Nonintrusive speech quality estimation using Gaussian mixture models

被引：25

作者：

Falk, TH ^{[1
]}

Chan, WY ^{[1
]}

机构：

[1] Queens Univ, Dept Elect & Comp Engn, Kingston, ON K7L 3N6, Canada

来源：

IEEE SIGNAL PROCESSING LETTERS | 2006年 / 13卷 / 02期

关键词：

Gaussian mixtures; quality assurance; quality measurement; quality of service; speech coding; speech quality; speech transmission; telephony;

D O I：

10.1109/LSP.2005.861598

中图分类号：

TM [电工技术]; TN [电子技术、通信技术];

学科分类号：

0808 ; 0809 ;

摘要：

An algorithm for nonintrusive speech quality estimation based on Gaussian mixture models (GMMs) is presented. GMMs are used to form an artificial reference model of the behavior of features of undegraded speech. Consistency measures between the degraded speech signal and the reference model serve as indicators of speech quality. Consistency values are mapped to an objective speech quality score using a multivariate adaptive regression splines function. When tested on unseen data, the proposed algorithm generally outperforms ITU-T standard P.563, which is the current "state-of-the-art" algorithm. The algorithm computes objective quality scores roughly twice as fast as P.563.

引用

页码：108 / 111

页数：4

共 16 条

[11]

ITU-T Rec. P.800, 1996, P800 ITUT

[12]

JIN C, 1996, P IEEE INT C AC SPEE, V1, P491

[13] ANIQUE: An auditory model for single-ended speech quality estimation [J].

Kim, DS .

IEEE TRANSACTIONS ON SPEECH AND AUDIO PROCESSING, 2005, 13 (05) :821-831

[14] Output-based objective speech quality measurement using continuous hidden Markov models [J].

Li, WX ;

Kubichek, RF .

SEVENTH INTERNATIONAL SYMPOSIUM ON SIGNAL PROCESSING AND ITS APPLICATIONS, VOL 1, PROCEEDINGS, 2003, :389-392

[15]

THORPE L, 1999, P IEEE WORKSH SPEECH, P144

[16]

ZHA W, 2005, EURASIP J APPL SIG P, P1410

← 1 2 →