Content in Context

[beginPage: Intro]

project 2002 - 2005


New technologies - distribution of video art, by  Gaby Wijers

The Netherlands Media Art Institute encourages the free development, application, distribution of and reflection on new technologies in visual art. Since it was founded in 1978, the Institute has brought together an extensive collection of video and media art, comprising more than 1500 works, from first experiments by now famous artists from The Netherlands and abroad to recent productions by rising new talents. These works are shown in-house during screenings and form the content for the exhibitions of the Netherlands Media Art Institute. In order to acquaint a wider audience with them, the Institute distributes both video works and media installations from its collection. These are to be seen at Dutch and international festivals and events, and exhibitions in galleries, museums, media centers and other art institutions. Internally, in the media library the public - students, curators, artists, lecturers, art historians and all other parties who are interested - can call up information on the artists and their work. All works from the collection can be seen on viewing sets. In addition to the video collection the mediatheque has a large collection of books, journals, and documentation on contemporary media and art.

End Report by Gaby Wijers [Download pdf]
[endPage]
[beginPage: Infrastructure streaming]
Because of the growing demand for access, availability and distribution for the large number of video works that are owned or administered by the Netherlands Media Art Institute, in 2002 we chose to move to the digital storage of the works. This storage had to fulfil a number of requirements. Capacity and speed are qualities that define accessibility. A high degree of accessibility to the works must be guaranteed by, for instance, making the works available for viewing both in-house, full screen in high quality (DVD) and by means of intranet streaming (for staff/mediatheque/exhibitions/presentations), and in high and low quality outside the Institute (internet streaming/experiments/research/distribution). This meant that the distribution and editing department had to change from the usual way of transcription from master to Umatic or VHS (sight copy) to transcription to DVD. In addition, all information had to be indexed and entered so it could be called up in a relational database. The MPEG-2 format was chosen for storage, because this fulfilled the specifications that are maintained for DVD productions and because of its relatively acceptable (i.e., broadcast) quality. Moreover, MPEG-2 met industry standards and appeared to guarantee the system.

High quality
In 2002 the Netherlands Media Art Institute investigated the various formats and systems for digital storage and established procedures. Since then works from the collection have also been distributed on DVD. Works seven or more years old are conserved on high quality Digital Betacams and are digitized from this source. The newer works are on various formats and are encoded on the same station. With orientation to available systems, price/performance/stability was an important focus, and for that reason the Netherlands Media Art Institute preferred realtime coding. The choice ultimately settled on a Windows workstation with a Mobius M720D * MPEG encoder board with composite, s-video, component and SDI inputs working under Windows NT 4. The management software is MPEGWorks Pro encoder with VTR operating system. Image and sound are encoded realtime. In general, the encoding of video takes place at a constant bit rate of 8 MB/sec. (in some cases, such as long works, a lower or variable bit rate is desirable), and in all cases for audio a sample rate of 48 kHz and bit rate of 224000 is employed. DVD/MPEG-2 production takes place in a DVD authoring station with various authoring software packets such as Sonic Reel DVD, Sonic DVDIT! PE and TMPG Authorer. An important point is a Realmagic Netstream2000 DVD decoder card that, combined with an RT video monitor, makes possible realtime preview. In addition to the production of DVDs, the system generates an MPEG-2 file which is stored on the file server for further use. The MPEG files are transported from the encoder to the file server. Since October 2003, a Snel & Wilcox compression pre-processor has been in standard use with the encoding process for noise suppression.

