SDFT 순환 구현 시 진동계수의 유한 비트 표현에 따른 오차영향 해석

Analytic Derivation of the Finite Wordlength Effect of the Twiddle Factors in Recursive Implementation of the Sliding-DFT

  • 김재화 (중앙대학교 전자전기공학부) ;
  • 장태규 (중앙대학교 전자전기공학부)
  • 발행 : 1999.11.01

초록

본 논문에서는 Sliding-DFT(sliding discrete Fourier transform : SDFT)를 순환구조로 구현할 때 복소수 계수를 유한한 비트로 근사 표현하여 생기는 오차영향을 해석적으로 구하는 방법을 제시하고 유도 과정을 기술하였다. 해석한 결과는 오차전력과 신호전력 비(noise-to-signal power ratio : NSR)의 식으로 얻었으며, DFT 대상신호가 평균이 ‘0’인 가우스 백색신호(zero-mean white Gaussian signal)인 것으로 가정하였다. NSR 식은 복소수계수를 표현하는 비트 수와 DFT 구간길이에 대한 식으로 구하였다. 유도 과정은 SDFF 순환 식(recursive equation)으로 유도한 오차방정식(error dynamic equation)과 계수근 사오차의 공간적인 확률분포특성에 근거하였다. 해석적으로 유도한 NSR 결과를 시뮬레이션 실험을 통해 얻은 결과와 비교하여 타당성을 확인하였다.

This paper presents an analytic derivation of the erroneous effect when the sliding-DFT is implemented in a recursive way with the finite-bit approximation of the twiddle factors. The analysis result is obtained in a closed form equation of the noise-to-signal power ratio(NSR) employing the zero-mean white Gaussian signal as the target input of the DFT. The parameters of the wordlength used in representing the twiddle factors and the blocklength of the DFT appear in the NSR explicitly as its function variables. The derivation is based on the error dynamic equation which is derived from the recursive SDFT, and on the analytic exploration of the statistical characteristics of the approximation coefficients treating them as random variables of having spatial distributions. The analytically derived results are verified through the comparison with the data actually measured from the computer simulation experiment.

키워드

참고문헌

  1. IEEE Trans. on Signal Processing v.44 no.2 A CORDIC-based unified systolic architecture for sliding window applications of discrete transforms Dulal C. Kar
  2. IEEE Trans. on Circuits Syst. Video Technol. v.4 Optimal unified architectures for the real-time computation of time-recursive discrete sinusoidal transforms K. J. R. Liu;C .T. Chiu;K. K. Kolagotla;J. F. Ja' Ja'
  3. IEEE Trans. on Acoust. Speech, Signal Processing v.ASSP-35 On the shift property of DCT's and DST's P. Yip;K. R. Rao
  4. IEEE Trans. on Signal Processing v.40 no.6 On the computation of running discrete cosine and sine transforms N. R. Murthy;M. N. S. Swamy
  5. IEEE Trans. on Signal Processing v.42 no.3 Generalized sliding FFT and its application to implementation of block LMS adaptive filters B. F. Boroujeny;S. Gazer
  6. IEEE Trans. on Circuits ans Systems-Ⅱ : Analog and Digital Signal Processing v.39 no.12 A comment on the computational complexity of sliding FFT B. F. Boroujeny
  7. Theory and Application of Digital Signal Processing L .R. Rabiner;B. Gold
  8. Proceedings of IEEE v.60 no.8 Effects of finite register length in digital filtering and the Fast Fourier Transform A. V. Oppenheim;C. J. Weinsten
  9. Digital Filter and Signal Processing L. B. Jackson