Font Size: a A A

Research On Fast Mathing Technologies Of Musical Melody

Posted on:2008-08-18Degree:MasterType:Thesis
Country:ChinaCandidate:L GaoFull Text:PDF
GTID:2178360212976049Subject:Speech recognition
Abstract/Summary:PDF Full Text Request
Query by humming, as a brand-new music retrieval method, can help user to locate the wanted piece within a huge music repository by simply sing several tones. Being a new style of human machine interface, the query by humming technology has important and practical meanings to content based music retrieval online. Query by melodic feature is one of the most important fields in query by content. In this field, how to fast match two pieces of melody is one of the most challenging questions. And the solution to this question is the key technology for the large scale QBH retrieval system to be practical.To query by content in large music library, in this paper we explore a new melody match method for digital music, called Layered Fast Matching Algorithm. It uses the N-Gram method to classify melodies, which is widely used in string matching. It works efficiently and robustly. Experiments are taken to test the new algorithm. 190 humming clips are searched in a music database comprising 3864 songs to test its performance compared with the LAM (Linear Alignment Matching). For same humming clips, cost time of our algorithm is approximately less than 1/3 of the LAM's. This experiment result shows the feasibility of using the...
Keywords/Search Tags:Content-based Music Retrieval, Query by Humming, Layered Fast Notes Matching Algorithm, N-Gram Algorithm for Notes Matching
PDF Full Text Request
Related items