MP3 encoding is based on a psychoacoustic model that uses frequency masking and temporal masking effects to remove frequency components imperceptible to the human ear, compressing PCM audio to approximately one-tenth of its original size.