Data homework 3: Background in Prosody and Acoustics

Distributed Oct 5; due in three parts:

• Part I due before class on Tuesday Oct 12.
• Part II will be a lab assignment to be given out and done in class, with your results emailed to us with Part III.
• Part III is due a few days later on the Friday of Week 4, i.e. Friday Oct 15 at 2:14pm

Part I: Praat practice (due before class on Tue Oct 12),

• Note that you should not turn in paragraphs on the reading for today
• Download Praat onto a laptop that you can bring to class. (If that's not possible, use another computer, and you will have to share a laptop with someone on Tuesday).
• Read the very brief tutorial here.
• Record yourself in Praat saying the following two sentences:
  • "That chapter is definitely too long"
  • "That cockroach is totally disgusting!"
• For each of the two sentences, mail us (before class) the following 6 pieces of information:
  1. The average pitch of the whole sentence in Hz
  2. The maximum pitch for the whole utterance in Hz
  3. The minimum pitch for the whole utterance in Hz
  4. The maximum intensity for the whole utterance in Hz
  5. The minimum intensity for the whole utterance in Hz
  6. Your rate of speech in words per second for the utterance.
• Write one more sentence based on these numbers describing the intonational/prosodic difference betweeen the two sentences and saying something about the cause.

Part II: More Praat practice (to be done as a lab in class on Tue Oct 12 with the writeup handed in together with Part III)

• Bring your laptop to class with Praat installed
• Get a partner; you can use both your laptops or share one.
• Smiling speech:
  1. Record yourself saying the sentence This sentence might be a smiling sentence in two ways: smiling and not smiling. Try for a genuine smile; use your partner to make sure you get a good example.
  2. Compare the two sentences on F0, energy and the first formant (F1). Discuss any patterns you see.
  3. Write up one pgraph with your numbers and the discussion. Make sure in your writeup you compare to the Robson paper for the October 19 reading.
• Sarcastic speech:
  1. Record yourself saying the sentence Yeah, right! in two ways: sarcastic and not sarcastic. Try for genuine sarcasm; use your partner to make sure you get a good example and the two utterances are distinct.
  2. Compare the two sentences on prosodic features (pitch, energy, duration). Discuss any patterns you see.
  3. Write up one pgraph with your numbers and the discussion. In your writeup, make sure you compare to the Tepperman et al. paper

Part III: Disfluency (due on Friday Oct 15 at 2:14pm)

1. Transcribe this southern-accented male speaker from his two conversations, 2301 and 2005 from the Switchboard corpus of telephone conversations between strangers. (You can use Praat or any software you want to transcribe). In conversation 2005 he is the second person (his speech starts with ``Well, of course''). Make sure you carefully transcribe the disfluencies, including filled pauses, word fragments, and restarts, since that's the focus of this question.
2. Discuss some of the differences between the disfluencies in the two passages (e.g., are there more or less of some disfluency types in one of the conversations?). Of course we don't expect any generalization based on one pair of examples to be correct; We just want you to think about possible hypotheses that might be testable on a larger dataset.