Sentences that start in a line and end in the other line (PROBLEM) - Any solution or post processing tool for that?

[NormalFall0]

Hi,
Sometimes a sentence is started in a line and is continued until the next line, for example:

"
Timestamp1
Hello everyrone, today were are going to show

Timestamp2
you ten great recipes from Australia. Let's start with

Timestamp3
...
"

I was there was a way to have full sentences within one line only.
So the timestamp for the srt, would include ALL the duration one ONE FULL SENTENCE, in one line/ one timestamp.

So it should be like this:

"
Timestamp1 (modified)
Hello everyone, today were are going to show you ten great recipes from Australia

Timestamp2 (modified)
Let's start with ..

Timestamp3
...
"

What solutions are there for this problem? Maybe I am not aware of existing tools/techs that could help somehow?

Thanks

[whicks1]

A: Use the python webvtt library to parse a saved vtt or srt output and loop over the cues. Keep track of periods and split/generate new cues based on the sentences. It’ll get weird with abbreviations like Mr. Dr. Etc so you’ll need a map of those to ignore. And you may want to recalculate a start/end timestamp for cues and figure out an avg # of words per cue you are dealing with. Outputs are going to be close but not perfect. You’ll also might run into oddities like the occasional casing/punctuation loss which would affect your output along with odd hallucinations loops.

B. Probably better option: use the JSON output with word timestamps and build your srt or vtt on your own.