How to bypass the lack of auto transcription in non-support auto transcription languages?

[Takeone.Digital]

I have 6 hours of raw footage from interviews for 2 documentaries films.
I want to edit the footage through the transcription window, an option that is supported by automatic transcription in English- it's much faster in this way. 
Since Adobe didn't bother * to support Hebrew auto transcript, I had to buy a third-party plugin that makes subtitles. I exported the subtitle file in SRT format but to my suprise the Primiere transcript can't imported it (even though it has a time code).

There's also no option to create the transcript manually. In addition, Adobe only supports importing .prtranscript formt files - a format it invented but didn't share with public, thus preventing the possibility of converting subtitles into transcription. 
Does anyone have an idea how to overcome this issue?
-
and as a side note:

* Why there is no ability to transcribe in Hebrew?
We are 10 million people, biased towards high-tech, art  and creativity. It's very strange that countries with equal or smaller populations received support for their local language. I have nothing against Sweden, Denmark, or Norway (great people)  but I'm convinced that Isral have an equal number and probably more Adobe and Premiere subscribers.

[mattchristensen]

Hi @Takeone.Digital we've just put into our Beta version support for a new JSON transcript format that is able to be imported or exported. It will be most useful when third-party transcription providers support it, but, in the mean time you can try it yourself: https://community.adobe.com/t5/premiere-pro-beta-discussions/now-in-premiere-pro-beta-import-your-ow...

If you can get an SRT, or even better a word-level-timecode export of your Hebrew transcript, then you need to convert it to the new JSON format and import it into Premiere Pro Beta and you'll be able to use the Hebrew transcript for Text-Based Editing or caption generation.

As mentioned in the forum post I linked to, take a shot at using an LLM to help with this. We've had pretty good success asking ChatGPT, for example, to take the transcript, the JSON spec (attached to the forum post), and to build and run a python script to translate the transcript into the JSON spec.

[Takeone.Digital]

Thanks for the detailed answer, Matt
I understand that you are excited about the upcoming solution, but let me explain what it means to me- the subscription paying client, who knows that the entire process below happens in other languages within minutes.
These are the steps in the process you suggesting, (as far as I understand- I'm not a Dev person):

1- Check which 3'rd party software converts Hebrew audio into subtitles with the fewest "typo" errors.
2- Pay $300~ for third-party software that will convert the 6 hours of audio into Hebrew subtitles.
3- Install and configure the plug-in and run 6 hours of material to produce subtitles.
4- Export the subtitle files to an SRT files.
5- Pay a software developer 500$-1000$ to convert the SRT files into a JSON files.
6- Convert the SRT files to a JSON files.
7- Import a JSON file (while in beta stage) and pray that all the money and time were not wasted for nothing.

A. Did I understand correctly or am I missing something about the process?
B. Why it's not possible to generate a manual transcription (in any preferred language including RTL) based on the cursor position in the timeline?
C. There are already many AI tools that can transcribe audio in Hebrew. Isn't it time for Adobe to provide proper support in the language of thousands of users in a slightly more respectful manner?

I have no complaints against you, Matt on the contrary- you're the one who bothered to answer me. But you certainly understand the level of frustration of Hebrew editors with Adobe. There are languages ​​that you have provided support for with a lower number of users. 

[Naor31994957vpyh · Accepted solution]

That doesn't help much, Matt. Today I can transcribe the video via services like ivrit.ai for a very small fee, and get a vtt subtitle file (btw, why can't I import vtt?) then I convert it to srt and import it. So it's an easier solution than yours. 

The only thing that would improve our lives is an actual support for Hebrew inside premiere.

[Stan Jones]

@Takeone.Digital,

Matt points to the new and powerful option for getting existing transcripts into the text-based editing workflow. Note that you can install the Beta, import your text, export the PR formatted .prtranscript file, and import that into the Release version of PR - latest 25.4.0, but any of the recent ones will take that PR transcript. So using the Beta to get your transcript does not mean switching to the Beta for your whole project.

My method before this was to create a nonsense transcript (let PR transcribe the spoken Hebrew into English) - it gives timecodes for "words." In your srt, remove all the timecodes and formatting, so you have a single paragraph with NO line breaks and one space between each sentence. One return at the end. Since you've created a transcript in PR, you can Import "as corrected." PR will happily use your words with its timecodes.

I have not tried that with Hebrew, and some languages did not work well. Let me know if you want me to give this a whirl. I'd just need a short audio clip and the Hebrew translation to go with it.

I used Notepad++ for the conversion, but other users have done similar things with ChatGPT or equivalent.

@mattchristensen, 

I downloaded LM Studio to give the new workflow a try with a free LLM, but it is daunting for those of us who are not actively playing with such scripts. Is there a tutorial for this new option?

Stan

[Takeone.Digital]

Thanks Stan,
I can't play with converters and syntax and zilion importing expoting proceeses without knowing it will work out- it's too technical and time consuming.  We need support for Hebrew transcription or at least the ability to produce a transcription manually (for non supported languages).