std.bitarray

[Saaa]

I understand a bitarray being faster than a boolean array as the first uses 
bitwise (hardware) operators.
Is this a correct understanding of the situation?
Why then is a boolean array not implemented in that way?
I expect there to be a good reason which I just don't know.

I'd like to be able to bit shift and test a large (1024) array of bits as 
fast as possible (runs thousands of times per iteration).
Should I go with the bitarray (I don't see the shift and test functions) or 
create my own bitarray somehow using intrinsics?
I'd rather use existing implementations, but if the speed is a factor off or 
something..