AUTO SYNC TRANSCRIPT TO SUBTITLES SOFTWARE

[filmboubd]

Hello, hope you are doing well.

 I am looking for a software that can auto sync transcript to timed subtitles, like YouTube 

Example transcipt https://www.pbs.org/wgbh/americanexperience/films/mlk/#transcript

Thanks in advance.

 

 

 

[filmboubd]

@kinglouisxx Can you help me out with this?

[kinglouisxx]

12 hours ago, filmboubd said:

@kinglouisxx Can you help me out with this?

Not sure we can help you out, my friend. If such a tool exists I'm affraid we don't have our hands on it.

For our part we sync subs manually, with a subtitle software. Such as Subtitle Edit or Visual Sub Sync.

Of course, most of the time, the raw subs we're using as a base are already timed, so what we do is make them fit the video version we're working on and adjust the sync when needed to match our standards.

We rarely sync a script from scratch, except when we do a transcription.

There are some softwares, such as SubSync (Subtitle Speech Synchronizer), claiming to auto sync subtitles based on the audio track. But I can't tell you how good it works, or if it works at all. Never tried it. Though I'm pretty sure it supports timed subtitles and not just a script as plain text without any timecodes.

But, hey, if you find some software that does it, let us know!

Take care!

 

 

 

[filmboubd]

@kinglouisxx Thank you for your answer! I know YouTube's Auto sync AI can put timings on transcript but you gotta verify your account to upload videos longer than 15 mins.  I have tried that by posting videos privately in short chunks and it works. So, I thought there would be something like this. But still, Thanks for your reply.  Regards.

[Pindiu]

I know this thread is a couple of years old, but I have the same question.

Just wondering if anything has changed - given AI has exploded since then.

I have a transcript of a 45 minute interview. Even if the result is inaccurate, I can easily tidy it up, but hand inserting time codes from scratch would be a lengthy and very boring job.

[felipios]

As of right now, I can recommend OpenAI's whisper. You should have some experience with Python or your Command Line though, otherwise navigating it will be a challenge. On the command line, whisper automatically creates txt, vtt and srt files containing the transcriptions. idk how to go about doing that with python and the documentation also seems thin at first glance.

The quality in English is pretty high (although it smoothes out stutters etc quite a lot), other languages can vary. I recently did some English-language interview transcription with it and there were very few errors for me to correct.

I am generally incredibly skeptical of AI stuff, but whisper works locally on your device and does not upload anything to OpenAI or anybody else, which is why I am comfortable with it. Depending on what material you seek to transcribe beware of any online services, be that YouTube or otherwise, you may very well breach confidentiality or privacy by uploading it onto remote servers.

Edited January 24 by felipios