A-4-9 DESIGN OF AN OVERSAMPLED-RATE FILTER FOR AMPLITUDE AND PHASE DATA WITH A SRFT APPROACH

概要

論文の詳細を見る
In this work we study the design of an oversampled-rate IIR digital filter in order to approximate an amplitude and phase spectrum which is given as an experimental data. In our case, the filter is a Delta-Sigma modulation-based 1-bit filter and the spectral data is that of an headphone noise cancellation system. We construct a prototype lossless two-port to approximate the gain and phase data, with lumped and/or distributed transmission-line elements. Accordingly, we have to describe the synthesized network in Laplace (p) or Richard (λ) depending on the prototype structure we use. We use the Simplified Real Frequency Technique (SRFT) [1] with a convergent fixed-point iteration algorithm to optimize the prototype two-port network. This algorithm gives the Scattering parameters as Belevitch polynomials in p or λ according to our choice, and from the S-parameters we compute the voltage transfer function. In case we use Lumped-element model, we obtain a transfer function in Laplace variable p, so we use Bilinear transformation to get the discrete-time transfer function in z variable. In case we use Distributed-element model, we obtain a transfer function in Richard variable λ, and then we use a Bilinear transformation on λ variable, in a Wave Digital Filter
社団法人電子情報通信学会の論文
2007-03-07