Output details
11 - Computer Science and Informatics
Goldsmiths' College
mHashup: fast visual music discovery via locality sensitive hashing
<17> SIGGRAPH demo of Mhashup developed through EPSRC OMRAS2 grant. This mode of working was seminal in defining Goldsmiths Computing research, bringing together Casey’s algorithms and Magas’s art, producing a system for visualising audio collections. Mhashup visualises content-based audio search via audio 'shingles', overlapping time-series features used in our Locality Sensitive Hashing algorithms. Work led to support from Google (Faculty Research Award in 2011 for Casey) for information retrieval; also Technology Strategy Board funding for Magas for Online Platform for Music Discovery; and to the formation of a company, Sonaris (somaris.org). See videos (including BBC coverage) at http://www.omras2.org/mHashup.