The Transcription Acceleration Project (TAP) is a Centre of Excellence for the Dynamics of Language (CoEDL) cross-disciplinary project aimed at identifying the workflows of linguists and language workers during transcription, and developing assistive tools to accelerate that process.
Tasks such as transcribing recorded audio can be very slow. Using contemporary software techniques such as speech recognition, we can improve the the transcription experience, resulting in practical and psychological benefits for people's work.
Technologies such as Google's Cloud Speech can transcribe 100+ of the world's languages, but these tools don't support any endangered languages. For languages with small quantities of data, or for research situations which prohibit the use of cloud technologies, TAP is developing Elpis, a tool to obtain a first-pass transcription on untranscribed audio. This best-guess is used as a canvas for the language worker to refine. Elpis brings cutting-edge speech recognition technology within reach of language workers and researchers who don’t have backgrounds in speech engineering. With Elpis, we hope to enable community members to make a significant impact on transcribing their own languages from new recordings, or breath life into archival, cultural heritage material.
This software is the product of academic research funded by the Australian Research Council Centre of Excellence for the Dynamics of Language. If you use the software or code in an academic setting, please cite it appropriately as follows:
Foley, B., Arnold, J., Coto-Solano, R., Durantin, G., Ellison, T. M., van Esch, D., Heath, S., Kratochvíl, F., Maxwell-Smith, Z., Nash, D., Olsson, O., Richards, M., San, N., Stoakes, H., Thieberger, N. & Wiles, J. (2018). Building Speech Recognition Systems for Language Documentation: The CoEDL Endangered Language Pipeline and Inference System (Elpis). In S. S. Agrawal (Ed.), The 6th Intl. Workshop on Spoken Language Technologies for Under-Resourced Languages (SLTU) (pp. 200–204). Available on https://www.isca-speech.org/archive/SLTU_2018/pdfs/Ben.pdf.