In October 2003, the set up and procedure for encoding at high quality was found to be operational, and since then encoding is proceeding several days per week. In 2005 the distribution collection (about 1.5 terabyte) was almost entirely encoded when the video server showed an unexpected defect. The server where the data is stored for further use is a Dual Pentium 1000 GHz machine (Dell Power Vault). Three cabinets with 14 disks = 42 disks are combined by a RAID 5 system into two virtual disks of 1.3 TB each. A customized Advanced Windows 2000 is installed in the server. The RAID system (Redundant Array of Inexpensive Disks) 'couples' multiple hard disks into virtual larger hard disks, and can minimize memory loss with mirror techniques and/or error restoration methods. But that did not prove sufficient. Two disks appeared unreadable after the RAID controller apparently refused to perform. RAID only works between the hard disks, and does not help if a lead becomes defective or a processor refuses to work. The only thing that can be done then is replacement on the repaired file server by a back-up file (stored on tapes). Unfortunately the back up also proved to be faulty and a large part of the encoding work had to be redone. Since then (obviously) a stringent back-up policy has been introduced. Furthermore, it appears that in the case of the breakdown of the file server it will be at least 30 hours before everything is operational again, and in that time the information can not be optimal consulted. For the problems which will arise with an expanding file server based network, in the future a solution will have to be sought in the form of Storage Area Networks (SANs). A SAN affords a high degree of reliability by making the hardware redundant (important components are always executed in duplicate). Further, the storage capacity of a SAN is almost infinitely expandable. For now a solution has been found in an extra server on which the collection is saved a second time.

Internal
Since March 2003, it has been possible to view at least a part of the works shown during exhibitions via a Video Jukebox. Selecting from a menu (for instance, by the artist's name, year and subject) the viewer or viewers can choose a particular work on computer monitors. With the aid of another computer it is possible to project the works with a beam. If they relate to the subject, a suitable selection from the Institute's collection can be presented with every exhibition. Furthermore, the interface is flexible. Records can be kept of what searches (and therefore, what works) are increasing popularity.

Beginning in May 2005, the staff of the Netherlands Media Art Institute (through their pc's) and the public (through the viewing sets in the mediatheque) have access to the files of the video server. With this we have opted for an integration of the functionality of the public cataloque. This is to say, the search and presentation function of the cataloque is used internally to present the video works full length and full screen via the server.

External
For the high quality external streaming, experiments were carried out in cooperation with GigaPort and the members of the Virtual Platform. The various media labs shared their knowledge about possible applications and the experience they had gained regarding what equipment, what structure, what software, what protocols and what standards, etc., could best be used. That all took some doing. The technology always proved much more complicated than expected, and the implementation more labor intensive. In comparison with fellow distributors, the Netherlands Media Art Institute now has a large head start in terms of knowledge about and experience in digital distribution, even in comparison with America. That is in part to be credited to our involvement with GigaPort. For instance, during the e-culture fair in 2003 e-culture TV was provided, in a live broadcast via www.centraalstation.net. The program was comprised of brief live interviews, impressions of the atmosphere, broadcasts of presentations and conference discussions, reports from the fair, and previously produced items. SURFnet took responsibility for the broadband broadcast. The first arrangement was that events would be streamed live from the Melkweg. Ultimately, because of practical problems, broadcast took place from Waag Society, where the GigaPort connection is found that is used for Waag Society, Paradiso, Melkweg, De Balie and the Netherlands Media Art Institute. The material, as recorded on tapes, had to be transported to Waag Society for this. There they could be sent out on broadband, but the stream could not be received simultaneously at the Netherlands Media Art Institute. This was the first time that Waag Society and the Netherlands Media Art Institute had provided a broadband broadcast.

In December 2003, an expert meeting took place at Waag Society under the auspices of GigaPort to get more insight into high quality broadband broadcasting.
The Netherlands Media Art Isntitute also has alliances with groups in Austria: the Kompetenznetzwerk Mediengestaltung consists of a long list of businesses and educational and art institutions. The Netherlands Media Art Institute furnishes content in the form of works from its collection and technical and legal expertise (the latter in regard to copyright) for Campus-TV, an art channel on the Campus.

In 2004 an application for a European subsidy for the exchange of media artworks by streaming was turned down. As a result of this, together with the cessation of our grant for the connection to GigaPort, developments stagnated. Unfortunately, because the knowledge and experience has been gained, the experimental stage is passed, the collection is available and the Netherlands Media Art Institute is ready for the next step to a wider public, it would be a shame if this next step cannot be taken. The Netherlands Media Art Institute has now focused its hopes on a future direct connection to the broadband in cooperation with the City of Amsterdam, XS4all and SURFnet, and new European requests.

