國立高雄大學圖資館 |

Language: English

Back

The shape of an instant: Measuring a...

Stanford University.

The shape of an instant: Measuring and modeling perceptual attack time with probability density functions (if a tree falls in the forest, when did 57 people hear it make a sound?).

Record Type:	Electronic resources : Monograph/item
Title/Author:	The shape of an instant: Measuring and modeling perceptual attack time with probability density functions (if a tree falls in the forest, when did 57 people hear it make a sound?).
Author:	Wright, Matthew James.
Description:	188 p.
Notes:	Adviser: Chris Chafe.
Notes:	Source: Dissertation Abstracts International, Volume: 69-02, Section: A, page: 0436.
Contained By:	Dissertation Abstracts International69-02A.
Subject:	Music.
Online resource:	http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3302886
ISBN:	9780549489351

The shape of an instant: Measuring and modeling perceptual attack time with probability density functions (if a tree falls in the forest, when did 57 people hear it make a sound?).
Wright, Matthew James.

The shape of an instant: Measuring and modeling perceptual attack time with probability density functions (if a tree falls in the forest, when did 57 people hear it make a sound?). - 188 p.

Adviser: Chris Chafe.

Thesis (Ph.D.)--Stanford University, 2008.

A sound's Perceptual Attack Time ("PAT") is its perceived moment of rhythmic placement. PAT is notoriously difficult to measure, because all methods measure a test sound's PAT in relationship to an action or second sound, which add uncertainty to the measurements. This work uses the ideal impulse, which is an ideal reference because it is perfectly isolated in time with a very clear attack. However, its perfectly broad frequency spectrum is problematic because it is harder to perceive the relative timing of sounds when their spectra differ greatly. This motivates Spectrally Matched Click Synthesis, the creation of arbitrarily short clicks whose spectra approximate those of arbitrary input sounds.

ISBN: 9780549489351Subjects--Topical Terms:

227185
Music.

The shape of an instant: Measuring and modeling perceptual attack time with probability density functions (if a tree falls in the forest, when did 57 people hear it make a sound?).
LDR:03382nmm _2200313 _450 001 206836
005 20090413125731.5
008 090730s2008 ||||||||||||||||| ||eng d
020 $a 9780549489351
035 $a 00372048
040 $a UMI $c UMI
100 $a Wright, Matthew James. $3 321770
245 1 4 $a The shape of an instant: Measuring and modeling perceptual attack time with probability density functions (if a tree falls in the forest, when did 57 people hear it make a sound?).
300 $a 188 p.
500 $a Adviser: Chris Chafe.
500 $a Source: Dissertation Abstracts International, Volume: 69-02, Section: A, page: 0436.
502 $a Thesis (Ph.D.)--Stanford University, 2008.
520 $a A sound's Perceptual Attack Time ("PAT") is its perceived moment of rhythmic placement. PAT is notoriously difficult to measure, because all methods measure a test sound's PAT in relationship to an action or second sound, which add uncertainty to the measurements. This work uses the ideal impulse, which is an ideal reference because it is perfectly isolated in time with a very clear attack. However, its perfectly broad frequency spectrum is problematic because it is harder to perceive the relative timing of sounds when their spectra differ greatly. This motivates Spectrally Matched Click Synthesis, the creation of arbitrarily short clicks whose spectra approximate those of arbitrary input sounds.
520 $a Subjects downloaded a listening test that presents a series of PAT measurement trials and allows them to adjust sounds' relative timing until they sound synchronous. This establishes perceptual "ground truth" for the PAT of a collection of 20 sounds compared against each other in various combinations. As hoped, subjects were able to align a sound more reliably to one of that sound's spectrally matched clicks than to other sounds of the same duration.
520 $a The representation of PAT with probability density functions provides a new perspective on the problem of predicting PAT directly from acoustical signals. Rather than choosing a single moment for PAT given a segment of sound known a priori to contain a single musical event, these regression methods estimate continuous shapes of PAT distributions from continuous (not necessarily presegmented) audio signals, formulated as a supervised machine learning regression problem whose inputs are DSP functions computed from the sound. This work concludes with some preliminary musical applications of the resulting models.
520 $a This work represents PAT as a function indicating the probability that a typical listener would hear the sound's PAT at each time, rather than as just one instant. The problem of deriving each sound's PAT from pairwise comparisons becomes one of estimating distributions of random variables for every sound's PAT given only observations of random variables corresponding to the difference two sounds' distributions plus noise. Methods to address this draw from maximum likelihood estimation and the shortest path problem.
590 $a School code: 0212.
650 $a Music. $3 227185
650 $a Psychology, Experimental. $3 212824
650 $a Computer Science. $3 212513
690 $a 0413
690 $a 0623
690 $a 0984
710 $a Stanford University. $3 212607
773 0 $g 69-02A. $t Dissertation Abstracts International
790 $a 0212
790 1 0 $a Chafe, Chris, $e advisor
791 $a Ph.D.
792 $a 2008
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3302886 $z http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=3302886