The Library of Congress, too, is investigating the feasibility of mass reformatting. The Library's collections include over 100,000 audio cassettes and 170,000 open-reel tapes. Library managers there suspect that the only hope for preserving the content of many of these recordings is to develop systems to extract the sound without real-time monitoring and adjustments by audio engineers. Several administrators at the Library believe that compromises, or fall-backs from traditional reformatting procedures, will be necessary to assure that audio, such as that on those cassettes, remains accessible for study and enjoyment.
Tape playback systems, such as those sold by Quadriga, monitor and transcribe technical metadata, including documentation of tape and signal deficiencies, as tape reformatting takes place. Many preservation specialists see the creation of these devices as positive developments but wish that these devices included more of the tools required for preservation reformatting, such as detailed documentation of and correction for tape variances. The Library of Congress is working with one firm now to create technical specifications for a mass duplication program. These specifications will be well documented and disseminated.
To accomplish mass reformatting more than one source must be duplicated at a time, which will preclude continuous, real-time monitoring of the recorded signal by an engineer. By its nature a mass reformatting system will compromise existing preservation standards. We don't know which compromises will be required, or whether they will be acceptable to archivists and users of audio collections. It will be interesting to observe whether and how these compromises can be agreed upon. What kinds of collections are appropriate for mass reformatting? How will those collections be identified and by whom?
In addition to the challenges and opportunities which digitization of sound presents for preservation, this digitization has produced many new methods of distribution of audio which give new responsibilities to archives. In light of these new formats and methods of distribution, many of us are in the process of rethinking what we acquire and subsequently preserve. MP3 file proliferation, and I mean music distributed legally and exclusively as MP3 files, has facilitated more "publishing" of greater amounts of music.
Direct marketing over the Internet by musicians challenges archives (especially those with subject or regional focuses) to find new and more thorough ways to identify, collect, and preserve music and other audio from these new, small business sources. Given the ephemeral nature of Web sites, archives will need to act quickly if the content of these sites is to be preserved.
The selection of a format in which to preserve Internet-distributed audio is another question archives must consider. The proliferation of audio on the Web has brought a diminution of the technical quality of much audio. The inherent compression of MP3 files or audio suitable for Web streaming implies a lowering of audio quality standards. If these new distribution models and businesses are unstable, as well as dependent upon compressed audio, might maintaining a collection of high-quality masters be a new responsibility for audio archives?
The present instability, if not disarray, of the music business has other effects upon archives concerned with preservation. Manufacturers no longer claim that compact discs are "permanent." If archives are committed to retaining their content eventually they will have to "rip," or copy the digital discs' content to bit streams. To thwart illicit ripping of CDs some manufacturers are encrypting, copy protecting, and/or watermarking their products. This will make legitimate duplication for preservation more challenging. It has been suggested that another response by the industry to illegal duplication of CDs is to attempt to eliminate compact discs and replace them with mixed-media DVDs or combination DVD/CDs. The hope is to make it more difficult to pirate music and at the same time, make it more appealing in the marketplace. No doubt these DVDs will be more difficult to preserve. The challenge will be to emulate their interactive behavior, in addition to preserving the bit streams.
There is a vast amount of sound that needs to be preserved, and the standards and new efficiencies are not adequate to assure that our audio heritage is secure for posterity. I believe that collaboration among archives is necessary. Given the seemingly perpetual limitation of financial resources available to archives for preservation and the significant number of duplicated holdings among archives, efforts should be made to reduce preservation redundancy as much as possible.
The Metadata Encoding and Transmission Standard is intended for more than the management and presentation of files. It has been designed to facilitate sharing of files. Utilization of that potential may be necessary to obtain adequate funding of preservation. It has been observed that to many people, digitization and access are synonymous. Griscom points out that while at one time access used to be secondary to preservation programs, with digitization the driving force is access.
Regardless of the legal obstacles to making file sharing between institutions a reality, the distinction between preservation and access has become blurred.13 Support by the Packard Humanities Institute to the Library of Congress for the creation of the National Audio-Visual Conservation Center has been motivated in part by an objective to improve access to older recordings held by the Library.
Archives should be exploring legal, as well as technical, methods to collaborate on preservation projects and share the products of those projects. This will not be simple. The piracy of recorded music is a great problem faced by the music industry, and convincing intellectual property holders in music to allow archives to share music files will be a significant challenge. In addition, the laws governing recorded sound are complex and often vague.
Digital audio reserve systems (the placement of listening assignments on servers to enable convenient and simultaneous use by students) are in common use today. Yet, copyright experts are not in agreement as to whether these are in strictest terms "legal." Since digital repositories result in more than the legally mandated maximum of three preservation copies of a recording, even they may be illegal under a strict interpretation of the law. Recordings of classic radio broadcasts will be particularly difficult to share legally. They are not protected by federal copyright law, but instead by often imprecise state copyright laws and various trade union contracts.
