In today's context, where 3D content is more abundant than ever and its acceptance by the public is probably
de_nitive, there are many discussions on controlling and improving the 3D quality. But what does this notion
represent precisely? How can it be formalized and standardized? How can it be correctly evaluated? A great
number of studies have investigated these matters and many interesting approaches have been proposed. Despite
this, no universal 3D quality model has been accepted so far that would allow a uniform across studies assessment
of the overall quality of 3D content, as it is perceived by the human observers.
In this paper, we are making a step forward in the development of a 3D quality model, by presenting the
results of an exploratory study in which we started from the premise that the overall 3D perceived quality is a
multidimensional concept that can be explained by the physical characteristics of the 3D content. We investigated
the spontaneous impressions of the participants while watching varied 3D content, we analyzed the key notions
that appeared in their discourse and identi_ed correlations between their judgments and the characteristics of
our database. The test proved to be rich in results. Among its conclusions, we consider of highest importance
the fact that we could thus determine three di_erent perceptual attributes ( image quality, comfort and realism
( that could constitute a _rst simplistic model for assessing the perceived 3D quality.