VideoHelp Forum
+ Reply to Thread
Results 1 to 5 of 5
Thread
  1. I've now used the Alphacephei models in Subtitle Edit a few times. There's no contact info I've seen where I could know when those models are updated and processing voice better. I'm using the vosk routine in SE pretty much just to get the timings of the speech and then doing all the corrections line by line. That's alright; it keeps me from being frustrated by You Tube Creator routines that fail to function. (I had one just stop on me today at minute 6.)
    Who knows what the release notifications might be for the vosk models?
    Quote Quote  
  2. The Alpha cephei models in Subtitle Edit - they can be a real lifesaver, right? As for updates, I'm in the same boat - it's a bit tricky to keep tabs on when those models get a boost. I've been relying on the vosk routine in SE for timing too, and hey, at least it's keeping our frustration levels in check compared to those wonky YouTube Creator routines!
    Quote Quote  
  3. Originally Posted by ogbonnayauche View Post
    The Alpha cephei models in Subtitle Edit - they can be a real lifesaver, right? As for updates, I'm in the same boat - it's a bit tricky to keep tabs on when those models get a boost. I've been relying on the vosk routine in SE for timing too, and hey, at least it's keeping our frustration levels in check compared to those wonky YouTube Creator routines!
    Thanks for the reply. I never got You Tube Creator to work consistently at all. I even made screen shots of each step. Regardless,
    it would look like it was writing the sub and then nothing. But output was pretty good the one lucky stroke I had. I'm hand correcting now with a printed play text since the content is historic.

    No doubt SE would be first in knowing about releases for updates. As I understand it there are multiple programmers or developers for it.

    Whisper AI is an alternative also in SE. Many swear by their install of that through Python programming. But Whisper fails often in SE-- seems like after it has run awhile, if that makes any sense. I'm no Python programmer so I have not been able to do the command line things needed.
    Quote Quote  
  4. I just came across your post and I totally feel you on the whole Alpha cephei models and vosk routine situation! Honestly, I'm just dipping my toes into the vast sea of machine learning myself, so I'm not exactly the go-to guru on updates. But hey, I've actually started exploring the realm of machine learning recently. I found this goldmine of training courses at selectcourses.com that I'm itching to dive into. Planning to grasp the basics, wrap my head around those algorithms, and see how machine learning spices up real-life processes and projects. Who knows, maybe this could even steer me into a whole new career path!
    Last edited by phillipevans; 22nd Aug 2023 at 04:47.
    Quote Quote  
  5. Member
    Join Date
    Mar 2021
    Location
    Israel
    Search Comp PM
    VOSK link
    https://alphacephei.com/vosk/models
    There is a contact section on this page if you want to contact them about anything.
    I was excited at first about Vosk, but then I discovered Whisper AI and never looked back at Vosk.
    But to get the best out of Whisper, you need a really good PC and a GPU of 8GB VRAM.
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!