It is of great importance to compare the relative merits of different cochlear-implant speech-processing strategies. Some groups have compared different strategies within single subjects, but usually the subject has prior experience with one strategy, and no allowance is made for this prior experience. We show in the present study that this is inappropriate. We tested one subject using the Melbourne (Cochlear Corp.) multichannel implant with the device set to process sounds in two different ways. In the first processing scheme, the device functioned normally, extracting information about voicing frequency, amplitude and second-formant frequency. This information activated the 21-channel device, determining pulse rate, pulse amplitude and electrode position (respectively). In the second processing scheme, a single electrode (with the largest dynamic range) was activated. This electrode coded overall amplitude and voicing frequency. The subject was tested on an audiovisual test of a 14-choice consonant recognition in the form /iCi/ over a period of over 4 months. During this time the subject used the 21-channel processor outside of the laboratory. Upon initial connection, there was little difference between the results obtained with the two schemes when tested in sound alone or in sound plus vision. However, after about 4 months, scores obtained with the 21-channel processor in sound plus vision were superior to the scores obtained with the one channel. This advantage came from a superiority in the features of voicing and nasality, but not place. Scores for sound-alone conditions between the two processing schemes remained similar for the 4-month period.(ABSTRACT TRUNCATED AT 250 WORDS)