With the rapid proliferation of low-cost acquisition and storage devices, anyone can easily create, store, transmit, modify, and manipulate multimedia objects. Throughout their lifecycle, a digital object may undergo several processing stages, including multiple analog-to-digital (A/D) and digital-to-analog (D/A) conversions, encoding and decoding, transmission, or editing that enhances quality, creates new content by mixing existing materials, or manipulates the content. The REWIND project aims to synergistically combine principles of signal processing, machine learning, and information theory to address relevant questions about the past history of these types of objects.