Ling 795.10 -- Group seminar

Summer 2006 Term 2 (Call No. 17439-4)

Contents

• Description
• Administration
• Schedule of group meetings
• Useful links

Description

This is a catch-bag seminar, aimed at three groups of people.

The first group is anyone who is working on a 2nd year paper, a 3rd year paper, a dissertation proposal, or whatever. With prior arrangment, anyone in this category can schedule a meeting with an audience of at least the instructor and as many as can be mustered in an e-mail advertisement to the relevant ling.osu community -- to read a paper together, to present an idea or data, to get feedback on a proposed data-gathering method or mathematical test, etc.

The second group this summer is a group of people who have already stated that they are interested in reading together some relevant papers in second-language speech perception and in second-language prosodic processing.

The third group this summer is a group of people who have a stake in learning how to manage and analyze large corpora of recordings that they want to use in analyses of speakers' vowel spaces. This group will work through a range of questions ranging from the nitty-gritty technical (e.g., "How can I write a Praat script to let me check the formant values I've extracted from all the vowels in this noisy recording I made in a barbershop.") to the profound (e.g., "In current models of sociophonetic perception, can there be even in theory any language-independent algorithm for normalizing vowel spaces across speakers?")

There will be N regularly scheduled group meeting times each week, to go over assignments made to all of the groups or to students in some specific one of the above groups. Students registered for credit will declare an affiliation with at least one (and as many as all three) of the above groups, and should sign up for as many credits as are appropriate for the affiliations that they declare. They will be expected to complete the assignments and attend the meetings that are relevant to the group(s) with which they are affiliated.

Administration

Common meeting times and Locations:

Group 1 -- ad hoc and TBA

Groups 2 & 3 -- Tuesdays, Wednesdays, Thursdays, and Fridays, 2:30-4:18, Derby 29.

Instructor:

Mary E. Beckman
office: room 21 Oxley, tel 292-0460
email: mbeckman@ling.osu.edu
office hours: Tuesdays, Thurdsays, Fridays at 1:00, & by appointment.

Schedule of meeting topics

• Monday, July 24 -- Introductions and planning goals.
• Tuesday, July 25 -- Praat scripting tutorial and R code for plotting vowels using the Hillenbrand et al. (1995) recordings.
  instructions for prep.
• Thursday, July 27 -- Adding noise to audio files to create stimuli; looking at Korean /tw/ versus /kw/; more on Hillenbrand et al. recordings.
  notes from Tuesday's class and instructions for prep.
• Friday, July 28 -- Using R to compare formants extracted using praat to formants reported in Hillenbrand et al. (1995).
  notes from Thursday's class and instructions for prep.
• First three meetings of week 2 will be in smaller groups, to get ahead on individual goals.
  notes from last Friday's class and individual assignments for this week's smaller meetings.
• Tuesday, August 1 -- Jeonghwa on KToBI and AME_ToBI tagging of stimuli and stimulus creation for L2 prosodic parsing.
  notes from this meeting
• Wednesday, August 2 -- David on tagging a continuous recording of narrative, ready to make a database from which it will be easier to identify and extract target words for analysis of the vowels for the vowel space study.
  notes from this meeting
• Thursday, August 3 -- Jocelyn on how to use the TextGrids made with tagSentences.praat to transcribe the key words in the six talkers' sentences that the listeners will hear as stimuli, and how to use these same TextGrids to extract the relevant formant measures for the vowel space analyses.
  
• Friday, August 4 -- Wrapup and comparing notes from the individual meetings. Notes from this class and instructions for preparing for next week's meetings, which will have the same organization.
• Tuesday, August 8 -- Jeonghwa on prosody in L2 sentence compression.
  notes from this meeting
• Wednesday, August 9 -- David on vowel normalization, with continuation on Thursday.
  notes from these meetings
• Thursday, August 10 -- Jocelyn on more stimulus preparation and decisions about methods.
• Friday, August 11 -- Reports to group from individual meetings, brief intro to design priniciples of ToBI annotation, looking at stop burst spectra, working on how to record directly to praat on a laptop.
• Tuesday, August 15 -- David on choosing word tokens to represent each vowel, and then extracting the tagged vowel's F1, F2, and F3 at three points.
  notes from this meeting
• Wednesday, August 16 -- Jeonghwa to work on practicing K_ToBI and AME_ToBI analyses using the ambiguous (and presumably possible to disambiguate by prosody) sentences that she recorded herself and Jocelyn reading on August 11. notes from this meeting
• Thursday, August 17 -- Jocelyn on how to do English ToBI and K-ToBI transcriptions of stimulus materials.
• Friday, August 18 -- Helen Riha presentation: "I will discuss the morphological structure and semantics of Mandarin hybrid words, which are composed of a roman letter component and a Chinese character component in their written form. Even though hybrid words are written in part with a foreign orthography, their structure and semantics reveal that they are actually types of Chinese words."
  wrapup of other three ongoing projcets
  SEIs

Back to top of page

Useful links

• Here is the homepage to go to in order to get R, the "GNU S" software package for statistical analysis and graphics.
• Here is the URL where you can download Praat and link to some user-provided tutorials.
• If you want to label audio files in Praat or EMU in the IPA but don't want to bother with the issues of how to embed an IPA font into a TextGrid file or in an EMU hlb file, you might want to look at the coventions for "transliterating" IPA symbols into the ASCII symbol described on the SAMPA home page. Alternatively, you might want to use WorldBet, which is what we are using to transcribe the children and adults participants in the paidologos project -- a cross-language study of phonological acquisition.
• You can get the recordings and the data from Hillenbrand et al. here.
• Here are some phonetics resources that are available from the OSU Linguistics Laboratory. This includes the tcl/tk code for the Method of Adjustment paradigm that you can use to test your speakers' perceptual vowel spaces.
• The home page for the ToBI framework for developing prosodic annotation tools links you to descriptions of prosodic systems for various languages, including Korean Links will be added here as the course progresses and it becomes clearer what different students' interests are.