Designing a temporomandibular joint (TMJ) total prosthesis requires the assessment of joint displacements for open/close movements. Current knowledge presents disc-condyle motions as involving both translation and rotation but there may be substantial variations from human to human. The aim of this study is to discriminate objective kinematic characters amongst thirty-two volunteers. The displacements are determined using 3D video analysis. The ratio between rotation and translation can be defined by introducing a coefficient. This coefficient varies relatively to the opening amplitude and presents the same dispersion rate whatever the variations. Then it allows to discriminate amongst volunteers, regardless of any jaw opening values. Three groups can be isolated relatively to three kinematic models: a translatory preponderant group, a common group and a rotatory preponderant group. All subjects in the first group present concomitant rotatory/translatory displacements up to maximal opening. The other two groups present variations due to different quasi-pure rotation phases at the end of the opening movement. These investigations will make it possible to establish a correlation between the kinematic characters and the disc-condyle trajectories. The disc-condyle glides along the temporal facet and the condyle centre describes the tubercular morphology. The temporal facet geometry, useful for the TMJ prosthesis design, will be studied in a next paper.