This is the class log, updated after every class meeting. It includes topics, assignments, announcements, Web links, and (if there is enough server space) the audio demonstrations from class.

Topics: Project writeups. Final-exam syllabus. Review date. Grade audit. Course evaluations.

Handouts:

• Final-exam syllabus
• Handing in the projects

• Continue working on final projects.
• Prepare in-class presentations (schedule). I will take attendance!
• Are you backing up your data?

Topics: Finish up infant phonetic acquisition.

Topics: Acquisition of phonetics by infants.

Web:

• Tables showing acquisition of segments by babies learning various languages. Thanks to Yu Li for finding this!
• Audio examples of infant vocalizations

Topics: Perception of speech by non-human animals.

Web:

• What's a chinchilla?

Topics: Categorical perception.

Web:

• Stimuli and results for a typical categorical-perception experiment with human listeners.

• Another mystery spectrogram

• Continue working on final projects. Start preparing your presentation.
• Back up your data, back up your data, back up your data!

Topics: Interlanguage differences in perception: Vowels, tones, fricatives.

Handouts:

• Presentations and writeups
• Sample handout (bogus!) for final presentation.

• Read this web page and do the categorical-perception experiments (both the "Labelling" and the "Discrimination" experiment). If you can't get it to work, try a different browser. Print out the graphs you get at the end of the experiment, and bring them to class on Wednesday.
• Continue working on final projects.
• Back up your data, back up your data, back up your data!

Topics: Finish up perceptual spaces. Audiovisual speech perception: The McGurk effect.

Web:

• Demo of McGurk Effect.
• More McGurk references.

Assignment:

• Read Johnson, Ch. 6.6 (cross-linguistic vowel perception), 7.5 (dimensions of fricative perception). We'll discuss those and the Mandarin tone stuff from 4.5 on Monday.
• Continue working on final projects.
  • Back up your data after recording! You don't want to lose those sound files if your disk crashes, or if someone steals your computer.
  • You'll be graded on how well you investigate the question, not whether you get "the right answer". Your study may not pan out. That's OK!
  • I'll have a handout for you on Monday about the in-class presentation and the writeup.

Topics: Speech perception. Perceptual space, perceptual maps.

Audio:

• Demonstration of confusion experiment with Hindi stops.

Assignment:

• Read Johnson, Ch. 4.4-4.5 for Friday the 12th.
• View movie of McGurk Effect.

Topics: Frequency perception. The Bark scale. Cochleagrams. Auditory representation of speech.

Audio:

• The same 50-Hz increment sounds smaller at some frequencies than others.
• A series of 1000-Hz increments: 500 Hz, 1500 Hz, ... 13500 Hz. Notice the change in perceptual spacing, and in spacing on the cochleagram.
• A series of 1-Bark increments, based on Zwicker (1961) JASA 33:248. Note even perceptual spacing, and compare the cochleagram.

• Animation of the cochlea.
• Another one.

• Read Johnson, Ch. 4.1-4.3 (speech perception and perceptual maps) for Wednesday the 10th.

2004 November 5 (F)

Topics: Acoustics vs. audition. Intensity and loudness. Anatomy and filtering effects of ear.

Audio:

• A tone rising from 70 Hz to 17000 Hz at constant pressure amplitude. Notice how the loudness is not constant, and compare Johnson's Figure 3.3.
• A 1000-Hz tone being reduced in loudness by 6-dB increments.
• Two tones, 82.23 dB and 88.25 dB (re some arbitrary level) are added; their sum has intensity 89.22 dB.

• Anatomy of the ear.

• View this animation of the cochlea.
• And this one too.
• Read Johnson, Ch. 6.5 (acoustic vs. auditory spectra of vowels), 7.4 (fricatives), and 8.4 (stops) for Monday the 8th..
• HW 5 is due Nov. 12 (Friday).

Topics: Clicks, trills, taps, flaps. Using the pitch and LPC formant trackers.

Audio:

• Nama clicks from the Ladefoged CD-ROM.
• X-ray video of a click, same source.
• Xhosa clicks with varying accompaniments, same source.
• French uvular trills, same source.
• Bilabial and alveolar trills in Kele and Titan, same source.

Assignments:

• Read Johnson, Ch. 3 ("Basic Audition") for Nov. 5 (Friday).

Topics: Laterals and semivowels.

Audio:

• Synthetic [l] made according to acoustic theory in Johnson Ch. 9.
• Same, without the antiformant.
• Natural English laterals and semivowels.
• Comparison of natural English [l] and [n].
• Mystery spectrogram: two more cities.

