Investigations of the multiplier error in digital filters

The multiplier error obtained in the quantization of the products in digital filters is studied. An analysis is conducted of the error in the case of a single multiplier and the error at the end of recursive blocks of the second degree. Theoretical and experimental investigations are carried out. A description is presented of the relation between the discrete distributions of the input signal and the error signal. It is shown that an error analysis utilizing the discrete model concept can be employed to explain deviations between the results of a theoretical error analysis with analog error models and the data of experimental error investigations.