Low quality
Since mid-2003 the formats, systems and procedures for encoding at low quality have been investigated and worked out. A test film was assembled for the streaming video test from frequently used video material which included sequences that were potentially problematic for conversion into streaming formats (for example, noise, rapidly blinking images, slow movements, graphic material, digital images, amateur material). The material was first encoded to MPEG-2 and subsequently transcoded to the four most popular streaming formats (Windows Media, Real, Quicktime, MPEG-4) in three qualities for various connection speeds. Not entirely unexpectedly, it emerged from this test that the quality of the source material is the most important factor for the quality of the stream. The differences in quality among the four formats were examined primarily for the highest resolution, because that is where the differences were most obvious. In general, all the formats were good. Quicktime and MPEG-4 were somewhat less successful in handling fast movements (especially blinking images), and Windows Media produced clearly smaller files. Leaving costs out of consideration, Windows Media turned out to be the best. But Windows Media needs Windows Media server software, and that runs only on a Microsoft Windows server. The Netherlands Media Art Institute prefers to work with an open system, and does not have Windows but Linux servers. Because of its somewhat better quality, and that fact that Real has a Surestream option (that is to say, if the connection deteriorates anywhere, the Real server automatically switches to a lower quality image), Real was initially chosen.

The Artlab of the Netherlands Media Art Institute produced thirty-second MPEG-2 fragments and stills of works in the distribution collection. These were converted to RealMedia. Everything was stored on the server. In the process the fragments were cropped, meaning that a part of the image frame was reduced or cut away. This is not necessary for showing on television sets or monitors for video presentation; these are made in such a way that they show no edges (underscan). But the edges are seen on computer monitors (overscan), and these edges would be intrusive when viewing the fragments on the Internet via the catalogue. At this time fragments from all single-channel works in the distribution collection are coded in RealMedia. Installations require a somewhat different approach, because there it is not a fragment from the signal but from the documentary record that must be coded. This documentation was not recorded and conserved in the same manner as the autonomous works.

In order to be able to convert the fragments easily to another format in the future, the MPEG-2 fragments were preserved as well as the RealMedia fragments. Unfortunately, during the crash in early 2005 about half of the MPEG-2 fragments were lost. These will be prepared again retroactively by students in 2005 and early 2006. Previously it was thought that MPEG-4 would provide the best results. MPEG-4 is a relatively new format with a standard set by a committee, which could quickly measure up with the other formats. The MPEG-4 codecs can be used freely, the source is available, and an active Internet community and a commission supervise the maintenance of the standard. Moreover, the necessary server, Darwin, is open source and thus free. In early 2004 the results of the MPEG-4 investigation were compared with the existing state of affairs. Because the automatic translation of MPEG-2 or RealMedia into MPEG-4 proved much more problematic, and the openness in the future less than was expected, the works have not yet been translated, and the investigation is continuing.
[endPage]
[beginPage: Content Management System]
Metadata research
Metadata are the rules and definitions that provide structure and consistency for databases and processes. Rules for the processing of data elements, layouts for reports, data dictionaries and object models: these are all examples of metadata. Metadata are important, because they provide the level of abstraction that is necessary for presenting large quantities of application data in an understandable form. The existing collection information system at the Netherlands Media Art Institute was based on MARC. The current standard for web-based catalogue and access systems is Dublin Core. Now that the development and use of metadata for digital documentation is taking more concrete forms, ways to link up with Dublin Core are being sought. The Dublin Core metadata provides definitions that are comparable with the 'old-fashioned' catalogue card system, with which the qualities of objects are described in web-based access systems. In addition, we are closely watching the developments and choices of the Capture Unstable Media project (at V2_) and the Open Archives System Information Sharing project, and others. As part of this project we will be reviewing the results of ongoing research, and probably applying them to further stimulate the exchange of information, in the Kompetenznetzwerk Mediengestaltung and OASIS (Open Archive System with Internet Sharing), among others.

