A tier is a set of annotations that share the same characteristics, e.g., one tier containing the orthographic transcription, or another tier containing the free translation.

A tier can be ‘independent’ and ‘time-alignable’, in which case it is directly linked to a time interval of the media file (e.g., the ‘orthographic transcription’ tier). Or it can be ‘referring’, in which case it is linked to another tier, its so-called parent tier (e.g., the ‘orthographic transcription’ tier is a parent tier to the ‘free translation’ tier). The referring tier shares its time alignment with its parent tier. Some referring tiers can be assigned to the time axis, but only to an interval that is contained within the interval of their parent annotation.

It is possible to build nested hierarchies, e.g., the ‘orthographic transcription’ tier is the parent tier to a ‘word’ tier, and the ‘word’ tier is the parent tier to a ‘morpheme break’ tier.

Tiers are assigned to tier types, which specify certain constraints. The following constraints exist: None (independent, time-alignable tiers), Time Subdivision (the annotation on the referring tier can be subdivided and linked to the time axis), Symbolic Subdivision (the annotation on the referring tier can be subdivided, but not linked to the time axis), Symbolic Association (one annotation on the referring tier corresponds to exactly one annotation on the parent tier).