During processing, LENA software partitions the recording audio stream into variable length segments that are matched statistically to sound source models (here referred to as “Speaker” models) based on the dominant audio information contained in each segment.  

Each segment is labeled using a 3-letter “Speaker ID” code to indicate the Speaker model matched that segment.  These 15 codes are displayed in the .trs export.

The Speaker label descriptions are listed below, as well as how these Speaker labels map to LENA Audio Environment categories and to Speaker ID labels in ADEX segment-level output (ADEX v1.1.3-5 r10725 and earlier).