Praat as phoneme detector [was: Re: Kyrgyz, &c.]
|From:||Dirk Elzinga <dirk_elzinga@...>|
|Date:||Monday, March 24, 2003, 14:52|
At 12:19 PM +0100 3/24/03, taliesin the storyteller wrote:
>* Dirk Elzinga said on 2003-03-12 17:53:38 +0100
>> At 9:38 AM +0100 2/27/03, taliesin the storyteller wrote:
>> >* Peter Clark said on 2003-02-26 14:20:57 +0100
>> >> Well, as I mentioned to Yitzik in a private email, my Kyrgyz-speaking
>> >> friend is coming in April, so I will be sure to pepper her with
>> >> questions. Now accepting suggestions!
>> >Record (as in sound) some saying in Kyrgyz, then map it by hand to IPA
>> >later, or feed it to a computer and a program like Praat[*], which can find
>> >the phones for you.
>> Praat is not a phon(em)e detector. You can do all sorts of things with
>> speech analysis using Praat (including segmenting and labelling sound
>> files -- it even includes an Optimality Theory tutorial and machine-
>> learning simulations), but phone identification still has to be done
>> the old-fashioned way.
>Then what program did I use during Phonetics 101 then, if it wasn't
>Praat? *sigh* *scratches head*
You may have used Praat, but I don't know of any function in the software package
which will automatically segment sound files and annotate them with phonetic
symbols. If there is such a function, I'd sure like to know about it! It seems
to be the sort of task best left to humans, though.
Dirk Elzinga Dirk_Elzinga@byu.edu
"It is important not to let one's aesthetics interfere with the appreciation of
fact." - Stephen Anderson