Doubt in portfolio management

In my SFM study mat the formula for standard deviation has been defined as 

std dev (rho) =  sq rt (X-Xbar)^2/n; which is correct if the we were looking at the population standard deviation.

But no where have I seen any distinctions being made with a sample standard deviation which is:

 sq rt (X-Xbar)^2/(n-1)

so my query is whether I should use a sample standard deviation when the representative data is a sample and a population std dev when the representative data is a population?

I feel I will be safe and go ahead with the formula in the book (i.e. the population std dev).

Comments?