Well, can you help me? That is, what does: "Assume that o 2 (over)1 not equal to o 2 over 2." imply?
Next, does u1=u2 imply that the two population means are equal?
In view of the above, I'm going to assume that the question is associated with a two-independent samples t-test.
t = (XBar1 - XBar2) / Sqrt [ ( Sp^2 * ( 1/N1 + 1/N2 )]
Sp^2 is the pooled variance and equal to
Sp^2 = ( (N1 - 1) * Var1 + (N2 - 1) * Var2 ) / (N1 + N2 - 2)
and where Var1 and Var2 are the variances associated with groups 1 and 2 where are equal to std1^2 and std2^2 (the standard deviations of the two groups)
I'll stop at this point.