Name
Abstract | ||
|---|---|---|
The fractional interval is used to determine the interval between the output sample and the previous input sample in interpolation filters. In this paper, the 1 effects of the quantization of this parameter are studied and the wordlength is derived for different kind of interpolation filters having anti-aliasing or anti-imaging properties. Fractional interpolation filters or fractional delay filters are used in various digital signal processing applications to evaluate sample values of a discrete-time signal at arbitrary points between the existing samples (Laa96). This paper concentrates on the interpolation filters having one continuous-valued input parameter µl which is called as the fractional interval. This parameter is used to determine the time interval between the interpolated output sample y(l) and the previous input sample x(nl). In most of the applications it is required that the fractional interval is adjustable during the computation. Historically, VLSI implementation of digital signal processing algorithms has required fixed-point arithmetic for the sake of cost, speed and power consumption, and it has been important to use the fewest number of bits possible to carry each signal in the system. The strong demand for wired/wireless communication services, among others, calls for low-cost high-volume IC solutions (ASIC or FPGA) that can perform real-time operations using as few chips as possible. Hence, specific designs that are to be implemented using ASIC or FPGA need to be optimised in terms of wordlength and architecture, in order to keep cost, size and power consumption to a minimum. In (Lop00), the effects of quantizing the fractional interval have been studied in theory. In this paper, the effects of finite wordlength representation of µ are further examined, with special attention on the different requirements for different kind of interpolation filters in terms of number of bits. |
Year | Venue | Keywords |
|---|---|---|
2002 | EUSIPCO | antialiasing,delay filters,interpolation,quantisation (signal),signal sampling,antiimaging property,fractional delay filter,fractional interval wordlength,interpolation filter,parameter quantization |
Field | DocType | ISSN |
Applied mathematics,Discrete mathematics,Multivariate interpolation,Interpolation,Bicubic interpolation,Stairstep interpolation,Quantization (signal processing),Mathematics,Bilinear interpolation | Conference | 2219-5491 |
Citations | PageRank | References |
1 | 0.36 | 1 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Francisco López | 1 | 1 | 0.36 |
| Jussi Vesma | 2 | 16 | 4.29 |
| markku renfors | 3 | 300 | 57.60 |