Minimizing Quantization Effects in Digital Filtering

Abstract

Quantization errors and errors due to finite word-length registers in processors result in undesirable responses in many cases. Along with these two errors, in some cases, data samples may not have exact representations in binary form. Hence it is very important to mitigate the effects of these errors. In this paper, we discuss the various errors arising in digital filters and the effect of using integer sequences for windowing techniques in FIR filtering to minimize these errors. We find that using integer sequences as window functions, the error in filter coefficients reduce by at least 30% in case of lower magnitude coefficients .

A Novel Method of Digital-to-Analog Converter Combination for Precise Digital Control in Closed Loop Systems

“…We can infer that dp and dn should be less than 1 (one) LSB for satisfying (5). If any/both of them is/are greater than or equal to 1 (one) LSB, l and u should be chosen based on INLr, as can also be seen from (7).…”

“…Digital control systems with analog input and output use analog-to-digital converters (ADCs), digital processors, and digital-to-analog converters (DACs). Coefficient wordlength can be optimized such that the digital processor is satisfying the performance requirements while the effects of quantization of arithmetic operations and saturation of the words can be mitigated by various digital signal processing techniques [4] [5]. However, the quantization of input due to ADC and output due to DAC becomes a cause of prime concern for precise closed-loop applications.…”