The Netherlands Media Art Institute conducts ongoing research into the development of models and methods for conserving media art and making it accessible. It is characteristic of video art that there is a vehicle carrying a signal which itself can only be made visible with the use of apparatus, a player, in a means of presentation which may or may not have been specified by the artist. The precise medium for storing this data or signal can vary and is at the discretion of the artist. It is also up to the artist to decide whether or not the carrier, playback equipment and manner of presentation are essential to the meaning of the work. The artist's perception of the video work is, in many cases, unknown. Whether changing the format or replacing the current equipment will affect the authenticity of the work of art is an open question. The codes that have traditionally applied to visual art, such as the uniqueness of the physical manifestation of the work, do not apply to video art. The artwork is indeed unique and authentic, but its physical form is not.

In contrast to traditional museum objects, video art and other media art forms need a different approach from storage oriented to what might be called classical objects. Proper storage of videotapes and equipment is needed, but that's only one part of the story. In contrast to the simple archiving approach of information and communication technology, authenticity and artistic intention, and information about it, can be very important. Context, user interaction and distributed authorship should be taken into account.

A responsible approach to documentation, presentation and preservation requires gaining insight into the artist's intention and the significance of the technology used. Video art, installations and live installation performances developed out of a multidisciplinary practice: conceptual art and performance. Sometimes the technology used was simply the vehicle which made it all technically possible. The technology used was always part of a larger social context, and is subordinate to the artist's intention, which can be conceptual or materialistic. There is a vast lack of information on this point. We seldom know if the technology used is a part of the intention behind the work and should be preserved as well. But for proper preservation we do have to research this, in order to know how to handle this in the long term.

To prevent of such problems in the future, together with SBMK (Foundation for the Conservation of Contemporary Art)and INCCA (International Network for the Conservation of Contemporary Art) the Netherlands Media Art Institute description models which have been implemented in the new collection management system that is based on Dublin Core. The Variable Media Network developed a database and questionnaire that are extremely useful for gaining insight into the artist's intention and into the significance of the technology used. Models and tools for description have been introduced in the Capturing Unstable Art project too. The OASIS project will also go into that direction.

The main question is what perspectives can be developed to make video art accessible, bearing in mind all its peculiar characteristics? This is a question not only for a research collection but for professional presentation. Real video art only works, so to speak, with respect to technical issues and artistic intent. Preservation and presentation of video art have different goals and different preferred formats. Although the technology is rapidly changing and offers more possibilities for digital distribution, this principle will not change in the near future. Accessibility lies in the research collections and documentation about video art works.
In preserving and presenting video art, the Netherlands Media Art Institute wants to keep the work alive with regard to all the aspects mentioned here. The starting point for this is knowledge of the artist's intentions and opinions, which is needed to make proper decisions, focussing on range of acceptable variability and documentation.

WatsNext
About ten years ago, in cooperation with Toxus, a collection information system called WatsOn was developed especially for the Netherlands Media Art Institute. Existing systems did not satisfy the specific requirements for describing the diverse forms and media involved in video art, performances and installations. The desired integration with moving images and diverse forms of documentation was also not an option with existing systems. In 2003 this collection information system was upgraded, under the new name WatsNext. This has become an integrated, relational collection database, still based on the previous WatsOn version (1994-2003), and developed for Explorer and Windows 2000, with the aid of MySQL, Delphi, PHP and XSLT. In WatsNext the work and the form it takes are distinguished. The core of the system is therefore the abstract description of the artworks, to which the various forms they take (video art, installations, performances), the media/vehicles (Digital Betacam, VHS, DVD, etc.) and artists (collectives, individuals, etc.) are coupled. Specific information is stored at each level. Thus there is, for instance, basic information to be found such as the title (in English and Dutch), date, length, a description of the content (in English and Dutch), color, sound, number of channels, etc., but there is also information about the rights, presentation and conservation.

Artists, codes (genres, subjects and key words, subtitling, language, etc.), pictures from or of the works, and documents and files regarding the artworks are also coupled to the works.

The same is true for media/vehicle. More technical information is provided for this latter category, including formats (Umatic, Betacam, etc.), systems (Pal, NTSC), conservation work performed, and quality, but there is also information about the collection to which the media belong, and the use permitted for the tapes/files/disks stored there. In general, works are present in multiple copies in one or another of the collections, on various vehicles, with different specifications and rights.

