Wow I never thought about that.

But it is always like this:

let there be any base "b"
That can represent a number by the sum of their positional digits:

number = sum(d_i * b ^ i)
   where i is the position index and d_i is the digit at this position. (note: index starts with 0, from the least digit farthest to the right)

So the (decimal) number 4 in base 4 is then

1×4¹ + 0×4^0 = 10

And (decimal) number 8 in base 8 is

1×8¹ + 0×8^0 = 10

And 10 in base 10:

1×10¹ + 0×10^0 = 10