Phonetic studies of nasal sounds pose numerous obstacles to researchers, there being no unequivocal parameters reflecting this type of articulation obtained by means of acoustic analysis. Conversely, special systems dedicated to the investigation of nasality are costly, often invasive, and do not examine naturally produced speech.
The present paper shows how a multi-channel recorder may be used for detecting nasality in pronunciation of the Polish vowel [ɔ̃] in a non-invasive manner. The described system records multi-channel audio and calculates spatial coordinates of sound propagation sources (3D acoustic field distribution), thus allowing the researcher to establish if the release of a segment is oral, oro-nasal or nasal.