ITU

FT Speech: Danish Parliament Speech Corpus

Research output: Conference Article in Proceeding or Book/Report chapterArticle in proceedingsResearchpeer-review

Standard

FT Speech: Danish Parliament Speech Corpus. / Kirkedal, Andreas Søeborg; Stepanovic, Marija; Plank, Barbara.

INTERSPEECH 2020. International Speech Communication Association (ISCA), 2020. (Annual Conference of the International Speech Communication Association).

Research output: Conference Article in Proceeding or Book/Report chapterArticle in proceedingsResearchpeer-review

Harvard

Kirkedal, AS, Stepanovic, M & Plank, B 2020, FT Speech: Danish Parliament Speech Corpus. in INTERSPEECH 2020. International Speech Communication Association (ISCA), Annual Conference of the International Speech Communication Association. https://doi.org/10.21437/Interspeech.2020-3164

APA

Kirkedal, A. S., Stepanovic, M., & Plank, B. (2020). FT Speech: Danish Parliament Speech Corpus. In INTERSPEECH 2020 International Speech Communication Association (ISCA). Annual Conference of the International Speech Communication Association https://doi.org/10.21437/Interspeech.2020-3164

Vancouver

Kirkedal AS, Stepanovic M, Plank B. FT Speech: Danish Parliament Speech Corpus. In INTERSPEECH 2020. International Speech Communication Association (ISCA). 2020. (Annual Conference of the International Speech Communication Association). https://doi.org/10.21437/Interspeech.2020-3164

Author

Kirkedal, Andreas Søeborg ; Stepanovic, Marija ; Plank, Barbara. / FT Speech: Danish Parliament Speech Corpus. INTERSPEECH 2020. International Speech Communication Association (ISCA), 2020. (Annual Conference of the International Speech Communication Association).

Bibtex

@inproceedings{592b099d76814c81bfc57b9ce5ac7d7b,
title = "FT Speech: Danish Parliament Speech Corpus",
abstract = "This paper introduces FT Speech, a new speech corpus created from the recorded meetings of the Danish Parliament, otherwise known as the Folketing (FT). The corpus contains over 1,800 hours of transcribed speech by a total of 434 speakers. It is significantly larger in duration, vocabulary, and amount of spontaneous speech than the existing public speech corpora for Danish, which are largely limited to read-aloud and dictation data. We outline design considerations, including the preprocessing methods and the alignment procedure. To evaluate the quality of the corpus, we train automatic speech recognition systems on the new resource and compare them to the systems trained on the Danish part of Spr{\aa}kbanken, the largest public ASR corpus for Danish to date. Our baseline results show that we achieve a 14.01 WER on the new corpus. A combination of FT Speech with in-domain language data provides comparable results to models trained specifically on Spr{\aa}kbanken, showing that FT Speech transfers well to this data set. Interestingly, our results demonstrate that the opposite is not the case. This shows that FT Speech provides a valuable resource for promoting research on Danish ASR with more spontaneous speech.",
keywords = "speech corpus, speech recognition, Danish language",
author = "Kirkedal, {Andreas S{\o}eborg} and Marija Stepanovic and Barbara Plank",
year = "2020",
doi = "10.21437/Interspeech.2020-3164",
language = "English",
series = "Annual Conference of the International Speech Communication Association",
booktitle = "INTERSPEECH 2020",
publisher = "International Speech Communication Association (ISCA)",

}

RIS

TY - GEN

T1 - FT Speech: Danish Parliament Speech Corpus

AU - Kirkedal, Andreas Søeborg

AU - Stepanovic, Marija

AU - Plank, Barbara

PY - 2020

Y1 - 2020

N2 - This paper introduces FT Speech, a new speech corpus created from the recorded meetings of the Danish Parliament, otherwise known as the Folketing (FT). The corpus contains over 1,800 hours of transcribed speech by a total of 434 speakers. It is significantly larger in duration, vocabulary, and amount of spontaneous speech than the existing public speech corpora for Danish, which are largely limited to read-aloud and dictation data. We outline design considerations, including the preprocessing methods and the alignment procedure. To evaluate the quality of the corpus, we train automatic speech recognition systems on the new resource and compare them to the systems trained on the Danish part of Språkbanken, the largest public ASR corpus for Danish to date. Our baseline results show that we achieve a 14.01 WER on the new corpus. A combination of FT Speech with in-domain language data provides comparable results to models trained specifically on Språkbanken, showing that FT Speech transfers well to this data set. Interestingly, our results demonstrate that the opposite is not the case. This shows that FT Speech provides a valuable resource for promoting research on Danish ASR with more spontaneous speech.

AB - This paper introduces FT Speech, a new speech corpus created from the recorded meetings of the Danish Parliament, otherwise known as the Folketing (FT). The corpus contains over 1,800 hours of transcribed speech by a total of 434 speakers. It is significantly larger in duration, vocabulary, and amount of spontaneous speech than the existing public speech corpora for Danish, which are largely limited to read-aloud and dictation data. We outline design considerations, including the preprocessing methods and the alignment procedure. To evaluate the quality of the corpus, we train automatic speech recognition systems on the new resource and compare them to the systems trained on the Danish part of Språkbanken, the largest public ASR corpus for Danish to date. Our baseline results show that we achieve a 14.01 WER on the new corpus. A combination of FT Speech with in-domain language data provides comparable results to models trained specifically on Språkbanken, showing that FT Speech transfers well to this data set. Interestingly, our results demonstrate that the opposite is not the case. This shows that FT Speech provides a valuable resource for promoting research on Danish ASR with more spontaneous speech.

KW - speech corpus

KW - speech recognition

KW - Danish language

U2 - 10.21437/Interspeech.2020-3164

DO - 10.21437/Interspeech.2020-3164

M3 - Article in proceedings

T3 - Annual Conference of the International Speech Communication Association

BT - INTERSPEECH 2020

PB - International Speech Communication Association (ISCA)

ER -

ID: 85246789