Free-Comment (FC) consists in panelists describing the products using their own terms. Despite its benefits, notably the circumvention of limitations inherent to pre-established lists of sensory descriptors, FC remains rarely used because its performances are not well documented and its analyses and range of application remain limited. This thesis aims to overpass these limitations, highlighting the benefits and the potency of FC and thus put it in the spotlight for sensory analysis with consumers.For the pretreatment of FC data, a semi-automatized procedure is proposed. It enables the practitioners to extract an a posteriori list of sensory descriptors with a compromise between minimizing the loss of information and maximizing the quicknes...