Assessing the quality of TTS audio in the LARA learning-by-reading platform

Elham Akhlaghi; Anna Bączkowska; Harald  Berthelsen; Branislav  Bédi; Cathy Chua; Catia  Cucchiarini; Hanieh Habibi; Ivana Horváthová; Pernille Hvalsøe; Roy Lotz; Christele Maizonniaux; Neasa Ní  Chiaráin; Manny  Rayner; Nikos Tsourakis; Chunlin  Yao

doi:10.14705/rpnet.2021.54.1299

Assessing the quality of TTS audio in the LARA learning-by-reading platform

Elham Akhlaghi, Anna Bączkowska, Harald Berthelsen, Branislav Bédi, Cathy Chua, Catia Cucchiarini, Hanieh Habibi, Ivana Horváthová, Pernille Hvalsøe, Roy Lotz, Christele Maizonniaux, Neasa Ní Chiaráin, Manny Rayner, Nikos Tsourakis, Chunlin Yao

College of Humanities, Arts and Social Sciences

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

48 Downloads (Pure)

Abstract

A popular idea in Computer Assisted Language Learning (CALL) is to use multimodal annotated texts, with annotations typically including embedded audio and translations, to support L2 learning through reading. An important question is how to create the audio, which can be done either through human recording or by a Text-To-Speech (TTS) synthesis engine. We may reasonably expect TTS to be quicker and easier, but humans to be of higher quality. Here, we report a study using the open-source LARA platform and ten languages. Samples of LARA audio totaling about three and a half minutes were provided for each language in both human and TTS form; subjects used a web form to compare different versions of the same item and rate the voices as a whole. Although human voice was more often preferred, TTS achieved higher ratings in some languages and was close in others.

Original language	English
Title of host publication	CALL and professionalisation
Subtitle of host publication	short papers from EUROCALL 2021
Editors	Naouel Zoghlami, Cédric Brudermann, Cedric Sarré, Muriel Grosbois, Linda Bradley, Sylvie Thouësny
Place of Publication	France
Publisher	Research-publishing.net
Pages	1-5
Number of pages	5
ISBN (Electronic)	9782490057979
DOIs	https://doi.org/10.14705/rpnet.2021.54.1299
Publication status	Published - Dec 2021
Event	EUROCALL 2021: CALL & Professionalisation - Online conference Duration: 26 Aug 2021 → 27 Aug 2021 https://research-publishing.net/book?10.14705/rpnet.2021.54.9782490057979

Conference

Conference	EUROCALL 2021
City	Online conference
Period	26/08/21 → 27/08/21
Internet address	https://research-publishing.net/book?10.14705/rpnet.2021.54.9782490057979

Keywords

Reading
Multimodality
Text-To-Speech (TTS)
Evaluation

Access to Document

10.14705/rpnet.2021.54.1299Licence: CC BY

Final published versionFinal published version, 443 KBLicence: CC BY

Cite this

Akhlaghi, E., Bączkowska, A., Berthelsen, H., Bédi, B., Chua, C., Cucchiarini, C., Habibi, H., Horváthová, I., Hvalsøe, P., Lotz, R., Maizonniaux, C., Chiaráin, N. N., Rayner, M., Tsourakis, N., & Yao, C. (2021). Assessing the quality of TTS audio in the LARA learning-by-reading platform. In N. Zoghlami, C. Brudermann, C. Sarré, M. Grosbois, L. Bradley, & S. Thouësny (Eds.), CALL and professionalisation: short papers from EUROCALL 2021 (pp. 1-5). Research-publishing.net. https://doi.org/10.14705/rpnet.2021.54.1299

@inproceedings{a53336d2a2c84fe99f40f802fbc191ad,

title = "Assessing the quality of TTS audio in the LARA learning-by-reading platform",

abstract = "A popular idea in Computer Assisted Language Learning (CALL) is to use multimodal annotated texts, with annotations typically including embedded audio and translations, to support L2 learning through reading. An important question is how to create the audio, which can be done either through human recording or by a Text-To-Speech (TTS) synthesis engine. We may reasonably expect TTS to be quicker and easier, but humans to be of higher quality. Here, we report a study using the open-source LARA platform and ten languages. Samples of LARA audio totaling about three and a half minutes were provided for each language in both human and TTS form; subjects used a web form to compare different versions of the same item and rate the voices as a whole. Although human voice was more often preferred, TTS achieved higher ratings in some languages and was close in others.",

keywords = "Reading, Multimodality, Text-To-Speech (TTS), Evaluation",

author = "Elham Akhlaghi and Anna B{\c a}czkowska and Harald Berthelsen and Branislav B{\'e}di and Cathy Chua and Catia Cucchiarini and Hanieh Habibi and Ivana Horv{\'a}thov{\'a} and Pernille Hvals{\o}e and Roy Lotz and Christele Maizonniaux and Chiar{\'a}in, {Neasa N{\'i}} and Manny Rayner and Nikos Tsourakis and Chunlin Yao",

year = "2021",

month = dec,

doi = "10.14705/rpnet.2021.54.1299",

language = "English",

pages = "1--5",

editor = "Naouel Zoghlami and Brudermann, {C{\'e}dric } and Cedric Sarr{\'e} and Muriel Grosbois and Linda Bradley and Sylvie Thou{\"e}sny",

booktitle = "CALL and professionalisation",

publisher = "Research-publishing.net",

note = "EUROCALL 2021 : CALL & Professionalisation ; Conference date: 26-08-2021 Through 27-08-2021",

url = "https://research-publishing.net/book?10.14705/rpnet.2021.54.9782490057979",

