The crack development in quasi-brittle granite and brittle gneiss under mode I loading condition was monitored using digital image correlation (DIC) technique during wedge splitting tests. The cracking behavior was studied on granite specimens that were split in one material direction, perpendicular to the rift plane, and gneiss specimens that were split in three different material directions, parallel and perpendicular to the foliation (along and across a lineation). The granite specimens had a saw cut 5 mm-wide notch and a blunt round notch in form of a 32 mm borehole. The gneiss specimens had a saw cut notch. The results from the DIC measurements revealed a meandering and branching crack path for the granite, whereas a smoother crack path for the gneiss with almost no branching. This behavior was confirmed by microscopy images from thin sections taken from specimens after testing. The thin sections showed that the fractures were prone to propagate across grains in the more coarse-grained granite than in the fine-grained gneiss, where the fractures propagated almost entirely along grain boundaries. The crack initiation occurred mainly in the corner of the saw cut notches and centrally in the round notch. However, the initiation locations in the granite were affected by the medium- to coarse-grained microstructure with grains preventing initiation and propagation which yielded displaced positions out from the ideal ones with respect to the highest stress in some cases. The crack opening displacement was determined along the crack path from the DIC measurements at 12 stages on each specimen of the advancing crack during the splitting progress. The critical crack opening displacement and length of the fracture process zone were assessed and the crack front position yielding the crack length along the tests was determined. The results showed a critical crack length when the deformations in the ligament (also called plastic hinge) affected the cracking process. The average crack velocity in gneiss during the test was more than twice as high as in the granite. This is attributed to a combined effect of the higher brittleness in the gneiss and the effect of a too large elastic energy in the specimen and test setup in relation to the dissipated fracture energy which made the initial crack propagation in the gneiss specimens nearly unstable. The strain energy release rate was calculated along the crack propagation and showed a lower value when the crack lengths were less than 40–60 mm. The calculation of the strain energy release rate was made on crack length measurements from DIC results. The results from the investigation were discussed in relation to the few other similar results found in the literature. The findings give an insight and understanding of the cracking process via both qualitative and quantitative results. Several used methods were novel or not used together in a single study as in this one.