Here simple candidate keys mean candidate keys having only one attribute.
So for m simple candidate keys, we will have 2m - 1 super keys(as each candidate key will either be included or not, so 2 choices for each simple candidate key and -1 when neither of candidate keys is selected)
Now, we're left with (n - m) (given that m≤nm≤n) attributes and we know that superset of super keys is also a super key. So, we either include them or don't include them. So, (n - m) elements have 2 choices each.
So, total super keys possible = (2m - 1) * 2n - m
Let's verify with an example. A relation R(A, B, C) with candidate keys as A and B.
So here, m = 2
Super keys possible with m candidate keys should be 2m - 1 = 3 which are {A, B, AB}
And now we're left with n - m = 3 - 2 = 1 element i.e. C which may be included or not to form super keys.
So total super keys possible = 3 * 21 = 6 {A, AB, AC, ABC, B, BC}