The performance of melody retrieval using a query-by-humming (QBH) system depends on different parameters. For the query, parameters like length of the query and possibly contained errors influence the success of the retrieval. But also the size of the melody database inside the QBH-system has a certain impact on the query. This paper describes how the statistical parameters of a random melody database are modelled to get the same behaviour like a database containing authentic melodies. Databases containing random melodies are a testing facility to QBH-systems.

Authors: Batke, Jan-Mark; Eisenberg, Gunnar
Affiliations: Fachgebiet Nachrichtenübertragung, Technische Universität Berlin, Germany; THOMSON, Corporate Research, Audio Processing Lab(See document for exact affiliation information.)
AES Convention: 120 (May 2006) Paper Number: 6727
Publication Date: May 1, 2006 Import into BibTeX
Subject: Audio Archiving, Storage, Restoration, and Content Management
Permalink: https://www.aes.org/e-lib/browse.cfm?elib=13531

Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!

This paper costs $33 for non-members and is free for AES members and E-Library subscribers.

Learn more about the AES E-Library

E-Library Location: (CD 120Papers)   /120/6727.pdf

Start a discussion about this paper!