We explore means to advance source camera identification based on sensor\nnoise in a data-driven framework. Our focus is on improving the sensor pattern\nnoise (SPN) extraction from a single image at test time. Where existing works\nsuppress nuisance content with denoising filters that are largely agnostic to\nthe specific SPN signal of interest, we demonstrate that a~deep learning\napproach can yield a more suitable extractor that leads to improved source\nattribution. A series of extensive experiments on various public datasets\nconfirms the feasibility of our approach and its applicability to image\nmanipulation localization and video source attribution. A critical discussion\nof potential pitfalls completes the text.\n