L590R: Sociophonetics

This page houses the tutorials for L590R Sociophonetics. The ultimate goal is for interested students to go from zero coding experience to being able to create a good vowel plot.

Series 1: Getting to a basic vowel plot

In this series, we’ll go from zero coding experience to creating a basic vowel plot in R.

Getting started with R and RStudio

This tutorial will give a little background on R and RStudio and will walk you through the (simple) installation process. It should take just a few minutes. You won’t need to do any coding.

As a supplement, you can look at this tutorial by Michael Hunt Gardner. It overlaps with mine a bit, but it’s nice to get a slid foundation.

R Basics

This week, you’ll get your hands dirty with R. You’ll learn the basics of a few commands, how to get data into R, some basic visuals, and where to go for help. If you’re stuck on getting data into R, please see this dedicated tutorial on that by Michael Hunt Gardner. Another one of his overlaps with mine, but you can work through it for additional reinforcement.

Tidyverse

Learn about an entire ecosystem of additional R packages that makes analysis and processing easier. This will based on Part 1 and Part 2 of a previous workshop. Michael Hunt Gardner has a nice tutorial that is similar. Fair warning that the code for the last part of my Part II tutorial is a little out of date, so you should switch to this blog post for the updated code. That post also has some specific guidance that is relevant for vowel data.

ggplot2

Learn the basics of plotting using ggplot2. This will be based on Part 1 and Part 2 (and supplement) of a previous workshop.

F1-F2 plots

Learn how to make a basic scatter plot of vowel formants. This tutorial is what you’ll need to know.

Plotting vowel trajectories

Learn how to work with vowel trajectories and the two main ways you can plot those. I have my own tutorial and this is another good one.

Series 2: Collecting Phonetic Data

In this series, we’ll go back to the beginning stages of the data collection process. We’ll see transcription, forced-alignment, and formant extraction—basically all the skills you need to go from raw audio to a spreadsheet of usable formants.

DARLA

Do the complete pipeline from start to finish.

Whisper

Automatically transcribe your audio with AI.

MFA

Align utterance-level transcriptions to audio to create phoneme-level transcriptions. You can start with Eleanor Chodroff’s tutorial but I’ll probably fill it in with more specifics.

Basic Praat

Learn the basics of Praat and Praat scripting. Based on this workshop on Praat generally and this one one Praat scripting.

Extracting vowel formants with Praat

Learn how to extract vowel formants using a Praat script. Based on this tutorial and supplement.

FastTrack

Extract vowel formant measurements.

New-Fave

Extract vowel formant measurements using the lattest algorithms.

Series 3: Data processing

In this series, we’ll look at the middle stage of data processing: going from a raw spreadsheet of formant measurements to cleaned up versions suitable for statistical analysis and data visualization.

Coding allophones

Not all allophones of a phoneme should be treated the same. Learn how to split your data up into various allophones so they can be handled individually.

Outliers

Learn various ways to detect and handle outliers and extreme measurements.

Normalizing

Every vocal tract is different, so their formant measurements will be different. Learn how to transform your data so you can compare speakers apples-to-apples. I’ll probably just have you read Joe Fruehwald’s tutorial. If we need to talk about Barks, I can bring up my barktools package.

Order of operaions

Learn how the order of these processing steps can affect your data and what the recommended order should be. I’ll probably mention this.

Series 4: Data visualization

In this series, we’ll just focus on making good plots. We’ll look at aesthetics and what makes for a good plot. We’ll also explore additional ways of visualizing data.

Color

Learn to use color appropriately. Based on this workshop.

Themes and aesthetics

Learn to make your plot look nice and how to consistently apply that look across all your plots. Based on this early blog post and supplement as well as this general data visualization workshop.

Other kinds of plots

[Curved text]/blog/geomtextpath/).

Animations

Learn to make really cool animations of vowel data! A sneak peek of animations can be found in this blog post and on 3D vowel plots in this blog post

Additional analysis

Measuring vowel overlap

My tutorial on Pillai scores and Bhattacharyya’s Affinity Part 1 and Part 2.

Statistical analysis

I defer to Michael Gardner’s suite of statistics tutorials on the Ling Methods Hub, starting with this one.