Log in

View Full Version : Software for auto subtitle generator from audio


MalickT
9th May 2022, 20:13
So..I have a Movie downloaded from streaming service but it has no subtitles and I could not find any subtitles for it online.
So I was thinking is there a software AI that would take the audio from the movie as input and transcript audio into text? .srt file for example. If it would be at least 80% accurate it would be great.

Is there any such software?

FranceBB
10th May 2022, 20:27
AWS Transcribe is what you're looking for. It can detect if multiple languages are spoken in the same movie and it can generate .srt subtitles from an audio track.
It's not a software, though, but rather a service in which you can upload the audio file and it will give you the transcript and subtitles back. It supports files as long as 4h and you can upload files directly from your browser.
It's not free, though, like everything offered by Amazon in the cloud you gotta pay for it.

VoodooFX
10th May 2022, 23:12
audio from the movie as input and transcript audio into text? .srt file for example. If it would be at least 80% accurate it would be great.

Is there any such software?
There are some (https://github.com/Picovoice/speech-to-text-benchmark) Speech to Text engines, I doubt that they would be much of use in Soundtrack to Text scenario.

StainlessS
11th May 2022, 05:02
I doubt that they would be much of use in Soundtrack to Text scenario.
yep, too much background noise.
at least 80% accurate it would be great.
Dream on. [in your average movie, 10% might be impressive - but maybe one day]

"Calling Macflono McFlooniloon":- https://www.youtube.com/watch?v=Y0hl1-06gOo

(OK, involves speech to text and then back to speech again [but hinted via your contacts])

EDIT: you could try this with a little snippet [Google Speech to Text, trial only, the real thing costs]
https://cloud.google.com/speech-to-text
Google have done wonders in spoken language translation [ie Android Google-Translate],
perhaps also good/best at speech to text in non ideal environment, ie not single person speaking directly into mic.
Is in the Cloud, so has the full power of Google Goblins trying to figure out whats being said.
(Pick a bit of audio that is not too silent in the background, or a bit varying)

VoodooFX
1st February 2025, 17:30
Just use Standalone Faster-Whisper-XXL (https://forum.doom9.org/showthread.php?t=184859)

PS:
PabloPablo and shailesen are spam bots.[post deleted by moderator]

therube
18th February 2025, 18:26
(I know nothing of such things, but it sounded related?)

LLPlayer (https://www.videohelp.com/software/LLPlayer) is a media player specialized for language learning. It has a very rich set of subtitle-related features, including dual subtitles, AI-generated subtitles, OCR subtitles, subtitles sidebar, real-time translation, word searches, and many other unique features. It also supports many video formats and real-time playback of online videos.