東華大學圖書館 |

Acoustic Measurement in Phonetics: Current Practices and Future Directions /

Record Type:	Electronic resources : Monograph/item
Title/Author:	Acoustic Measurement in Phonetics: Current Practices and Future Directions // Emily Grabowski.
Author:	Grabowski, Emily,
Description:	1 electronic resource (87 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 86-03, Section: B.
Contained By:	Dissertations Abstracts International86-03B.
Subject:	Linguistics. -
Online resource:	https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31294095
ISBN:	9798384452577

Acoustic Measurement in Phonetics: Current Practices and Future Directions /
Grabowski, Emily,

Acoustic Measurement in Phonetics: Current Practices and Future Directions /Emily Grabowski. - 1 electronic resource (87 pages)

Source: Dissertations Abstracts International, Volume: 86-03, Section: B.

Acoustic analysis is a key component of scientific investigation in phonetics, and offers a flexible approach with significant freedom for researchers to decide what and how to measure. In some cases, methodological variation affects results to the point of qualitatively different conclusions (as in Roettger et al. 2019 and Coretta et al. 2023), which highlights a need for further study of the quality and consistency of acoustic analysis. Current approaches are typically focused on specific phonetic phenomena, and the field lacks a holistic understanding of the usage and development of phonetic methodology in the field.This dissertation will focus on understanding and evaluating methods of acoustic measurement in phonetics. The first goal is to understand the current state of the field, the second is to quantify the relative performance of acoustic measurements, and the third is to outline key considerations for future development of acoustic measurements and point to promising future directions. Chapter 1 describes a broad overview of the types of measurement that are possible in acoustic phonetics. This chapter is organized based on how the measurement is derived from speech and outlines key factors influencing the formulation, usage, and adoption of these methods. In particular, this chapter highlights the incredible variation possible when it comes to choosing acoustic measures. Chapter 2 investigates the practical usage of acoustic measurement strategies via a quantitative methodological study. This chapter illustrates a decrease in studies regarding consonants and an increase in those pertaining to vowels, and other results indicate a general lack of confidence in acoustic measurement of consonants. In addition, a relatively small number of acoustic measurement strategies make up most of the total strategies in the study, despite the wide range available. Taken together, these chapters highlight a fundamental challenge in phonetic methodology - a pattern where a small subset of methods dominate the field despite a wide variety of options for acoustic measurement. This aligns with general intuitions in the field of best practices that are robust but cover a small set of phonetic phenomena of interest, and supports the conclusion that further development of acoustic measurement is of benefit to the field.The second section of the dissertation quantifies the performance of acoustic features in phonetics via two experiments. Chapter 3 evaluates how much incorporating context improves acoustic representation. The results of this experiment are two-fold: firstly context does seem necessary for capturing at least some phonetic categories, and secondly due to long-range context effects, selecting a model should be done with care and attention. Chapter 4 broadens this methodology to create a general benchmark for quantifying the performance of acoustic features. This benchmark and experimental design provides a framework for quantitatively evaluating the performance of acoustic representations, and is positioned to streamline acoustic parameter selection methods.Chapter 5 concludes this dissertation by presenting an informed discussion of the future of acoustic features in phonetics. I begin by conceptualizing the key properties of a good phonetic feature: performance, interpretability, and accessibility, and discuss operationalization of these properties for evaluating features in the field. Finally, I use this conceptualization to identify one candidate for phonetic feature representation, a neural network with potential to have high performance, interpretability, and accessibility, as an example of one potential avenue of implementing the findings described in the previous sections to tangibly improve acoustic representations in phonetics.In summary, this dissertation presents an argument for continued development of acoustic measurement in the field and provides a broad, concrete framework to assist in future development, testing, and selection of acoustic representations in phonetics.

