Error spectrum shaping quantizers with non-ideal reconstruction filters and saturating quantizers

It is pointed out that previous analyses of ESS (error spectrum shaping) quantizers did not consider either the limitations of realizable reconstruction filters or the effects of quantizer saturation. Consequently, predictions of ESS distortion were often optimistic. The authors extend the original...

Full description

Saved in:
Bibliographic Details
Main Authors: Stacey, N.D., Frost, R.L., Ware, G.A.
Format: Conference Proceeding
Language:English
Subjects:
Bandwidth
Computer errors
Electronic switching systems
Feedback
Finite impulse response filter
Frequency
Image reconstruction
Nonlinear distortion
Passband
Quantization
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:It is pointed out that previous analyses of ESS (error spectrum shaping) quantizers did not consider either the limitations of realizable reconstruction filters or the effects of quantizer saturation. Consequently, predictions of ESS distortion were often optimistic. The authors extend the original analysis to include the use of nonideal reconstruction filters. Simulations show that this extended analysis predicts accurately the distortion of nonsaturating ESS quantizers. An attempt to account analytically for saturation effects led to a difficult nonlinear design problem. A constraint on the energy feedback filter impulse response was effective in suppressing saturation and led to a straightforward linear optimization problem.< >
ISSN:1520-6149
2379-190X
DOI:10.1109/ICASSP.1991.150746