MUSHRA (MUlti Stimulus with Hidden Reference and Anchor, ITU-R BS.1534) is a subjective listening-test methodology built for comparing several versions of the same audio at once — for example, the same clip through different codecs. Listeners rate all versions on a 0-100 scale in a single screen, with two safeguards hidden among them: an unprocessed reference (which should score near 100) and one or more deliberately degraded anchors (low-pass filtered), which calibrate the scale and let you spot unreliable listeners. Presenting everything together makes MUSHRA more sensitive than single-stimulus tests at the high-quality end, where codecs are near transparent and small differences matter, which is why it is the go-to method for serious codec evaluation.
