Title :
High-Accuracy Length-6 Cubic Interpolators
Author_Institution :
Dept. of Inf. Sci., Toho Univ., Chiba, Japan
Abstract :
This letter proposes a frequency-domain weighted least-squares (WLS) method for designing a length-6 cubic interpolation kernel constructed by connecting three piecewise polynomials of third-degree, and the optimal coefficients are found through minimizing the weighted squared error between the desired and actual frequency responses of the length-6 cubic. This frequency-domain approach can design various cubics with different frequency responses by adjusting the weighting functions in different frequency bands, and even ignore “don´t care” bands. An interpolation example is given to illustrate that the length-6 cubic can achieve much higher accuracy interpolations than the other interpolators with fixed frequency characteristics.
Keywords :
frequency-domain analysis; interpolation; least squares approximations; piecewise polynomial techniques; signal processing; frequency-domain weighted least-squares method; length-6 cubic interpolation kernel; piecewise polynomial; signal interpolation; weighted squared error; Accuracy; Frequency domain analysis; Frequency response; Interpolation; Kernel; Polynomials; Speech processing; Catmull–Rom cubic; frequency-domain WLS design; interpolator; length-6 cubic; signal interpolation;
Journal_Title :
Signal Processing Letters, IEEE
DOI :
10.1109/LSP.2010.2072780