Copy link to clipboard
Copied
Hi, I have been surfing the Internet searching for a solution for offline speech recognition in Air for android and ios and have just come across this code. I'm wondering how can I modify this code to create an application that is actually a 'Voice Recognition Game'.
I should Define/add a "sentence" to each frame and check if users voice input matches that "sentence" or not...
For example we add this sentence to frame 10 : "This is a hello world test." if timeline reaches frame 10 then :
1) Speech recognition automatically starts.
2) User's voice convert to text.
3) If user's input matches "This is a hello world test." then we continue to frame 11
4) if user's input does not match, then speech recognition starts again (To let the user his second try..his last try)
5) This time if user's input match our sentence, continue to frame 11 and if it doesn't match go to frame 1
I don't need any text to speech that is included in the code or any of controllers and I just want English as my speech to text engine.any help or suggestions is greatly appreciated.
here is the link to the code and source files : https://fabricemontfort.com/voice-recognition-and-speech-synthesis-in-as3-with-air/
Copy link to clipboard
Copied
what are you testing? whether there's any voice that says 'this is a hello world test', or whether there's a particular voice/user?
if the former, what's the problem? montfort's code works surprisingly (imo) well in a quiet environment and needs no modification.
if the later, i think that could be done using a spectral analysis of the sounds but i don't see how montfort's code plays a role.
Copy link to clipboard
Copied
Totally hopless with as3 right now ...:(
I thought that code works offline when I asked this question but I tested it and unfortunately it only works online ...
Montfort's code IS TOTALLY USELESS if it works only online....
Did you use that code before ?
Copy link to clipboard
Copied
an internet connection is needed and yes, i've used his code.
Copy link to clipboard
Copied
"On both systems, this ANE use the Speech Recognition Cloud API (juste like Siri and Google Agent). You have to ensure you have a connexion to internet.
If you want to do offline Speech Recognition, you have to ensure that appropriate language packs are installed (In Settings > Languages > Speech Recognition > Offline). There is actually no way to do this programmaticaly." I found this in his docs but I'm Sure that I've downloaded offline voices and I've used offline recognition with native android apps...
Do you know if there is an offline solution or not ?
Copy link to clipboard
Copied
i don't know of any offline solution.
Copy link to clipboard
Copied
I guest you should work on an ANE that embed a whole speech recognition like transiflex or orther.
but it's a tough work.
Copy link to clipboard
Copied
someone would have to hire me to do that.
Copy link to clipboard
Copied
I can't find this code, could you please help me