Qt5.x IVI Speech Recognition

[Ramakanthreddy Kesireddy]

I downloaded the simon code from the repository that you shared.

If am not mistaken, I need to build the same using build.sh  script.
Am not sure about build time dependencies though. We are planning to use Julius installed on board.

1. Create a Julius RecognitionConfiguration that points to your speech model 2. Create a JuliusRecognizer and call init(yourConfig) (returns a boolean for success; if it fails, call getLastError() and / or review the log returned by getLog()) 3. Call recognize(pathToDataFile) which returns a list of RecognitionResult (The sentence() getter method of that returns the recognized text) 4. After you're done, call uninitialize()


>The recognition will be explicitly triggered through some control.
Good, that makes the task a lot easier.

Incase if VR is expected to be running all time onboard for some usecases, what would be the additional tasks?

Best Regards,
Ramakanth

-----Original Message-----
From: Peter Grasch [mailto:peter at grasch.net]
Sent: Tuesday, April 15, 2014 6:33 PM
To: Ramakanthreddy Kesireddy
Subject: Re: Qt5.x IVI Speech Recognition

Hi,

On Tuesday, April 15, 2014 09:43:41 AM you wrote:
> Am looking for a Qt API to easily add the features and enable to work
> on other IVI platforms as well. However, the current platform is
> embedded linux i.e. Yocto on IMX6 with framebuffer as backend.
You need more than an API for speech recognition. You also need a speech model (data file(s)).

If your dictionary is static, I advise you to do this offline (you can build a speech model with Simon and export it; all of this is documented in the Simon
manual) and then just deploy the model with your product.
Please note that you will need to build different models for every language you aim at supporting.

> The recognition will be explicitly triggered through some control.
Good, that makes the task a lot easier.

> Could you please provide sample Qt4 app which uses libsimonrecognizer
> and the prerequisites/dependencies to execute the application?
No, there really is none if you discount Simond which is not a great example application because it does other, quite complex stuff. If you look at the source, though, you will see that it's extremely straight forward.
Assuming you use SPHINX:
1. Create a SphinxRecognitionConfiguration that points to your speech model 2. Create a SphinxRecognizer and call init(yourConfig) (returns a boolean for success; if it fails, call getLastError() and / or review the log returned by
getLog())
3. Call recognize(pathToDataFile) which returns a list of RecognitionResult (The sentence() getter method of that returns the recognized text) 4. After you're done, call uninitialize()

The steps for using Julius are the same, just substitude Julius for Sphinx in the class names.

If you look at the implementation of those classes, you can see that this is also not substantially harder if you use the C API directly if you'd prefer that over introducing an additional dependency (I'd consider it as we make no guarantees for e.g. binary compatibility for libsimonrecognizer).

Best regards,
Peter

________________________________

DISCLAIMER:
This email (including any attachments) is intended for the sole use of the intended recipient/s and may contain material that is CONFIDENTIAL AND PRIVATE COMPANY INFORMATION. Any review or reliance by others or copying or distribution or forwarding of any or all of the contents in this message is STRICTLY PROHIBITED. If you are not the intended recipient, please contact the sender by email and delete all copies; your cooperation in this regard is appreciated.


============================================================================================================================
Disclaimer:  This message and the information contained herein is proprietary and confidential and subject to the Tech Mahindra policy statement, you may review the policy at http://www.techmahindra.com/Disclaimer.html externally http://tim.techmahindra.com/tim/disclaimer.html internally within TechMahindra.
============================================================================================================================