This paper proposes to evaluate the method used in Chlébowski and Ballier [1] for the annotation of F₀ variations in nasal grunts. We discuss and test issues raised by this kind of approach exclusively based on visual inspection of the F₀ tracking in Praat [2]. Results tend to show that consistency in the annotation depends on acoustic features intrinsic to the grunts such as F₀ slope and duration that are sensitive to display settings. We nonetheless acknowledge the potential benefits of such a method for automation and implementation in IA and in this respect, we introduce Prosogram [3] as an alternative material-maker.