언어 설정

Menu
Sites
Language
Add external Voice recognition to Input Method Editor for Wearables

Hi,

Like many people I'm in shock that samsung is not able to add Voice-to-text (VTT) processing for more then the standard languages. Even after all these years it has been using S-voice. Is I speak dutch, i'd like to use my gear s3 to send text messages in Dutch. And currently this is not working and very frustrating to have spend 400€ on it. 

So I began lookin into some options, but Tizen applications are very scarce and i found nothing that could help. Only a very interesting app called Travel translator seems to catch my eye. It's able to translate voice from 40+ languages (hello SAMSUNG, why can't we get simple VTT like that!). I looked further into it and seems in uses Nuance VTT for it's speechprocessing. (It is also the only provider that supports AMR recordings (as the gear S3 can only record that format) and suport HTTP rest. 

Further into researching tizen I found some templates about Input Method Editor. If I understand well, this is an input method you make that will appear everytime an input field is used on the device aka keyboard. I've also looked at the voice recording template which can atleast record

My question would be: is it possible to have within the IME an option to record audio, send it with HTTP to Nuance and get the resulting text as input? 

I thought I would be able to atleast fiddle a bit myself, but as I'm not familiar with HTML nor Java nor CSS and as there are near to no guides for basic Tizen developing it's out of my league. All the building blocks are already made in the templates, but I can't seem to decipher even the basics.

I wonder why no one before seems to have tought about the idea?

As a side note I've developped an application in Unity  which does about what I say, but currently the microphone is not supported in Tizen api so as last resort I'd post in here. 

Responses

3 댓글
Armaan-Ul- Islam

Seems possible from your findings.

If Nuance VTT is providing complete solution for Voice-to-Text Services (may be for fee / free of cost ?), 

- You have to develop your own IME application (you may start from IME template) with a button.

- Action of that button would be to record voice (May try Native Recorder API )

- and transfer that recording to Nauce VTT server & recieve the text returned. (XMLHttpRequest could be an option)

- and display the text on message board (Some HTML,CSS tricks).

 

I guess you have to make yourself familiar with JavaScript for that.

Hollanders

To be clear, this means I should go for the Native application api and not the web application? Or can both be used together

Armaan-Ul- Islam

You may switch for Native as the Recorder API is on Native platform.