SoundSoftware 2013: Workshop on Software and Data for Audio and Music Research

The second SoundSoftware.ac.uk one-day workshop on “Software and Data for Audio and Music Research” will include talks and discussions on issues such as robust software development for audio and music research, reproducible research in general, management of research data, and open access.

Details

• Location: Queen Mary University of London (Maths Lecture Theatre)
• Date: Wednesday 26 June 2013
• Registration: closed (deadline was 24 June 2012)

Programme

Oral session 1

Ten reasons you must publish your source

Nick Barnes, Climate Code Foundation
[Video]

A quick survey of the changing landscape for publishing software developed for research purposes. Since Nick Barnes created the Science Code Manifesto in 2011, the shifts that were already under way have accelerated, and the future of science as a dynamic, open, collaborative enterprise is becoming clear. There is an increasingly powerful case for open sharing and reuse.

Caveat Emptor: academic software and user experience

Jamie Bullock, Birmingham Conservatoire
[ Video | Slides ]

The role of programming in academic research is diverse, ranging from short experimental scripts to large-scale applications serving as a independent research outputs. Regardless of scale, user experience (UX) is often given a low priority. Researchers commonly develop software for themselves, and if it is released publicly, the code is provided “as is” with little thought given to the experience of the end user. Another scenario is the institutional “flagship project” involving software arising from years of development, with a proportionate user and developer community. In these cases basic details of user experience are also often neglected, and the assumption is made that other researcher-users will “make it work”. If software and its surrounding ecosystem do not provide good user experience, it will ultimately be less widely adopted and the productivity of users will be encumbered. This translates into less effective research with lower impact. In this talk I will discuss the state of the art in user experience design, and suggest ways in which the UX of academic software can be improved.

Programming Gloves and Debugging the Brooklyn Bridge: Lessons Learned from Developing for Interactive Installations

Becky Stewart, Codasign
[ Video | Slides ]

In 2011, Adam Stark and Becky Stewart ventured outside the Centre for Digital Music where they had both completed their PhDs. For the past two years they have been developing software and hardware for creative projects, often collaborating with artists, designers, and musicians. They will highlight what they have learned as they have taken academic research and directly applied it to commercial projects, what differs from software development for an academic setting versus deployment in public settings, and what tools commonly used by the creative technology community may be useful to academic research.

Oral session 2

British Library Labs

Mahendra Mahey, British Library Labs
[Video]

Mahendra Mahey, manager of a new initiative called British Library Labs, will give an overview of the project which is about engaging researchers and developers to do innovative research and development with British Library Digital collections through various activities such as competitions, hack events, virtual activities, meetings. Mahendra will give an overview of some of the Digital Collections that are being used with Labs including audio collections and possible ideas on how they could be used in research.

Tunebot: Software development, data sharing and data protection in an academic environment.

Bryan Pardo, Northwestern University
[Video]

Building reproducible, sharable, robust software and data collections of music in an academic environment presents many challenges. These include high turnover rate of workers; an inexperienced workforce; a culture that places low value on robust code, documentation or reproducibility; copyright concerns for music collections; privacy concerns for study participants. Professor Pardo will discuss how these challenges have been faced in the context of the Tunebot online music search engine, an NSF-funded research project developed in his lab at Northwestern University in Chicago, IL.

The BBC Desktop Jukebox: A large-scale music recommendation system and user trial

Mathieu Barthet, Queen Mary, University of London
[Video]

Recommending music for professional use in media and broadcast productions presents challenges that are substantially different from those faced by systems adapted for recreational listeners. Television and radio producers often need tracks which sound similar to other tracks ("something that sounds like…") or which suggest specific emotions or moods (e.g. "some happy commercial music for…").

In this talk, Mathieu Barthet will present some content– and metadata– based recommendation technologies addressing these issues which were developed as part of the 'Making Musical Mood Metadata' project (Technology Strategy Board) in collaboration with BBC R&D and the music provider I Like Music. He will show how the code.soundsoftware platform was used to collect user feedback data to assess a prototype software integrated to the online music service 'BBC Desktop Jukebox'.

Oral session 3

Industrial Strength Software: Research-Driven Software in a Commercial Environment

Angus Hewlett, FXpansion
[Video]

The end-point for audio and music research is very often a commercial or consumer product - but there are many ways it can get there. FXpansion has been deploying audio R&D to music production professionals & hobbyists for more than a decade, and I will be discussing some of the possible models for doing so and the technical challenges encountered on the way.

Learning how to learn, with Software Carpentry

Luis Figueira, SoundSoftware.ac.uk
[ Video | Slides ]

Software development is almost ubiquitous in audio and music research. However, researchers are often self trained in software development, which may be a cause of the lack of confidence in their code that has been reported as a significant obstacle to software publication and public collaboration.

Although software development is a deep subject, our belief is that worthwhile improvements to normal working practice can follow relatively small amounts of training. As a consequence we have been organising short bootcamps on software development good practice for scientists and researchers, in collaboration with Software Carpentry, a voluntary organisation funded by Mozilla and the Sloan Foundation.

In this presentation we will describe how these bootcamps are structured, review their impact and benefits, discuss which aspects of the bootcamps are most (and least!) effective, and talk about lessons learned while organising and delivering them.

Registration

Registration is now closed. Please drop us an email to info@soundsoftware.ac.uk if you'd like to attend the workshop.

Workshop Venue

The Workshop will take place in the Maths Lecture Theater, (Mathematical Sciences building, Queen Mary University of London's Mile End campus).

Please download the following campus map (pdf), with the location of the Mathematical Sciences building.

For further information on the campus, and how to get to Queen Mary University, please follow this link: Mile End Campus.

Additional info

This workshop follows the successful SoundSoftware 2012 held last June; see the programme and footage from last year's event.

This workshop is the day after the Listening in the Wild workshop, also at Queen Mary.

Support

This workshop is supported by the UK Engineering and Physical Sciences Research Council (EPSRC).