Mesh saliency, the process of detecting visually important regions in 3D meshes, is a significant component in computer graphics, that can be used in various applications such as denoising and simplification. In this paper, we propose a new 3D mesh saliency measure that can identify sharp geometric features in meshes. A local normal-based descriptor is built for each vertex thanks to a spiral path within a k-hop neighborhood. First, a geometric-based saliency is computed as the mean local alignment between the spiral descriptors. Second, a spectral-based saliency is computed from the structure tensor total variation of each vertex structure tensor with the gradient defined from the spiral descriptor alignments. The final saliency is then defined as the vertex roughness weighted by the sum of geometric ans spectral saliencies. This single-scale saliency can be extended to a multi-scale saliency by decimating the mesh at several scales. The approach presents competitive results with the state-of-the-art.