In computational homogenization, the considered volume element must be statistically representative of the overall infinite microstructure ensemble up to a prescribed accuracy. To reduce the simulation effort, these volume elements need to be as small as possible. In particular, microstructure generation algorithms need to accurately reproduce the ensemble’s statistical characteristics for low sample sizes. For polycrystalline microstructures, the relevant microstructure descriptor is the orientation distribution function, which admits a Fourier decomposition with tensorial coefficients as discussed by Guidi et al. [1]. A texture-sampling algorithm based on the first few coefficients was introduced by Kuhn et al. [2]. We improve on this texture-coefficient-based microstructure generation algorithm by using the harmonic basis [3] to formulate efficient higher-order texture tensor constraints. Using the novel algorithm, we investigate the influence of higher-order texture constraints on the mechanical behavior of polycrystalline microstructures. In particular, we discuss whether texture sampling accuracy is relevant only for accurate simulation of materials with pronounced texture, or also for efficient sampling of statistically isotropic (texture-free) polycrystals. REFERENCES [1] Guidi M., Adams B. L., Onat E. T., Tensorial Representation of the Orientation Distribution Function in Cubic Polycrystals. Texture, Stress, and Microstructure, 19(3), 147-167, 1992. [2] Kuhn J., Schneider M., Sonnweber-Ribic P., Böhlke T., Generating polycrystalline microstructures with prescribed tensorial texture coefficients. Computational Mechanics, Vol. 70 (3), pp. 639-659, 2022. [3] Krause M., Böhlke T. Tensorial harmonic bases of arbitrary order with applications in elasticity, elastoviscoplasticity and texture-based modeling. Mathematics and Mechanics of Solids, (2024).