14 DAY TRIAL // Scientists were just given the go-ahead from sponsors and funders to embark on the largest music-related endeavor that a group has ever undertaken online. The goal of the Structural Analysis of Large Amounts of Music Information (SALAMI) project is to allow experts to analyze more than 350,000 songs at the same time, and then extract all relevant information from them. This will allow them to construct a large database featuring this knowledge, which is destined to be used by music scholars. The latter will thus be provided with an invaluable resource to learn why good music sounds good.
The project is aimed at gathering information about a wide variety of music genres, specific to many countries or regions around the globe. The leaders of the investigation, experts at the University of Southampton, in the UK, and international partners, say that diversity is the key to such endeavors. Analysts agree, saying that the boundaries between types of music are slowly beginning to fade away, as more and more elements are being adopted out of their original genres. A large number of bands are cataloged at this point as “cross-over,” which means precisely that they use elements not characteristic to their respective styles.
“To date, musical analysis has been conducted by individuals on a small scale,” explains University of Southampton School of Electronics and Computer Science professor of Computer Science David De Roure. The expert is a musician himself, and has a rich background in music retrieval, AlphaGalileo reports. “This innovative project means that the range of music analysed will be of far larger variety than anything previously done. Previous analytic research work focused primarily on Western popular and 'classical' music. Our vast dataset includes a wide variety of music from all over the world, from many time periods, and includes folk music, 'classical' music, contemporary music, improvised music, and live music,” he adds.
According to the team, the set of algorithms and research tools that will be used to extract and analyze the 350,000 songs has already been set in place. Group experts estimate that the entire endeavor will last for about 15 months, during which time most of the relevant songs will be gathered. All the money necessary for the investigation was awarded yesterday (November 3). The main funders are the Joint Information Systems Committee (JISC) in the United Kingdom, the National Science Foundation (NSF) in the United States, and the Social Sciences and Humanities Research Council (SSHRC) in Canada.