Contents - Previous - Next
7.7 Strategies for the safeguarding of audio and video materials in the long term
Dietrich Schüller, Phonogrammarchiv, Vienna
Ever since PICKETT and LEMCOE's report (1959), which marked the beginning of systematic approaches towards audio preservation, audio, and later video, archivists have been looking for measures to preserve their holdings in the best possible way - this meant the artifacts: the cylinders, the disks, the tapes. Considerable headache and research has been, and is still, directed towards the question of the life-expectancy of audio and video carriers and the measures necessary to retain their chemical and physical integrity for as long as possible.
Film was the first audiovisual medium that attracted systematic activity towards its preservation. This is most understandable as the cellulose nitrate, which was used as filmbase until the early fifties, is inherently unstable and extremely flammable. This situation also lead to the foundation of FIAF, the Fédération Internationale des Archives du Film, as the first audiovisual archives association as early as 1938.
The crucial problem has always been the fact that audio and video carriers have never really been produced with longevity in mind. They are part of the entertainment industry and are in daily use - both in the private and in the professional domain. The products have proved to be more or less stable enough for the purposes they were originally intended for. Archivists, however, are thinking in different dimensions of time. They had to carry all the burden of preservation including expensive and elaborate research into the stability of the many materials used. There was little systematic support from or co-operation with the producers of tapes and disks. The situation, however, changed gradually during the second half of the eighties as a result of two initiatives:
Members of the Association of Recorded Sound Collections (ARSC) in the United States approached the Audio Engineering Society (AES) to include topics of archival matters in a workshop during the AES Convention in Los Angeles, November 1986. This workshop raised great interest from the side of the audio engineers and brought about a Sub-Committee on Audio Restoration and Preservation within the AES Standards Committee. William D. Storm, a member of both ARSC and IASA Technical Committees and the then Director of the Belfer Audio Laboratories and Archives of Syracuse University, was appointed Chairman; the Committee has since met regularly at the fringe of the American AES Conventions. The core of this group consists of audio archivists - members of the Technical Committees of ARSC and IASA - and of interested manufacturers - mainly of tape and tape equipment - who realized that archiving obviously was a yet undiscovered and possibly profitable market. The Subcommittee later also joined forces with ANSI (American National Standards Institute) and formed the Joint Technical Commission AES/ANSI IT9-5 which is actively engaged in working out standards and recommended practices for the preservation, re-recording etc. of audiovisual materials.
In parallel, and in close co-operation, with this AES Group, and with several personalities in common, the technical bodies within FIAF, FIAT, IASA and ICA joined forces in 1987 by forming, with the help of UNESCO, the Technical Co-Ordinating Committee (TCC). One of its main activities was the organization, with the support of UNESCO, of a dialogue between audiovisual archivists and manufacturers of technical equipment.
In the course of the activities of these groups it became apparent that manufacturers were not aware of the fact that a large part of all professional tapes produced and sold are being kept and put into an archive for at least a mid-term period. The audiovisual archivists, on the other hand, were unable to make any statements on the order of magnitude of their holdings.
In this situation the IASA Technical Committee started a project to provide an estimate of how many hours of recorded audio material is kept worldwide. It was intended to scan the situation in selected countries and to extrapolate from there on the worldwide situation. The results of the United States, Sweden, Austria, and Germany were been presented at the IASA Conference in Ottawa 1990. A publication of these results, however, was withheld in view of the proposed worldwide survey by the Library of Congress, the results of which were given in a paper by Gerald D. Gibson. The idea behind both surveys is the same: it is an attempt to assess the worldwide storage requirements for audiovisual materials and to draw conclusions about the likely future technical and financial requirements for the preservation of the audiovisual materials which have become an important part of the cultural and intellectual heritage of mankind.
In looking at the audio and video domaines, the results can be summarized as this: the reported holdings embrace 14 million hours of audio, and 2.7 million hours of video. The worldwide projection of the holdings is 45 million hours of audio and 8.6 million hours of video. The growth rate of these holdings (and this is known from the IASA TC study) is between 5 and 10 percent per year.
This is an impressive amount of holdings; the preservation of which in the long term raises considerable technical, logistic and financial problems.
In developing possible scenarios for preservation of these holdings into the future, let us first stick to traditional archivist's thinking of safeguarding the artifacts they have in their collections. Lifetime expectancy of the various carriers plays the predominant role in this kind of thinking. It is not the purpose of this paper to go into a lengthy debate but let us assume an average lifetime of 50 years of a carrier - an assumption which will be regarded as over-optimistic by many, especially for some modern formats. Even in this rather optimistic scenario, the transmission of the contents of such carriers over a millennium implies the necessity of 20 generations of subsequent preservation or master copies. If this is done in the analogue domaine - and most of our holdings are still analogue - everyone familiar with the audible and visible degradation of signals after only a few generations of copies will agree that such a strategy cannot be called "archival".