Works from the various collections are integrated in WatsNext:
- Distribution (video art and installations that are actively distributed >1600);
- Documentation (video documentation, interviews, works for reference > 4000)
- Conservation (conserved works from the Institute's own collection and from third parties > 3500)
- Masters and footage (artist's examples of their own work and records of events > 2500)
- Administered works (works by others administered by the Institute, excluding conservation > 500)
- Video server (MPEG-2 files of the distribution collection > 1500)

Thus it is possible to tell at a glance what works and vehicles with that work are present in the collection, and what the differences are in terms of version, edition, length, subtitling, editing (for instance, loops or signals), presentation quality and rights. WatsNext not only provides information on the works, but physical files such as photographs, stills, video fragments, and, for instance, contacts, are also coupled with the information and entities. Of course, distribution data (who, what, where, and when shown) and royalty payments are also kept up to date. The existing, well-functioning bilingual brief listing (with descriptions) of genres, key words and subjects including scope notes remains functional and will continue to be maintained.
The first beta version of WatsNext was installed in November 2003; in mid-2004 the existing collection database was converted. After that conversion control took place and the files were cleaned up. In early 2005 the conservation database was converted, and in mid-2005 the addresses followed. The biographical entries and information on installations will be checked and supplemented in the second half of 2005.

Public catalogue
Similarly, about ten years ago Cyclope, the terminal for the public, was developed in cooperation with Toxus. This also later became partially accessible by Internet. In 2004, on the basis of information for internal and administrative use from WatsNext, an English-language Internet version was developed for the public under the title Catalogue.
The information from WatsNext to the Catalogue is updated daily. Works from the distribution collection are presented in context, provided with descriptions, key words, stills and short fragments. Information on the artists, works from other collections, and also the book collection will be added retrospectively in the future. The catalog was operational in mid-2004. As of September 2005, all works from the distribution collection will be provided with a video fragment, and can be looked up by Internet. With that, justice will finally be done to the essence of video art, by integrating the description directly with moving images and sound.

The user has access to a large, if not the most important, part of the collection via the website www.montevideo.nl or direct by catalogue.montevideo.nl. The start page gives random access to various works from the collection. The catalog of the Netherlands Media Art Institute has a simple but powerful basic search system. One can search using the artist's name, genres, subjects and key words (including scope notes). Other possibilities are the artist main entry, date, title (or words from a title), description of contents, or a combination of the above.

A list with all the matches is given as the search result for each search entry. These provide the basic description, with the artist or artists, title, content description, length, date, and an illustration for the preview.

Users have access to the basic description, and can click on through to a comprehensive description or screening of a video fragment in RealMedia. Information regarding what formats are available and in what way these should be presented is deliberately not reported at this point. Users now have the information to determine what works can be viewed in the media library. Those wishing to screen a work can contact the distribution department.

In April 2005, during the Dutch Open, our activities with respect to Content in Context were presented in an international symposium and workshops regarding the new distribution system.

Videoregistration Seminar
There are plans to further publicize the collection at the national and international level in the near future. As a part of these, there are projects under way for the development of Digital Rights Management systems and simultaneous search actions in various media art databases. These will also enable educational institutions and professional screening locations to obtain information and material from the collection of the Netherlands Media Art Institute directly through the Internet. For the Netherlands Media Art Institute, this will be an opportunity to test the Catalogue and other systems that have been developed. Moreover, this will be an especially good opportunity to improve the streaming media facilities. Thus, in the future high quality, full length, full screen viewing of the collection, with immediate handling of administrative matters and payment of royalties will be offered to educational institutions, artists, museums, festivals, curators, and other interested parties. Through this the availability and accessibility of the cultural heritage of media art, and the reputation and name recognition of the Netherlands Media Art Institute will both be increased. Moreover, such service will attract a new, international audience.
[endPage]
[beginPage: Dealing with rights]
The Internet offers the Netherlands Media Art Institute new possibilities for distributing its collection by means of streaming video techniques. With them, in the long run distribution will become easier, quicker and less expensive. There are however two problems, both involving copyright, which stand in the way:

1. How the works will be presented is more or less beyond our control; that it to say, it depends on the situation of the 'receiver'. Moreover, for the time being, anyone can download an artwork and show it when and where they wish, without sanctions. In that case, professional presentation is not to be guaranteed. Many artists object to that, and will not pass control over their works to others.
2. A second problem involves the financial aspect. When works are offered on the net for free, the artist derives no direct income, and one can even speak of depriving artists of income. On the other hand, this does provide greater visibility, which will build the artist's reputation and, in the long run, could increase income.
 

