Jont B. Allen, L. R. Rabiner

Research output: Contribution to journalConference articlepeer-review


A short-time Fourier transform analysis technique is proposed in which the influences of the window on a spectral estimate can essentially be removed entirely (an unbiased estimator) by linearly combining biased estimates. As a result, section (FFT) lengths for analysis may be made small, thereby increasing the speed of the algorithm without sacrificing accuracy. The method assumes that the system to be identified is a finite impulse response (FIR) system. The proposed algorithm has the property that as the number of samples used in the estimate increases, the solution rapidly approaches the least squares analysis (LSA) (theoretically optimum) solution. The method also uses a fixed discrete Fourier transform (DFT) length independent of the amount of data being analyzed allowing the estimate to be recursively updated as more data is made available.

Original languageEnglish (US)
Pages (from-to)166-172
Number of pages7
JournalNational Conference Publication - Institution of Engineers, Australia
StatePublished - 1979
Externally publishedYes
EventProc, Jt Autom Control Conf - Denver, CO, USA
Duration: Jun 17 1979Jun 21 1979

ASJC Scopus subject areas

  • General Engineering


Dive into the research topics of 'UNBIASED SPECTRAL ESTIMATION AND SYSTEM IDENTIFICATION BY DFT.'. Together they form a unique fingerprint.

Cite this