Tracking the Tunes at the Cobblestone
A new project by a team based in Dublin Institute of Technology and University College Dublin seeks to document the tunes played at the Cobblestone pub’s traditional music sessions. TuneTracker is a software system developed by Dr Norman Makoto Su and Dr Bryan Duggan that will keep a record of the tunes played in any session. The list of tunes played will be then posted on the project website.
The organisers believe this trial system will be of use to both professional and amateur musicians. ‘Perhaps the most daunting task for anyone starting in traditional sessions is the sheer number of tunes everyone seems to know by heart,’ they say. ‘By making the tunes played in a session explicit, amateurs can intelligently practice certain (e.g. most popular) tunes in a session.
‘We also feel that professionals will gain by seeing an explicit record of the tunes being played. They might consider the repertoire of their session, adjust the tunes they play, or be able to recommend their session to others. TuneTracker moreover has the potential to attract newcomers to sessions. In the future, we might imagine TuneTracker deployed in sessions all over the world, unveiling the process of tradition globally.’
The TuneTracker system makes use of the Tunepal application, developed by Duggan previously. Tune sources used are Henrick Norbeck’s collection of tunes, and Johnny O’Leary of Sliabh Luachra, a collection edited by Terry Moylan.
The system is located near the ceiling in the area of the Cobblestone by the front entrance where traditional music sessions take place. Musicians concerned about their privacy and copyright issues might be reassured to know that all recordings are deleted after a tune is recognised; also, no musician names are recorded or published with the data. ‘We recognise still that musicians may feel that such a system intrudes on tradition,’ say the organisers, however. ‘Because of this, we will be running this on a trial basis — as an interesting social experiment to see how this might benefit or prove a detriment to sessions. We ask that people keep an open mind!’
How accurate is the TuneTracker? ‘TuneTracker will try its best to determine whether music is being played (versus chatter) and then detect the tune name from the music,’ say Makoto Su and Duggan. ‘However, factors such as ambient noise, false starts, microphone placement, power outages, the lack of the transcribed tune in the collections above, etc., can all effect the recognition algorithm.’
For the latest list of tunes played at the Cobblestone, see here.