Get only 1 big transcript

[gerardbsc]

Hello,

I am facing a problem everyday, that maybe is possible to solve with a simple parameter, but I never found it.

When I do a transcript, Premiere is always separating the text into segments automatically. I dont want that as I constantely need to then use the "Merge segments" command a hundred of times to get everything merged (I need to merge everything because its easier to edit the text then).

Is there a hidden parameter in Premiere to generate a transcript, but without merging it into segments?

Thank you very much

[gerardbsc]

EDIT : *Or is there a command to "Merge all segments together"?

[Clare_Major]

Similar over here! I'm currently wishing for a way to merge multiple transcript segments at once—a "batch merge"? I had Premiere separate speakers while transcribing but it went overboard and has separated out individual words from one speaker's sentences into separate segments, in some cases. I've tried selecting more than two segments in the Text panel and clicking the "Merge segments" button, but the result is either no change or the top segment merging with the one above it, with no change to the lower segments.

Premiere Pro 23.4.0 (Build 56)

MacOS Ventura 13.4

[Stan Jones]

@gerardbsc and @Clare_Major,

Correct; there is no batch merge function. That would be a feature request.

To use the merge segment function, you can select segments in various ways. However, it performs most consistently if you just click in the second of two segments in order to merge that one and the one before. For example, you can Ctrl+A to select all segments and merge segments repeatedly. But it only merges two segments at a time, and at some point it stops doing so.

The easiest workaround I see is to click in the last segment in the transcript, and execute the merge, using the merge icon or a keyboard shortcut. After the merge, you will be in the merged segment, which is now the last one. Repeat until all segments are merged.

The keyboard shortcut for the Text Panel is "Merge Segments." The default is Alt+M which does not work for me on windows, so I reassign it to Alt+Shft+B.

This also works even if the segments to merge are different speakers.

Yes, if it is a lengthy transcript, this is a pain.

Stan

[gerardbsc]

I am currently using this workaround (click on the last segments and click on merge 40 times), but like you said it's a pain.

I thought about "batch merge" function because it's a simple way to explain.

But I think for dev teams a simpler way to solve our problem would be to "not split" the transcript by default (or to offer a checkbox parameter that could let us choose to "not split the transcript by default").

To explain a little bit more, I have a different usage of transcript than Clare_Major. On my side I already have the script of my videos on word documents, and auto-transcript is not doing a good job (I don't blame Premiere, our scripts are very hard to understand). Where Premiere is doing a very good job is synchronizing the correction of its auto-transcript with the audio.

As far as I know, there is no function called "import transcript". So my only goto is to do this:
1) auto-transcript
2) Merge all segments
3) Replace all the text present in the "1 big segment" copying the text from Word, and let Premiere do the magic. If you never done that, try it, you will be astonished how well Premiere can associate words from your text with words from the auto-transcript it heard.

Let me know what you think. What could be the best option to ask as a feature request?

[Stan Jones]

@gerardbsc,

You say "no function called "import transcript" and "let Premiere do the magic. If you never done that, try it, you will be astonished how well Premiere can associate words from your text with words from the auto-transcript it heard."

I think you are saying there is no import of [a user-created] transcript, and I agree. And I think you must be referring to the new "Import corrected transcript" function. I discovered this in the testing I did for my post above, but did not mention it, because I think there is more to test. It is in fact rather amazing.

My notes (which assumed that we needed the PR transcription to edit):

Create transcript in PR. Export as csv so you have beginning and ending timecode. Create a fully merged transcript by editing or using an exported csv to combine segments. (See Excel comment below.) Add a space between each segment. Use a single timecode pair with first segment beginning timecode and last segment ending timecode. Make edits (I only used a couple simple word changes). The exported/import .txt form is simply line 1 begin timecode space dash space end timecode, line 2 Speaker name, and line 3 word-wrapped entire (merged) text.

Result: When imported as corrected transcript, word edits are kept, but all timecodes remain as created in the original PR transcript. Speaker name is not changed whether the edited name exists or not.

For a process that does not require merging in PR, just create a text file with the timecodes as line 1 and speaker as line 2. Then paste YOUR entire transcript word-wrapped as line 3.

Excel comments. When you open the csv export in Excel, B3 is the beginning timecode and C(last line) is the ending timecode. The Speaker is A3 - the speaker name can't be edited here, but just be sure the line is filled. The first caption segment text is D3, then D5, etc. So paste the first formula below in E5, and then the second formula in E7.

=D3&" "&D5
=E5&" "&D7

Select E7 and E8 (to include the blank line), copy, then select from E9 to the last entry in column E plus the blank line, and paste. The last entry in column E will be a word-wrapped version of the entire transcript - i.e. merged.

Is that the gist of what you were doing? And it works?

@Kerstin Ebert Any other suggestions for extending this functionality?

Stan

[Kerstin Ebert · Accepted solution]

Hi Gerard,

a solution would be to select all the text in your source transcript (use cmd+a or select everything with your mouse) and then click on the Merge button in the tool bar. This should merge all text segments into one big segment.

Hope that helps!

Best,

Kerstin

[gerardbsc]

Hello, unfortunatlety it does not, like Stan said with cmd+a it only merge 2 segments maximum.

I just tried again and to be more precise there is 2 behaviors :
 - 1 segment was selected before pressing cmd+a  >>  2 segments are merged

 - nothing selected other than the text panel, then press cmd+a  >>  merge button is greyed out

[Kerstin Ebert]

Can you select all the text with your mouse, and when all the text is highlighted blue, you click on the Merge button? Does this still not merge everything that you selected?

In which build are you working?

[Stan Jones]

@Kerstin Ebert,

Thanks for the response. Always appreciated.

Wow! This works in the Beta - I'm in 23.6.0.18.

In Release 23.4.0, you click in the last segment before selecting all and then must repeat the command. I tried the merge button and shortcut, and also the shortcut to merge selected caption segments - that last one one does not work at all.

Stan

[gerardbsc]

Indeed, in beta it works perfectly fine ! I work in 23.4.
These little improvements are really saving us a lot of time, like "upgrade captions to graphics" did a little while ago.

Thank you very much to both of you.

[Stan Jones]

Sorta related: I just posted a Beta bug report related to the imported transcript adding extra line spacing when the import includes the individual timecodes. I list importing merged as a workaround, because the extra spaces are not added. If you merge in PR, the timecodes are changed to just the beginning and end. If you merge outside PR and import corrected, PR keeps the original timecodes.

https://community.adobe.com/t5/premiere-pro-beta-bugs/import-corrected-transcript-adds-spaces/idi-p/...

Great additions to functionality.

Stan