Title :
Multirate filter banks and transform coding gain
Author :
Kok, C.W. ; Nguyen, T.Q.
Author_Institution :
Dept. of Electr. & Comput. Eng., Wisconsin Univ., Madison, WI, USA
fDate :
7/1/1998 12:00:00 AM
Abstract :
An analytical derivation of quantization error in filter banks using additive and signal dependent noise models is presented. The coding gain over PCM is derived for both noise models. Interesting properties of coding gain in optimized filter banks with an AR(1) input signal model, asymptotic performance of multirate systems, and the analytical constraints required to achieve such performance are discussed. Connections to the recent developments of optimal subband systems are presented
Keywords :
autoregressive processes; band-pass filters; filtering theory; high-pass filters; noise; pulse code modulation; quantisation (signal); signal reconstruction; transform coding; AR(1) input signal model; PCM; additive noise models; analytical constraints; asymptotic performance; multirate filter banks; optimal subband systems; optimized filter banks; perfect reconstruction filter banks; quantization error; signal dependent noise models; transform coding gain; Additive noise; Channel bank filters; Filter bank; Image coding; Image reconstruction; Phase change materials; Quantization; Shape; Signal analysis; Transform coding;
Journal_Title :
Signal Processing, IEEE Transactions on
