RESUMO
Two methods of encoding speech for tactile displays were compared in discrimination experiments using speech segments. One display represented the short-term speech spectrum in time-swept mode and used vibration amplitude to encode spectral amplitude. The other represented the linear predictive coding (LPC)-derived vocal tract shape as a filled bar graph in which the number of active vibrators was used to encode cross sectional area. The displays were applied to the thigh via a matrix of vibrators. The vibrators were driven at 250 Hz during voiced segments, and by random noise during unvoiced segments. Overall results show a slight superiority for the spectral display in vowel discrimination. Detailed results were analyzed in terms of an articulatory description of the speech stimuli, a multidimensional scaling (MDS) analysis of confusions, and an ideal receiver analysis. The results of these analyses suggest that the detailed characteristics of the tactile patterns were only crudely discriminated.