Dividing into segments

[ammish-tandon]

Whisper-timestamped divides the audio into segments, and for each word in every segment, we're getting the word-level timestamps. I have observed that there is a pause/gap where the end time of the last word of the previous segment and the start time of the first word of the next segment are not the same. There may be a genuine pause between the 2 words, but there may be cases where there isn't, and still, the timestamps don't match. The audios are < 30 seconds in duration (duration of whisper chunks). Is this the case, that whisper-timestamped adds a pause between segments? If yes, then how to avoid it?