What does the transcription line in the .trs contain?

LENA .trs output shows the audio stream partitioned into "Speaker" (sound source) segments.  Each segment is labeled using the 3-letter “Speaker ID” code for the Speaker model that was statistically matched to that segment during processing, based on the dominant audio signal.  

The transcription line in the LENA .trs contains the following information, explained in this article:

  1. CHN and CHF segments display the type of vocal output
  2. MAN, FAN, CXN, CHN, and CHF segments contain conversational turn data.  These map to ADEX output data at the Vocalization Activity Block + Segment Details level.

1. CHN and CHF segments: Vocal output classification

CHN and CHF segments: Vocal output classification

VOC = vocalization

A vocalization is a speech-related sound from the key child.

VFX = vegetative/fixed-signal sounds

Vegetative sounds have to do with digestion and respiration, e.g., breathing, coughing, burping, hiccoughing, etc.

Fixed-signal sounds have to do with responses to the environment, e.g., cries and laughter.

CRY = crying

Crying is a separately categorized subset of VFX.

SIL = “interstitial” silences occurring between VOCs, VFXs, and CRYs

The software counts as one vocalization any speech-related sound bounded by >=300 ms of something else.  The minimum duration for key child is 600 ms, so small amounts of silence may be included.

2. CHN, CHF, FAN, MAN, and CXN segments: Conversational turn data

Between pipes || are codes related to conversation blocks.

CHN, CHF, FAN, MAN, and CXN segments: Conversational turn data

Here is a sample string, color coded and broken out to show each variable that is being reported on:

|RC|4|79|13|XIC|TIMI|FI|

|RC                          |4                   |79                   |13                    |XIC            |TIMI         |FI             |

|Block_Info_Segment|Block_Number|Block_Info_RTF|Block_Info_RTC|Block_Type|Turn_Type|Floor_Type|

These variables map to ADEX output extracted from the ITS file at the Segment Details level. Following is a description of each of the variables. For definitions of the code options within each variable, please see the Reference Codes document in the Help menu of the ADEX software.

 

Block_Info_Segment

  • Where does this segment fall within the context of its "conversation block" - beginning, continuation, or end?

Block_Number

  • As of this segment, how many conversation blocks have been counted in this file?
  • Value is an integer. In the example, this segment is part of the 4th block in the file.

Block_Info_RTF

  • As of this segment, how many turns have been counted in this file?
  • RTF = running turns in file

Block_Info_RTC

  • As of this segment, how many turns have been counted in this block?
  • RTC = running turns in conversation

Block_Type

  • What type of block is this segment part of?
  • The block type classification encodes information about who initiated the block and who is involved.
  • The classification codes are explained in ADEX Reference Codes document, accessed via Help menu in the ADEX software itself.

Turn_Type

  • Is this segment part of a turn?  What part?
  • The classification codes are explained in ADEX Reference Codes document, accessed via Help menu in the ADEX software itself.

Floor_Type

  • Is the speaker "holding the floor" or "taking the floor" in this segment?