there is a vid on YouTube link
https://www.youtube.com/watch?v=UYUyiloS-yY
does anyone know how to programme voice attack to do this?
https://www.youtube.com/watch?v=UYUyiloS-yY
does anyone know how to programme voice attack to do this?
does anyone know how to do speech to text using voice attack?
- Thread starter Jake Savage
- Start date
Accuracy renders it pretty much useless, sadly,
you use the {DICTATION} tag and quick input or windows clipboard. Doing a fair amount of speech recog trainning first increases the accuracy, but full dictation is hard for the algorithms because the computer doesn't hear "words", but analyzes the sound waves for matching words.
Basically it has to guess what the next word could be and then see if the sound file matches the pre-recorded state of said word. That's why sometimes it hears "pen" instead of "ben" or in extreme cases "tower" or something like that.
to answer your question though:
make a command (let's call it "I'm talking")
give it the "start dictation" command (Other > Dictation > Start Dictation) **Should also have it clear the buffer or you will get multiple dictation sessions in one**
then make another command (let's call it "done talking")
give it the command "Stop Dictation" (Other > Dictation > Stop Dictation) **Note: do not clear buffer
now you could also give the "Stop Dictation" command the quick input command after the stop dictation command to have it put into (as an example) comms panel. Or you could make a third command to do it. (Personally I use a third command).
(Other > Voice Attack Action > Quick Input) and use the {DICTATION} tag as it's text input.
**Note: when you use dictation the words "stop dictation" will not go into the buffer IF you give a pause between the last word you want in the buffer and the command "Stop Dictation"
Hope this helps
EDIT: it can also help if you going into speech recognition's dictionary and add any words (with recordings) that you have trouble getting the system to recognize.
Basically it has to guess what the next word could be and then see if the sound file matches the pre-recorded state of said word. That's why sometimes it hears "pen" instead of "ben" or in extreme cases "tower" or something like that.
to answer your question though:
make a command (let's call it "I'm talking")
give it the "start dictation" command (Other > Dictation > Start Dictation) **Should also have it clear the buffer or you will get multiple dictation sessions in one**
then make another command (let's call it "done talking")
give it the command "Stop Dictation" (Other > Dictation > Stop Dictation) **Note: do not clear buffer
now you could also give the "Stop Dictation" command the quick input command after the stop dictation command to have it put into (as an example) comms panel. Or you could make a third command to do it. (Personally I use a third command).
(Other > Voice Attack Action > Quick Input) and use the {DICTATION} tag as it's text input.
**Note: when you use dictation the words "stop dictation" will not go into the buffer IF you give a pause between the last word you want in the buffer and the command "Stop Dictation"
Hope this helps
EDIT: it can also help if you going into speech recognition's dictionary and add any words (with recordings) that you have trouble getting the system to recognize.
Last edited:
I found the VA profile in the first post attatchment very useful to help me send standard messages. I used to have to type in single keys, but i notice it allows one to paste a whole text strigng.
As for Dictation, i'll play with it and see how it goes... but the more the engine learns your speech the better.
As for Dictation, i'll play with it and see how it goes... but the more the engine learns your speech the better.