Recording — one audio file. Identified by its actual audio content,
not its filename or metadata — so re-classifying a recording on the
detector later, which renames the file and rewrites its metadata,
doesn't create a duplicate.
Identification — one source's claim about a recording (a species, or
a No ID/Noise verdict). Multiple sources can each have their own claim
on the same recording at once; they never overwrite each other. See
The classifier model for how
Fledermap picks which one wins.
Verdict — what an identification actually asserts: a species, a
No ID (a bat call, but not identifiable further), or Noise (no
bat call at all). Both No ID and Noise are answers in their own right,
not missing data.
Source — where an identification came from: a specific field in the
detector's own metadata (e.g. its GUANO or wamd auto-ID), or a manual
correction made in Fledermap.
Taxon — a species, genus, or broader group (e.g. a frequency-class
group like HiF/LoF) that an identification can point at.
Session — a group of recordings from one detector with no large time
gap between them — roughly, one outing.
Site — a place, derived by clustering the locations of
species-identified recordings — roughly, a spot worth remembering, not
every single GPS point.