|Abstract||In this paper we describe our submission for the audio melody extraction task of the Music Information Retrieval Evaluation eXchange (MIREX) 2011 campaign. The system presented here is an updated version of the one submitted to last year's campaign. Following a detailed analysis of each step of our method, system parameters have been optimised for melody extraction and the implementation is now more efﬁcient. Two variants of the system have been submitted, each making use of a different spectral transform, allowing us to asses whether the difference between them is significant for overall performance.
Following the description of the system, we describe the data-sets and metrics used for evaluation. This is followed by a summary of the results and some conclusions.