Copy link to clipboard
Copied
The transcript formats in PP are very limited. Is there no fast workaround to taking a corrected transcript from Premiere and getting a text doc that looks like a dialogue with speaking turns? Instead, we are only given repetitive names with timecode and sentence parts broken up for every edit like below.
Person A
00:00:00:00-00:00:00:00
Words in a
Person A
00:00:00:00-00:00:00:00
sentence are
Person A
00:00:00:00-00:00:00:00
not broken up when
Person A
00:00:00:00-00:00:00:00
read silently or out loud.
Does someone have a way to take the above and quickly change a long interview like that and turn it into like below:
Person A
Words in a sentence are not broken up when read silently or out loud.
Copy link to clipboard
Copied
You could run it through an LLM, like ChatGPT, to help with formatting.
For example, you could use a command like: "Remove the timecodes from this transcript and group by speaker name."
Copy link to clipboard
Copied
@PaulMurphy , I have tried ChatGPT in the past. Accuracy is a big problem when handling large amounts of text, especially with long interviews. It repeated sections or refused to do the rest of the interview. And buying into ChatGPT for testing is not a cheap experiment. A reliable way to do this without worrying if the AI is advanced enough would be appreciated.
Get ready! An upgraded Adobe Community experience is coming in January.
Learn more