Spectral modification for recognition of children’s speech under mismatched conditions

Hemant Kathania; Sudarsana Kadiri; Paavo Alku; Mikko Kurimo

Spectral modification for recognition of children’s speech under mismatched conditions

Hemant Kathania, Sudarsana Kadiri, Paavo Alku, Mikko Kurimo

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

63 Lataukset (Pure)

Abstrakti

n this paper, we propose spectral modification by sharpening formants and by reducing the spectral tilt to recognize children’s speech by automatic speech recognition (ASR) systems developed using adult speech. In this type of mismatched condition, the ASR performance is degraded due to the acoustic and linguistic mismatch in the attributes between children and adult speakers. The proposed method is used to improve the speech intelligibility to enhance the children’s speech recognition using an acoustic model trained on adult speech. In the experiments, WSJCAM0 and PFSTAR are used as databases for adults’ and children’s speech, respectively. The proposed technique gives a significant improvement in the context of the DNN-HMM-based ASR. Furthermore, we validate the robustness of the technique by showing that it performs well also in mismatched noise conditions.

Alkuperäiskieli	Englanti
Otsikko	Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa)
Julkaisupaikka	Sweden
Kustantaja	Linköping University Electronic Press
Sivut	94–100
Sivumäärä	7
ISBN (elektroninen)	978-91-7929-614-8
Tila	Julkaistu - 31 toukok. 2021
OKM-julkaisutyyppi	A4 Artikkeli konferenssijulkaisussa
Tapahtuma	Nordic Conference on Computational Linguistics - Reykjavik, Islanti Kesto: 31 toukok. 2021 → 2 kesäk. 2021

Julkaisusarja

Nimi	Linköping electronic conference proceedings
Numero	178
ISSN (painettu)	1650-3686
ISSN (elektroninen)	1650-3740

Conference

Conference	Nordic Conference on Computational Linguistics
Lyhennettä	NoDaLiDa
Maa/Alue	Islanti
Kaupunki	Reykjavik
Ajanjakso	31/05/2021 → 02/06/2021

Pääsy asiakirjaan

2021.nodalida-main.10Final published version, 1,4 MBLisenssi: CC BY-NC-SA

https://www.aclweb.org/anthology/2021.nodalida-main.10

-: Elokuvien muovaama Suomi: suomalainen näytelmäelokuva audiovisuaalisena big datana 1907-2017
Kurimo, M. (Vastuullinen tutkija), Virkkunen, A. (Projektin jäsen), Moisio, A. (Projektin jäsen), Porjazovski, D. (Projektin jäsen) & Kathania, H. (Projektin jäsen)
01/01/2020 → 31/12/2022
Projekti: Academy of Finland: Other research funding

Science-IT
Hakala, M. (Manager)
Perustieteiden korkeakoulu
Laitteistot/tilat: Facility

Siteeraa tätä

Kathania, H., Kadiri, S., Alku, P., & Kurimo, M. (2021). Spectral modification for recognition of children’s speech under mismatched conditions. teoksessa Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa) (Sivut 94–100). (Linköping electronic conference proceedings; Nro 178). Linköping University Electronic Press. https://www.aclweb.org/anthology/2021.nodalida-main.10

@inproceedings{f8d4629ad7ea4a3db10fce0cbed26c35,

title = "Spectral modification for recognition of children{\textquoteright}s speech under mismatched conditions",

abstract = "n this paper, we propose spectral modification by sharpening formants and by reducing the spectral tilt to recognize children{\textquoteright}s speech by automatic speech recognition (ASR) systems developed using adult speech. In this type of mismatched condition, the ASR performance is degraded due to the acoustic and linguistic mismatch in the attributes between children and adult speakers. The proposed method is used to improve the speech intelligibility to enhance the children{\textquoteright}s speech recognition using an acoustic model trained on adult speech. In the experiments, WSJCAM0 and PFSTAR are used as databases for adults{\textquoteright} and children{\textquoteright}s speech, respectively. The proposed technique gives a significant improvement in the context of the DNN-HMM-based ASR. Furthermore, we validate the robustness of the technique by showing that it performs well also in mismatched noise conditions.",

