Such quantization is a nonlinear phenomenon and can be further subdivided into two types of effects, called limit-cycle oscillations and roundoff noise. The second type of quantization is due to signal rounding the internal signals in a digital filter, which take part in the filtering process, are invariably subject to quantization, causing an error in the computed output. Once the quantization has been done, this error is a fixed, well-determined quantity. The result of parameter quantization is that the actual implemented transfer function H q( z) is different from H( z). The first is due to parameter quantization, where the term “parameter” refers to the fixed digital-filter coefficients (or multipliers). Two basic types of quantization effects should be distinguished in any implementation. As a result, the overall input-output behavior is not ideal. When one implements a digital filter transfer function using a digital machine, it invariably involves quantization of signals and coefficients in the system. VAIDYANATHAN, in Handbook of Digital Signal Processing, 1987 I INTRODUCTION Low-Noise and Low-Sensitivity Digital Filters
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |