LINGUIST List 17.218
Mon Jan 23 2006
Qs: Category Rating Survey; Measuring Vowel Duration
Editor for this issue: Jessica Boynton
<jessicalinguistlist.org>
Directory
1. Raluca
Budiu,
Category Rating Survey
2. Roy
Becker,
Measuring Vowel Duration from Spectrogramme
Message 1: Category Rating Survey
Date: 23-Jan-2006
From: Raluca Budiu <budiuparc.com>
Subject: Category Rating Survey
Hello, all,
We are conducting an investigation about how people navigate through webpages (and similar hierarchic structures). We are especially interested infinding out how the label of a link affects navigation choices.
We would be very grateful if you could spend a few minutes to fill in thiscategory rating survey. It's located at:
http://glsa.parc.com/ratings/
It should take at most 15 minutes.
We will not be recording any information about who you are, so yourresponse will be completely anonymous.
Thank you in advance for your help.Raluca
-- Raluca Budiu, Ph.D. Palo Alto Research Center User Interface Research Group
Linguistic Field(s):
Cognitive Science
Psycholinguistics
Semantics
Message 2: Measuring Vowel Duration from Spectrogramme
|
Date: 22-Jan-2006
From: Roy Becker <roybeckerhumnet.ucla.edu>
Subject: Measuring Vowel Duration from Spectrogramme
Physical duration of segments is relevant for phoneticians and experimentalphonologists, among others, both as a stand-alone phenomenon and as acorrelate of other speech phenomena (speech-rate, accent, articulatortrajectory etc.) as well as linguistic phenomena (intrinsic phonemeproperty, realization of morae, compensatory lengthening, phrasing etc.)Hence, measuring duration of segments using waveform and spectrogramme is acommon practice among phoneticians and experimental phonologists working onfield- or lab-collected corpora. It is well-known, however, that measuringsegmental duration is not always straightforward, mostly due to gesturaloverlap.
In particular, measuring duration of vowels by determining the temporalboundaries with adjacent segments poses certain methodological problems,for example:
1. Voiceless stops and fricatives can be pre-/post-aspirated.2. Stop-release may include multiple bursts and/or slight frication.3. Periods of articulator approximation without true contact, e.g. thetransition into/out-of voiced fricatives, when there is no frication butvoicelessness would have made a perfect fricative.4. When the adjacent segment is an approximant, and any point across theboundary into the approximant would have probably counted as part of thevowel if it were part of a transition into/out-of a contact-involvingconsonant. For example, the acoustics of a dorsal glide are no differentfrom the acoustics of the transition after the release of a homorganicstop, and whatever counts as consonantal in the case of a glide would countas vocalic in the case of a stop (e.g. palatal glide vs. palatal stop).Similarly, in the case of a vocalized velarized lateral, whatever counts asthe consonant would have probably counted as part of the vowel if theconsonant weren't velarized.
All these, and many other situations, are sources for methodologicalinconsistency in determining segmentation of vowels or any kind of sound.While it is perfectly legitimate to use any methodology if it is designedwith common-sense, strictly obeyed and explicitly described, the absence ofstandard may render related studies mutually incompatible, and may alsoresult in methodological slacking in the case of the less experiencedresearcher.
I would like to know if there is any written standard(s) for making suchmeasurements, and whether researchers adhere to this/these standard(s).I am not interested in personal opinions or in methods practiced in aparticular stduy, because these can be found in any experiment report.
Thanks,Roy Becker.
Linguistic Field(s):
Phonetics
|