Automatic voice onset time detection for unvoiced stops (/p/,/t/,/k/) with application to accent classification

Hansen, John H. L.; Gray, Sharmistha; Kim, Wooil

doi:10.1016/j.specom.2010.05.004

Cited by 36 publications

(5 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In recent years, several systems have been proposed to measure VOT automatically (e.g., Das & Hansen, 2004; Hansen, Gray, & Kim, 2010; Kazemzadeh, et al, 2006; Lin & Wang, 2011; Sonderegger & Keshet, 2012; Stouten & Van hamme, 2009). Building on the greatly improved success of speech segmentation in general, these programs have been reported to have a reasonable degree of success.…”

Section: Automatic Measuresmentioning

confidence: 99%

Voice Onset Time (VOT) at 50: Theoretical and practical issues in measuring voicing distinctions

Abramson

Whalen

2017

Journal of Phonetics

138

View full text Add to dashboard Cite

Just over fifty years ago, Lisker and Abramson proposed a straightforward measure of acoustic differences among stop consonants of different voicing categories, voice onset time (VOT). Since that time, hundreds of studies have used this method. Here, we review the original definition of VOT, propose some extensions to the definition, and discuss some problematic cases. We propose a set of terms for the most important aspects of VOT and a set of Praat labels that could provide some consistency for future cross-study analyses. Although additions of other aspects of realization of voicing distinctions (F0, amplitude, duration of voicelessness) could be considered, they are rejected as adding too much complexity for what has turned out to be one of the most frequently used metrics in phonetics and phonology.

show abstract

Section: Automatic Measuresmentioning

confidence: 99%

Voice Onset Time (VOT) at 50: Theoretical and practical issues in measuring voicing distinctions

Abramson

Whalen

2017

Journal of Phonetics

138

View full text Add to dashboard Cite

show abstract

“…There are mainly three possible reasons for damage or beam degradation in fiber systems operated at high peak power levels: self-focusing, surface or bulk damage. Concerning selffocusing, the critical peak power threshold commonly used for silica fibers is 4 MW [34,35] The manufacturer reported values in the order of several hundred W/µm 2 for the surface damage threshold peak intensity of similar Yb-doped cores without an end-cap [36].…”

Section: Further Scaling Of These Systems: Damage and Self-focusing Cmentioning

confidence: 99%

Pulse compression of a high-power thin disk laser using rod-type fiber amplifiers

Saraceno¹,

Heckl²,

Baer³

et al. 2011

Opt. Express

View full text Add to dashboard Cite

We report on two pulse compressors for a high-power thin disk laser oscillator using rod-type fiber amplifiers. Both systems are seeded by a standard SESAM modelocked thin disk laser that delivers 16 W of average power at a repetition rate of 10.6 MHz with a pulse energy of 1.5 μJ and a pulse duration of 1 ps. We discuss two results with different fiber parameters with different trade-offs in pulse duration, average power, damage and complexity. The first amplifier setup consists of a Yb-doped fiber amplifier with a 2200 μm2 core area and a length of 55 cm, resulting in a compressed average power of 55 W with 98-fs pulses at a repetition rate of 10.6 MHz. The second system uses a shorter 36-cm fiber with a larger core area of 4500 μm2. In a stretcher-free configuration we obtained 34 W of compressed average power and 65-fs pulses. In both cases peak powers of > 30 MW were demonstrated at several μJ pulse energies. The power scaling limitations due to damage and self-focusing are discussed.

show abstract

“…The second box in Figure 1 is a structured prediction algorithm for measurement of VOT (Sonderegger & Keshet, 2012; see Ryant, Yuan, & Liberman, 2013, for an alternative approach). Many standard approaches measure parameters based on pre-programmed rules developed in consultation with expert annotators (Boyce, Fell, MacAuslan, & Wilde, 2010; Hansen, Gray, & Kim, 2010; Prathosh, Ramakrishnan, & Ananthapadmanabha, 2014; Stouten & van Hamme, 2009). In contrast, the algorithm utilized here was designed to minimize the error in the predicted measurement and had a unique feature set.…”

Section: Introductionmentioning

confidence: 99%

Automatic analysis of slips of the tongue: Insights into the cognitive architecture of speech production

et al. 2016

View full text Add to dashboard Cite

Traces of the cognitive mechanisms underlying speaking can be found within subtle variations in how we pronounce sounds. While speech errors have traditionally been seen as categorical substitutions of one sound for another, acoustic/articulatory analyses show they partially reflect the intended sound. When “pig” is mispronounced as “big,” the resulting /b/ sound differs from correct productions of “big,” moving towards intended “pig”—revealing the role of graded sound representations in speech production. Investigating the origins of such phenomena requires detailed estimation of speech sound distributions; this has been hampered by reliance on subjective, labor-intensive manual annotation. Computational methods can address these issues by providing for objective, automatic measurements. We develop a novel high-precision computational approach, based on a set of machine learning algorithms, for measurement of elicited speech. The algorithms are trained on existing manually labeled data to detect and locate linguistically relevant acoustic properties with high accuracy. Our approach is robust, is designed to handle mis-productions, and overall matches the performance of expert coders. It allows us to analyze a very large dataset of speech errors (containing far more errors than the total in the existing literature), illuminating properties of speech sound distributions previously impossible to reliably observe. We argue that this provides novel evidence that two sources both contribute to deviations in speech errors: planning processes specifying the targets of articulation and articulatory processes specifying the motor movements that execute this plan. These findings illustrate how a much richer picture of speech provides an opportunity to gain novel insights into language processing.

show abstract

Automatic voice onset time detection for unvoiced stops (/p/,/t/,/k/) with application to accent classification

Cited by 36 publications

References 39 publications

Voice Onset Time (VOT) at 50: Theoretical and practical issues in measuring voicing distinctions

Voice Onset Time (VOT) at 50: Theoretical and practical issues in measuring voicing distinctions

Pulse compression of a high-power thin disk laser using rod-type fiber amplifiers

Automatic analysis of slips of the tongue: Insights into the cognitive architecture of speech production

Contact Info

Product

Resources

About