Jonathan and Tristan's Big Phonetic Project

This page is here for collaboration on Jonathan and Tristan's Big Phonetic Project. Research questions, plans for proceeding, progress, and results will all be posted here.

Research Questions

1. What are the subjects' phonetic vowel spaces like?

Jonathan-specific

1. What are [ə] and [ɨ] like in various environments (in open syllables, and closed syllables, closed syllables before labials, etc)?
2. What is the vowel space like before /ɻ/ codas (does /Vɻ/ count as a nucleus, or a nucleus and a coda? i.e., does it count as a diphthong or not?)

Tristan-specific

1. What is the vowel space like before /l/?
2. What is /k/ like before each vowel? (i.e. is there a categorical distinction of front vs back /k/, or is it gradient?)

What needs to be done

Processing

• divide recordings into files by page
• choose which repetitions of what to measure (middle 3 of total 5 tokens per word?)

Phonetic Vowel Space

• Things we want to know about each vowel:
  • F1, F2, length, diphthongisation
  • and on a whim: F0 (voicing?), F3 (ɻ/ø)

• So we need to measure:
  • length
  • F0 at start
  • F1 at start
  • F2 at start
  • F3 at start
  • F0 at middle
  • F1 at middle
  • F2 at middle
  • F3 at middle
  • F0 at end
  • F1 at end
  • F2 at end
  • F3 at end

Random Decisions

• Jonathan's /Vɻ/ clusters get measured as both VV and VC

Fields for storage of extracted data

• unique id; e.g. "JNWroses3"
• word in use; e.g. "roses"
• thing being measured (with phonemic transcription?); e.g. "ɨ (/ɻoz+ɨz/?)"
• all measured fields above; e.g. F0start: "150Hz"

Progress

To do

• (both) Decide on data storage medium
  • test gnumeric's python API
• (both) Divide recordings into pages and standardise numbering; e.g. JNWp2, TAMp5
• (JNW) Write praat script
• (both) measure!
• (JNW) Write gnuplot graphing script