Artists take various positions with regard to copyright and royalties. There are artists who give permission to have their work (generally made for a different medium) accessible on the Internet. But there are, of course, also artists who absolutely don't want this. The latter believe that their work and the medium represent a unity, and accordingly under no circumstances can they be seen as separate from each other. There are also artists who find the reception quality for most users too poor, so that, at least for the time being, streaming appears to be ruled out as a new platform for their artworks. Further, the context of the presentation of the work also plays a role in this.

Because the Netherlands Media Art Institute acts as a distributor, the copyright situation of the works to be distributed was clear, and the focus of its investigations into new techniques for distribution and legal rights lay not so much on copyright, but more on protection, safeguarding data, registration of use and the payment of royalties in the case of streaming of existing media artworks. One of the first research tasks undertaken, as part of the Content in Context project an inventory was prepared of the relevant initiatives in the field of Digital Rights Management (abbreviated DRM). DRM refers to 'electronic systems for making creative material available in digital form, and for its use, with which protection against illegal use is possible, and through which legal use can be monitored and paid for' (Arkenbout, et al., 2002, p. 4). The systems most thoroughly investigated (for instance, by Beeld en Geluid) work with material for which different copyright problems play a role, and therefore generally with material specifically for educational use, which is more or less free of legal restrictions on its use.

In addition, copyright strategies for the distribution/streaming of works made for the Internet have been researched. The starting point here has been the licensing model developed in 'Playing Field'. From the research it appeared that works, which are made for Internet, involve different copyright problems, and generally are advantaged in open source/open content solutions. The knowledge and experience with regard to digital rights, for instance, which the Netherlands Media Art Institute gained from its placing projects online, and the exchanges regarding this, proved essential in the investigation.
This system has already been implemented for the development of new media art works. We are now seeking a system with which both DRM and open source/open content solutions can be combined in one databank.

Artists' views
In 2003 and 2005, the artists whose work is distributed by the Netherlands Media Art Institute were informed of the digitization program and its results. The reactions were diverse, from extremely positive to defensive. In addition to written correspondence, during the course of the project an information day and expert meetings were held, and a seminar was organized.

The responses from the artists were highly diverse, but they were in any case overwhelmingly positive about storage on the server, and the in-house screening from it. Technical aspects such as quality, security and royalty payment played a role in their response. A selection of their remarks:

- I'm very pleased that you are doing this and certainly give you permission to do so with my work.
- It seems fine to me but ... until you have decided how to treat the financial part of it, it is not possible to give you a definite answer. If you are going to distribute videos for free it is of course not interesting for anyone to pay for them, they (institutions, galleries etc.) will be able to show the works without any cost, just by connecting to your server. And this is of course not what I at least think is fair.
- Yes, I would like my work to be included in your digitization plans.
- With regard to the proposal involving digitizing video material - fine!
- It seems to me a good plan to research distribution via Internet further for its workability, etc. and thus I agree to make my work available for this.
- I agree under the following conditions:

- Storage on digital media of course, but showing in small format or on a monitor is not a good idea as far as I am concerned.
- I am not willing to cooperate in this case. I do NOT agree with any of points from this offer.
- I do not want my work put up on the Internet.
- I received your letter concerning the project of putting tapes onto a server - and you have my support. I do agree to the proposed procedure. Good luck.
- I in no way want any part of … [being] streamed on the Internet or any other way outside of my contract.

In September 2005, all the artists will once again be informed, and invited to the presentation of this publication. During 2005 and 2006 the distribution contracts will be adjusted. Presentation by Internet will them become an integral part of the distribution contracts of the Netherlands Media Art Institute.
[endPage]