On Reliability of Log-Spectral Distortion Measure in Speech Quality Estimation

被引：0

作者：

Prodeus, Arkadiy ^{[1
]}

Kotvytskyi, Igor ^{[1
]}

机构：

[1] NTUU Igor Sikorsky KPI, Fac Elect, Acoust & Acoustoelect Dept, Kiev, Ukraine

来源：

2017 IEEE 4TH INTERNATIONAL CONFERENCE ACTUAL PROBLEMS OF UNMANNED AERIAL VEHICLES DEVELOPMENTS (APUAVD) | 2017年

关键词：

speech quality; log-spectral distortion; speech signals limited in the frequency band; signal model; signal spectrum;

D O I：

暂无

中图分类号：

TP [自动化技术、计算机技术];

学科分类号：

0812 ;

摘要：

In this paper, the results of experimental studies of the reasons for the low reliability of the log-spectral distortion (LSD) measure for estimating the quality of speech signals limited in the frequency band are presented. It is shown that the LSD measure has an increased sensitivity not only to the speech formants, but also to the shape of frequency response of low-frequency filters limiting the frequency band. It is also shown that normalization by variance of band-limited signals leads to LSD measure anomalous behavior. Simple modification of the LSD estimate is proposed, which makes it possible to increase its statistical stability.

引用

页码：121 / 124

页数：4

共 50 条

[31] Log-Rayleigh distribution: A simple and efficient statistical representation of log-spectral coefficients
Rivet, Bertrand
Girin, Laurent
Jutten, Christian
IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2007, 15 (03): : 796 - 802
[32] Evaluation of speech quality based on wavelet spectrum distortion measure
School of Electronics and Information Engineering, Xi'an Jiaotong University, Xi'an 710049, China
不详
Binggong Xuebao/Acta Armamentarii, 2008, 29 (01): : 33 - 36
[33] Maximum a posteriori noise log-spectral estimation based on first-order vector Taylor series expansion
Ding, Guo-Hong
IEEE SIGNAL PROCESSING LETTERS, 2008, 15 : 158 - 161
[34] New parameter of spectral distortion for predicting subjective quality of speech
Yang, Z.
Bi, H.J.
Dianzi Yu Xinxi Xuebao/Journal of Electronics and Information Technology, 2001, 23 (07):
[35] Assessment of Disordered Voices Using Empirical Mode Decomposition in the Log-Spectral Domain
Kacha, A.
Grenez, F.
Schoentgen, J.
13TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION 2012 (INTERSPEECH 2012), VOLS 1-3, 2012, : 66 - 69
[36] Log-Spectral Magnitude MMSE Estimators under Super-Gaussian Densities
Hendriks, Richard C.
Heusdens, Richard
Jensen, Jesper
INTERSPEECH 2009: 10TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION 2009, VOLS 1-5, 2009, : 1351 - +
[37] MINIMUM GENERATION ERROR TRAINING BY USING ORIGINAL SPECTRUM AS REFERENCE FOR LOG SPECTRAL DISTORTION MEASURE
Wu, Yi-Jian
Tokuda, Keiichi
2009 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING, VOLS 1- 8, PROCEEDINGS, 2009, : 4013 - 4016
[38] Objective speech quality evaluation based on weighted bark spectral distortion
Institute of Communication Engineering, PLA University of Science and Technology, Nanjing 210007, China
Shu Ju Cai Ji Yu Chu Li, 2006, 3 (302-306):
[39] Multiband vocal dysperiodicities analysis using empirical mode decomposition in the log-spectral domain
Kacha, Abdellah
Grenez, Francis
Schoentgen, Jean
BIOMEDICAL SIGNAL PROCESSING AND CONTROL, 2015, 17 : 11 - 20
[40] Microphone Array Post-filter in Frequency Domain for Speech Recognition Using Short-Time Log-Spectral Amplitude Estimator and Spectral Harmonic/Noise Classifier
Salishev, Sergey
Klotchkov, Ilya
Barabanov, Andrey
SPEECH AND COMPUTER, SPECOM 2017, 2017, 10458 : 525 - 534

← 1 2 3 4 5 →