Here is another alternative for speech to text, after VL.Speech and Azure Speech Recognizer Demo
This is using Coqui.ai which was formerly Mozilla DeepSpeech.
Advantages:
- offline
- many models to choose from
- instructions on how to train your own acoustic and language models
Note: this is really rather wip: just a demopatch getting it all to work, not a library to use.
CoquiSTT.zip (1.3 MB)
NOTE: requires VL.Audio
As for the included .dll and .so files, here is where they are from.