Assignments:

• Read Ladefoged, Ch. 13.4, 13.5, 13.6 for Friday the 5th.
• Final-project proposals are due on Friday, November 5.
• There will be a short HW on Friday the 5th, due back on Wednesday the 10th.

Topics: Nasals.

Audio:

• Natural nasals before schwa.
• Swapping nasal murmurs:
  • [m] murmur with [n] transitions
  • [n] murmur with [m] transitions
• Synthetic nasals made from the simplified acoustic theory in Johnson Ch. 9, with closures at 100%, 90%, ..., 50% of the distance from the glottis to the lips.
• Synthetic demo of widening the F1 bandwidth.
• Mystery spectrograms: two more cities.

Assignments:

• Bring mirrors on Monday!
• Read Johnson, Ch. 9.3 (laterals), for Monday the 1st.
• Read Ladefoged, pp. 153-154 (laterals), for Wednesday the 3rd.
• Final-project proposals are due on Friday, November 5th.

Topics: Check up on project ideas. Non-English places of articulation.

Audio:

• Hungarian palatal stops from the Ladefoged CD-ROM.
• Nunggubuyu stops from the Ladefoged CD-ROM.

• Read Johnson 9.1 and 9.2 for Monday, November 3. "FFT spectrum" is just an ordinary spectral slice. "LPC spectrum" is an estimate of the vocal-tract response curve when certain assumptions are made. (The red dots in Praat show the peaks in the LPC response curve, i.e., the estimated formants.) If you're curious, see Johnson 5.5.

Topics: Phonation types for stops. Airstream mechanisms (pulmonic/glottalic, ingressive/egressive). Time permitting: practice project design.

Audio:

• Hindi stops from the Ladefoged CD-ROM.
• Quechua stops from the Ladefoged CD-ROM.
• Sindhi stops from the Ladefoged CD-ROM.

Topics: Formant transitions for alveolars in non-schwa vowels. Spectrogram reading! Comments on HW 4 and partnership questionnaires.

Audio:

• Names of four familiar cities.
• Assorted English monosyllables.

• Read Ladefoged 12.4 (breathy voice), 12.7 (ejectives), and 12.8 (implosives).

Topics: Formant transitions and bursts for consonants in non-schwa vowels (velars and alveolars only). Transitions and bursts as cues to place of articulation.

Audio:

• Natural VgV syllables for various vowels.
• Natural VdV syllables for various vowels.
• Natural b, d, and g before wedge.

Topics: Formant transitions in stop-plus-schwa syllables. Partnership evaluation questionnaires.

Audio:

• Sound file for HW4
• Synthetic stop-plus-schwa syllables, with constrictions 100%, 90%, ... 10%, 0% of the way from the glottis. (Best to download it and listen to it in Praat, so you can hear the syllables separately.)

• Article report due Monday the 25th.
• Read Johnson 8.2 and 8.3 for Wednesday the 20th.

Topics: Catch-up day. VOT. Final projects.

Assignments:

• Article reports are due Monday the 25th.
• Start thinking about final-project proposals, which are due Friday, November 5th!

Topics: Sound sources and sequence of events in stops.

Handouts:

• Schematic illustration of events in stop production. Uses copyrighted material; not on web.

• Natural English labial stops, illustrating voicing and aspiration.

• Homework 4 due Monday the 18th. (Should take at most one hour to solve. Hint: Johnson Figure 6.12.)

Topics:Fricatives. Field trip to the Odum Institute

Audio:

• Synthetic frication source
• Artificial frication unfiltered, then filtered by 2.5-cm, 5-cm, and 9-cm half-open tubes.
  For best view, in the Edit window go to View > Show Analyses and set "longest analysis" to 30s, use a narrow-band spectrogram, and set it so you can see up to 8000 Hz. The lowest fricative formants are at about 3400 Hz (2.5-cm tube), 1600 Hz (5-cm tube), and 1100 Hz (9-cm tube). Is this what you expect? What happens at higher frequencies?
• Synthetic fricative noise with filter center frequency falling.

• Article selections due Wednesday the 13th!
• Read Ladefoged 6.5, and learn IPA symbols for English fricatives (for next time.)
• Read Johnson, through the end of 8.1 (source functions for stops and affricates) (for next time).
• Homework warning: HW 4 will be assigned on Monday, due the following Monday.

Topics: Consonant place-of-articulation. Article reports.

