00796297

a

00796297 Ansari, Rashid Liu, Bede Multirate signal processing. Mitra, Sanjit K. (ed.) et al., Handbook for digital signal processing. New York, NY: John Wiley \& Sons. 981-1084 (1993). 1993 New York, NY: John Wiley \& Sons EN multirate signal processing sampling rate conversion filter design multirate filter banks Zbl 0832.94001 Zbl 0843.93078 [For a review of the entire Handbook see Zbl 0832.94001.] Multirate signal processing (MSP) is one of the most interesting domain of digital signal processing (DSP). This stems from the fact that by the MSP approach one can get results that cannot be thought of in the realm of analog signals. In the simple formulation, MSP means the following fact: Suppose we consider a signal of finite duration (which may change but always finite). We can sample this signal to get its sampled form. The celebrated sampling theorem says that one can recover the initial signal if the sampling rate obeys the condition $T< 1/2f_M$, where $f_M$ is the highest spectral component. It is understood that the recovering is obtained when the samples are provided with the same rate. But, one could provide the samples slower or faster, what is a sampling alteration. In applications, one is faced with two situations: to insert zero samples between the previously existing ones (this is the so-called up-sampling) or to discard some samples (decreasing the rate) and this constitutes the so-called down sampling. To get an idea on the aspects treated here, we give the contents of the chapter: (14-1) Introduction; (14-2) Sampling rate conversion; (14-3) Filter design for sampling rate alteration; (14-4) Multistage implementation of rate conversion; (14-5) Multirate filter banks; (14-6) Applications; (14-7) Summary. The goals of the MSP approach are the following: to decrease as much as possible the computing task by use of the spectral or temporal particularities of the processed signal (voice, music or image). It must be emphasized that the contemporary achievements of the transmitting systems using three main procedures, time division multiplexing (TDM), frequency division multiplexing (FDM), and transmultiplexing (allowing to pass from TDM to FDM) were not possible without the MSP approach. The chapter provides the block schemes of the MSP systems and refers continuously to the content of the preceding chapter (see the review above). We think the chapter is of high interest for providing the bases of MSP, discusses the design of the entering block and gives at the end some interesting applications (transmultiplexer design, spectrum estimation and subband coding of images). D.Stanomir (Bucure\c{s}ti)