Issue contents
This article discusses the points made in the presentation made to the Cataloguing and Indexing Group in Scotland entitled "There's no place like home: YouTube and the National Library of Scotland".
The National Library of Scotland recently engaged with a number of Web 2.0 services online, which is discussed at greater length elsewhere. In the Scottish Screen Archive, the newest division of the Library, the idea of placing film clips online is not a new one. However, YouTube offers a far more public interaction with viewers of archive material. The interface allows users to comment, and post short films of very wide provenance and standard, indeed as wide as the audience can imagine, within certain legal caveats as laid out in its Community Guidelines.
The following article will explore how we have approached the description and management of digital moving image on YouTube. Film is being shared in an increasingly digital, online environment, the tools and methods of describing content is changing, often it seems, outwith our control.
YouTube describes itself as "The home for video online" where you are encouraged to "Broadcast yourself".
Founded in February 2005, purchased by Google in November 2006, YouTube carries a large amount of commercial content as well as amateur material, from companies such as Sony Music Group and Universal. Its Wikipedia entry (8 Apr 2008) notes it as currently having around 70 million user videos and approximately 5 million user channels - uploading a video somewhere in the world every second. Content can cover everything imaginable, from video diaries to serious political issues of the moment, from the amateur to the commercial. Beginning in December 2007, 14 clips have been added to the NLS User Channel. These have now attracted an encouraging response, more than 6000 views and a number of positive comments but this was not necessarily our expectation. Being initially rather cautious of exposing the material in this way, it was most important to identify risks concerning copyright and also the loss of control, and even the quality of presentation. The whole exercise raised a variety of questions.
To compare the quality of films please view the clips of "The Kirkwall Ba'" (YouTube version) and the equivalent clip on the National Library's new site.
Partial screenshot (8 Apr 2008) of "Dunfermline Gala Day 1932" on YouTube.
This clip is an indication of what may happen when an enthusiastic amateur adapts archive film for an online audience. An added soundtrack and an inauthentic sepia tone have been created – the film has been changed in a way that the Archive has always been careful to prevent. Although the intent is not malicious, the film appears as a curiosity rather than representing the intention of the film maker. The poster has given a rather idiosyncratic synopsis, making no comment on the context of the film, and giving only four tags, of which, only two are (arguably) useful to the searcher. There is a deep link to the old Scottish Screen site: now out of date, it leads only to a file 404 "page not found", formerly a useful opportunity to discover further information about this particular film. This raises an interesting question about how fully NLS wishes to delve into Web 2.0. Does the library contact all those who use our material? In this case copyright has expired; however, should NLS make it policy to ask that altered material be removed or simply accurately signposted?
Information describing content on YouTube is deliberately simple, encouraging as much completion as possible. From the point of view of the cataloguer, the "Video Info" page seems to have a few obvious omissions: YouTube assumes that the uploader is the creator of content but this often not the case, particularly when posting archive film. There is no official field identifying the 'edition' of the film or clip; a user may run through a number of incarnations in a found list before finding a high quality original. For example, try entering "Cat Man Do" into YouTube's search; users have added copies, tributes and even posted "improvements" to the original film by Simon Tofield.
Together with encouraging input with a simple "Video Info" interface, YouTube imposes information on the creator. Related films are automated by the site, as it crawls for matching tags and keywords. Some control is therefore lost regarding interpretation of content, and what it is related to. Tagging well and appropriately encourages people to respond positively to films, as they will have found what they want and hopefully a community of interest on the web around that content. However, text may not always remain as tightly connected to the common search experience if visual methods of browsing gain function and popularity (see the warp demo). If so, tagging will need to be as relevant as possible in order to create useful relationships - at the moment this form of exploration feels fairly uncontrolled and often lacks relevance.
Certain basic information is mandatory: title, description, one category and tagging - this depends entirely on the motivation and even honesty of channel users, who might be motivated to use tags which will increase hits. Consequently, the richness and usefulness of the metadata varies a lot. The identity of the film is further outlined by choosing to broadcast your video or keep it private for up to 25 viewers with a log-in. The film may be dated and identified by a geographical location - both being limited to one entry which is not always adequate - films can have a range of dates or an unknown date, for example. As for information which might be added to the film, sharing options may be controlled to allow comments, ratings, video embedding, and permit content to be used freely on mobile phones or TV.
Experimenting in order to enrich the available metadata, the team played with the position of information within fields. Titling including information like dates or "silent" or "sound" took advantage of the freedom allowed. Titles were made as simple and descriptive as possible, hoping to pick up hits in a search. For example "OUR THREE R's" [our capitals] was renamed "First Day at School", providing context for the clip, rather than keeping the title associated with the whole film, which may not be as useful. In lists of results and on film pages, YouTube does not allocate much space to written information - synopses are displayed in a shortened form by default. In order to list the film usefully, particularly within this framework, retitling may be necessary. Categories given by YouTube are not adequate, offering only the following; Autos & Vehicals, Comedy, Education, Entertainment, Film & Animation, Howto & Style, Music, News & Politics, People & Blogs, Pets & Animals, Science & Technology, Sports and Travel & Events. This leaves no obvious place for agricultural film, for example, which forms an important part of our collection.
Tagging appears to become more cohesive when shared amongst social groups, such as people interested in trains for example; users are their own guides to the resource. Questions are fairly common amongst posts; it is still easier to ask another person to assess the meaning of an enquiry, rather than search - even though responses are not always helpful. In the experience of the YouTube team, encouragement of relevant interaction by forging a social group echoes the role of front of house staff in the library. Just as the Library's Customer Charter lays out rules and privileges, related users appear to self-regulate, creating useful comments in order to inform and co-operate with their social network.
