Area-efficient VLSI implementation of digital filters via multiple product intercoding

A novel approach to computing scalar-vector products based on digit coding of the vector elements is proposed. Many signal-processing algorithms involve multiplications that can be organized as a scalar-vector product. A several-fold reduction in the number of additions required to implement a scalar-vector product is obtained. This method is applied to the VLSI design of a FIR (finite impulse response) digital filter, and is shown to be competitive with other methods in use