As a result, these broadcasts are among the most legally restricted recordings held in archives. Archives will need to work together to establish copyright licenses if they are to share any audio files. Constituencies, both within and outside our institutions, will need to be built, and potential collaborators will have to advocate for legal solutions.
In addition to establishing the legal means to share files, storage and server networks must be established and administered. Congress has charged the Library of Congress with building the National Digital Information Infrastructure and Preservation Program. The NDIIPP was created to help provide the legal and technical blueprint for these networks. The NDIIPP has begun this work by enlisting collaborators and obtaining the counsel of information technology experts, legal authorities, and representatives of content industries. The program will not be the sole province of the Library of Congress. To be successful, it will need to be a national effort that includes participants from the private as well as public sector.
Further assistance will come from the work of the National Recording Preservation Board, created by Congress in late 2000. The board is comprised of 22 individuals, representing archives, the recorded sound and music industries, and relevant professional organizations. It advises the Librarian of Congress on the annual selection of sound recordings to a National Recording Registry of culturally, historically, or aesthetically significant sound recordings. The board, in collaboration with the NDIIPP and the Council on Library and Information Resources, is commissioning studies on a number of audio preservation issues.
The recording preservation legislation also directs the Librarian of Congress to "implement a comprehensive national sound recording preservation program, in conjunction with other sound recording archivists, educators and historians, copyright owners, recording industry representatives, and others involved in activities related to sound recording preservation..."14 The legislation also establishes a National Recording Preservation Foundation, a federally chartered, but independent, corporation to raise funds and award grants for the preservation of audio collections.
We have entered a very promising era for the preservation of audio. There is a broader constituency for preservation than ever before and ever-increasing resources, but of course we need more. We must be careful not to throw money at problems. Many archivists are hopeful that enormous strides will be made in the next ten years: research completed, programs established, and thousands of recordings preserved and made available to the public. This conference is an opportunity to collaborate on the development of an agenda for research and action to address the challenges seriously. I look forward to looking back on this Sound Savings conference as another landmark in a new era of professional sound preservation practice.
Two more recent best-practice guidelines are a recommendation to package acetate-based recording tape in boxes which are not airtight, in order to enable necessary off-gassing; and a retraction of the directive to "exercise" (slow-wind) tapes periodically. The latter is no longer a recommended practice. Unfortunately, word of the change in recommended practice has not been adequately circulated to archivists in charge of audio tape collections .
"RIT Studies Increasing Shelf Life for History Stored on Tape." Rochester Institute of Technology press release, June 25, 2003.
Association for Recorded Sound Collections, Associated Audio Archives, Audio Preservation: A Planning Study: Final Performance Report. (Silver Spring, MD: Association for Recorded Sound Collections, 1988).
Pressed vinyl discs have been proposed as a preservation medium but vinyl, too, degrades eventually, and the cost of creating master discs and pressings for every hour of audio to be saved would be prohibitively expensive. The Church of Scientology commissions platinum analog discs of recordings of founder L. Ron Hubbard and these may well a permanent medium. But if such a solution is viable, it is so only for a limited body of work as each disc must cost the Church thousands of dollars to produce.
Committee on an Information Technology Strategy for the Library of Congress, Computer Science and Telecommunications Board, Commission on Physical Sciences, Mathematics, and Applications, National Research Council. 2001. LC21: a digital strategy for the Library of Congress . (Washington, D.C.: National Academy of Sciences).
ISO Archiving Standards--Reference Model Papers. http://ssdoo.gsfc.nasa.gov/nost/isoas/ref_model.html
Metadata Encoding and Transmission Standard. http://www.loc.gov/standards/mets/
National Academy of Sciences. LC21.
Richard Griscom, "Distant Music: Delivering Audio Over the Internet," Notes (March 2003).
Anne R. Kenney and Deirdre C. Stam, The State of Preservation Programs in American College and Research Libraries: Building a Common Understanding and Action Agenda. (Washington: Council on Library and Information Resources, 2002). http://www.clir.org/pubs/reports/pub111/introsum.html
"Sound Reproduction R & D Home Page." http://www-cdf.lbl.gov/~av/
Ottar Johnson, Frédéric Bapst, etc., "VisualAudio: An Optical Technique to Save the Sound of Phonographic Records." IASA Journal 21 (July 2003): 38-47.
Griscom, Notes (March 2003).
National Recording Preservation Act of 2000, Public Law 104-474.
This paper represents work carried out for a federal government agency and is not protected by copyright.