“…The quantization error effect on digital filters resulting from the finite precision using floating-point arithmetic has been fairly extensively studied over the last four decades (for example, Sandberg 1967, Liu and Kaneko 1969, Kan and Aggarwal 1971, Kaneko and Liu 1971, Liu 1971, Zeng and Neuvo 1991, Smith et al 1992, Rao 1996, Bomar et al 1997, Tsai 1997, Ko and Bitmead 2004, see Kontro et al (1992) for a review. The effect of coefficient rounding in floating-point arithmetic seems first to have been considered by Kaneko and Liu (1971) (see also Liu 1971), who analysed the sensitivity of the filter poles and the sensitivity of the frequency response to multiplicative perturbations on the coefficients for several filter structures.…”