Lauler commited on
Commit
4731821
·
verified ·
1 Parent(s): e9d82cc

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +1 -1
README.md CHANGED
@@ -14,7 +14,7 @@
14
  | district | The district the person represents (if an MOP) |
15
  | role | Role(s) the speaker has had over their career. |
16
  | gender | Gender. |
17
- | start_segment | Start time of the speech within the media file (filename can be found in `audio_file` variable). Obtained by running speaker diarization on top of fuzzy string matching in order to refine existing timestamps from fuzzy string matching (the values in `start_text_time`). This is generally the most accurate estimate of when the speech was given. |
18
  | end_segment | End time of the speech within the media file (according to diarization method). |
19
  | duration_segment | Duration of speech (diarization method). |
20
  | text | Original protocol transcript of speech from SWERIK. |
 
14
  | district | The district the person represents (if an MOP) |
15
  | role | Role(s) the speaker has had over their career. |
16
  | gender | Gender. |
17
+ | start_segment | Start time of the speech within the media file (filename can be found in `audio_file` variable). Obtained by running speaker diarization on top of fuzzy string matching in order to refine existing timestamps from fuzzy string matching (the values in `start_text_time`). This is generally the most accurate estimate of when the speech was given. Time in seconds. |
18
  | end_segment | End time of the speech within the media file (according to diarization method). |
19
  | duration_segment | Duration of speech (diarization method). |
20
  | text | Original protocol transcript of speech from SWERIK. |