}

Akhlaghi, E, Bączkowska, A, Berthelsen, H, Bédi, B, Chua, C, Cucchiarini, C, Habibi, H, Horváthová, I, Hvalsøe, P, Lotz, R, Maizonniaux, C, Chiaráin, NN, Rayner, M, Tsourakis, N & Yao, C 2021, Assessing the quality of TTS audio in the LARA learning-by-reading platform. in N Zoghlami, C Brudermann, C Sarré, M Grosbois, L Bradley & S Thouësny (eds), CALL and professionalisation: short papers from EUROCALL 2021. Research-publishing.net, France, pp. 1-5, EUROCALL 2021, Online conference, 26/08/21. https://doi.org/10.14705/rpnet.2021.54.1299

Assessing the quality of TTS audio in the LARA learning-by-reading platform. / Akhlaghi, Elham; Bączkowska, Anna; Berthelsen, Harald et al.
CALL and professionalisation: short papers from EUROCALL 2021. ed. / Naouel Zoghlami; Cédric Brudermann; Cedric Sarré; Muriel Grosbois; Linda Bradley; Sylvie Thouësny. France: Research-publishing.net, 2021. p. 1-5.

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

TY - GEN

T1 - Assessing the quality of TTS audio in the LARA learning-by-reading platform

AU - Akhlaghi, Elham

AU - Bączkowska, Anna

AU - Berthelsen, Harald

AU - Bédi, Branislav

AU - Chua, Cathy

AU - Cucchiarini, Catia

AU - Habibi, Hanieh

AU - Horváthová, Ivana

AU - Hvalsøe, Pernille

AU - Lotz, Roy

AU - Maizonniaux, Christele

AU - Chiaráin, Neasa Ní

AU - Rayner, Manny

AU - Tsourakis, Nikos

AU - Yao, Chunlin

PY - 2021/12

Y1 - 2021/12

N2 - A popular idea in Computer Assisted Language Learning (CALL) is to use multimodal annotated texts, with annotations typically including embedded audio and translations, to support L2 learning through reading. An important question is how to create the audio, which can be done either through human recording or by a Text-To-Speech (TTS) synthesis engine. We may reasonably expect TTS to be quicker and easier, but humans to be of higher quality. Here, we report a study using the open-source LARA platform and ten languages. Samples of LARA audio totaling about three and a half minutes were provided for each language in both human and TTS form; subjects used a web form to compare different versions of the same item and rate the voices as a whole. Although human voice was more often preferred, TTS achieved higher ratings in some languages and was close in others.

AB - A popular idea in Computer Assisted Language Learning (CALL) is to use multimodal annotated texts, with annotations typically including embedded audio and translations, to support L2 learning through reading. An important question is how to create the audio, which can be done either through human recording or by a Text-To-Speech (TTS) synthesis engine. We may reasonably expect TTS to be quicker and easier, but humans to be of higher quality. Here, we report a study using the open-source LARA platform and ten languages. Samples of LARA audio totaling about three and a half minutes were provided for each language in both human and TTS form; subjects used a web form to compare different versions of the same item and rate the voices as a whole. Although human voice was more often preferred, TTS achieved higher ratings in some languages and was close in others.

KW - Reading

KW - Multimodality

KW - Text-To-Speech (TTS)

KW - Evaluation

U2 - 10.14705/rpnet.2021.54.1299

DO - 10.14705/rpnet.2021.54.1299

M3 - Conference contribution

SP - 1

EP - 5

BT - CALL and professionalisation

A2 - Zoghlami, Naouel

A2 - Brudermann, Cédric

A2 - Sarré, Cedric

A2 - Grosbois, Muriel

A2 - Bradley, Linda

A2 - Thouësny, Sylvie

PB - Research-publishing.net

CY - France

T2 - EUROCALL 2021

Y2 - 26 August 2021 through 27 August 2021

ER -

Akhlaghi E, Bączkowska A, Berthelsen H, Bédi B, Chua C, Cucchiarini C et al. Assessing the quality of TTS audio in the LARA learning-by-reading platform. In Zoghlami N, Brudermann C, Sarré C, Grosbois M, Bradley L, Thouësny S, editors, CALL and professionalisation: short papers from EUROCALL 2021. France: Research-publishing.net. 2021. p. 1-5 doi: 10.14705/rpnet.2021.54.1299

Assessing the quality of TTS audio in the LARA learning-by-reading platform

Abstract

Conference

Keywords

Access to Document

Fingerprint

Cite this