Hypertext Atlas of Speech Sounds

We present a project concerning phonetics teaching and learning. The user will have at his/her disposal a computerized text book with hypertext links with a speech data base, whose samples are directly accessible by a speech signal editor, Snorri. The link between the explanatory text book, the data base and Snorri allows the teacher to illustrate each statement of the course. Among the main functions of Snorri, we can mention editing spectrogram calculation, visualisation of spectral slices, formant tracking, projection of formant frequencies in the vocalic triangle, as well as the display of the variation of the fundamental frequency. The user can record his/her own voice and compare it with a model. Of particular interest for the teacher is the possibility to store results in the form of ascii text files, search for specific sequences of phonemes and specific coarticulatory phenomena such as devoicing, nasalisation... in the data base, and store them in a new speech file.