Perception of frequency contours via temporal and spatial tactile transforms

Two tactile coding schemes of voice fundamental frequency were compared in terms of the detection of terminal frequency changes in simple syllable-like frequency contours. The coding schemes were: (1) temporal, single-channel--in which input frequency is represented as rate of vibration; and, (2) spatial, multichannel--in which input frequency is represented as location of vibration. An adaptive, three-interval, forced choice oddity procedure was used. The temporal, single-channel coding scheme provided a frequency resolution between 0.2 and 0.3 octaves at the fingertip. The spatial, multichannel scheme provided a spatial resolution, on the forearm, of 1-channel, which, for this 16-channel display, translates into a frequency resolution of 0.14 octaves. More learning was required with the temporal, single-channel coding scheme, than with the spatial, multichannel scheme.