Elliptic bulge tests are conducted on liver capsule, a fibrous connective membrane, associated with a field measurement method to assess the global geometry of the samples during the tests. The experimental set up is derived from a previous experimental campaign of bulge tests under microscope. Here, a stereoscopic Digital Image Correlation (DIC) system is used to measure global parameters on the test and investigate some assumptions made on the testing conditions which could not been assessed with microscopic measurements. In particular, the assumption of an ellipsoidal shape of the inflated membrane is tested by comparing the actual sample shape measured by stereoscopic DIC with an idealized ellipsoidal shape. Results indicate that a rather constant gap exists between the idealized and actual position. The approximation in the calculation of a macroscopic strain through analytical modeling of the test is estimated here. The study of the liver capsule case shows that important differences can be observed in strain calculation depending on the method and assumptions taken. Therefore, analytical modeling of mechanical tests through ellipsoidal approximation needs to be carefully evaluated in every application. Here the field measurement allows assessing the validity of these modeling assumptions. Moreover, it gives precious details about the boundary conditions of the bulge test and revealed the heterogeneous clamping, highlighted by strain concentrations.