An interesting example of seeking "Stakeholders" might be that of Gadget Show presenter Jason Bradbury, who has collected more than 4m (mostly appreciative) views for his film "Extreme Caterpillar Breakdance" using his online social networks. In his blog he recounts his experience of "viral marketing": very simple marketing; for the most part he publicised his film through those areas in which he already had a platform and a relationship to others in the group. NLS, by extending its online presence to Facebook, Flickr and YouTube, opens up opportunities in marketing through word of mouth.
Our user comments range from:
Would such comments increase in the future, if we promoted videos to stakeholders or "friends" on YouTube? This interaction may help us understand the information they would prefer to find, perhaps by making video responses to our films, and we could tap into existing online communities to spread the word about the Archive.
In comparison, WeShow, a "human-powered guide to the online video universe" was launched in January 2008. This service employs teams of web editors, harvesting content from YouTube and other online film depositories, "labelling" and organising this content for different nationalities. They had already provided links to Scottish Screen Archive material before CIGS' seminar in March. This new website underlines a real need for more usefully organised, vetted content.
It is a fluid situation, moving towards equality between the cataloguer and the user, with distance between public and private metadata becoming ever closer. However the most important input remains that of the channel; user metadata, if it is allowed, is less controlled; even if social groups appear to contribute in a more constructive and mutually beneficial manner. We aim to stand out and create a good first impression on the user, by being relevant to their search and helpful - in order to inspire good feedback. It's easy to email a link to a clip, or to subscribe to our Channel and receive alerts when new content is added.
It is certainly a future goal that the whole Scottish Screen Archive collection would be available to view online, accessible by everyone. In joining YouTube, we face challenges in terms of cost, content and copyright.
We only considered titles that were already on tape, because in reality film is a very expensive medium to duplicate. For example, the recently launched Heritage Lottery funded project averaged GBP 1000 per one hour of running time to transfer Film to Video. That digital tape must then be encoded in order to create files that can be used online. We chose not to reinvent the wheel - where digitised content existed from other projects, even if in a different format we chose to re-use that pre-existing material. Clip selection is time consuming – someone has to view and select a good starting frame and clip, consider running time and audio, administration of copyright clearance letters. How do you manage that? Cost can also affect the approach. Do you invest in equipment and skills in-house or do you outsource quickly but at a higher cost?
When it comes to choosing material to be published, ideally the Scottish Screen Archive wants to illustrate the breadth and depth of the collection, by geography, date, and genre. This is certainly true on the Scottish Screen Archive web catalogue. However the YouTube experiment asked questions which would not occur when considering the new website. For example, do we have the same duty to represent the collection fully on YouTube? Or should we pick our favourites, tailoring playlists to certain user communities?
Personal films and commercial films are often what users want to see but we don’t own the rights - so copyright can be a barrier to the very content that might be popular. At the moment we have uploaded clips infrequently on YouTube and not as a core activity. If the response to Archive material continues, and demand is high, it is possible that this might change in future.
Copyright is a very important consideration; it curtails our choice of material for use on YouTube and similar sites. The Archive has recently undertaken a project to add 700 new short clips to its website bringing the total to approximately 1000. Apart from collections which are Scottish Screen Archive copyright, every single use of every single clip has had to be cleared for this specific project. There are a wide range of copyright holders - from big business to local authorities to special interest clubs to families and their home movies. So far 99 percent have agreed to use of material – but this percentage is high because there were entire collections or companies we knew to avoid from start, mostly because they would charge for use.
Over 90 percent agreed to use of entire programmes at a later date - it is necessary to future-proofed to avoid further correspondence. These results are for use on a reputable website of a national organisation, the National Library of Scotland. We will have to approach copyright holders again to seek permission to use extracts of their films on YouTube.
We used the recent project as a pilot to define a standard specification for clips that we put on our own website - receiving input from Digital Library colleagues, the BBC and people in the industry. Every film is watermarked with the NLS logo so no matter where it appears, it is identifiable. However this does not deter all - for example Pathe watermarked content can be found on private channels. Sites such as YouTube or Facebook compress files further during the upload process but we set the benchmark for the minimum quality of our images online, aiming to establish and follow best practice for the presentation of film online.
We see YouTube acting as a kind of "filmshow" which the Archive can decide to promote and develop but with the rich metadata and quality delivery of online content still remaining within our own catalogue. YouTube's metadata works when users are within their networks of human contacts enjoying the content we spend time and money cataloguing and putting online. We should engage with our users, allow comment and ratings to help spread the word about the film archive, enable voluntary and independent metadata, entering into a partnership with our catalogue. But we cannot allow original film's intent or purpose to be changed by allowing users to edit it or change its meaning. If YouTube can enable the uploading of High Definition quality files to be uploaded this will have huge implications for the Archive of the future - users are becoming broadcasters.
In conclusion, there really is no place like home to us at the moment - our catalogue is in much better shape than YouTube's. But we think YouTube is a great place to take our filmshows, to visit and meet our users, to build relationships and to interact with them, and perhaps in this aspect, YouTube is better than any online service we yet offer.
Addendum (14 Jul 2008): This article is based on the presentation There's no place like home: YouTube and the National Library of Scotland created by Ann Cameron, Liam Paterson, and Eilidh MacGlone.
All rights reside with the author(s). WIDWISAWN recognises the author(s) as the copyright holder(s) unless specifically directed otherwise.
WIDWISAWN is delivered via the SAPIENS e-publishing service and is hosted by the Centre for Digital Library Research at the University of Strathclyde
Design and layout © CDLR 2008
Last updated: Apr 2008