F. Itakura and S. Saito. Analysis synthesis telephone based upon the maximum likelihood method. In Conf. Rec. 6th Int. Congress Acoust., Tokyo, 1968. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Variable-Precision Arithmetic for Vector Quantization - Dionysian (1994) (Correct)
....problems can be reduced to detecting sign of an inner product. The most widely use application is in Mean Square Error (MSE) classification [Wha71, Gra84] MSE can also be generalized to weighted MSE criteria [DH73] Moreover, there is a less apparent application: Itakura Saito measure [IS68] which approximates a waveform with the output of a recursive linear filter. All of these three applications have been implemented with inner product similarity measure: sign(x t y b) sign( K X i=1 x i y i b) The inner product of the point s coordinates with the plane s normal ....
....precision increases from n to (n m) VPC, however, will be shown to deal with the increased codevector precision without modification, as it can evaluate the codevector at any precision. Linear Prediction s Similarity Measure For completeness we will also discuss Itakura Saito measure [MRG85, IS68] It is used, when compressing speech with Linear Predictive Coding (LPC) In LPC, an all pole linear predictor estimates a segment of speech. The predictor is given as input a sinusoid whose pitch is extracted from the speech segment. The predictor s coefficients are then chosen to minimize the ....
F. Itakura and S. Saito. Analysis synthesis telephone based upon the maximum likelihood method. In Conf. Rec. 6th Int. Congress Acoust., Tokyo, 1968.
Adaptive Image Coding Using Spectral Classification - Hamid Jafarkhani (1997) (3 citations) (Correct)
....coefficients, we need to use a distortion measure that ultimately minimizes the mean squared prediction error in the original signal domain not in the LP coefficient domain. The Itakura Saito distortion measure, first developed for quantization of speech spectral parameters, is such a measure [5] [6]. For each block, the VQ designed based on the Itakura Saito distortion measure selects the best AR model from among a finite number of AR models in the codebook (all having a fixed and predefined order) The codebook contains a set of codevectors each described by a vector of LP coefficients; ....
F. Itakura and S. Saito, "Analysis synthesis telephone based upon the maximum likelihood method," in Conf. Rec., 6th Int. Congr. Acoust., Y. Yonasi, Ed., Tokyo, Japan, 1968.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC