Pushing the envelope - Aside

Nelson Morgan; Qifeng Zhu; Andreas Stolcke; Kemal Sönmez; Sunil Sivadas; Takahiro Shinozaki; Mari Ostendorf; Pratibha Jain; Hynek Hermansky; Dan Ellis; George Doddington; Barry Chen; Özgür Çetin; Hervé Bourlard; Marios Athineos

doi:10.1109/MSP.2005.1511826

Pushing the envelope - Aside

Nelson Morgan, Qifeng Zhu, Andreas Stolcke, Kemal Sönmez, Sunil Sivadas, Takahiro Shinozaki, Mari Ostendorf, Pratibha Jain, Hynek Hermansky, Dan Ellis, George Doddington, Barry Chen, Özgür Çetin, Hervé Bourlard, Marios Athineos

Research output: Contribution to journal › Article › peer-review

72 Scopus citations

Abstract

The Defense Advanced Research Projects Agency (DARPA) sponsored a project known as Novel Approaches which is a component of the Effective Affordable Reusable Speech-to-text (EARS) program. The main goal for this effort was to push the spectral envelope away from being the sole source of acoustic information incorporated by the statistical models of modern speech recognition systems (SRSs). To date, much of the effort has gone towards the design of new features and experimentation with their incorporation in a modern speech-to-text system. The new features have already provided significant improvements in such a system in the 2004 NIST evaluation of recognizers of conversational telephone speech.

Original language	English (US)
Pages (from-to)	81-88
Number of pages	8
Journal	IEEE Signal Processing Magazine
Volume	22
Issue number	5
DOIs	https://doi.org/10.1109/MSP.2005.1511826
State	Published - Sep 2005
Externally published	Yes

ASJC Scopus subject areas

Signal Processing
Electrical and Electronic Engineering
Applied Mathematics

Access to Document

10.1109/MSP.2005.1511826

Fingerprint Dive into the research topics of 'Pushing the envelope - Aside'. Together they form a unique fingerprint.

Cite this

Pushing the envelope - Aside. / Morgan, Nelson; Zhu, Qifeng; Stolcke, Andreas; Sönmez, Kemal; Sivadas, Sunil; Shinozaki, Takahiro; Ostendorf, Mari; Jain, Pratibha; Hermansky, Hynek; Ellis, Dan; Doddington, George; Chen, Barry; Çetin, Özgür; Bourlard, Hervé; Athineos, Marios.

In: IEEE Signal Processing Magazine, Vol. 22, No. 5, 09.2005, p. 81-88.

Research output: Contribution to journal › Article › peer-review

@article{024804faeb1d4d71b54efb311daecc99,

title = "Pushing the envelope - Aside",

abstract = "The Defense Advanced Research Projects Agency (DARPA) sponsored a project known as Novel Approaches which is a component of the Effective Affordable Reusable Speech-to-text (EARS) program. The main goal for this effort was to push the spectral envelope away from being the sole source of acoustic information incorporated by the statistical models of modern speech recognition systems (SRSs). To date, much of the effort has gone towards the design of new features and experimentation with their incorporation in a modern speech-to-text system. The new features have already provided significant improvements in such a system in the 2004 NIST evaluation of recognizers of conversational telephone speech.",

author = "Nelson Morgan and Qifeng Zhu and Andreas Stolcke and Kemal S{\"o}nmez and Sunil Sivadas and Takahiro Shinozaki and Mari Ostendorf and Pratibha Jain and Hynek Hermansky and Dan Ellis and George Doddington and Barry Chen and {\"O}zg{\"u}r {\c C}etin and Herv{\'e} Bourlard and Marios Athineos",

note = "Funding Information: This work was made possible by funding from the DARPA EARS “Novel Approaches” Grant MDA972-02-1-0024. We are also grateful to Jeff Bilmes for his work on GMTK, which was extensively used in the multirate/variable-rate experiments. The anonymous reviewers also made many useful suggestions, many of which we actually used. Finally, we thank the many other contributors in the speech groups at our five sites who contributed to our methods over the last three years. Copyright: Copyright 2017 Elsevier B.V., All rights reserved.",

year = "2005",

month = sep,

doi = "10.1109/MSP.2005.1511826",

language = "English (US)",

volume = "22",

pages = "81--88",

journal = "IEEE Acoustics, Speech, and Signal Processing Newsletter",

issn = "1053-5888",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

number = "5",

}

TY - JOUR

T1 - Pushing the envelope - Aside

AU - Morgan, Nelson

AU - Zhu, Qifeng

AU - Stolcke, Andreas

AU - Sönmez, Kemal

AU - Sivadas, Sunil

AU - Shinozaki, Takahiro

AU - Ostendorf, Mari

AU - Jain, Pratibha

AU - Hermansky, Hynek

AU - Ellis, Dan

AU - Doddington, George

AU - Chen, Barry

AU - Çetin, Özgür

AU - Bourlard, Hervé

AU - Athineos, Marios

N1 - Funding Information: This work was made possible by funding from the DARPA EARS “Novel Approaches” Grant MDA972-02-1-0024. We are also grateful to Jeff Bilmes for his work on GMTK, which was extensively used in the multirate/variable-rate experiments. The anonymous reviewers also made many useful suggestions, many of which we actually used. Finally, we thank the many other contributors in the speech groups at our five sites who contributed to our methods over the last three years. Copyright: Copyright 2017 Elsevier B.V., All rights reserved.

PY - 2005/9

Y1 - 2005/9

N2 - The Defense Advanced Research Projects Agency (DARPA) sponsored a project known as Novel Approaches which is a component of the Effective Affordable Reusable Speech-to-text (EARS) program. The main goal for this effort was to push the spectral envelope away from being the sole source of acoustic information incorporated by the statistical models of modern speech recognition systems (SRSs). To date, much of the effort has gone towards the design of new features and experimentation with their incorporation in a modern speech-to-text system. The new features have already provided significant improvements in such a system in the 2004 NIST evaluation of recognizers of conversational telephone speech.

AB - The Defense Advanced Research Projects Agency (DARPA) sponsored a project known as Novel Approaches which is a component of the Effective Affordable Reusable Speech-to-text (EARS) program. The main goal for this effort was to push the spectral envelope away from being the sole source of acoustic information incorporated by the statistical models of modern speech recognition systems (SRSs). To date, much of the effort has gone towards the design of new features and experimentation with their incorporation in a modern speech-to-text system. The new features have already provided significant improvements in such a system in the 2004 NIST evaluation of recognizers of conversational telephone speech.

UR - http://www.scopus.com/inward/record.url?scp=85032751546&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85032751546&partnerID=8YFLogxK

U2 - 10.1109/MSP.2005.1511826

DO - 10.1109/MSP.2005.1511826

M3 - Article

AN - SCOPUS:85032751546

VL - 22

SP - 81

EP - 88

JO - IEEE Acoustics, Speech, and Signal Processing Newsletter

JF - IEEE Acoustics, Speech, and Signal Processing Newsletter

SN - 1053-5888

IS - 5

ER -

Pushing the envelope - Aside

Abstract

ASJC Scopus subject areas

Access to Document

Other files and links

Cite this