DISTORTION MEASURES FOR SPEECH PROCESSING.

Robert M. Gray, Andres Buzo, Augustine H. Gray, Yasuo Matsuyama

Research output: Contribution to journalArticle

236 Citations (Scopus)

Abstract

Several properties, interrelations, and interpretations were developed for various speech spectral distortion measures. The principal results are 1) the development of notions of relative strength and equivalence of the various distortion measures both in a mathematical sense corresponding to subjective equivalence and in a coding sense when used in minimum distortion or nearest neighbor speech processing systems; (2) the demonstration that the Itakura-Saito and related distortion measures possess a property similar to the triangle inequality when used in nearest neighbor systems such as quantization and cluster analysis; and 3) that the Itakura-Saito and normalized model distortion measures yield efficient computation algorithms for generalized centroids or minimum distortion points of groups or clusters of speech frames, an important computation in classical cluster analysis and in algorithms for optimal quantizer design. It is argued that the Itakura-Saito and related distortions are well-suited computationally, mathematically, and intuitively for such applications.

Original languageEnglish
Pages (from-to)367-376
Number of pages10
JournalIEEE Transactions on Acoustics, Speech, and Signal Processing
VolumeASSP-28
Issue number4
Publication statusPublished - 1980 Aug
Externally publishedYes

Fingerprint

Speech processing
Cluster analysis
Demonstrations

ASJC Scopus subject areas

  • Signal Processing

Cite this

Gray, R. M., Buzo, A., Gray, A. H., & Matsuyama, Y. (1980). DISTORTION MEASURES FOR SPEECH PROCESSING. IEEE Transactions on Acoustics, Speech, and Signal Processing, ASSP-28(4), 367-376.

DISTORTION MEASURES FOR SPEECH PROCESSING. / Gray, Robert M.; Buzo, Andres; Gray, Augustine H.; Matsuyama, Yasuo.

In: IEEE Transactions on Acoustics, Speech, and Signal Processing, Vol. ASSP-28, No. 4, 08.1980, p. 367-376.

Research output: Contribution to journalArticle

Gray, RM, Buzo, A, Gray, AH & Matsuyama, Y 1980, 'DISTORTION MEASURES FOR SPEECH PROCESSING.', IEEE Transactions on Acoustics, Speech, and Signal Processing, vol. ASSP-28, no. 4, pp. 367-376.
Gray RM, Buzo A, Gray AH, Matsuyama Y. DISTORTION MEASURES FOR SPEECH PROCESSING. IEEE Transactions on Acoustics, Speech, and Signal Processing. 1980 Aug;ASSP-28(4):367-376.
Gray, Robert M. ; Buzo, Andres ; Gray, Augustine H. ; Matsuyama, Yasuo. / DISTORTION MEASURES FOR SPEECH PROCESSING. In: IEEE Transactions on Acoustics, Speech, and Signal Processing. 1980 ; Vol. ASSP-28, No. 4. pp. 367-376.
@article{4ce7c7b7f6ca43eba87938e5ab43f41d,
title = "DISTORTION MEASURES FOR SPEECH PROCESSING.",
abstract = "Several properties, interrelations, and interpretations were developed for various speech spectral distortion measures. The principal results are 1) the development of notions of relative strength and equivalence of the various distortion measures both in a mathematical sense corresponding to subjective equivalence and in a coding sense when used in minimum distortion or nearest neighbor speech processing systems; (2) the demonstration that the Itakura-Saito and related distortion measures possess a property similar to the triangle inequality when used in nearest neighbor systems such as quantization and cluster analysis; and 3) that the Itakura-Saito and normalized model distortion measures yield efficient computation algorithms for generalized centroids or minimum distortion points of groups or clusters of speech frames, an important computation in classical cluster analysis and in algorithms for optimal quantizer design. It is argued that the Itakura-Saito and related distortions are well-suited computationally, mathematically, and intuitively for such applications.",
author = "Gray, {Robert M.} and Andres Buzo and Gray, {Augustine H.} and Yasuo Matsuyama",
year = "1980",
month = "8",
language = "English",
volume = "ASSP-28",
pages = "367--376",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "4",

}

TY - JOUR

T1 - DISTORTION MEASURES FOR SPEECH PROCESSING.

AU - Gray, Robert M.

AU - Buzo, Andres

AU - Gray, Augustine H.

AU - Matsuyama, Yasuo

PY - 1980/8

Y1 - 1980/8

N2 - Several properties, interrelations, and interpretations were developed for various speech spectral distortion measures. The principal results are 1) the development of notions of relative strength and equivalence of the various distortion measures both in a mathematical sense corresponding to subjective equivalence and in a coding sense when used in minimum distortion or nearest neighbor speech processing systems; (2) the demonstration that the Itakura-Saito and related distortion measures possess a property similar to the triangle inequality when used in nearest neighbor systems such as quantization and cluster analysis; and 3) that the Itakura-Saito and normalized model distortion measures yield efficient computation algorithms for generalized centroids or minimum distortion points of groups or clusters of speech frames, an important computation in classical cluster analysis and in algorithms for optimal quantizer design. It is argued that the Itakura-Saito and related distortions are well-suited computationally, mathematically, and intuitively for such applications.

AB - Several properties, interrelations, and interpretations were developed for various speech spectral distortion measures. The principal results are 1) the development of notions of relative strength and equivalence of the various distortion measures both in a mathematical sense corresponding to subjective equivalence and in a coding sense when used in minimum distortion or nearest neighbor speech processing systems; (2) the demonstration that the Itakura-Saito and related distortion measures possess a property similar to the triangle inequality when used in nearest neighbor systems such as quantization and cluster analysis; and 3) that the Itakura-Saito and normalized model distortion measures yield efficient computation algorithms for generalized centroids or minimum distortion points of groups or clusters of speech frames, an important computation in classical cluster analysis and in algorithms for optimal quantizer design. It is argued that the Itakura-Saito and related distortions are well-suited computationally, mathematically, and intuitively for such applications.

UR - http://www.scopus.com/inward/record.url?scp=0019050955&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0019050955&partnerID=8YFLogxK

M3 - Article

VL - ASSP-28

SP - 367

EP - 376

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 4

ER -