Which command option enables tighter segment timing for VTT cue output?

[prarsena]

I'm having a hard time understanding some of the command options from the help file.

I have been experimenting with max length and found a pretty good value at 50.

./main -t 6 -ml 50 -sow true -m models/ggml-small.en.bin -f "${script_dir}/${wav_file}"\ -ovtt true -of "${script_dir}/${transcription_file}"

But I'm looking to have the VTT Cues be more tightly aligned with when they are spoken. Currently, the timing of the caption starts at 00:00, and the second caption begins immediately after, regardless of any pauses in the dialogue:

WEBVTT

00:00:00.000 --> 00:00:03.010
 *Singing* And I learned to get along and so you're

00:00:03.010 --> 00:00:06.220
 back. From outer space. I just walked in to find

00:00:06.220 --> 00:00:09.040
 you here with that sad look upon your face. I

00:00:09.040 --> 00:00:12.110
 should have changed that stupid lock. I should

00:00:12.110 --> 00:00:15.070
 have made you leave your key. If I had known for

Is there a command option to align the value of the cues with the words as they are spoken? This would allow the captions to disappear when dialogue is not being spoken.

[mrfragger]

there is no empty time or pauses in whisper