Title :
Design of efficient M-band coders with linear-phase and perfect-reconstruction properties
Author :
Alkin, Oktay ; Caglar, Hakan
Author_Institution :
Dept. of Electr. Eng., Southern Illinois Univ., Edwardsville, IL, USA
fDate :
7/1/1995 12:00:00 AM
Abstract :
This paper presents a new design technique for obtaining M-band orthogonal coders where M=2i. The structures obtained using the proposed technique have the perfect reconstruction property. Furthermore, all filters that constitute the subband coder are linear-phase FIR-type filters. In contrast with conventional design techniques that attempt to find a unitary alias-component matrix in the frequency domain, we carry out the design in the time domain, based on time-domain orthonormality constraints that the filters must satisfy. The M-band design problem is reduced to the problem of finding a suitable lowpass filter h0(n). Once a suitable lowpass filter is found, the remaining (M-1) filters of the coder are obtained through the use of shuffling operators on the lowpass filter. This approach leads to a set of filters that use the same numerical coefficient values in different shift positions, allowing very efficient numerical implementation of the subband coder. In addition, by imposing further constraints on the lowpass branch impulse response h0(n), we are able to construct continuous bases of M-channel wavelets with good regularity properties. Design examples are presented for four-, eight-, and 16-band coders, along with examples of continuous wavelet bases that they generate
Keywords :
FIR filters; delay circuits; encoding; filtering theory; low-pass filters; signal reconstruction; wavelet transforms; 16-band coders; M-band orthogonal coders; M-channel wavelets; continuous bases; continuous wavelet bases; eight-band coders; four-band coders; linear-phase FIR filters; linear-phase coders; lowpass branch impulse response; lowpass filter; numerical coefficient values; perfect-reconstruction properties; regularity properties; shuffling operators; subband coder; time domain design; time-domain orthonormality constraints; Continuous wavelet transforms; Discrete cosine transforms; Discrete transforms; Discrete wavelet transforms; Filter bank; Finite impulse response filter; IIR filters; Image coding; Image reconstruction; Signal resolution;
Journal_Title :
Signal Processing, IEEE Transactions on