Saving the Sounds of the UK in the UK SoundMap
The impact of the digital age upon libraries has been profound, changing not only the back office, services, and the range of materials available to users, but also the public face of libraries and the relationship between the library and its users. Within this changed relationship, collaboration, participation, and online social networks play an increasingly important role in the user experience, especially in large university and national libraries. At the same time, a shift is taking place in the type of collection items held in libraries, and the percentage of born-digital materials acquired is increasing on a daily basis.
The British Library is no exception, making use of a wide range of online services and tools to engage with users and enhance access to the collections, both digitised and born-digital. Numerous initiatives are currently taking place across the Library to engage with users and address these changes, and one in particular has sought to capitalise on both the increase in participatory networks and the opportunities afforded by born-digital material. This initiative is the UK SoundMap, an online crowd-sourcing activity driven by the British Library in partnership with the Noise Futures Network to engage and build the community in development of a new born-digital audio-visual research resource [1][2].
Unlocking and Integrating Audio-Visual Content at the British Library
The UK SoundMap is being carried out as part of a wider project by the British Library’s Sound & Vision Department: Unlocking & Integrating Audio Visual Content (UIAVC). The UIAVC project seeks to address changing user needs in a multi-media research environment by establishing the building blocks for a redefined and integrated sound and moving image service within the Library. Other, complementary initiatives in the project include:
- Modernising and enhancing interactive features in the existing Archival Sound Recordings (ASR) portal, which currently provides access to over 45,000 selected recordings of music, spoken word, and human and natural environments [3].
- Establishing a New Music Network to select and capture content from musicians whose work is produced outside the usual commercial channels
- Exploring and piloting new R&D Tools to improve resource discovery through new search and analysis tools for speech and music [4]
- Increasing the amount of digital audio and video content accessible to users at the Library (i.e. onsite), and remotely
Overall, the project is key to meeting the Library’s audio-visual strategy, which aims to unlock and integrate audio-visual content across the library according to user needs. The initiatives interrelate to a significant degree as they each follow the content path from acquisition to curation to integrated delivery. They each focus on digital content (both born-digital and digitised analogue content), they embrace both onsite and remote (Web) access, and collectively they express the commitment the Library now has towards integrating audio-visual media within the research experience.
The UK SoundMap: Mobile, Crowdsourced Curation
The UK SoundMap uses innovative methods to engage users in recording and contributing relevant born-digital field recordings using mobile technology for public access both onsite and online. The ability to share information in real time and at the right place is changing the way we run our lives and organise our thoughts. Mobile social networking has become a commonplace 21st-century activity, currently accounting for 50% of all mobile Internet traffic in the majority of developed markets, and over 50% in a number of emerging markets [5]. Whilst opinions may vary about the value of academic and scientific research generated over mobile social networking platforms, no one can deny the potential offered by low-cost, even free applications for generating and recording data in a range of formats ‘in the wild’, particularly for those who are motivated to extend the corpus of human knowledge.
UK SoundMap breaks new ground through the deployment of several technologies at once. Maps are a powerful visual alternative to traditional text search and browse, particularly where geo-tagged files are concerned. UK SoundMap uses crowd-sourcing to acquire soundscape recordings, images and metadata, and social networking to stimulate engagement with users. It uses the ubiquitous mobile phone and the free Audioboo app, coupled with realtime curation and GoogleMaps browsable interface – all at a very low cost – to deliver a valuable dataset of everyday sounds for preservation and future reuse.
UK SoundMap project home page
The project has three main objectives:
- To explore the potential for mash-ups using data accumulated by cutting edge applications to build significant resources for digital scholarship
- To map the evolution of the national landscape and how people feel about it
- To involve the public in contributing to British Library acquisitions of research material.
We are now roughly halfway through the project and aim to complete by the end of June 2011.
