This is not a new issue.
The idea of transcription is great, but the implementation has one tiny bug which would be trivial to fix.
The text is properly chunked into sentences. HOWEVER, the audio cuts off before the end of the sentence. This may be a rounding error, not sure.
Any way, this is so easy to fix, just add 500ms to the end of each audio chunk. It is better to overlap a bit then to cut off the end of the sentence!
I can’t believe this hasn’t been fixed yet. Who would want to pay for transcription when the implementation is so shoddy… Please have someone spend one hour on this, it’s not rocket science.
If someone wants to try, import this mp3 file (link expires in 3 days)
Go to sentence view and listen to the first five or six phrases
Even if you don’t know Turkish, looking at the text you can see that for many sentences the audio cuts off too fast
It’s worse than I thought
Look at this video
First you’ll see some completely out of sync text vs audio
Then at the end you’ll see that hundreds of audio have the same text line
Please share this with your team, something really big is broken in “audio transcription”
It’s unusable…
That link says Filen but if you click on it it’s a video, okay?
Okay, thank you Zoran. Looking forward to this being fixed.
It will be absolutely amazing.
In theory it’s a terrific feature, that’s one of the reasons I signed up for premium.
@balat thank you for clarification. I tried my MP3 file and it is fine. Could you please share your file, so I can check it? Also please try to import this file again - will there be the same mistakes? Thank you for your help!
Hi, sorry about the delay.
It is worse than ever. Just tried it again. Opened a new ticket.
If you’d like to try it, I uploaded the file.
The steps to replicate are on this thread (on my screen there’s no contrast, but the words “this thread” are a link you can click on).
