This poster presents a beginning Ph.D. research. The subject focuses on an apparent paradox between different approaches to wood: from the viewpoint of engineering sciences (as from industrial uses), "good" wood is orthotropic, well oriented and homogeneous, singularities are considered as "defects". In standing tree, as the living organism, homogeneity can be considered as a singularity. And from the viewpoint of artisanal usage, either heterogeneity or singularities or both can be considered as "defects", on the contrary, as "qualities" according to the considered uses. The subject, therefore, concerns the permeability between the concepts of "defect" and "quality". It aims to fulfil the gaps in knowledge of heterogeneous woods (figured woods, grain deviations, pieces including gradients, etc.). It also concerns usage-function adequacy and should open the way to a better identification of "precious" wood (e.g., figured) and to a better valorisation of wood types that are usually underestimated. The scientific question is posed, different from most current research work on "clear wood", for seldom addressed in wood mechanics: How to explain the mechanical behaviour of wood with strong gradients in its properties or with grain deviations or both? Knowing that such types of wood are sought after and deliberately selected in several fields of woodcrafts. The “figured wood” is a particular concern here, because of the trade-offs between mechanical and aesthetical criteria for the choices. The main orientation of the cells does not follow the axis of the trunk, but instead the cells’ direction present more or less periodic reorientations, according to different schemes and material axes (Beal and Davis, 1977).