When magnetic properties are analysed in a TEM using the technique of electron magnetic circular dichroism (EMCD), one of the critical parameters is the sample orientation, and this, independently on the chosen acquisition geometry. Since small orientation changes can have a strong impact on the EMCD measurement, it is experimentally non trivial to measure the EMCD signal as a function of sample orientation. The classical EMCD experimental setup requires to tilt the crystal in a 2 beam orientation and to acquire two electron energy loss spectra at two conjugate scattering angles. The effect of a mistilt from the perfect 2-beam orientation on the measured EMCD signals has not been explored yet due to different experimental constraints. In order to maintain the exact sample location and orientation for the acquisition of the EMCD signal, we have developed a methodology to simultaneously map the quantitative EMCD signals and the local orientation of the crystal. We analyse, both experimentally and with simulations, how the measured magnetic signals evolve with a change in the crystal tilt from the exact 2-beam orientation. Based on this analysis, we establish an accurate relationship between the crystal orientations and the EMCD signals. Our results demonstrate that a small crystal tilt away from the 2-beam orientation can significantly alter the strength and the distribution of the EMCD signals. From an optimisation of the crystal orientation, we obtain quantitative EMCD measurements.