It becomes clear that traditional thinking, in its attempts to minimize the shortcomings of the analogue world, is the source for the idea of the "eternal carrier". Though technically possible in the form of metal matrices, this possibility has never succeded on a large scale because of its enormous costs. There are few examples where metal masters, mainly of the LP format, have been produced just for preservation purposes. The BBC did so in their earlier days, political events of special importance in the former communist countries have been preserved in this way, sometimes also most valuable research materials like historical holdings of the St. Petersburg Phonogrammarchiv. Durable carriers are also available in parts of the digital world. There are metal masters used in the production of optical discs (CD) and the French manufacturer DIGIPRESS has developed the "Century Disc", a compact disc of tempered glass, with a life time expectation of centuries. It must be stressed, however, that "eternal" carriers also require the "eternal" availabilty of the required playback equipment.
With the advent of digital storage techniques, however, the situation has, in principle, changed radically. In the digital domain it is possible to copy information without any loss from generation to generation. The content of each copy is identical with the original and, most justifiably therefore, the term "cloning" is associated with digital copying. Digital storage offers also another most important feature over the analogue: the integrity of the information can be checked precisely, and, most important, errors - up to a certain threshold - can also be fully corrected.
Many archivists will remember the optimistic view especially the audio archivists (including the author) took in the late seventies and early eighties of the digital future. Professional digital formats have been envisaged, and accepted as an universal standard, by the late eighties or early nineties. Similar thoughts may have lent wings to our colleagues in the video archives. The market, however, developed differently. Various formats have been developed, none - except the compact disc - has gained universal acceptance so far.
Unlike analogue times, when the consumers were blessed with by-products of a professional market, the digital audio development became driven by the consumer formats rather than the other way round. A good example is R-DAT which was intended to replace the compact cassette and now survives in the professional world.
The video scene is even more complex: six digital formats have been developed so far, none of which is in a dominant position in the market. It is not unrealistic to say that the first digital video formats will become obsolete before they will have gained any sizeable market penetration. This adds another, hitherto unknown, problem to audiovisual preservation: the obsolescence of hardware or - to define it from another side - the physical survival of the software over the economic survival of the hardware. While it is expensive, but still affordable, to construct replay machines for Edison cylinders from scratch using the latest technology and exceeding by far the performance of the historical machines, it is unaffordable if not impossible to rebuild a complex digital audio or video player once mass production has ceased.
Though obsolescence of hardware plays an ever increasing role, emphasis has to be given to the fact that none of the hitherto developed commercially available audio and video formats have been developed with longevity in mind. None of these formats have so far been accepted by the community of archivists as "archival".
The IASA Technical Committee in reviewing the situation drafted, during the Annual Conference in Canberra 1992, the following resolution: "During its meeting during the Canberra IASA/ASRA Conference in September, 1992, the IASA Technical Committee discussed its ongoing review of the progress of digital formats in view of its suitability for archival purposes. Although decided progress has been made, the IASA TC is still not in a position to predict a better life-expectancy for a digital system and format than for the known analogue formats in present use. Hence the recommendation that analogue tape is the preferred master storage medium still has validity. However, archives should actively play a role in the development of digital systems optimized for long-term storage of our cultural heritage" (Minutes of the General Assembly II, Canberra, 29th September 1992).
In this complex and desperate situation the solution lies in the use of digital mass storage systems for audio and video signals. Such systems would hold audio and video information as computer files. As in the world of "traditional" computers it would be possible record, read and transfer data without necessarily personally handling a physical carrier. The enormous amount of data required by audiovisual information, however, can only be managed adequately by jukebox-like robot systems which automatically load the required data carrying medium. Such systems, of course, are remotely accessible and this feature is of special interest for radio, television, and national audiovisual archives: it would radically improve and dramatically reduce the cost of access to their holdings. Automated access also permits the self-checking of the integrity of the holdings. Without any human intervention a copy (clone) can be made of a carrier found to be at risk of losing information because the error rate is likely to rise beyond full error correction capability. If, after a given time, the system becomes technically outdated, transmigration, i.e. the transfer of the total collection into a new system, can also be made automatically. The idea itself is not new, of course. Ever since the use of computers the very data carriers and their longevity have never played any significant role. We all have personally witnessed, for example, continous bank services throughout the decades without noticing the many transmigration processes performed on our accounts - from punch cards to various magnetic tape and disk formats - which took place behind the scenes. The difference lies only in the order of magnitude of storage space once audio or, even more, video signals should be stored digitally. This new way of thinking can be traced back to May 1989, when - on the occasion of the 90 anniversary of the Phonogrammarchiv - the Second UNESCO Consultation took place in Vienna.
The first debate on such a scenario in the sound archives world was associated with the author's presentation of the paper "Towards the Automated `Eternal' Sound Archive" at the Joint Technical Symposium (JTS) in Ottawa, May 1990. This presentation was not generally hailed and scepticism was raised at that time, especially with regards to the authors view, that longevity of carriers of such systems is of minor importance: as lifetime of the system of more than 10-15 years would only unneccessarily freeze outdated technology.
