This paper contributes to a multimodal EMCA approach to sensoriality and to a reflection about how video can support it. First, it discusses how the intersubjectivity and accountability of sensorial practices are locally and endogenously achieved by and for the participants. This accountability is implemented through various multimodal resources, which make sensorial practices accessible for the co-participants. Second, it shows how the visual, verbal and sometimes co-tactile orientations of the participants are also the very basis on which researchers and other professionals build the videographability of the activity. The paper articulates these two aspects by studying activities dealing with food, in which the participants engage in touching food as a relevant sensorial practice within their ongoing course of action.