Single-channel vibro-tactile discrimination and identification of vowels was assessed using a scheme which maintains the spectral shape of the complex speech waveform. A series of vowels and diphthongs were synthesized using the Klatt algorithm. Unlike conventional synthesis, the formant frequencies were set such that the first four formants fell under 660 Hz and the formant bandwidths were narrowed. This synthesis effectively generated 5:1 frequency compression. The synthesized segments sound speech like. Naive subjects were presented both discrimination and identification tasks. The stimuli were felt with a Audiological Engineering V1220 transducer. With no prior exposure discrimination exceeded 7%, and within a few hours of testing discrimination levels of 95% were achieved. Subjects were able to accurately identify many of the vowel segments. Furthermore examination of the confusions reveals patterns similar to those reported for auditory confusions of natural vowels. These findings appear to contradict the premise underlying the development of multichannel tactile aids and cochlear implants that the tactile senses and the impaired ear are incapable of extracting the appropriate information from a single complex speech waveform.