Fixed-point optimization utility for C and C++ based digital signal processing programs

Abstract

Two fixed-point optimization utility programs, the range estimator and the fixed-point simulator, are developed for scaling and wordlength determination of digital signal processing algorithms written in C or C++ language. By exploiting the operator overloading characteristics of C++ language, range estimation and fixed-point simulation can be conducted just by modifying the variable declaration of the original floating-point digital signal processing program. Since this utility evaluates the range and the fixed-point performance by simulation, not by analytical methods, it is easily applicable to nearly all type of digital signal processing algorithms including non-linear and time-varying systems. In addition, this utility software can be used for comparing the fixed-point characteristics of different implementation architectures.