The document discusses light field and coded aperture cameras. It describes the Stanford plenoptic camera which uses a microlens array to sample individual rays of light, capturing 14 pixels per lens. An alternative approach is a mask-based light field camera that uses a narrowband cosine mask to sample a coded combination of rays. This heterodyne approach captures half the brightness but avoids wasting pixels and issues with lens array alignment. The document outlines how such cameras can digitally refocus images and increase depth of field. It also discusses using the Fourier transform to compute a 4D light field from 2D photos captured with a mask.