Metode za inteligentnu analizu glazbenih zapisa

master's thesis

• Title (english): Methods for intelligent analysis of music recordings
• Author: Ena Tomić
• Mentor: Antonio Petošić (mentor)
• Committee member: Antonio Petošić (predsjednik povjerenstva)
• Committee member: Krešimir Pripužić (član povjerenstva)
• Committee member: Kristian Jambrošić (član povjerenstva)
• Granter: University of Zagreb; Faculty of Electrical Engineering and Computing; Zagreb
• Defense date and country: 2018-07-16, Croatia
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Computing
• Scientific / art field, discipline and subdiscipline: TECHNICAL SCIENCES; Electrical Engineering

Abstract

U ovom radu, razmatrani su osnovni koncepti inteligentne analize glazbenih zapisa.
Najprije su predstavljeni definicija i ciljevi ove discipline, kao i povijesni razvoj. Opisani
su primarni deskriptori, relevantni za analizu (spektar, kepstar, CHROMA, MFCC
koeficijenti, prolasci kroz nulu, spektralni centroid, i dr.), te metode za predprocesiranje
signala, uključujući i pretvorbu zapisa u digitalni oblik. Zatim su navedene dostupne
baze glazbenih zapisa, kao i alati pogodni za analizu glazbe. Opisane su glavne
metode za ekstrakciju odabranih glazbenih obilježja: ritma, tonaliteta i teksture. Prikazani
su praktični primjeri ekstrakcije značajki dva glazbena zapisa različitog žanra
pomoću MIRToolbox alatnog okvira, te su napravljene analiza i usporedba.

Abstract (english)

The thesis discusses the basic concepts of intelligent analysis of music records. First,
the definition and objectives of this discipline as well as historical development are presented.
Low level descriptors, relevant for analysis (spectrum, cepstrum, CHROMA,
MFCC coefficients, zero crossing rate, spectral centroid, etc.) are described, as well
as methods for signal preprocessing, including digitalization process. Furthermore,
there are freely available music datasets as well as analysis tools described. The main
methods for extraction of selected musical features are discussed: rhythm, tonality and
texture. Practical examples of feature extraction of two music tracks from different
genres are presented using the MIRToolbox, and an analysis and comparison are made.

• Keywords: inteligentna analiza glazbe; digitalizacija; predprocesiranje; MIRToolbox; baze glazbenih zapisa; ekstrakcija glazbenih značajki; ritam; tonalitet; tekstura; kromagram; primarni deskriptori; MFCC koeficijenti
• Keywords (english): intelligent music analysis; digitalization; preprocessing; MIRToolbox; music databases; music features; feature extraction; rhythm; tonality; texture; chromagram; low level descriptors; MFCC coefficients
• Language: croatian
• URN:NBN: urn:nbn:hr:168:179162
• Study programme: Title: Information and Communication Technology; Study programme type: university; Study level: graduate; Academic / professional title: magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije (magistar/magistra inženjer/inženjerka informacijske i komunikacijske tehnologije)
• Type of resource: Text
• File origin: Born digital
• Access conditions: Closed access
• Created on: 2019-03-14 01:36:31