Praat Tutorial

I. Basic Software Stuff—How to Create a Sound Object

1. Close the "Pratt Picture" window…you won’t need it. Go to the "Pratt Objects" window and hit "New", then "Record Mono Sound".
2. In the window that pops up, hit "Record" and make your sound. When finished, hit "Stop".
3. Type a distinctive name for your sound in the space at the bottom of the window and hit "To List".
4. Close "SoundRecorder" window.
5. Select the sound object you just created and hit "Label & Segment" on the right side of the "Pratt Objects" window.
6. Hit "To TextGrid", then when the "To TextGrid" window pops up, hit "Ok" to get out of it. Ignore the "Mary John Bell" stuff. We won’t be using that function.
7. Select your Sound and its Textgrid simultaneously by clicking on the Sound, hitting the Ctrl key, and then clicking on the Textgrid.
8. Hit "Edit" on the right side of the Praat Objects window. This opens your Textgrid window, which is what you will be using to analyze your sounds.

II. How to Use Your Sound Object

A. Pitch

1. Make a sound object by first saying the statement "John hit Mary" twice (enunciating clearly), then asking the question "John hit Mary?" twice. Name this sound and send it to your list, then make a textgrid of it as explained above.

2. In your textgrid window, go to "View", then "Formant Settings". There should be an option to "Show"—make sure this option is off by unclicking the little box next to it. Hit Ok.
3. Look at the light blue lines below the acoustic diagram of your sound object. These show the pitch of your voice throughout the speech act. Compare the pitch in statements and questions, making up statements and questions of your own if you wish.



 

4. Create your own sound objects using different tones of voice to create different implications (about your emotional state, level of sarcasm, etc.) in your statements. Look at how the pitch changes from one to another and see if you can find certain pitch patterns that tend to create certain assumptions about the meaning of a statement (or the emotional state, etc.).
5. What do these data imply about how we interpret speech acts? How could acoustic software like this help linguists study pragmatics?
6. Chinese is a tonal language. In other words, the pitch of one’s voice is a phonemic, not merely a phonetic, distinction. The following is an acoustic diagram of the syllable "ma" spoken (by a non-native speaker) in 5 different tones: high, rising, low, falling, and neutral. The meanings are, respectively, "mother", "hemp", "horse", "curse", and an interrogative particle.

B. Vowels

1. In the textgrid window, go to "View", then "Formant Settings" and select "Show". Increase the "Maximum duration (s)" to 10 if it is not there already, so that formants will be shown for your entire speech object and not just 2 seconds of it.
2. Notice that each vowel has 4 lines of red dots, and the relative distances of these lines are different for each vowel. These are the formants. Basically, each vowel has several pitches at which it tends to vibrate most strongly. This combination of pitches is different for each vowel because your oral cavity has a distinct shape for each vowel (affected by tongue height, lip roundness, etc.), and so the way it resonates is also distinctive for each vowel. This has nothing to do with the pitch of your voice itself. If you whisper that list of vowels, you can hear the pitch go down through the list, even though your vocal cords are not vibrating.



3. Notice the second formant from the bottom. As you move down this list of vowels, it should gradually decrease. This formant corresponds (somewhat loosely because of individual peculiarities) to the degree of "backness" of a vowel…that is, as you move your tongue further and further back to form a vowel, that formant goes down.
4. Now look at the first formant, the one on the bottom. This one should increase through the first four vowels in that list and decrease through the second four. This formant corresponds to the height of a vowel. It increases as you go from the high vowel in "heed" to the low one in "had", and decreases as you go from the low vowel in "hod" to the high one in "who’d".
5. The higher formants correspond in part to the degree of lip rounding. In general, as sounds become more rounded, the third and fourth formants tend to decrease. However this effect is not quite as tidy.

3. Consonants

1. Make a sound object by saying "cat" three times and then saying "scat" three times. Enunciate clearly. Make another sound object in the same way with "top" and "stop".
2. Notice the difference in the acoustic signature of aspirated consonants (the initial consonants of "cat" and "top") and unaspirated consonants (the "c" of "scat" and the "t" of "stop").

 

3. Make a sound object by saying the words "caught", "pot", and "taught", then "got", "bought", and "dot". Compare the acoustic signatures of voiced and unvoiced consonants.



4. Make a sound object by repeating the word "desire" , then the word "attire", three times each. Compare the signatures of stops and other consonants in the middle of a word. You can see that the air actually stops briefly in a stop, hence the name.

Information sources for this tutorial:

Van Lieshout, Pascal, Ph.D. Pratt Short Tutorial: A basic introduction. V. 3.0

© January 7, 2002