It's the same as a normal t-test. You just have to keep in mind that you're testing against the null hypothesis H0: μD = 0, where μD is the population mean of differences between pairs of observations. So it's
t0 = d(bar) / (sd / √n)
where d(bar) and sd are the sample mean and sample standard deviation for the differences between each pair.
edit: And just to be clear, n is the number of pairs of observations.
Basically, you treat this like a regular t-test problem, except your random variable is the difference between each pair of observations.