Version Control for Speech Corpora

Vlad-Andrei Dumitru; Matthias Boehm; Martin Hagmüller; Barbara Schuppler

Version Control for Speech Corpora

Vlad-Andrei Dumitru^*
, Matthias Boehm
, Martin Hagmüller
, Barbara Schuppler

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference paper › peer-review

Abstract

While the audio recordings of a corpus represent the ground truth, transcriptions are – in the case of manual annotations – subject to human error, and subject to changes related to technology improvements underpinning automated annotation methods. In order to facilitate the dynamic extension of speech corpora, we introduce Speechcake, a tool for centralized version control for speech corpora, enabling the automatic check-in and merging of annotations. It considers typical workflows of phoneticians, linguists and speech technologists, and enables the development of dynamic, collaborative, and perpetually-improving speech corpora.

Original language	English
Title of host publication	Proceedings of the 20th Conference on Natural Language Processing (KONVENS 2024)
Publisher	Association for Computational Linguistics (ACL)
Pages	303-308
Publication status	Published - 2024
Event	20th Conference on Natural Language Processing, KONVENS 2024 - Vienna, Austria Duration: 10 Sept 2024 → 13 Sept 2024

Conference

Conference	20th Conference on Natural Language Processing, KONVENS 2024
Country/Territory	Austria
City	Vienna
Period	10/09/24 → 13/09/24

Access to Document

https://aclanthology.org/2024.konvens-main.30.pdfLicence: Other

FWF - Spontansprache - Cross-layer language models for conversational speech
Schuppler, B. (Consortium manager resp. coordinator with external organisations) & Schuppler, B. (Project manager on research unit)
1/11/19 → 31/10/24
Project: Research project
FWF - CLCS_2 - Cross-layer prosodic models for conversational speech
Schuppler, B. (Project manager on research unit)
1/10/18 → 30/11/21
Project: Research project

Elise Richter
Schuppler, B. (Recipient), Nov 2017
Prize: Fellowship awarded competitively

Cite this

Dumitru, V-A, Boehm, M, Hagmüller, M & Schuppler, B 2024, Version Control for Speech Corpora. in Proceedings of the 20th Conference on Natural Language Processing (KONVENS 2024). Association for Computational Linguistics (ACL), pp. 303-308, 20th Conference on Natural Language Processing, KONVENS 2024, Vienna, Austria, 10/09/24. <https://aclanthology.org/2024.konvens-main.30.pdf>

@inproceedings{b84895e138964ead8efd4a87606b0e8c,

title = "Version Control for Speech Corpora",

abstract = "While the audio recordings of a corpus represent the ground truth, transcriptions are – in the case of manual annotations – subject to human error, and subject to changes related to technology improvements underpinning automated annotation methods. In order to facilitate the dynamic extension of speech corpora, we introduce Speechcake, a tool for centralized version control for speech corpora, enabling the automatic check-in and merging of annotations. It considers typical workflows of phoneticians, linguists and speech technologists, and enables the development of dynamic, collaborative, and perpetually-improving speech corpora.",

author = "Vlad-Andrei Dumitru and Matthias Boehm and Martin Hagm{\"u}ller and Barbara Schuppler",

year = "2024",

language = "English",

pages = "303--308",

booktitle = "Proceedings of the 20th Conference on Natural Language Processing (KONVENS 2024)",

publisher = "Association for Computational Linguistics (ACL)",

note = "20th Conference on Natural Language Processing, KONVENS 2024 ; Conference date: 10-09-2024 Through 13-09-2024",

}

TY - GEN

T1 - Version Control for Speech Corpora

AU - Dumitru, Vlad-Andrei

AU - Boehm, Matthias

AU - Hagmüller, Martin

AU - Schuppler, Barbara

PY - 2024

Y1 - 2024

N2 - While the audio recordings of a corpus represent the ground truth, transcriptions are – in the case of manual annotations – subject to human error, and subject to changes related to technology improvements underpinning automated annotation methods. In order to facilitate the dynamic extension of speech corpora, we introduce Speechcake, a tool for centralized version control for speech corpora, enabling the automatic check-in and merging of annotations. It considers typical workflows of phoneticians, linguists and speech technologists, and enables the development of dynamic, collaborative, and perpetually-improving speech corpora.

AB - While the audio recordings of a corpus represent the ground truth, transcriptions are – in the case of manual annotations – subject to human error, and subject to changes related to technology improvements underpinning automated annotation methods. In order to facilitate the dynamic extension of speech corpora, we introduce Speechcake, a tool for centralized version control for speech corpora, enabling the automatic check-in and merging of annotations. It considers typical workflows of phoneticians, linguists and speech technologists, and enables the development of dynamic, collaborative, and perpetually-improving speech corpora.

M3 - Conference paper

SP - 303

EP - 308

BT - Proceedings of the 20th Conference on Natural Language Processing (KONVENS 2024)

PB - Association for Computational Linguistics (ACL)

T2 - 20th Conference on Natural Language Processing, KONVENS 2024

Y2 - 10 September 2024 through 13 September 2024

ER -

Version Control for Speech Corpora

Abstract

Conference

Access to Document

Fingerprint

Projects

FWF - Spontansprache - Cross-layer language models for conversational speech

FWF - CLCS_2 - Cross-layer prosodic models for conversational speech

Prizes

Elise Richter

Cite this