A Fixed-Point Implementation of the Unnormalized Least-Squares Lattice Using Scaled-Integer Arithmetic

David Carl Swanson, Frank W. Symons

Research output: Contribution to journalComment/debate

1 Citation (Scopus)

Abstract

The unnormalized least-squares lattice algorithm is modified for fixed-point hardware by appropriate scaling of the crosscorrelation, covariance, and likelihood parameters. The scaling reveals the dependence of the unnormalized parameter magnitudes on the number of error signal observations used in their respective updates. Comparing fixed-point and floating-point simulations shows no significant loss of precision in the estimated PARCOR coefficients in the fixed-point environment with scaled parameters. Scaling requires seven more multiplies per lattice stage but provides a least-squares lattice algorithm suitable for a wide variety of fast low-power signal processing chips without square-root normalization.

Original languageEnglish (US)
Pages (from-to)1781-1782
Number of pages2
JournalIEEE Transactions on Acoustics, Speech, and Signal Processing
Volume35
Issue number12
DOIs
StatePublished - Jan 1 1987

Fingerprint

Signal processing
Hardware

All Science Journal Classification (ASJC) codes

  • Signal Processing

Cite this

@article{7b4fcce65ea54697a542be9c9678de41,
title = "A Fixed-Point Implementation of the Unnormalized Least-Squares Lattice Using Scaled-Integer Arithmetic",
abstract = "The unnormalized least-squares lattice algorithm is modified for fixed-point hardware by appropriate scaling of the crosscorrelation, covariance, and likelihood parameters. The scaling reveals the dependence of the unnormalized parameter magnitudes on the number of error signal observations used in their respective updates. Comparing fixed-point and floating-point simulations shows no significant loss of precision in the estimated PARCOR coefficients in the fixed-point environment with scaled parameters. Scaling requires seven more multiplies per lattice stage but provides a least-squares lattice algorithm suitable for a wide variety of fast low-power signal processing chips without square-root normalization.",
author = "Swanson, {David Carl} and Symons, {Frank W.}",
year = "1987",
month = "1",
day = "1",
doi = "10.1109/TASSP.1987.1165104",
language = "English (US)",
volume = "35",
pages = "1781--1782",
journal = "IEEE Transactions on Signal Processing",
issn = "1053-587X",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "12",

}

A Fixed-Point Implementation of the Unnormalized Least-Squares Lattice Using Scaled-Integer Arithmetic. / Swanson, David Carl; Symons, Frank W.

In: IEEE Transactions on Acoustics, Speech, and Signal Processing, Vol. 35, No. 12, 01.01.1987, p. 1781-1782.

Research output: Contribution to journalComment/debate

TY - JOUR

T1 - A Fixed-Point Implementation of the Unnormalized Least-Squares Lattice Using Scaled-Integer Arithmetic

AU - Swanson, David Carl

AU - Symons, Frank W.

PY - 1987/1/1

Y1 - 1987/1/1

N2 - The unnormalized least-squares lattice algorithm is modified for fixed-point hardware by appropriate scaling of the crosscorrelation, covariance, and likelihood parameters. The scaling reveals the dependence of the unnormalized parameter magnitudes on the number of error signal observations used in their respective updates. Comparing fixed-point and floating-point simulations shows no significant loss of precision in the estimated PARCOR coefficients in the fixed-point environment with scaled parameters. Scaling requires seven more multiplies per lattice stage but provides a least-squares lattice algorithm suitable for a wide variety of fast low-power signal processing chips without square-root normalization.

AB - The unnormalized least-squares lattice algorithm is modified for fixed-point hardware by appropriate scaling of the crosscorrelation, covariance, and likelihood parameters. The scaling reveals the dependence of the unnormalized parameter magnitudes on the number of error signal observations used in their respective updates. Comparing fixed-point and floating-point simulations shows no significant loss of precision in the estimated PARCOR coefficients in the fixed-point environment with scaled parameters. Scaling requires seven more multiplies per lattice stage but provides a least-squares lattice algorithm suitable for a wide variety of fast low-power signal processing chips without square-root normalization.

UR - http://www.scopus.com/inward/record.url?scp=0023589085&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0023589085&partnerID=8YFLogxK

U2 - 10.1109/TASSP.1987.1165104

DO - 10.1109/TASSP.1987.1165104

M3 - Comment/debate

AN - SCOPUS:0023589085

VL - 35

SP - 1781

EP - 1782

JO - IEEE Transactions on Signal Processing

JF - IEEE Transactions on Signal Processing

SN - 1053-587X

IS - 12

ER -