An audio-intonation calibration method in which an audio signal emitted by a subject is reproduced to the auditory organs of the subject after real time processing, which comprises the following steps: acquisition of a model audio signal to be imitated; spectral analysis of the model audio signal; acquisition of an imitation audio signal emitted by the subject; spectral analysis of the imitation audio signal; comparison of the spectra of the model audio signal and the imitation audio signal; correction of the imitation audio signal as a function of the result of said comparison; and reproduction to the auditory organs of the subject of the corrected audio signal.