A study of the dimensionality of the Face Authentication problem using Principal Component Analysis (PCA) and a novel dimensionality reduction algorithm that we call Support Vector Features (SVF) is presented. Starting from a Gabor feature space, we show that PCA and SVF identify distinct subspaces with comparable authentication and generalisation performance. Experiments using KNN classifiers and Support Vector Machines (SVMs) indicate that the number of PCs or SVF required for the authentication performance to saturate heavily depends on the choice of the classifier. SVMs appear to be vulnerable to excessive PCA-based compression.