scientific article; zbMATH DE number 4052967

From MaRDI portal
Publication:3789031





This paper is concerned with the error analysis of a floating point FFT (fast Fourier transform processor. A statistical model of the error propagation in each pass is developed, which takes into account three different sources of errors: - input error, due to the finite resolution of the analog-digital converter,- roundoff errors, associated with the computation of the transform, - coefficient quantization errors. Each of the errors is investigated in details and compared with the experimental results. It is concluded that for the particular structure of the processor realised, the input error has the maximum effect.











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q3789031)