Paper and paperboard are commonly used for writing, printing, and packaging. In Germany alone, more than 12 million tons of paper and board were used for packaging purposes in 2022. However, paper's outstanding mechanical properties and excellent recyclability make it a viable substitute for conventional engineering materials. Paper exhibits a microstructural fiber network, which has a significant influence on the macrostructural properties, including in-plane and out-of-plane anisotropy and tear strength. Macrostructural issues such as folding problems or package instability are often related to this network, in particular to fiber orientation. While the composition of paper is the subject of much research, the distribution of fiber orientation is still poorly understood. While these distributions are intriguing in their own right, they are of great value as input data for network-level computational models. It is worthwhile to gain a deeper understanding of the microstructure using non-destructive imaging methods. These methods have been proven in the past in numerous applications with different materials. We studied the fiber distribution of a 240 g/m^2 three-layer liquid packaging board. Forty samples were taken from the edges and center of a single roll of paper to observe microstructural differences within the roll. We then performed micro-CT scans that provided data on the total thickness of the paper. After segmentation and identification of individual fibers, their orientations were described by orientation tensors for cell sizes of 500x500x2 microns. The analysis calculated the principal orientations for over 20,000 cells and showed that the fibers are predominantly in-plane oriented. The parameters of several periodic and non-periodic probability density functions were determined using a maximum likelihood method. The goodness of fit of the parameterized probability density functions was evaluated and the best performing candidates were identified. These functions will be used to improve our network models for computational analysis.