In a sensing task, particularly imaging, significant benefits are obtained by considering the physical and computational aspects in a joint manner. By this we mean that image acquisition is explicitly affected by the presence of subsequent digital computing, while data processing is tailored to this acquisition. We show examples of such joint methods,dealing with all the setup components: illumination, media, cameras and, finally, computer. Specifically, noise in measurement arrays is reduced by a multiplexing approach which considers saturation, photon noise and natural crosstalk. In addition, images obtained under camera motion can self-calibrate radiometric imperfections, such as flare. Finally, with the aid of simple calculations, polarization can dehaze scenes and recover their 3D structure.