Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Factoring Perfect Reconstruction Filter Banks into Causal Lifting Matrices: A Diophantine Approach

Christopher M. Brislawn

TL;DR

This work reframes two-channel FIR PR filter-bank factorization as a causal lifting problem grounded in linear Diophantine equations over polynomial rings. It introduces the Causal Complementation Algorithm (CCA), which replaces the noncausal Laurent-Euclidean division with a Gaussian-elimination–driven, degree-reducing approach, enabling multiple, causally realizable factorizations and diagonal-delay extractions. Through case studies on LeGall–Tabatabai and cubic B-spline banks, the paper demonstrates that CCA can reproduce EEA results while also producing additional, causally valid factorizations via generalized division (SGDA). Theoretical development includes a rigorous treatment of LDEs in polynomial rings, a generalized division framework, and a constructive algorithmic path, with practical implications for causal, memory-efficient filter-bank implementations and systematic enumeration of factorization options. Overall, the CCA advances the design space for causal, realizable lifting factorizations and paves the way for broader applicability in wavelet and multirate systems engineering.

Abstract

The elementary theory of bivariate linear Diophantine equations over polynomial rings is used to construct causal lifting factorizations (elementary matrix decompositions) for causal two-channel FIR perfect reconstruction transfer matrices and wavelet transforms. The Diophantine approach generates causal factorizations satisfying certain polynomial degree-reducing inequalities, enabling a new factorization strategy called the Causal Complementation Algorithm. This provides a causal (i.e., polynomial, hence realizable) alternative to the noncausal lifting scheme developed by Daubechies and Sweldens using the Extended Euclidean Algorithm for Laurent polynomials. The new approach replaces the Euclidean Algorithm with Gaussian elimination employing a slight generalization of polynomial division that ensures existence and uniqueness of quotients whose remainders satisfy user-specified divisibility constraints. The Causal Complementation Algorithm is shown to be more general than the causal version of the Euclidean Algorithm approach by generating additional causal lifting factorizations beyond those obtainable using the polynomial Euclidean Algorithm.

Factoring Perfect Reconstruction Filter Banks into Causal Lifting Matrices: A Diophantine Approach

TL;DR

This work reframes two-channel FIR PR filter-bank factorization as a causal lifting problem grounded in linear Diophantine equations over polynomial rings. It introduces the Causal Complementation Algorithm (CCA), which replaces the noncausal Laurent-Euclidean division with a Gaussian-elimination–driven, degree-reducing approach, enabling multiple, causally realizable factorizations and diagonal-delay extractions. Through case studies on LeGall–Tabatabai and cubic B-spline banks, the paper demonstrates that CCA can reproduce EEA results while also producing additional, causally valid factorizations via generalized division (SGDA). Theoretical development includes a rigorous treatment of LDEs in polynomial rings, a generalized division framework, and a constructive algorithmic path, with practical implications for causal, memory-efficient filter-bank implementations and systematic enumeration of factorization options. Overall, the CCA advances the design space for causal, realizable lifting factorizations and paves the way for broader applicability in wavelet and multirate systems engineering.

Abstract

The elementary theory of bivariate linear Diophantine equations over polynomial rings is used to construct causal lifting factorizations (elementary matrix decompositions) for causal two-channel FIR perfect reconstruction transfer matrices and wavelet transforms. The Diophantine approach generates causal factorizations satisfying certain polynomial degree-reducing inequalities, enabling a new factorization strategy called the Causal Complementation Algorithm. This provides a causal (i.e., polynomial, hence realizable) alternative to the noncausal lifting scheme developed by Daubechies and Sweldens using the Extended Euclidean Algorithm for Laurent polynomials. The new approach replaces the Euclidean Algorithm with Gaussian elimination employing a slight generalization of polynomial division that ensures existence and uniqueness of quotients whose remainders satisfy user-specified divisibility constraints. The Causal Complementation Algorithm is shown to be more general than the causal version of the Euclidean Algorithm approach by generating additional causal lifting factorizations beyond those obtainable using the polynomial Euclidean Algorithm.

Paper Structure

This paper contains 38 sections, 11 theorems, 131 equations, 3 figures, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background and Relation to Other Work
  3. Degree-Reducing Causal Complements and Degree-Lifting Factorizations
  4. Overview of the Paper
  5. Case Study: The LeGall-Tabatabai Filter Bank
  6. Causal Factorization via the Causal Extended Euclidean Algorithm
  7. EEA in Column 0
  8. EEA in Column 1
  9. EEA in Row 0
  10. EEA in Row 1
  11. Causal Factorization via the Causal Complementation Algorithm
  12. CCA With Division in Column 0
  13. CCA With Division in Column 1
  14. CCA With Division in Row 0
  15. CCA With Division in Row 1
  16. ...and 23 more sections

Key Result

Lemma 3.1

Let $\mathcal{R}$ be a unique factorization domain and let $A$ be a finite subset with common divisior $d$: $a_i = \tilde{a}_i d$ for all $a_i\in A$. Then $d=\gcd(A)$ if and only if the set of all $\tilde{a}_i$ is coprime.

Figures (3)

  • Figure 1: A two-channel multirate analysis/synthesis filter bank.
  • Figure 2: The polyphase-with-delay (PWD) filter bank representation.
  • Figure 3: Standard causal lifting form for a FIR PR filter bank. In this example the initial lifting matrix, $\mathbf{U}_0(z)$, corresponding to the left-most lifting filter, $U_0(z)$, is lower-triangular, and the number of lifting steps, $N$, is odd.

Theorems & Definitions (24)

  • definition 1
  • definition 2: cf. Bris:10:GLS-I, eq. (27)
  • definition 3
  • Lemma 3.1
  • Lemma 3.2
  • definition 4
  • Lemma 3.3
  • Theorem 3.4: Homogeneous LDE Theorem
  • proof
  • Corollary 3.5: Abstract Lifting Theorem; cf. Herley93VetterliHerley:92:Wavelets-filter-banksSweldens96
  • ...and 14 more