The performance of spectral quality measures

Abstract

Two different classes of quality measures are discussed and compared: absolute and relative measures. The relative class to which the prediction error belongs has many different approximations and equivalents, like the spectral distortion and the likelihood ratio. This measure is based on time series theory and can be written as a relative error in the frequency domain. It is useful in many applications. It will be compared to some absolute measures. To that class belong a squared difference measure on the integrated spectrum, that gives equal weights to all frequencies, and also a measure that is based on the squared difference between impulse responses. The absolute class has only a few practical applications, mainly in speech.

Keywords

This publication has 11 references indexed in Scilit:

Facts and fiction in spectral analysis
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2002
Estimation of the accuracy of mean and variance of correlated data
Published by Institute of Electrical and Electronics Engineers (IEEE) ,2002
The quality of models for ARMA processes
IEEE Transactions on Signal Processing, 1998
Model parameter estimation from non-equidistant sampled data sets at low data rates
Measurement Science and Technology, 1998
Estimation of turbulence power spectra from randomly sampled data by curve-fit to the autocorrelation function applied to laser-Doppler anemometry
Measurement Science and Technology, 1998
Reconstruction error distortion measurefor quantisation of LPC models
Electronics Letters, 1996
Equivalent distortion measures forquantisation of LPC model
Electronics Letters, 1995
A general lower bound for parametric spectrum estimation
IEEE Transactions on Acoustics, Speech, and Signal Processing, 1984
Some recent advances in time series modeling
IEEE Transactions on Automatic Control, 1974
Factorization of the Covariance Generating Function of a Pure Moving Average Process
SIAM Journal on Numerical Analysis, 1969