A digital lossless two-pair terminated at both ends with "passive" multipliers is studied. Conditions for low sensitivity of the transfer-function magnitude with respect to the digital multiplier coefficients are derived. It is shown that low sensitivity property can be achieved by forcing certain "matching" conditions, at the terminations. The application of these results to the understanding of some well-known digital filter structures is outlined. In particular, it is shown that the coupled-form biquad can be interpreted as a doubly terminated lossless digital two pair, and that it satisfies the "termination matching conditions" for almost all pole locations. All results derived in the paper are based on independent

-domain arguments.