Data collection
Our data consist of audio recordings including:
- Legacy recordings from Heart of the Storm, made available to us by the Braidwood Museum.
- Interviews we recorded in 2024-2025 with Heart of the Storm contributors to detail their demographic background and degree of local attachment.
- Over 120 interviews (typically 1-1.5hrs long) we recorded in 2024-2025 with men and women from the Braidwood and Trentham regions.
Participants range in age from 18 to 88, and work in diverse sectors including farming, building, trades, emergency services, administration, health, transport, business management, hospitality, education, the arts, and more.
Community Research Assistants recorded interviews with friends, family members, neighbours, and others. Interviews were conducted in a relaxed setting, typically in the participant’s home, and were recorded with a handheld ZOOM recorder.
Interviewers received training in ‘psychological first aid’ delivered by Phoenix Australia
to help them recognise and manage any emotional distress that might arise.
Data collection procedure was reviewed and approved by ANU Human Ethics Committee (Protocol #2023_068) and informed consent was given by participants prior to the recording.
View PDF: Demographic Form & Local Attachment Questionnaire
View PDF: Sociolinguistic Interview Topics
Transcription
Automatic Speech Recognition (Microsoft Azure AI Speech)
A transcription tool which automatically converts speech to text.
ASR transcriptions have varying degrees of accuracy.
We produce a first-pass transcription using Microsoft Azure AI Speech, via a dedicated speech-to-text transcription platform developed at Griffith University, using Microsoft technologies, with the Locale set to Australian English.
https://azure.microsoft.com/en-us/products/ai-services/ai-speech/
We import our first-pass transcriptions into Elan (see below) for manual correction.
View PDF publication: Weiss, M., Gnevsheva, K., Travis, C. & Docherty, G. (2024). Application of ASR to a sociolinguistic corpus of Australian English. Australasian Speech Science and Technology Association.
Transcription correction (Elan)
An annotation tool for audio and video recordings, where transcript and audio are aligned and can be manually corrected.
Free and open source.
Developed by The Language Archive, Max Planck Institute for Psycholinguistics.
Sample Elan file, showing Automatic Speech Recognition tier, and subsequent tiers for manual correction
Analysis
Language Brain and Behaviour Corpus Analysis Tool (LaBB-CAT)
A browser-based linguistics research tool, with a range of corpus management and analytic capabilities, including forced alignment and search functions.
Free and open source.
Developed by New Zealand Institute of Language Brain and Behaviour.
Sample search options in LaBB-CAT
Acoustic Analysis
Praat
Software package used for acoustic analysis (estimating formant frequencies, pitch, intensity, voice quality) and producing spectrograms (visual representations of sound, like the one below).
Free and open source.
Created by Boersma & Weenink, University of Amsterdam.
Sample transcription segment in Praat, downloaded from LaBB-CAT, showing some of the LaBB-CAT tiers
Statistical Analysis
R software environment
A free software environment for statistical computing.