Open TheAtomicNord opened 9 years ago
Hey @TheAtomicNord,
I've been looking into that recently actually, but the compatibility, dependency, and networking issues mean that this probably won't happen soon, It's definitely something I'd eventually like to have though.
Hey @Uberi, If we pretended that network speed wasn't an issue and that the recognizer service didn't kick us out for too many queries, what exactly is stopping this from happening? I'm asking out of pure curiosity because I'd really like to help make this happen. It would greatly benefit some of my projects.
Time, mostly - it's definitely possible, but it will take time to implement/test/refine. If anyone is interested in doing so now, here are some nice resources:
Otherwise, this is something that I have planned for the next major release.
Gotcha. Thanks for the great explanation! This module is fantastic and I've loved it since the beginning. Thank you so much for all the hard work you've put into it. I can't wait to see where it goes! I'll start counting the days until 4.0! :smiley:
Why not implement for just pocketsphinx? or have a few api keys to send requests with?
Hey Uberi, I love the work you've done with this and just want to say keep up the good work!
I do have one tiny problem though. I'm trying to get this to recognize speech in real time (or as close as possible) and I'm having a bit of trouble. The idea is to have it print each word as I say it instead of waiting for the designated amount of silence. The "offset" parameter does help with this, but like your documentation says, this can lead to wildly inaccurate results.
I was wondering if you had any advice on how to make this work better or if you were working on it. Thanks!