The Technology
The popular Audioboo service provides the underlying audio file creation and sharing technology [6]. Users need an Audioboo account, which can be created independently at the Audioboo Web site or using Twitter credentials. Most audio files (so-called ‘Boos’) are created using a smart phone application – Audioboo apps for both the iPhone and Android phones are currently available free of charge - though a Web-based interface is also available [7]. A particular advantage of smart phone use is the automatic inclusion of geo-location data, enabling files to be accurately plotted on the GoogleMap interface.
Guidance on creation of Boos is available on the UK SoundMap Web site, including how to avoid unwanted audio interference when using recording devices outdoors. Users are asked to title their Boos appropriately, to provide location details if these have not been automatically added, to tag entries with relevant keywords, and to add the all-important #uksm tag, which enables all Boos submitted to the UK Sound Map to be monitored and moderated by Library staff. Once selected, recordings are given an additional tag (a ‘magic tag’) that allows Library staff to list all files destined for inclusion in the SoundMap in an RSS feed.
Boos are converted from their original upload formats (usually FLAC) to MP3 format and plotted onto a Google Maps interface, embedded into the UKSM Web site. On accessing the map in a browser, red pins identify locations against which Boos have been uploaded. Clicking on a single pin generates a pop-up bubble with details of the Boo, namely the title, duration of the recording, username and tags. A simple audio player is embedded in the page that enables the audio file to be played and heard by the user without leaving the Web site or requiring any other technology. A search facility is available and results are highlighted on the map using large blue pointers.
Screenshot of UK SoundMap
All submissions are acknowledged by moderators and promoted over the Audioboo and Twitter networks via the @uk_soundmap account.
The Crowd
There are currently more than 300 different contributors to the UK SoundMap. The basic profiling data we have suggests that the crowd is not heavily dominated by one particular gender group, and that a reasonable spread of age ranges is represented. However, little hard demographic detail is available as there is no requirement for users to provide complete profiles and background details. As a result, there are currently no plans to harvest user profiles from Audioboo, though this may change as the project matures and we gain a better understanding how people’s interests and personal circumstances impact on the type of sounds they choose to record and submit.
As the success of the project depends on the participation of users, the technologies used in the project were selected not least because of their suitability for the task at hand, but also their user-friendliness. We want submissions from as many contributors as possible, regardless of their background, experience, age, profession, and interests. The registration and participation process is quick and simple, in order to encourage contributions, and new contributors are welcome to join the project at any time [8].
The Content
Since the project launched in mid-2010, over 1,200 files have been successfully uploaded to the SoundMap. The types of sounds represented on the map are a fascinating mix of old and new – though admittedly, perhaps less of the new than one may expect. The typical UK SoundMap recording is made during the daytime, in a built-up urban area, in the street, with one or more voices present, and the sound of traffic in the background. More than 90% of all recordings have been made during the day, and half of them in cities and large towns [9]. However, 27% were also made in villages, small towns and suburbs; 12% in the countryside; and 11% by the sea and in seaside resorts. Within these broad geographical categories are more precise settings, ranging from fields and footpaths to homes and gardens, transport hubs, commercial establishments, and leisure or heritage locations. Smaller numbers have come from settings as diverse as boats, places of worship, street markets, public demonstrations, and festivals.
Settings in the UK SoundMap
The voice in some form is the most common sound, appearing in over half of all recordings. In 40% it consists of a single voice (other than that of the recordist) or several voices, and in 10% a hubbub of many voices. 9% include the sound of an electrically amplified voice, including loudhailers, PA or tannoy systems, radios and televisions. Female voices predominate in public announcements. A disproportionate number of street recordings are made in pedestrianised areas too, where the speech of passers-by can rise above the background growl of traffic, though sounds of transport dominate the London recordings. In another 9% of recordings the recordist’s own voice is heard, often describing the scene for the benefit of the listener. But this rarely occurs in built-up areas. More often the recordist speaks at home or when in the countryside, sometimes with humour, sometimes to express pleasure at what they are hearing.
