Transform from floating point to fixed point for digital filters design
Tong Qing-qiao · Journal of Tianjin University of Technology and Education · 2010
Digital filter algorithms can be transformed to run on fixed point hardware to reduce implementation complexity,which needs a transform form floating algorithm to fixed point algorithm.During fixed point transform,proper ranges at each variable are estimated by statistical approaches,and word length optimization of fixed pointed data is found by the search method based on gradient information.Considering the low pass filter example with Matlab,the transform from floating point to fixed point for digital filters design is discussed,and a fixed pointed filter can be acquired.