Harry Woodhouse

My blog :^)

About me...

I am an undergraduate student conducting research into articulatory phonetics. Currently developing MOSAIC - a research tool used to observe how articulation correlates to speech-sounds. This research is being conducted at the University of York, department of Language and Linguistic Science.

Currently

Developing MOSAIC - A computational system which is used to run statistical analysis on correlation studies analysing articulatory phonetics.

LinkedIn GitHub Google Scholar

Research & Projects

Carstens AG501 Articulograph interface for EMA research
AG501 Sensor Interface
Active Development

MOSAIC: Motor-Oral Speech Analysis & Integrated Computation

A computational framework under the GLPv3 license designed to analyze physical articulation and acoustic output. MOSAIC can perform statistical analysis on provided datasets, allowing for more precise observations of how speech-sounds are physically produced.

  • Built using Rust, Python & R
  • Will nateively run OpenFace and all common EMA datasets
  • Current release: v0.2.0 alpha
Example of Praat software
Praat Example
Active Development

RustMouth

RustMouth is a Praat API for the Rust programming language. It was started on 02/02/2026 and is released as version 0.0.3 as a Rust Crate. It is currently heavily experimental and is for MacOS only (soon to be multi-platform). It is made up of the low level praat-sys package which compiles Praat into a library and crawls for all cpp headers and rustmouth, the actual API.

  • Built using Rust and C++
  • You can use all Praat commands in a Rust script!
  • RustMouth Current release: v0.0.3
  • Praat-sys Current release: v0.0.3
Carstens AG501 Articulograph sensor applied to participants upper lip.
AG501 Sensor Application
Active Research

Correlational Study of PWS Oromotor Activity to Known Speech Disorder Diagnostic Metrics

A study attempting to correlate left-sided articulation with disfluent speech in people with stutters by comparing the articulation of people with and without stutters under specific speech conditions.

  • Analysed using MOSAIC
  • Uses 2D facial tracking data from OpenFace for preliminary pilot-study data.