Quick links: Current | Past | Upcoming | Requirements | Term project
Meetings: Tu Th 10:00-11:15, ILC N451
Instructor: Kristine Yu (krisyu@linguist.umass.edu / ILC N422)
Office hours: by appointment
This is a graduate-level phonetics seminar. There are two goals for this course. The first is empirical: to expose students to the phonetics of the voice source, e.g., phonation contrasts of the world's languages; articulatory, acoustic, and perceptual models of voice quality, and how to study them. The second is methodological. The challenge of understanding the phonetic spaces of voice quality forces us to confront two methodological issues that are common in phonetic research: massive collinearity in high-dimensional spaces and time series analysis. In this course, students will be introduced to current methods to tackle these two challenges and they will be implement them on voice quality acoustic data sets. In addition, we will work together to design a perceptual experiment starting around Weeks 5-6 that we will start running in the second half of the semester. The work emerging from acoustic analysis and the perceptual experiment is expected to result in eventual journal publications that students will be encouraged to participate in as collaborators.
| Week | Date | Topic | Class | HW to do |
|---|---|---|---|---|
| 13 | Tu 12/03 | Final project: VoiceSauce processing/FDA | Reports on exploratory data analysis, discussion of next steps | Do second VoiceSauce processing assignment (due by 11:59PM Friday 12/06) |
| Th 12/05 | FDA, clustering, GAMMs and write-up discussion | Analysis/write-up planning | ||
| Tu 12/10 | Final project: write-up | |||
| Th 12/20 | Speech Prosody submission due |
| Week | Date | Topic | Class | HW to do |
|---|
| Week | Date | Topic | Class | HW to do |
|---|---|---|---|---|
| 12 | Tu 11/19 | Final project: exploratory analysis | Exploratory analysis | Do exploratory analysis in R notebook, due by 11/22 |
| Th 11/21 | KY away for RecPhon | |||
| 11 | Tu 11/12 | Final project: parameter extraction (Speech Prosody 2020) | Finish up GAMMs tutorial, intro to final project | |
| Th 11/14 | Parameter extraction | Do first VoiceSauce processing assignment (due by class time Tues 11/19) | ||
| 10 | Tu 11/05 | Intro to generalized additive mixed modeling | Student presentation: Wieling 2018, Michael Clark GAMs introduction, Soskuthy tutorial | |
| Th 11/07 | ||||
| 09 | Tu 10/29 | Intro to Functional Data Analysis | Student presentation: Gubian materials | |
| Th 10/31 | ||||
| 08 | Tu 10/22 | Intro to collinearity | Tomaschek et al. (2018) | Work through Tomaschek et al. paper through section 4 and code up through SupplementaryMaterial_P2_Inspect_collinearity.html |
| Th 10/24 | ||||
| 07 | Tu 10/15 | NO CLASS (UMass Monday) | ||
| Th 10/17 | VS benchmarks, FDA discussion | Be ready to present FDA next week, start looking at GAMMs too | ||
| 06 | Tu 10/08 | VS benchmarks, FDA, teaching | class notes | Prep FDA tutorial |
| Th 10/10 | KY away for NWAV | |||
| 05 | Tu 10/01 | Semi-automated acoustic methods: introduction to VoiceSauce | class notes | Determine which stats method you will present by start of class 10/03 |
| Th 10/03 | ...continued | Go through FDA tutorial, check VoiceSauce with hand-measured benchmarks | ||
| 04 | Tu 09/24 | Student presentations | Measures of the glottal source spectrum. Kreiman, Gerratt and Antonanzas-Barroso 2007: Seoyoung, Modeling the voice source in terms of spectral slopes. Garellek, Samlan, Gerratt and Kreiman (2015)., see also POMA proceedings and POMA poster: Alex | |
| Th 09/26 | Student presentations | Perceptual sensitivity to first harmonic amplitude in the voice source. Kreiman and Gerratt (2010), see also follow-up article: Claudia | Due by beginning of class Tues 10/01: See class notes for assignment. Install Matlab and download VoiceSauce | |
| 03 | Tu 09/17 | Psychoacoustic model overview / how to review a paper | notes | Choose one of the papers to review/present next week |
| Th 09/19 | More psychoacoustic model discussion, articulatory modeling | High speed video: Chen (2014), Glottaltopogram, Body-cover model: NCVS tutorial, Story (2002) Two-layer model: Zhang (2009), Zhang et al. (2013) | ||
| 02 | Tu 09/10 | Spectrum review | exercises (pdf) | |
| Th 09/12 | Discussion of Garellek (2019) | plan | Read Kreiman et al. 2014 and post at least 5 questions and/or comments in the Box discussion thread by Monday 09/16 12pm | |
| 01 | Tu 09/03 | Voice quality around the world | notes | Due by class 09/05: Read Kreiman and Sidtis Chapters 1-2. |
| Th 09/05 | Discussion of K&S reading | notes | Due by class 09/10: Read Garellek (2019) and answer reading questions |
Readings and Exercises
Readings will be assigned approximately weekly, particularly in the first half of the course, linked from this page. Reading questions will be assigned for each reading and are due before class when the reading is due, accounting for a total of 15% of the grade.
Each student must also choose one of the readings to do a review on, i.e., a review that you would submit as part of the peer review process in journal article publication; the review will count for 15% of the grade.
Class presentation
Each student will study one of the methods we are covering in the course and be responsible for teaching it to the class, for 20% of grade.
Experimental design
Student participation in the design of the perceptual experiment---including acoustic data analysis that will help guide the design of the experiment---will be organized into a few different milestones (TBA) during the semester, the completion of which will account for a total of 50% of the grade.