Distance measures for speech processing
A. Gray, John D. Markel · IEEE Transactions on Acoustics Speech and Signal Processing · 1976
The properties and interrelationships among four measures of distance in speech processing are theoretically and experimentally discussed. The root mean square (rms) log spectral distance, cepstral distance, likelihood ratio (minimum residual principle or delta coding (DELCO) algorithm), and a cosh measure (based upon two nonsymmetrical likelihood ratios) are considered. It is shown that the cepstral measure bounds the rms log spectral measure from below, while the cosh measure bounds it from above. A simple nonlinear transformation of the likelihood ratio is shown to be highly correlated with the rms log spectral measure over expected ranges. Relationships between distance measure values and perception are also considered. The likelihood ratio, cepstral measure, and cosh measure are easily evaluated recursively from linear prediction filter coefficients, and each has a meaningful and interrelated frequency domain interpretation. Fortran programs are presented for computing the recursively evaluated distance measures.
Cited by 639 works. doi:10.1109/tassp.1976.1162849
Authors
References
- Linear Prediction of Speech
- Minimum prediction residual principle applied to speech recognition
- Speech Analysis Synthesis and Perception
- Nonlinear filtering of multiplied and convolved signals
- A statistical method for estimation of speech spectral density and formant frequencies
- Multivariate Statistical Methods.
- Effectiveness of linear prediction characteristics of the speech wave for automatic speaker identification and verification
- On autocorrelation equations as applied to speech analysis
- A Difference Limen for Vowel Formant Frequency
- Residual energy of linear prediction applied to vowel and speaker recognition