Voice Recognition Implementation

[Naveen_D]

Hi All,

I wanted to know is it possible to implement voice recognition using qt and what is the procedure for the same.

thanks in advance

[QtAndrew]

@Naveen_D you must use an external voice recognizer, or write it by your own (i think thats not possible...)

[mrjj]

Hi
as @Naveen_D says you will need an external lib and which one also depends on what platforms you want to support?
https://en.wikipedia.org/wiki/List_of_speech_recognition_software

For Text to voice, Qt has
https://doc-snapshots.qt.io/qt5-5.8/qtspeech-index.html

[Naveen_D]

@QtAndrew ok i will use external library such as pocketsphinx.
Using this how i can do. Since em new to this em asking...if small example is available that would help me alot.

[mrjj]

@Naveen_D said in Voice Recognition Implementation:

pocketsphinx

hi
for any external lib , one should go and look at build info for the platform that is wanted.

Lets see for windows:

https://github.com/cmusphinx/pocketsphinx

Seems to use on windows, you must have V Studio 2010 or newer installed for this lib.
So for it to work for you , you should have Qt for VS version installed. exact version. (2010,2012,2013,2015)

Then after you made the LIB to build. That is, it produces .DLL or .LIB files

Then you will include in your own Qt project
http://doc.qt.io/qtcreator/creator-project-qmake-libraries.html

Check all paths in the .pro file. If it gives errors, check forum. plenty of posts about using a lib/dll.

Now you are ready to try use the functions it gives. :)

So a key thing to understand is that the external lib and the Qt version your are using - should been made by same compiler.
So if you are using mingw compiler, using a Visual Studio DLL won't work and reverse.

[Naveen_D]

@mrjj Thanks alot.

[Naveen_D]

@mrjj https://github.com/cmusphinx/pocketsphinx in this which file or lib i need to download for windows and also for linux ?

[mrjj]

@Naveen_D
If you mean the pocketsphinx-master.zip , its the source code.

On windows, you will use the pocketsphinx.sln to compile.
Then you get DLLS etc. ( the resulting library)

There seems to be no precompiled binaries so its up to you to make them.

Note. you must compile both on windows and linux. You cannot
use libs from windows in linux and reverse.

[Naveen_D]

@mrjj that means i should run the complete source code in qt creator and build the binary and use the required libraries ???

[Naveen_D]

@mrjj for linux also i need to use pocketsphinx.sln ?

[mrjj]

@Naveen_D
Yes, first step is to get it to compile. Either in Creator or in Visual Studio .
That will produce DLL/LIB file.

Then you will make new project ( your project) and add this DLL/LIB to it.

• linux also i need to use pocketsphinx.sln ?
  No. for linux , there is other build instructions.
  $ ./autogen.sh
  $ ./configure
  $ make clean all
  $ make check
  $ sudo make install
  --
  SLN files are for visual studio.

There is no visual studio on linux.

[Naveen_D]

@mrjj Okay thanks...i wanted to know is it possible to develop an desktop app using this which accepts voice recognition? or we need a pos device with some microphone?

[mrjj]

Well the pc or device
must have soundcard and a microphone but other than that,
there should be nothing stopping you to run as a Desktop app.

I used this on pc
http://www.nuance.com/for-individuals/by-product/dragon-for-pc/index.htm

Its brilliant.

[Naveen_D]

@mrjj ya but nuance is not open source rite so em trying with pocketsphinx

[mrjj]

@Naveen_D
Oh, no. its as commercial as it gets.

it was as example of voice Recognition on the desktop :)

The best i ever tried. It worked flawless even when multiple people speaking!

They even do allow others to use
http://www.nuance.com/for-developers/dragon/index.htm
but its not open source or gratis.

So i just mentioned it for a sample of VR that truly works :)

[Naveen_D]

@mrjj Is there any algorithm for voice recognition that i can use ?

[jsulm]

@Naveen_D Yes, pocketsphinx for example which @mrjj already suggested. Or do you want to implement it by yourself?

[Naveen_D]

@jsulm no i will use pocketsphinx only

[SGaist]

Hi,

Just in case, there's a speech recognition branch in the QtSpeech module that's current work in progress but might be interesting for you.

[Naveen_D]

Hello everyone

my question is do I need to install CMU-SPHINX first and then pocketsphinx ? or there is any other way for installing pocketsphinx in ubuntu?