Ultraspeech-aquisition allows the synchronous acquisition of high-speed ultrasound and video images of the vocal tract together with the audio speech signal.
Ultraspeech-aquisition was originally designed for the portable ultrasound machine Terason T3000 (laptop/OEM version) but is now also compatible with Telemed systems. Ultraspeech is compatible with the WDM-compliant industrial cameras provided by Imaging Source and ASIO / DirectX compatible soundcards.
Ultraspeech-aquisition is “user friendly”. The entire recording procedure is fully automatic (push-button like) and no a posteriori human check is needed. After each acquisition, data are directly available as series of bitmaps for both image streams and WAV files for the audio stream, in the specified directory. Ultraspeech provides also convenient tools for large database recording, such as an automatic file naming system, the automatic display of the text stimuli and finally, an interactive inter-session re-calibration mechanism that allows recording of large audiovisual speech databases in multiple acquisition sessions.
Related publications :
Hueber, T., Chollet, G., Denby, B., and Stone, M. (2008). “Acquisition of ultrasound, video and acoustic speech data for a silent-speech interface application,” Proceedings of International Seminar on Speech Production (Strasbourg, France),
Hueber, T., Bailly, G. (2016), Statistical Conversion of Silent Articulation into Audible Speech using Full-Covariance HMM, Computer Speech and Language, vol. 36, pp. 274-293
Fabre, D., Hueber, T., Girin, L., Alameda-Pineda, X., Badin, P., (2017) “Automatic animation of an articulatory tongue model from ultrasound images of the vocal tract“, Speech Communication, vol. 93, pp. 63-75