I want to develop a software which can recognize the voice and convert it to text and initiate a website like google(passing the text as a parameter). And this text(which was converted from voice) should be searched in google(or anyother website of choice)
If you have any idea how to do it …please reply
thanks
You’ll need an external service for this. Voice recognition in javascript only is not going to work. And to embed some algorithms to keep track of that inside the app itself is not going to work either.
I don’t know if there are API’s you can send snippets of sound files to, but my best bet is to google if services like this are offered by f/e google.
Then still I’m not sure if it would be able to recognize the user, maybe just translate to text.
I saw that thyere are some libraries for vioce recognition…so is there a way we can use that in our project?
Maybe you can use https://api.ai/, they offer also an cordova plugin.
I’d say start with using the browser’s built in speech recognition. I’m not sure if the native browser that ionic runs on supports this, but my guess would be that it does, at least in newer OS versions. I taught a workshop on this at a Code Day over the summer, so I actually have a really simple walkthrough available here: http://www.roblouie.com/article/222/code-day-summer-2016-voice-driven-web-apps/
If using the native browser api doesn’t work out for you, this plugin looks to provide a nice wrapper to both the browser speech recognition and the phone’s speech recognition: https://github.com/macdonst/SpeechRecognitionPlugin
Somewhere between those two things you should be good to go.
Another thing you can do is keep an eye out for Lex. This is the technology Amazon uses on Echo.
They’ll probably have some kind of javascript SDK.
other service its Google Speech . https://cloud.google.com/speech/
As already mentioned you should send your recordings to the online services. Which one depends on you - we’ve only looked at Google and Microsoft, and they’re both a payed service (there’s basically two options - translating word-for-word or entire sentences/recordings).
Google offers speach to text via their cloud services (i.e. you have to have an app id running in their cloud) and Microsoft’s version was simply buying a key to call the service. I’m not sure about the billing amount so you should check that with the provider.
As for the quality, surprisingly, none of them proved as good as their assistants do it (i.e. Google search is recognizing much better than through the offered API, and Microsoft Cortana is doing awesome job compared to theirs API which is right next to non-operational). I’m not sure why is this the case, but you should have this in mind before you go too far away in your project…
Also worth of mention is the free Chrome browser plugin which works awesome and can be used in web apps, but I’m not sure if there’s some plugin like it for cordova…
Have you tried integrating amazon lex in your ionic app ? I am using amazon mobile hub but wondering how can I integrate Lex with aws-amplify … any idea ?
Bro, I am in the same boat as you, struggling to get the amazon lex into ionic using amplify.
Amplify documentation is really not helpful. Please help if you have got any solution for this.