Hedgepig
Active member
- Messages
- 152
- Reactions
- Points
- 43
I'm developing a speech recognition app for VAM, and I've noticed a few VAMers have problems configuring speech recognition. This means they can't use either MacGruber's Speech Recognition nor VeeRifter's Voice Control. This is a shame because these plugins have so much potential and they are tremendous fun.
Please let me know what problems you're having with speech recognition and let's see if we can fix the issue.
Please post the problem here.
Quick tips.
* One basic principal to grasp is that Windows Speech Recognition was never built to provide speech services directly in games, it has to be adapted. It was built for apps like Narrator, in which the keystrokes are for doing things like speech to text and turning on and off Windows apps. Confusingly perhaps, third party apps like Voice Attack use these keystrokes in Games, but this is not how speech recognition in VAM works, so please just absolutely forget that mode of thinking.
** One issue is that the VAM window will loose focus, this means Windows Speech Recognition has shifted its focus from VAM to your desktop and thinks you want it to activate or deactivate something on your desktop. This is not really a problem, because its only doing what it was intended for, and totally fixable. Just click on the VAM window and it should start listening. If it doesn't try saying 'Start listening'. works best for me form activating in Sleep Mode, because in " Listen to Everything" mode, if picks up a sound that it thinks is a word, it loses VAM window focus , and then it will start clicking on and off desktop stuff like calculators and the tool bar. Which can be somewhat frustrating if you don't understand what's happening and why.
**A further issue is the actual set up of Windows Speech Recognition, You do need to read and reread the cards and train, and retrain, Speech recognition. If not you will have linguistic and comprehension insane bot that will confuse semi-homophones like "boots" and "boobs". Once trained it will easily recognize the difference between " so/sow" and "soo/ sue/Sue".
If you want to work with Speech Recognition, you will only get so far with code, and then you will need to start thinking about the sounds of words and variations in semantic meaning.
If you need a' how-to' video, about setting up and training Speech Recognition please just ask.
Please let me know what problems you're having with speech recognition and let's see if we can fix the issue.
Please post the problem here.
Quick tips.
* One basic principal to grasp is that Windows Speech Recognition was never built to provide speech services directly in games, it has to be adapted. It was built for apps like Narrator, in which the keystrokes are for doing things like speech to text and turning on and off Windows apps. Confusingly perhaps, third party apps like Voice Attack use these keystrokes in Games, but this is not how speech recognition in VAM works, so please just absolutely forget that mode of thinking.
** One issue is that the VAM window will loose focus, this means Windows Speech Recognition has shifted its focus from VAM to your desktop and thinks you want it to activate or deactivate something on your desktop. This is not really a problem, because its only doing what it was intended for, and totally fixable. Just click on the VAM window and it should start listening. If it doesn't try saying 'Start listening'. works best for me form activating in Sleep Mode, because in " Listen to Everything" mode, if picks up a sound that it thinks is a word, it loses VAM window focus , and then it will start clicking on and off desktop stuff like calculators and the tool bar. Which can be somewhat frustrating if you don't understand what's happening and why.
**A further issue is the actual set up of Windows Speech Recognition, You do need to read and reread the cards and train, and retrain, Speech recognition. If not you will have linguistic and comprehension insane bot that will confuse semi-homophones like "boots" and "boobs". Once trained it will easily recognize the difference between " so/sow" and "soo/ sue/Sue".
If you want to work with Speech Recognition, you will only get so far with code, and then you will need to start thinking about the sounds of words and variations in semantic meaning.
If you need a' how-to' video, about setting up and training Speech Recognition please just ask.
Last edited:
