Re-reading this bit, I agree my wording is very clunky. What I meant by "whether we know the bias of the input bit stream or not" was:
[For p≠1/2] whether you know the value of p or not, this is the best approach. Of course, p=1/2 is the trivial case where we can just pass through..
Re-reading this bit, I agree my wording is very clunky. What I meant by "whether we know the bias of the input bit stream or not" was:
[For p≠1/2] whether you know the value of p or not, this is the best approach. Of course, p=1/2 is the trivial case where we can just pass through..