26.515, Qs: Help needed to write/adjust a Praat-script

The LINGUIST List via LINGUIST linguist at listserv.linguistlist.org
Mon Jan 26 15:39:17 UTC 2015


LINGUIST List: Vol-26-515. Mon Jan 26 2015. ISSN: 1069 - 4875.

Subject: 26.515, Qs: Help needed to write/adjust a Praat-script

Moderators: linguist at linguistlist.org (Damir Cavar, Malgorzata E. Cavar)
Reviews: reviews at linguistlist.org (Anthony Aristar, Helen Aristar-Dry, Sara Couture)
Homepage: http://linguistlist.org

Editor for this issue: Damir Cavar <damir at linguistlist.org>
================================================================


Date: Mon, 26 Jan 2015 10:38:56
From: Benina Knothe [benina.knothe at uni-muenster.de]
Subject: Help needed to write/adjust a Praat-script

 
For my PhD-project, I am looking for someone to write or adjust a Praat-script
in order to automatically analyse the F0-F3-frequencies of vowels. For each
vowel, the duration as well as the frequencies should be measured at (at
least) 25%, 50% and 75% of the time in order to allow the detection of
diphthongization. In return for a serviceable and effective script
remuneration is possible.

For several Dutch sound files, TextGrids have been automatically generated
using the BAS Web Service WebMAUS Basic. The TextGrids each contain three
Tiers: the first contains the word boundaries and an orthographic
transcription, the second word boundaries with a canonical pronunciation
transcription using SAMPA and the third phone boundaries again transcribed in
SAMPA. All three Tiers contain errors in the spelling and/or boundaries which
may be neglected.

Within each of these TextGrids a fourth Tier is manually being generated
marking the accented vowels with corrected phone boundaries. The Praat-script
should therefore use this fourth Tier to generate the formant frequencies and
the durations. This fourth Tier contains different vowel types each using the
SAMPA symbol. Within the Tier, the vowels differ in their number of tokens.

For each vowel type, the Praat-script should measure the duration per token.
In addition, the F0-, F1-, F2- and F3-frequencies should be measured in Hertz
at (at least) three points for each token. All of these results should be
automatically saved in a table compatible with Microsoft Excel 2010,
preferably ordered by vowel type. Preferably, the script should be
written/adjusted for a computer using Windows 7.

I'm happy to send a TextGrid example to anyone who wishes to see one.
 

Linguistic Field(s): Applied Linguistics
                     Computational Linguistics
                     Phonetics

Subject Language(s): Dutch (nld)

Language Family(ies): Germanic






----------------------------------------------------------
LINGUIST List: Vol-26-515	
----------------------------------------------------------







More information about the LINGUIST mailing list