Naming celebrities from visual input (i.e., face presentation) was examined in 98 healthy participants. Normative data are provided both for a 12-and a 63-item test. The first one is a brief screening tool, while the latter is to be administered whenever a pathological performance turns out in the former. Age adversely affected both performances. In both tests the most frequent errors were no responses, followed by semantic errors. The study makes explicit reference to current serial models of familiar person processing and naming. Findings allow to complement previously reported testing procedures, with the specific aim of qualifying proper name anomia by identifying the cognitive locus of lesion.