The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. ex. Some numerals are expressed as "XNUMX".
Copyrights notice
The original paper is in English. Non-English content has been machine-translated and may contain typographical errors or mistranslations. Copyrights notice
Dans cet article, nous proposons de nouvelles approches d'amélioration de la parole basées sur la décision douce. Afin d'améliorer la fiabilité statistique de l'estimation de l'activité vocale, nous introduisons le concept de probabilité globale d'absence de parole (GSAP). Tout d’abord, nous calculons la probabilité d’absence de parole (SAP) conventionnelle, puis nous la modifions en fonction du GSAP nouvellement proposé. La modification est faite de telle sorte que le SAP ait la même valeur de GSAP en cas d'absence de parole alors qu'il est maintenu à sa valeur d'origine lorsque la parole est présente. De plus, pour améliorer les performances des SAP au niveau des queues vocales (périodes de transition de la parole au silence), nous révisons les SAP en utilisant un schéma de gueule de bois basé sur le modèle de Markov caché (HMM). De plus, nous proposons un algorithme robuste de mise à jour du bruit dans lequel la puissance du bruit est estimée non seulement pendant les périodes d'absence de parole mais également pendant l'activité vocale sur la base d'une décision douce. En outre, pour améliorer les routines de détermination SAP et de mise à jour du bruit, nous présentons un nouveau concept de rapport signal sur bruit (SNR), appelé SNR prédit dans cet article. De plus, nous démontrons que la transformée en cosinus discrète (DCT) améliore la précision de l'estimation SAP. Un certain nombre de tests montrent que la méthode proposée, appelée algorithme d'amélioration de la parole basée sur la décision douce (SESD), donne de meilleures performances que les approches conventionnelles.
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Copier
Joon-Hyuk CHANG, Nam Soo KIM, "Speech Enhancement: New Approaches to Soft Decision" in IEICE TRANSACTIONS on Information,
vol. E84-D, no. 9, pp. 1231-1240, September 2001, doi: .
Abstract: In this paper, we propose new approaches to speech enhancement based on soft decision. In order to enhance the statistical reliability in estimating speech activity, we introduce the concept of a global speech absence probability (GSAP). First, we compute the conventional speech absence probability (SAP) and then modify it according to the newly proposed GSAP. The modification is made in such a way that the SAP has the same value of GSAP in the case of speech absence while it is maintained to its original value when the speech is present. Moreover, for improving the performance of the SAP's at voice tails (transition periods from speech to silence), we revise the SAP's using a hang-over scheme based on the hidden Markov model (HMM). In addition, we suggest a robust noise update algorithm in which the noise power is estimated not only in the periods of speech absence but also during speech activity based on soft decision. Also, for improving the SAP determination and noise update routines, we present a new signal to noise ratio (SNR) concept which is called the predicted SNR in this paper. Moreover, we demonstrate that the discrete cosine transform (DCT) enhances the accuracy of the SAP estimation. A number of tests show that the proposed method which is called the speech enhancement based on soft decision (SESD) algorithm yields better performance than the conventional approaches.
URL: https://global.ieice.org/en_transactions/information/10.1587/e84-d_9_1231/_p
Copier
@ARTICLE{e84-d_9_1231,
author={Joon-Hyuk CHANG, Nam Soo KIM, },
journal={IEICE TRANSACTIONS on Information},
title={Speech Enhancement: New Approaches to Soft Decision},
year={2001},
volume={E84-D},
number={9},
pages={1231-1240},
abstract={In this paper, we propose new approaches to speech enhancement based on soft decision. In order to enhance the statistical reliability in estimating speech activity, we introduce the concept of a global speech absence probability (GSAP). First, we compute the conventional speech absence probability (SAP) and then modify it according to the newly proposed GSAP. The modification is made in such a way that the SAP has the same value of GSAP in the case of speech absence while it is maintained to its original value when the speech is present. Moreover, for improving the performance of the SAP's at voice tails (transition periods from speech to silence), we revise the SAP's using a hang-over scheme based on the hidden Markov model (HMM). In addition, we suggest a robust noise update algorithm in which the noise power is estimated not only in the periods of speech absence but also during speech activity based on soft decision. Also, for improving the SAP determination and noise update routines, we present a new signal to noise ratio (SNR) concept which is called the predicted SNR in this paper. Moreover, we demonstrate that the discrete cosine transform (DCT) enhances the accuracy of the SAP estimation. A number of tests show that the proposed method which is called the speech enhancement based on soft decision (SESD) algorithm yields better performance than the conventional approaches.},
keywords={},
doi={},
ISSN={},
month={September},}
Copier
TY - JOUR
TI - Speech Enhancement: New Approaches to Soft Decision
T2 - IEICE TRANSACTIONS on Information
SP - 1231
EP - 1240
AU - Joon-Hyuk CHANG
AU - Nam Soo KIM
PY - 2001
DO -
JO - IEICE TRANSACTIONS on Information
SN -
VL - E84-D
IS - 9
JA - IEICE TRANSACTIONS on Information
Y1 - September 2001
AB - In this paper, we propose new approaches to speech enhancement based on soft decision. In order to enhance the statistical reliability in estimating speech activity, we introduce the concept of a global speech absence probability (GSAP). First, we compute the conventional speech absence probability (SAP) and then modify it according to the newly proposed GSAP. The modification is made in such a way that the SAP has the same value of GSAP in the case of speech absence while it is maintained to its original value when the speech is present. Moreover, for improving the performance of the SAP's at voice tails (transition periods from speech to silence), we revise the SAP's using a hang-over scheme based on the hidden Markov model (HMM). In addition, we suggest a robust noise update algorithm in which the noise power is estimated not only in the periods of speech absence but also during speech activity based on soft decision. Also, for improving the SAP determination and noise update routines, we present a new signal to noise ratio (SNR) concept which is called the predicted SNR in this paper. Moreover, we demonstrate that the discrete cosine transform (DCT) enhances the accuracy of the SAP estimation. A number of tests show that the proposed method which is called the speech enhancement based on soft decision (SESD) algorithm yields better performance than the conventional approaches.
ER -