Traditional blind identification methods deal with the estimation of an unknown system which can be described by a transfer function H(z). In some scenarios, the unknown system is a cascade of two systems with transfer functions H(z) and G(z) and both channels have to be separately estimated. This paper presents a new scrambling aided blind identification method that permits these systems to be unraveled. The identification algorithm is derived. It is shown the identification is unique up to a nonzero scalar and a delay for both channels.