This paper considers the problem of blind estimation of multiple FIR channels. When a subspace algorithm is applied to the blind identification problem, incorporating information about the symbol constellation is in general not possible. However, by exploiting special properties of one dimensional symbol constellations (BPSK), it is shown that it is possible to improve or simplify a class of algorithms for blind channel identification. It is also shown that in the case of one dimensional symbol constellations there is a third way, apart from multiple antennas and oversampling, of arriving at a multichannel representation of the communication system.