Perceptual similarity in input–output mappings: A computational/experimental study of non-native speech production

Lingua ◽  
2011 ◽  
Vol 121 (8) ◽  
pp. 1344-1358 ◽  
Author(s):  
Jason A. Shaw ◽  
Lisa Davidson
Keyword(s):  
Experimental Study ◽  
Speech Production ◽  
Perceptual Similarity ◽  
Input Output ◽  
Native Speech

He who Hesitates: Hesitation Phenomena as Quality Control in Speech Production, Obstacles in Non-Native Speech Perception

2000 ◽  
Vol 182 (3) ◽  
pp. 72-97 ◽  
Author(s):  
Marnie Reed
Keyword(s):  
Quality Control ◽  
Speech Production ◽  
Speech Processing ◽  
Native Speakers ◽  
Spontaneous Speech ◽  
Classification Framework ◽  
Native Speech ◽  
Control Devices ◽  
Content Processing ◽  
Hesitation Phenomena

This study was designed to determine the nature and occurrence of hesitation phenomena in spontaneous speech of native and non-native speakers, and to determine whether and to what extent the hesitation phenomena normal in spontaneous speech pose perception problems for non-native speakers. A quantitative analysis reveals that hesitation phenomena are ubiquitous in both native and non-native speech production. A qualitative analysis based on a content-processing classification framework reveals the function of hesitations. Hesitations act as overt traces of prospective and retrospective speech-processing tasks which function to forestall errors, and to permit detection and repair of errors once they are committed. Hesitations are quality control devices; native and non-native speakers are highly successful utilizing them to forestall errors. However, hesitation phenomena clearly pose perception problems for non-native speakers who show little evidence of recognizing them as such. Like native speakers, non-native speakers produce hesitation phenomena. Unlike native speakers, who edit and filter out the hesitations they hear, non-native speakers attempt to assign meaning to speakers' faulty output or to parenthetical remarks. Hesitations are unpredictable in their frequency or occurrence; failure to provide training in these oral discourse features of connected speech may result in non-native speakers whose speech production vastly outstrips their perception.

Passivity and Noncollocation in the Control of Flexible Multibody Systems

1999 ◽  
Vol 122 (1) ◽  
pp. 11-17 ◽  
Author(s):  
Christopher J. Damaren
Keyword(s):  
Experimental Study ◽  
Multibody Systems ◽  
Flexible Structures ◽  
Robotic Manipulation ◽  
Control Problems ◽  
Vibration Modes ◽  
Stabilization Problem ◽  
Input Output ◽  
Mass Properties ◽  
Adaptive Feedforward

Collocation of actuation and sensing in flexible structures leads to the desirable input-output property of passivity which greatly simplifies the stabilization problem. However, many control problems of interest such as robotic manipulation are noncollocated in nature. This paper examines the possibility of combining collocated and noncollocated outputs so as to achieve passivity. An appropriate combination is shown to depend on the interplay between collocated and noncollocated mass properties. Tracking problems are also studied and a controller with adaptive feedforward elements is introduced. An experimental study using a simple flexible apparatus with one rigid degree of freedom and two vibration modes is used to validate the analysis. [S0022-0434(00)01701-9]

scholarly journals Individual phonological attrition in Albanian–English late bilinguals

2017 ◽  
Vol 21 (2) ◽  
pp. 278-295 ◽  
Author(s):  
ESTHER DE LEEUW ◽  
AURELA TUSHA ◽  
MONIKA S. SCHMID
Keyword(s):  
Second Language ◽  
United Kingdom ◽  
Speech Production ◽  
Learning Model ◽  
L2 Acquisition ◽  
British English ◽  
Speech Learning Model ◽  
Native Speech ◽  
Acoustic Analyses ◽  
Speech Learning

The purpose of this study was to investigate phonological attrition in 10 native Albanian speakers who acquired Standard Southern British English (SSBE) as a second language (L2) in London, United Kingdom. A contrast was examined which is phonemic in Albanian but allophonic in SSBE, namely the production of light and dark lateral approximants. Impressionistic and acoustic analyses revealed that one late bilingual completely neutralized the phonemic contrast in her native Albanian speech. Furthermore, two other bilinguals neutralized the phonemic contrast between light /l/ and dark /ɫ/ in coda position, and overall there appeared to be a stronger trend for light /l/ to become dark in coda position than for dark /ɫ/ to become light in onset position. The findings are discussed in relation to the Speech Learning Model (Flege, 1995) and indicate that phonological attrition in native speech production is possible in late L2 acquisition, although not inevitable.

scholarly journals Crosslinguistic perceptions of /s/ among English, French, and German listeners

2021 ◽  
pp. 1-27
Author(s):  
Zac Boyd ◽  
Josef Fruehwald ◽  
Lauren Hall-Lew
Keyword(s):  
Sexual Orientation ◽  
Speech Production ◽  
English Listener ◽  
Show Evidence ◽  
Native Speech

Abstract This study reports the results of a crosslinguistic matched guise test examining /s/ and pitch variation in judgments of sexual orientation and nonnormative masculinity among English, French, and German listeners. Listeners responded to /s/ and pitch manipulations in native and other language stimuli (English, French, German, and Estonian). All listener groups rate higher pitch guises as more gay- and effeminate-sounding than lower pitch guises. However, only English listeners hear [s+] guises as more gay- and effeminate-sounding than [s] or [s−] guises for all stimuli languages. French and German listeners do not hear [s+] guises as more gay- or effeminate-sounding in any stimulus language, despite this feature's presence in native speech production. English listener results show evidence of indexical transfer, when indexical knowledge is applied to the perception of unknown languages. French and German listener results show how the enregistered status of /s/ variation affects perception, despite crosslinguistic similarities in production.

Performing Multi-Target Regression via a Parameter Sharing-Based Deep Network

2019 ◽  
Vol 29 (09) ◽  
pp. 1950014 ◽  
Author(s):  
Oscar Reyes ◽  
Sebastián Ventura
Keyword(s):  
Neural Network ◽  
Experimental Study ◽  
Network Model ◽  
Neural Network Model ◽  
State Of The Art ◽  
Experimental Results ◽  
Input Output ◽  
Deep Architecture ◽  
Proposed Model ◽  
Input Variables

Multi-target regression (MTR) comprises the prediction of multiple continuous target variables from a common set of input variables. There are two major challenges when addressing the MTR problem: the exploration of the inter-target dependencies and the modeling of complex input–output relationships. This paper proposes a neural network model that is able to simultaneously address these two challenges in a flexible way. A deep architecture well suited for learning multiple continuous outputs is designed, providing some flexibility to model the inter-target relationships by sharing network parameters as well as the possibility to exploit target-specific patterns by learning a set of nonshared parameters for each target. The effectiveness of the proposal is analyzed through an extensive experimental study on 18 datasets, demonstrating the benefits of using a shared representation that exploits the commonalities between target variables. According to the experimental results, the proposed model is competitive with respect to the state-of-the-art in MTR.

Sign in / Sign up

Export Citation Format

Share Document