Since then, however, the change of paradigm has been generally accepted: If a considerable part of the audio and video holdings is to be preserved for the future, it is the information that has to kept rather than the carrier. It has to be recognized, however, that in the case of mass-produced carriers, namely disks, the carrier itself and its associated materials like labels, covers, booklets, etc. are cultural objects worthy of preservation. The preservation of these will the tasks of museum-like collection, which undoubtedly will do this in a very selective manner, once the survival of the contents is guaranteed in a quantitively satisfactorily way.
In the light of the inherent instability of audiovisual data carriers any legal deposit legislation that demands the preservation of the very artifact should be reconsidered. It seems an undue financial - if not technically impossible - burden for future generations to preserve, for example, every single mass-produced cassette in the very form it has been manufactured.
Currently, several attempts are made to enter this new domaine of audiovisual preservation. Most advanced is the project of the Bundesverband der deutschen phonographischen Industrie (Federal Association of the German Phonographic Industry). A Working Group on Archival Systems, of which the author is a member, is currently finalizing the specification of a mass storage system suitable for the requirements of digital multi media management on a large scale. Such systems would be capable of holding several tens of Terabytes.
The rather exotic numerical magnitudes mentioned in this paper suggest a table of the prefixes standardized in the decadic numerical system:
|1 k (kilo)||103||1 000 one thousand|
|1 M (Mega)||106||1 000 000 one million|
|1 G (Giga)||109||1 000 000 000 one billion|
|1 T (Tera)||1012||1 000 000 000 000 one trillion|
|1 P (Peta)||1015||1 000 000 000 000 000 one thousand Tera|
|1 E (Exa)||1018||1 000 000 000 000 000 000 one thousand Peta|
In addition to the recorded sound, written documents, label and cover layouts, photos and other information associated with the recordings will be stored. The first implementation of such a system is expected for 1995.
Beside this most advanced project, several radio stations, especially within ARD, the community of German radio stations, are considering similar solutions.
There is a noteworthy development with runs parallel to the endeavours of the world of archives: the so-called information or data highways. These new communication tools are associated with video-on-demand, teleshopping, electronic libraries, multi media data banks, etc. All these new and attractive services have a common prerequisite which is not part of the daily discussion: mass storage systems. This convergence will undoubtedly lead to intensive further development of such devices.
Returning to the Library of Congress survey, it is possible for the first time to roughly estimate the worldwide digital storage requirement of audiovisual archives. As film preservation is not yet economically feasible in the digital domain, let us calculate the storage space requirement of the projected worldwide audio and video holdings only:
The results are impressive: Uncompromising digitization of 35 mm color film requires 40 MB (Megabyte) storage space per frame and 3.6 TB (Terabyte) per hour. Audio is estimated to require 30 PB. Video, depending on the chosen format, 825-1050 PB or around 1 EB.
Even in the unlikely event that only 10 % of the worldwide holdings are kept "to eternity", the result is still impressive. Furthermore: photographic stills and film preservation would vastly increase this amount, once more powerful mass storage system come within financial reach. And, finally, if only the most important recent literature is put into full text data banks, another enormous increase of storage requirement has to be added.
All these requirements are in the immediate future: they will undoubtedly lead to a boom of digital mass storage systems in the very near future. Obviously, there will be various competing systems on the market and in principle there is - as with audio and video formats - the danger of obsolescence of hardware. But there is an aspect which gives rise to optimism. Although some mass storage systems are extremely attractive in terms of costs per stored Byte, the initial investment in hardware is considerable. It is unlikely, therefore, that too many competing systems will turn up, leading to the phenomenon of the obsolescence of hardware as we know it from the audiovisual formats. The financial sums necessary call for cautious decisions. It can be assumed that the first succesfully operating systems will have an influence on subsequent decisions, possible thus establishing a quasi-standard.
From the high costs of hardware it also becomes clear that small institutions like academic units, etc., which do not hold great amounts of audio and video data, would probably not attempt to take responsibility for the preservation of their holdings themselves. Safeguarding could be passed on to greater, central units which are able to offer professional data storage at economically attractive conditions.
The big problem to be solved in the forthcoming years is the question of the transfer of the hitherto accumulated materials in such storage systems. There is yet no serious estimate on the duration of this transfer but, as a preliminary rugh guess, it may be assumed that this procedure may last at least 20 years. This perspective brings the aspect of longevity back into the debate. Undoubtedly the transfer has to be structured according to priorities: frequent demand and endangered original materials will be the most important parameters in setting such priorities. Therefore, research into the life-expectancy of existing carriers and measures to prolong their life are still top issues on the agenda; it would be unwise to start the transfer with relatively stable materials while others rot away.
In summarizing, it must be stressed that all technical prerequisites for a new strategy in audio and video preservation are available. Preservation of the information rather than the carriers is the only feasable solution to preserve in the long term what is considered of archival importance. Undoubtedly, a start into this new era of preservation has to be made soon and the implementation of the first systems is very close.
Contents - Previous - Next