Web:

• Some low-tech ways to observe your articulators

• Naming conventions for consonant articulations

• Read Ladefoged, Ch. 6.5 (English fricatives) and learn the IPA symbols for English fricatives for October 8 (Friday)
• Article report
  • Choose an article by October 13 (Wednesday)
  • Hand in writeup by October 25 (Monday)

Topics: Vowel inventories and adaptive dispersion theory. Go over HW 3.

Web:

• Crothers (1978) on vowel-system typology

• Read Ladefoged, Ch. 10.2 and 10.3 (this is review from Ling 30 on consonant place of articulation).
• Read Johnson, Ch. 7.1 and 7.2 (introduction to fricatives).
• Bring mirrors!

Topics: Midterm. HW3. Two-tube theory of /i/ and "quantal vowels".

Web:

• Comments on the midterm

MIDTERM

Topics: Vowel typology. Two-tube models and the quantal theory. Vowel midterm.

Assignment:

• Prepare for the vowel midterm on Wednesday, September 29. (midterm syllabus (.pdf)).

Topics: Nasalization and nasalized vowels. Two-tube theory and nomograms.

Audio:

• French oral and nasal vowels (on Ladefoged CD-ROM).

• Video of French speaker producing oral and nasal vowel (on Ladefoged CD-ROM).

• Read Johnson, sections 5.2 and 6.3 (quantal theory and adaptive dispersion).
• Homework 3 is due Friday, October 1. (There is a hand-drawn page that goes with it, not on the Web.)

2003 September 24 Wed

Topics: English and non-English vowels. The IPA.

Handouts:

• What you need to know for this course about using the IPA for vowels (pdf).

• The IPA vowel chart (this is also on your CD-ROM).
• Three phoneticians perform the IPA vowel chart.
• Vowel-feature demos made from the Ladefoged IPA examples (not linked from this page, because of copyright -- I'll email the URL).
• From Sept. 17: Synthetic demo of how changing F1 changes perceived height.
  • F1 varying through schwa.
  • F1 varying through [i].
• From Sept. 19: Synthetic demo of how changing F2 changes perceived backness.
  • With F1 low (high vowels).
  • With F1 middling (mid vowels).
  • With F1 high (low vowels).
• Demos of oral (= non-nasal) vowels from Ladefoged's Chapter 14:
  • French.
  • Swedish.
  • German.
  • Scottish Gaelic.

• Homework warning: HW3 goes out Friday the 24th, due back after midterm.
• Midterm is on Wednesday the 29th, one week from today; syllabus is here.
• Reminder: Human-subjects training is due Wednesday the 29th.

Topics: HW2. Laryngeal vowel features. Review English vowels.

Audio:

• Synthetic demo of soprano vowel effect.
• Synthetic voice-quality demos (creaky, normal ("modal"), breathy)
  • Glottal source output
  • Schwa

• Listen to the IPA vowel chart (this is also on your CD-ROM).
• Read Ladefoged Ch. 14, "Vowels around the world"

Topics: English vowels, IPA symbols, and vowel features. Perceptual effect of varying F3.

Audio:

• Synthetic demo of F3 being lowered.

• Read Ladefoged 12.4-12.6 ("Actions of the Larynx")
• Listen to Ladefoged's audio examples from Gujarati, Jalapa Mazatec, and Mpi. (Chapter 12 is misidentified on the web page as Chapter 13.)

Topics: Vowel feature backness. Perturbation theory and the backness-F2 relation. Vowel charts (F1-F2 plots).

Audio:

• Synthetic demo of F2 being varied:
  • With F1 low (high vowels).
  • With F1 middling (mid vowels).
  • With F1 high (low vowels).

• Didier Demolin's vowel articulations (Ladefoged)

• Finish up Homework 2.
• The online human-subjects training is due Wednesday the 29th.

Topics: Vowel feature height. Perturbation theory and the height-F1 relation.

Audio:

• Pinching the tube near the open end raises F1; pinching it near the closed end lowers F1.
• Perceived height changes when F1 is changed.
  • F1 varying through schwa.
  • F1 varying through [i].

• Didier Demolin's vowel articulations (Ladefoged)

• Read Ladefoged Ch. 3, "Vowel Contrasts". This will introduce the English vowel inventory and the IPA symbols that go with it.
• Bring mirrors again on Wednesday.
• Human-subjects certification deadline is hereby extended one week, i.e., to Sept. 29 (W).

Topics: Perturbation theory. First non-schwa vowel feature: Rounding.

Audio:

• Lowering schwa formants by 20%.
• Lowering [i] formants by 20%.

