WhisperLiveKit/whisperlivekit/result_diarization.md
2025-11-19 21:10:28 +01:00

60 lines
No EOL
1.5 KiB
Markdown

########### WHAT IS PRODUCED: ###########
SPEAKER 1 0:00:04 - 0:00:06
Transcription technology has improved so much in the past
SPEAKER 1 0:00:07 - 0:00:12
years. Have you noticed how accurate real-time speech detects is now?
SPEAKER 2 0:00:12 - 0:00:12
Absolutely
SPEAKER 1 0:00:13 - 0:00:13
.
SPEAKER 2 0:00:14 - 0:00:14
I
SPEAKER 1 0:00:14 - 0:00:17
use it all the time for taking notes during meetings.
SPEAKER 2 0:00:17 - 0:00:17
It
SPEAKER 1 0:00:17 - 0:00:22
's amazing how it can recognize different speakers, and even add punctuation.
SPEAKER 2 0:00:22 - 0:00:22
Yeah
SPEAKER 1 0:00:23 - 0:00:26
, but sometimes noise can still cause mistakes.
SPEAKER 3 0:00:26 - 0:00:27
Does
SPEAKER 1 0:00:27 - 0:00:28
this system handle that
SPEAKER 1 0:00:29 - 0:00:29
?
SPEAKER 3 0:00:29 - 0:00:29
It
SPEAKER 1 0:00:29 - 0:00:33
does a pretty good job filtering noise, especially with models that use voice activity
########### WHAT SHOULD BE PRODUCED: ###########
SPEAKER 1 0:00:04 - 0:00:12
Transcription technology has improved so much in the past years. Have you noticed how accurate real-time speech detects is now?
SPEAKER 2 0:00:12 - 0:00:22
Absolutely. I use it all the time for taking notes during meetings. It's amazing how it can recognize different speakers, and even add punctuation.
SPEAKER 3 0:00:22 - 0:00:28
Yeah, but sometimes noise can still cause mistakes. Does this system handle that well?
SPEAKER 1 0:00:29 - 0:00:29
It does a pretty good job filtering noise, especially with models that use voice activity