English

ISBN: 9798384452577Subjects--Topical Terms:

524476
Linguistics.
Subjects--Index Terms:

Phonetics

Acoustic Measurement in Phonetics: Current Practices and Future Directions /
LDR:05526nmm a22004453i 4500 001 2434328
005 20250915135049.5
006 m o d
007 cr|nu||||||||
008 261011s2024 miu||||||m |||||||eng d
020 $a 9798384452577
035 $a (MiAaPQD)AAI31294095
035 $a AAI31294095
040 $a MiAaPQD $b eng $c MiAaPQD $e rda
100 1 $a Grabowski, Emily, $e author. $3 3824278
245 1 0 $a Acoustic Measurement in Phonetics: Current Practices and Future Directions / $c Emily Grabowski.
264 1 $a Ann Arbor : $b ProQuest Dissertations & Theses, $c 2024
300 $a 1 electronic resource (87 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 86-03, Section: B.
500 $a Advisors: Johnson, Keith Committee members: Begus, Gasper; Sande, Hannah; Anumanchipalli, Gopala.
502 $b Ph.D. $c University of California, Berkeley $d 2024.
520 $a Acoustic analysis is a key component of scientific investigation in phonetics, and offers a flexible approach with significant freedom for researchers to decide what and how to measure. In some cases, methodological variation affects results to the point of qualitatively different conclusions (as in Roettger et al. 2019 and Coretta et al. 2023), which highlights a need for further study of the quality and consistency of acoustic analysis. Current approaches are typically focused on specific phonetic phenomena, and the field lacks a holistic understanding of the usage and development of phonetic methodology in the field.This dissertation will focus on understanding and evaluating methods of acoustic measurement in phonetics. The first goal is to understand the current state of the field, the second is to quantify the relative performance of acoustic measurements, and the third is to outline key considerations for future development of acoustic measurements and point to promising future directions. Chapter 1 describes a broad overview of the types of measurement that are possible in acoustic phonetics. This chapter is organized based on how the measurement is derived from speech and outlines key factors influencing the formulation, usage, and adoption of these methods. In particular, this chapter highlights the incredible variation possible when it comes to choosing acoustic measures. Chapter 2 investigates the practical usage of acoustic measurement strategies via a quantitative methodological study. This chapter illustrates a decrease in studies regarding consonants and an increase in those pertaining to vowels, and other results indicate a general lack of confidence in acoustic measurement of consonants. In addition, a relatively small number of acoustic measurement strategies make up most of the total strategies in the study, despite the wide range available. Taken together, these chapters highlight a fundamental challenge in phonetic methodology - a pattern where a small subset of methods dominate the field despite a wide variety of options for acoustic measurement. This aligns with general intuitions in the field of best practices that are robust but cover a small set of phonetic phenomena of interest, and supports the conclusion that further development of acoustic measurement is of benefit to the field.The second section of the dissertation quantifies the performance of acoustic features in phonetics via two experiments. Chapter 3 evaluates how much incorporating context improves acoustic representation. The results of this experiment are two-fold: firstly context does seem necessary for capturing at least some phonetic categories, and secondly due to long-range context effects, selecting a model should be done with care and attention. Chapter 4 broadens this methodology to create a general benchmark for quantifying the performance of acoustic features. This benchmark and experimental design provides a framework for quantitatively evaluating the performance of acoustic representations, and is positioned to streamline acoustic parameter selection methods.Chapter 5 concludes this dissertation by presenting an informed discussion of the future of acoustic features in phonetics. I begin by conceptualizing the key properties of a good phonetic feature: performance, interpretability, and accessibility, and discuss operationalization of these properties for evaluating features in the field. Finally, I use this conceptualization to identify one candidate for phonetic feature representation, a neural network with potential to have high performance, interpretability, and accessibility, as an example of one potential avenue of implementing the findings described in the previous sections to tangibly improve acoustic representations in phonetics.In summary, this dissertation presents an argument for continued development of acoustic measurement in the field and provides a broad, concrete framework to assist in future development, testing, and selection of acoustic representations in phonetics.
546 $a English
590 $a School code: 0028
650 4 $a Linguistics. $3 524476
650 4 $a Language. $3 643551
650 4 $a Acoustics. $3 879105
650 4 $a Sociolinguistics. $3 524467
653 $a Phonetics
653 $a Acoustic analysis
653 $a Consonants
653 $a Phonetic methodology
653 $a Vowels
690 $a 0290
690 $a 0679
690 $a 0636
690 $a 0986
710 2 $a University of California, Berkeley. $b Linguistics. $e degree granting institution. $3 3824279
720 1 $a Johnson, Keith $e degree supervisor.
773 0 $t Dissertations Abstracts International $g 86-03B.
790 $a 0028
791 $a Ph.D.
792 $a 2024
856 4 0 $u https://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31294095