author = "Hemant Kathania and Sudarsana Kadiri and Paavo Alku and Mikko Kurimo",

year = "2021",

month = may,

day = "31",

language = "English",

series = "Link{\"o}ping electronic conference proceedings",

publisher = "Link{\"o}ping University Electronic Press",

number = "178",

pages = "94–100",

booktitle = "Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa)",

address = "Sweden",

note = "Nordic Conference on Computational Linguistics, NoDaLiDa ; Conference date: 31-05-2021 Through 02-06-2021",

}

Kathania, H, Kadiri, S, Alku, P & Kurimo, M 2021, Spectral modification for recognition of children’s speech under mismatched conditions. julkaisussa Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa). Linköping electronic conference proceedings, Nro 178, Linköping University Electronic Press, Sweden, Sivut 94–100, Nordic Conference on Computational Linguistics, Reykjavik, Islanti, 31/05/2021. <https://www.aclweb.org/anthology/2021.nodalida-main.10>

Spectral modification for recognition of children’s speech under mismatched conditions. / Kathania, Hemant; Kadiri, Sudarsana; Alku, Paavo et al.
Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa). Sweden: Linköping University Electronic Press, 2021. s. 94–100 (Linköping electronic conference proceedings; Nro 178).

Tutkimustuotos: Artikkeli kirjassa/konferenssijulkaisussa › Conference article in proceedings › Scientific › vertaisarvioitu

TY - GEN

T1 - Spectral modification for recognition of children’s speech under mismatched conditions

AU - Kathania, Hemant

AU - Kadiri, Sudarsana

AU - Alku, Paavo

AU - Kurimo, Mikko

PY - 2021/5/31

Y1 - 2021/5/31

N2 - n this paper, we propose spectral modification by sharpening formants and by reducing the spectral tilt to recognize children’s speech by automatic speech recognition (ASR) systems developed using adult speech. In this type of mismatched condition, the ASR performance is degraded due to the acoustic and linguistic mismatch in the attributes between children and adult speakers. The proposed method is used to improve the speech intelligibility to enhance the children’s speech recognition using an acoustic model trained on adult speech. In the experiments, WSJCAM0 and PFSTAR are used as databases for adults’ and children’s speech, respectively. The proposed technique gives a significant improvement in the context of the DNN-HMM-based ASR. Furthermore, we validate the robustness of the technique by showing that it performs well also in mismatched noise conditions.

AB - n this paper, we propose spectral modification by sharpening formants and by reducing the spectral tilt to recognize children’s speech by automatic speech recognition (ASR) systems developed using adult speech. In this type of mismatched condition, the ASR performance is degraded due to the acoustic and linguistic mismatch in the attributes between children and adult speakers. The proposed method is used to improve the speech intelligibility to enhance the children’s speech recognition using an acoustic model trained on adult speech. In the experiments, WSJCAM0 and PFSTAR are used as databases for adults’ and children’s speech, respectively. The proposed technique gives a significant improvement in the context of the DNN-HMM-based ASR. Furthermore, we validate the robustness of the technique by showing that it performs well also in mismatched noise conditions.

M3 - Conference article in proceedings

T3 - Linköping electronic conference proceedings

SP - 94

EP - 100

BT - Proceedings of the 23rd Nordic Conference on Computational Linguistics (NoDaLiDa)

PB - Linköping University Electronic Press

CY - Sweden

T2 - Nordic Conference on Computational Linguistics

Y2 - 31 May 2021 through 2 June 2021

ER -

Spectral modification for recognition of children’s speech under mismatched conditions

Abstrakti

Julkaisusarja

Conference

Pääsy asiakirjaan

Sormenjälki

Projektit

-: Elokuvien muovaama Suomi: suomalainen näytelmäelokuva audiovisuaalisena big datana 1907-2017

Laitteet

Science-IT

Siteeraa tätä