Previous investigations found that loudness-contours within individual Mandarin monosyllables can drive categorical perception of Mandarin tone for cochlear implant (CI) users, while in normal hearing (NH) subjects the pitch contour is phonologically acknowledged to be the dominant cue. Here we further examine the weighting strategy of pitch induced and loudness induced contour identification on Mandarin tone perception by CI users. Twenty-seven versions of the disyllabic utterance /Lao3 Shi/ with orthogonally manipulated loudness-contour and pitch-contour of the voiced portion of the second monosyllable /Shi/ served as the stimuli to both CI and NH subjects. In Mandarin, if /Shi/ is pronounced with high-flat-pitched Tone 1 the word means “teacher”, with rising Tone 2 it means “well-behaved”, or with falling Tone 4 it means “always”. CI users generally had poorer word-recognition scores and their inter-subject variance was large. While NH subjects recognized tone reliably based on pitch-contour, half of the CI users relied on pitch-contour, the other half on loudness-contour, implying systematic differences in pitch coding in their CI processing. This paradigm of orthogonal manipulation of pitch and loudness contours could be developed into improved audiometric tests of Mandarin tone perception and pitch coding with CIs.