Unable to beautify timecodes in subtitle edit|Need actual good sync app/OCR

[Yosho]

The "Use FFprobe to extract exact frame time codes" does not work on any video with subtitles.
Is there something blocking it and subtitle edit is not telling me?
I used it a couple times on super old avi files and it worked, but not new media, at all.
What could I be doing wrong?

I have tried so so many programs and git projects and nothing actually syncs the subs to the spoken words..
I have to do it via waveform every single time.
It takes me a very very long time to get through timing each line via waveform.
I'm looking for something to help cut down the time for me. To sync lines for me.
Subsync has never worked for me. I tried github stuff based on subsync, but I can't get them to work for some reason.

I am also looking for an OCR like subrip was, but focus only on English lines and that I can build a dictionary for letter by letter just like vobsub was with .sub and .idx was for dvds, allow it to be extended at a line that's got two subtitle lines, or 3, even 4 lines per subtitle.

[Subtitles]

The "Use FFprobe to extract exact frame time codes" does not work on any video with subtitles.
Is there something blocking it and subtitle edit is not telling me?
I used it a couple times on super old avi files and it worked, but not new media, at all.
What could I be doing wrong?

I have tried so so many programs and git projects and nothing actually syncs the subs to the spoken words..
I have to do it via waveform every single time.
It takes me a very very long time to get through timing each line via waveform.
I'm looking for something to help cut down the time for me. To sync lines for me.
Subsync has never worked for me. I tried github stuff based on subsync, but I can't get them to work for some reason.

I am also looking for an OCR like subrip was, but focus only on English lines and that I can build a dictionary for letter by letter just like vobsub was with .sub and .idx was for dvds, allow it to be extended at a line that's got two subtitle lines, or 3, even 4 lines per subtitle.

I use Whisper Audio to Text software and it works very well.
You can upload an audio or video file that you are having trouble with and then PM with the download link and I will test it with my system.

You can also try uploading the video files to your YouTube channel (if they are not copyrighted stuff) and then wait a while depending on how long is the file in minutes and then you will get the timestamps. Works only on few languages but English is included.
Of course you keep the files private.

[Yosho]

The "Use FFprobe to extract exact frame time codes" does not work on any video with subtitles.
Is there something blocking it and subtitle edit is not telling me?
I used it a couple times on super old avi files and it worked, but not new media, at all.
What could I be doing wrong?

I have tried so so many programs and git projects and nothing actually syncs the subs to the spoken words..
I have to do it via waveform every single time.
It takes me a very very long time to get through timing each line via waveform.
I'm looking for something to help cut down the time for me. To sync lines for me.
Subsync has never worked for me. I tried github stuff based on subsync, but I can't get them to work for some reason.

I am also looking for an OCR like subrip was, but focus only on English lines and that I can build a dictionary for letter by letter just like vobsub was with .sub and .idx was for dvds, allow it to be extended at a line that's got two subtitle lines, or 3, even 4 lines per subtitle.

I use Whisper Audio to Text software and it works very well.
You can upload an audio or video file that you are having trouble with and then PM with the download link and I will test it with my system.

No sorry, Whisper is not reliable and can come out as garbage text.
It'd be great if it could capture everything, but often times it comes out as mostly garbage text for me.

[Subtitles]

The "Use FFprobe to extract exact frame time codes" does not work on any video with subtitles.
Is there something blocking it and subtitle edit is not telling me?
I used it a couple times on super old avi files and it worked, but not new media, at all.
What could I be doing wrong?

I have tried so so many programs and git projects and nothing actually syncs the subs to the spoken words..
I have to do it via waveform every single time.
It takes me a very very long time to get through timing each line via waveform.
I'm looking for something to help cut down the time for me. To sync lines for me.
Subsync has never worked for me. I tried github stuff based on subsync, but I can't get them to work for some reason.

I am also looking for an OCR like subrip was, but focus only on English lines and that I can build a dictionary for letter by letter just like vobsub was with .sub and .idx was for dvds, allow it to be extended at a line that's got two subtitle lines, or 3, even 4 lines per subtitle.

I use Whisper Audio to Text software and it works very well.
You can upload an audio or video file that you are having trouble with and then PM with the download link and I will test it with my system.

No sorry, Whisper is not reliable and can come out as garbage text.
It'd be great if it could capture everything, but often times it comes out as mostly garbage text for me.

I don't know which version of Whisper have you used. I am using the Whisper CMD with GPU of 8GB VRAM and 32GB memory and I get excellent results every time in any language.

[Yosho]

I've used them before and the timing is often off by 1000ms or more. Sometimes 2-3 seconds.
If you point me to an executable I'd appreciate it.
I'll check it out, but every whisperer I've ever used comes out wrong like Tim for Jim, or Kim for Tim, or Jim kinda thing as well as being oos.
I have subs, but I want to sync them, not transcribe new subs from audio, and not have to keep using waveform with subtitle edit editing 2000 lines in my spare time.
If anyone has a better suggestion than a whisperer, or the lie that is subsync, that is also a free unlimited use software I'd sincerely appreciate it.

[Subtitles]

You can try this
https://github.com/openai/whisper
You need a good computer with a GPU in order to get accurate transcription. With a GPU of 8GB VRAM you can use the model: medium or model: medium.en (for English).
The result will also depend on the audio quality of your videos and the spoken language.
Transcription is inevitable in order to get the correct timestamps.
Other options available today for speech to text is in this link:
https://listenmonster.com/blog/whisper-alternatives/#:~:text=1.-,Deepgram,more%20affor...0Whisper%20API.