Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for evaluating the quality of a communication session or of a communication path used for the communication session. One of the methods includes initiating a communication session between a first communications device and a second communications device, wherein initiating the communication session comprises routing session data for the communication session along a first communication path between the first communications device and the second communications device; generating, at the first communications device, a plurality of reference content samples; generating a recording of the communication session as received at a first destination along the first communication path; and evaluating a quality of the communication session or of the first communication path by comparing the plurality of reference content samples with the recorded communication session.