The sounds of water occur in nearly 15% of recordings, including rain, dripping taps, fountains, the guttural noises of drains, and the sound of the sea. Water in its gassy state is represented with nearly a dozen recordings of steam-driven vehicles, and the sounds of coffee machines and boiling kettles. We are also gathering disappearing sounds; as the landscape and society changes, so do the sounds associated with them. Whistling, for example, is hardly represented at all. As more contributions come in, so the dataset will grow richer, and more in-depth statistical explorations will become possible.
Technical quality is typically lower than in professional sound recordings. This is not surprising given the use of commonplace recording devices rather than professional audio recording equipment, as well as the lack of professional training typically experienced by contributors. For the purposes of this project however, a lower technical quality baseline is acceptable and only around 7% of submissions are rejected, often for reasons other than quality (for example, inclusion of copyrighted music/speech or obscenities).
Metadata is a challenge for this project. Whilst all contributors include the #uksm tag, few add keywords distinctly different to the basic information already contained in the title. As a result, we are not getting the additional value that the research community requested and may have to consider options for adding this retrospectively before the files are archived for long-term preservation.
Risks and Issues
Most of the risks in the initiative derive from the crowd-sourcing element – for example, special interest groups may exploit the initiative for political gain, whilst other content may be submitted that is deemed inappropriate because it infringes personal privacy or copyright. This can be mitigated with appropriate moderation and intervention.
The dependence of the initiative on third parties is also a risk – particularly given that we rely not just upon external contributors for the content but also the Audioboo service and Google Maps for the underlying technology. Despite this, the project is considered to be, by and large, a low-risk project, due to the strength of the Audioboo app and the Google Maps technology, as well as the established SoundMap community. Perhaps a larger risk is that we do not engage with the critical mass of contributors needed to gather a truly representative soundscape of the UK; for this reason we plan to embark on a new publicity drive for the remaining term of the project and stimulate new interest.
We are currently working to solve two technical issues. The primary issue at the moment is the collection of files that have had the #uksm tag added retrospectively. The current harvesting method does not gather these and the team is working to develop a new, more inclusive ‘catch-all’ approach. The other concerns the file formats – the MP3s are generated for access purposes only and we will archive the original audio files with their original metadata. As the format of the original files may vary, we must explore the impact this may have on our long-term preservation approach and our options for normalisation.
Conclusion
The UK SoundMap is an innovative and unique example of crowd-sourced curation, using mobile apps to generate a new corpus of audio research data tied to specific geographical locations over a 12-month period. The coupling of moderation and support from British Library staff with third-party applications and contributors has proven to be a rapid and cost-effective way to collect large amounts of data that augment existing research collections. Though selection (in the guise of moderation) remains a manual activity, the process has proven an efficient and enjoyable way to engage the public with our collection and ‘democratise the curator’s role’.
The success of the initiative has already been clear – the rate of submissions is steady, and the team was delighted to win the Public Sector category at the 2010 some comms UK Social Media Awards in December [10]. Submissions to the UK SoundMap will be accepted up until the end of June 2011 – why not add your sounds whilst you can and contribute to the UK SoundMap?
References
- UK SoundMap http://sounds.bl.uk/uksoundmap/
- Noise Futures Network is an UK-based interdisciplinary research group http://www.noisefutures.org/
- Archival Sound Recordings portal http://sounds.bl.uk/
- The New Music Network pilot project was recently carried out in the West Midlands prior to wider rollout
http://www.pilot-project.co.uk/ - Perez, Sarah ‘Social Networking More Popular Than Voice, SMS by 2015’, ReadWriteWeb, 15 November 2010 (Accessed 4 January 2011)
http://www.readwriteweb.com/mobile/2010/11/social-networking-more-popular.php - Audioboo Web site http://audioboo.fm/
- Between July – October 2010, 82% of recordings were made with mobiles.
- Information on how to participate is available at http://sounds.bl.uk/uksoundmap/takepart.html
- Figures relate to the month of November 2010.
- More information on the somecomms awards is available from http://www.dontpanicprojects.com/somecomms/