Sorry your browser is not supported!

You are using an outdated browser that does not support modern web technologies, in order to use this site please update to a new browser.

Browsers supported include Chrome, FireFox, Safari, Opera, Internet Explorer 10+ or Microsoft Edge.

Raspberry Pi / Voice Recognition?

Author
Message
Anglerte
User Offline
Joined: 20th Jan 2017
Location:
Posted: 14th Feb 2017 00:23
Would it be possible to implement some sort of voice recognition in an AppGameKit Pi app, and if so, how would one go about this? The question stands for Tier 1 and Tier 2, although I imagine the latter could be integrated with some other C++ library.
CJB
Valued Member
13
Years of Service
User Offline
Joined: 10th Feb 2004
Location: Essex, UK
Posted: 15th Feb 2017 09:52
Unfortunately AppGameKit doesn't even allow us to record sound in Tier 1. Using a 3rd party lib in Tier 2 would be the only way at the moment.

If we COULD record sound, or access the audio input stream in Tier 1, then we'd have the possibility of sending audio snippets to a voice-to-text API. That is already on the wish-list for additional features.
Anglerte
User Offline
Joined: 20th Jan 2017
Location:
Posted: 23rd Feb 2017 04:09
Thank you for the information. Here's to hoping they get around to it when they can.
Jack
13
Years of Service
User Offline
Joined: 4th Oct 2004
Location: [Germany]
Posted: 28th Feb 2017 14:17
Actually you can achieve something like this by using a seperate program.
Your target should be to find a Voice to Text lib for linux.
Modify it, so it creates .txt files with the last input. This file can be loaded into agk with
the opentoread commands.

[/url]
smerf
12
Years of Service
User Offline
Joined: 24th Feb 2005
Location: nm usa
Posted: 28th Feb 2017 22:49
first thing I ever wanted to do with dbpro was an ai with voice rec had high hopes before I knew it could barley do anything with audio at all
A child's dream never dies.





Alex Deef
3
Years of Service
User Offline
Joined: 8th Apr 2014
Location: eastern europe
Posted: 6th Mar 2017 18:02
Your target should be to find a Voice to Text lib for linux.
http://www.pishop.de/Cialis.html
http://www.achetermed.fr/Levitra.html
janbo
8
Years of Service
User Offline
Joined: 10th Nov 2008
Location: Germany
Posted: 19th Mar 2017 00:22 Edited at: 19th Mar 2017 00:30
I looked into it two month ago I think and asked for the sound recording again.
Also I singed in several free voice recognition API's out there and tried my best.
I thought that it could be possible to dictate into a hidden Editbox with focus or as I mentioned somewhere to use another script to record and making the request as we even can't change the header for the HTTP packets.
But I managed to make a separate script which does the recording and gives the right text back.
And I tried to create my own HTTP packets via the socked commands but as long as we cant record sound we rely on a second script/program.

Ah here it is ...I used curl.exe to make the requests and the ffmpeg.exe to record sounds then I wrote batch which creates a JSON file with the needed data to send to the google speech API... the rest did Jack mention before
Just ask If you want the files

Using AGKv2 Tier1
Jack
13
Years of Service
User Offline
Joined: 4th Oct 2004
Location: [Germany]
Posted: 14th Apr 2017 23:26 Edited at: 14th Apr 2017 23:29
Quote: "Just ask If you want the files"

I would really like to see it, since I have a google speech to text api key but never used it before.
I have a simple AI that can already talk.
Voice recognition would be awesome, so I could learn my AI without writing any text!

[/url]
janbo
8
Years of Service
User Offline
Joined: 10th Nov 2008
Location: Germany
Posted: 16th Apr 2017 04:21 Edited at: 15th May 2017 13:20
Ok ehm... here is something ... ...since we have plugins we could do better

Using AGKv2 Tier1

Attachments

Login to view attachments

Login to post a reply

Server time is: 2017-10-22 23:52:53
Your offset time is: 2017-10-22 23:52:53