# D-brane tension of type I D9-branes after T-duality

Type I contains branes with a brane tension which is $\frac{1}{\sqrt 2}$ times that of type II branes. The reason is that, e.g., in the computation of the D9-D9 amplitude the open string is unoriented. If one compactifies type I on a torus and performs a T-duality along the $\mathbb T^2$ one ends up with type IIB$/\{\mathbb 1, \Omega \sigma (-1)^F\}$, where $\sigma: z \mapsto -z$ on the torus.

On the one hand, by applying the usual rules of T-duality, I would expect that the resulting 16 D7-branes have a tension which is smaller by a factor of $\frac{1}{\sqrt 2}$ as compared to D7-branes in pure type IIB. On the other hand, the local physics around a (or two) D7-branes is that of type IIB, as the orientifold group only relates the brane and its mirror image at a different position, and thus it should be the tension of a D7-brane of pure type IIB.

Which answer is correct and why?