What is the difference between each of them?
this wiki page - http://en.wikipedia.org/wiki/Latin_hypercube_sampling says that: "In Orthogonal Sampling, the sample space is divided into equally probable subspaces", which to me sounds like a property of low discrepancy sequence.