• Simulation of air resonating in half-open pipe (Java)
• Explanation of perturbation theory from Scientific American (explains why pinching nodes and antinodes has the effect it does)
• Some ways to observe your articulators.

• Homework 2 (due Friday the 17th).
• Read Ladefoged Ch. 5, "Charting Vowels".
• The online human-subjects training is due on September 22 (W). NOTE: Click the link to the "CITI Program Website". We are in Group 2, Social and Behavioral Research. The certification program has recently changed and may take longer than I thought (up to 4 hours).
• Bring mirrors again on Monday.

Topics: Nodes and antinodes. How to record in Praat.

Audio:

• Examples of good and bad recording. The difference isn't easy to hear, but you can see it clearly in Praat.

• Simulation of air resonating in half-open pipe (Java)
• Sign up here for the soundproof studio in Dey 103.

• Johnson Ch. 6, Sections 1 and 2 (for Friday the 10th)
• Bring hand mirror to class on Friday.
• Start thinking about when you're going to find time for the online human-subjects training from the National Cancer Institute. (It takes about 2 hours total, but you can do in pieces).

Topics: Resonances of half-open tube. Formants vs. harmonics
Audio:

• Inversion of a pulse bouncing back and forth in a half-open pipe. (How long is the pipe?)

• Inversion of a pulse reflected off a "fixed" end. (Note: These examples are of transverse waves on strings. The open end of a pipe is "fixed" as far as the pressure wave is concerned -- do you see why?)
  • Demo 1
  • Demo 2
• Simulation of air resonating in half-open pipe (Java)

• Read Ladefoged, Chapter 11 (for Wednesday the 8th)
• Read Johnson, Chapter 6, sections 1 and 2 (for Friday the 10th)
• One member of your pair should bring a computer to class on Wednesday.

Topics: Source/filter theory. Resonances of a half-open pipe(i.e., schwa). Harmonics versus formants. Narrow- versus wide-band spectrograms.

Audio:

• Glottal waveform, F0 = 100 Hz (similar to Johnson Figures 5.1 and 5.2)
• Schwa made from that glottal wave.
• Schwa glissando: Fundamental and harmonics change, formants stay the same.
• Diphthong: Formants change, fundamental and harmonics stay the same.
• Low-tech source-filter synthesis: Buzzing lips, then buzzing lips into 17.7-cm tube (as done in class).

Assignment (for September 3 Fri):

• Read Johnson 5.4 (resonance and schwa)

Topics: HW 1; power spectra (spectral slices). Vowels. Source/filter theory. Voicing source (glottal wave). Partners.

Audio:

• Glottal waveform, F0 = 100 Hz (similar to Johnson Figures 5.1 and 5.2)
  • Schwa made from that glottal wave.
  • [i] made from that glottal wave.
• Glottal waveform, F0 = 200 Hz
  • Schwa made from that glottal wave.
  • [i] made from that glottal wave.
• Glottal glissando
• Schwa glissando

• Larynx atop trachea
• Larynx in throat and an explanation of phonation
• Phonation in slow motion (Ladefoged)
• Phonation at different frequencies (Ladefoged)

• Read Johnson 5.1 (voicing), 5.3 (vocal-tract filtering)
  • p. 79, "1 second/0.0066 seconds" should be "1 cycle/0.0066 seconds".
  • "Power spectrum" means "spectral slice".
  • Definition of bandpass filter

Topics: Using Praat and the Ladefoged CD-ROM. Sound: waveforms, periodic vs. aperiodic, amplitude, frequency, wavelength, phase. Fundamental and harmonics. Fourier analysis. Spectrograms. Partnership questionnaires.

Audio:

• Illustration of Johnson's Figure 1.6.
• Adding two sine waves
• Illustration of frequencies in musical scales and speech.
  

• Why are they called sine waves?
• Illustration of Fourier analysis

• Read Ladefoged Ch. 2.
• Do Homework 1. The questions refer to these sound files:
  • Question 1
  • Question 2
  • Question 3 (to find the point 80 ms into the vowel, put the red cursor so that the number at the top of it is about 0.0800)

Topics: Course organization. Praat and the Ladefoged CD-ROM. Intro to sound.

Handouts: Syllabus. Installing Praat. Installing the Ladefoged CD-ROM.

Web: Sound waves. How microphones work.

Assignments (to be done by 8/27 Fri):

• Read Ladefoged Ch. 1 and Johnson Ch 1 (sections 1-3 only).
• Install Praat on your computer and check that it runs.
• Check that you can listen to sound files and watch movies on the Ladefoged CD-ROM.
• Bring laptops to class on Friday with Praat installed!