The paper presents a technique for compensating error associated with partial approximation of simulating filter coefficients. This technique enables us to preserve the model quality while reducing the required volume of data to be stored or transmitted and essentially reducing the computational costs of simulation.
展开▼