MUSI-6201 | Computational Music Analysis...inference meta data feature extraction dimensionality...

overview intro content ACA summary course outline

MUSI-6201 — Computational Music AnalysisPart 2: Introduction

alexander lerch

November 4, 2015

introductionoverview

text bookChapter 1: Introduction (pp. 1–6)

sources: slides (latex) & Matlab

github repository

lecture contentwhat is audio content analysis?what are typical applications?what is audio content?what are the processing blocks of a typical ACA system?

github repository

introductionaudio content analysis — terminology

goalextract information about the content of audio data

terminologymusic information retrieval (MIR):

analysis and retrieval of music databoth audio and symbolic data

machine listening & computer audition

focus on the recognition and understanding of music

computational auditory scene analysis (CASA)

focus on human perception & cognition, understanding of theauditory scene

introductionaudio content analysis — terminology

goalextract information about the content of audio data

terminologymusic information retrieval (MIR):

analysis and retrieval of music databoth audio and symbolic data

machine listening & computer audition

focus on the recognition and understanding of music

computational auditory scene analysis (CASA)

focus on human perception & cognition, understanding of theauditory scene

introductionaudio content analysis — research field

interdisciplinarydigital signal processingmachine learning / data miningmusicologypsycho-acoustics. . .

communityISMIR: ismir.net

annual conferencescumulative list of conference papersISMIR-Community mailing listMIREX: MIR Evaluation eXchange

related publicationsconferences: ICASSP, ICME, SMC, DAFx, ACM MM, . . .journals: TASLP, Computer Music, JNMR, JAES, . . .

introductionapplications

organization in large databases

search & retrieval, classification, similarity

interfaces to search and retrieval

fingerprinting, query-by-humming systems

music visualizationsymbolic (bars, harmony, score, . . . ), similarity mappings

adaptive processingadaptive effect parametrization or algorithm selection

adaptive interactionplaylist generation, recommendation

introduction(commercial) examples

recommendation, playlist generation

fingerprinting

score following

(multi-) pitch detection

fingerprinting

score following

fingerprinting

score following

fingerprinting

score following

audio contentsources

what are the sources of (musical) audio content?

1 score:definition of musical ideas“blue-print” of the musicexamples: melody, key, harmony, rhythmic patterns, . . .

2 performance:unique acoustic renditioninformation in the score is interpreted, modified, added toexamples: (micro-)tempo, dynamics, intonation, . . .

3 production:aesthetic choicesediting & processingexamples: sound quality (EQ, microphone positioning),changes in timing and pitch

audio contenttechnical categories

audio content can be structured into 5 technical fundamentalcategories:

1 timbral: related to sound quality

examples: instrument(ation), playing technique, venue, audioprocessing, . . .

2 intensity-related: related to musical dynamics

examples: accents, loudness, . . .

3 tonal: related to pitch

examples: melody, chords, intonation, vibrato, . . .

4 temporal: related to rhythm and tempo

examples: timing, meter, rhythmic patterns, . . .

5 statistical & technical: related to signal properties

examples: amplitude distribution, number of zero crossings,. . .

audio content analysissystem overview

audiosignal

featureextraction

decision,interpretation,classification,

inference

metadata

feature extractiondimensionality reductionmeaningful representation

classificationmap or convert feature tocomprehensible domain

audiosignal

featureextraction

inference

metadata

audiosignal

featureextraction

inference

metadata

summarylecture content

what is audio content?

what are the technical categories of interest?

what are the typical processing blocks of an ACA system?

course outlineoverview 1/2

1 fundamentalsdigital audio signalsconvolution & block based processingFourier transform and filterscorrelation

2 instantaneous featuresaudio pre-processingstatistical and spectral featuresfeature post-processing

3 intensitylevel & loudness

4 tonal analysisfundamental frequencytuning frequencykey and chords

5 temporal analysisonset detectiontempo & beatdownbeat & time signature

6 genre, similarity & mood

7 alignmentaudio-to-audioaudio-to-score

8 audio fingerprinting

9 structural segmentation

MUSI-6201 | Computational Music Analysis...inference meta data feature extraction dimensionality...

Documents

Gui Contab i 6201

High dimensionality

Singer Samba 6201/6211

Dasar Teori RBS 6201

ISyE 6201: Manufacturing Systems Instructor : Spyros ... · ISyE 6201: Manufacturing Systems Instructor : Spyros Reveliotis ... ISYE 6201 Spring 2007 Homework 4 Solution 2 A. Question

Convolution Fourier Convolution - MIT OpenCourseWare · Convolution Fourier Convolution Outline • Review linear imaging model • Instrument response function vs Point spread function

QlikView Dimensionality

Dimensionality Reduction - University of Pittsburghpeople.cs.pitt.edu/~iyad/DR.pdf · Dimensionality Reduction Problems of learning in high dimensional spaces: • Curse of dimensionality

rbs 6201 fam 01_03813

JIS A 6201 1991

RBS 6201/6202 & BBS/BBU 6201 Installationdocshare01.docshare.tips/files/31145/311455292.pdf · © Ericsson AB 2010 2/038 13-LZU 108 7592 Uen Rev C 2 RBS 6201/6202 & BBS/BBU 6201 Installation

DT Convolution (1A) - Wikimedia · 19.12.2013 · DT Convolution 13 (1A) Convolution Sums in FIR Systems Computing Convolution Sums of an FIR Flip and Slide Form LTI Form Convolution

rbs 6201 02_03813

Dimensionality reduction Usman Roshan CS 675. Dimensionality reduction What is dimensionality reduction? –Compress high dimensional data into lower dimensions

6201 Cambridge Drive

STAT 6201 Midterm Exam II November 6, 2015 Nameacademics.rherbei.com/.../6201/PDF/STAT6201_midterm2_AU_2015_Solutions.… · STAT 6201 Midterm Exam II November 6, 2015 Name

22.058 - lecture 4, Convolution and Fourier Convolution Convolution Fourier Convolution Outline Review linear imaging model Instrument response function

Active Convolution, Deformable Convolution ―形状・スケールを学習可能なConvolution―

The Convolution Operationsrihari/CSE676/9.1 ConvOperation.pdf · 4.Discrete convolution viewed as matrix multiplication 3. Deep Learning Srihari What is convolution? •Convolution

Objectives: Convolution Definition Graphical Convolution Examples Properties