Help - Search - Members - Calendar
Full Version: ULRTVR - Skype4COM Voice Recognition IE Example
Skype Community > English > Development, Betas and Skype Garage > Skype Public API
TheUberOverlord
Note: Do NOT leave this running without keeping an eye on it, voice recognition can be very easily confused, and this has the ability to make calls


The ULRTVR For Skype (Beta 1.0) - Universal Language Real-Time Voice Recognition for Skype example shows once again the power of the Skype4COM lib.

This is a simple PoC "Proof of Concept" to show how easy it is to create your own voice recognition interface to the Skype client.

These concepts could be used to do many other things for both home and business, and can interface to sound devices other than the sound card as well.

NOTE: If you do not have the Skype4COM lib already installed you will need to download it from here, unzip it, and run the register.bat file:

https://developer.skype.com/Download?action...0.0.21.beta.zip

Before even attempting to play with this you will need 3 files supplied by Microsoft, this PoC will ONLY work with Internet Explorer. Please download and install these files in the order listed below:

SAPI 4.0 runtime support from Microsoft

http://activex.microsoft.com/activex/contr...api/spchapi.exe

Speech control panel from Microsoft

http://download.microsoft.com/download/c/9...439/SpchCpl.exe

Microsoft Speech Recognition Engine

http://activex.microsoft.com/activex/contr...ent2/actcnc.exe

Follow the instructions for the speaker and microphone test in last download and install above. Make sure you are in a quiet room when doing this.

Once the software above has been downloaded and installed on your system you can begin to use the example.

There are several commands supported, all of the large "Buttons" names that are displayed can also be spoken and supply the same actions as if they were pressed. Also you can speak the name of a status and your user status will change to the status spoken.

There are two additional commands which are not buttons called "Sleep" if you say "Sleep" then ALL commands besides "Wake" will not be processed until you speak "Wake".

Again, do not leave this running and not watching it, because even with it sleeping, it could confuse noise as the word "Wake" and possibly make calls, if you are NOT using it, close it!

This is important, because you may wish to adjust the sensitivity of the speech recognition if you see there is a high error rate in determining your spoken words.

Adjustments to the sensitivity can be made by going to your control panel, and selecting the Speech Icon, select the "Other" tab, click on the SAPI 4 control panel button, select the "Microsoft Speech Recognition 4.0 English, click on properties.

If you need to change your microphone settings, you can click on the Adjust Microphone settings button as well. After you make any changes, click the OK button.

You are more than welcome to look at the HTML source code, however PLEASE do NOT distribute this, it is a protected work.

The Commands:

Away - Changes User Status to Away.
Do Not Disturb - Changes User Status to Do Not Disturb.
Down - Will move down a list one item at a time.
Focus - Will Open the Skype Client from a Minimized state.
Invisible - Changes User Status to Invisible.
Minimize - Will Minimize the Skype Client.
No - Will end a call started by yes.
Not Available - Changes User Status to Not Available.
Offline - Changes User Status to Offline.
Online - Changes User Status to Online.
Page Down - Will page down one page of a list of items.
Page Up - Will page up one page of a list of items.
Skype Me - Changes User Status to Skype Me.
Sleep - Will not take any commands other than wake, see notes above.
Tab - Will change to the next tab on the Skype client.
Up -Will move up a list one item at a time.
Wake - Will now take other commands if was sleeping.
Yes - Will make a call to the list item selected from Contacts or History.

These concepts have the ability to do many things, not just process voice, but also DTMF.

As stated before this concept can interface to other audio devices besides your sound board, such as Skype phones, USB devices and other hardware.

This is just a simple example of interfacing to the Skype client using the SKype4COM lib and Voice Recognition. The reason Internet Explorer is used for most of my PoC's to date is it is a generic way to show these concepts.

There is a "You Said?:" text area on the page that displays what command was thought to be said, you can use the adjustments to fine-tune the recognition to be better for your voice by watching this.

Here is the ULRTVR web page, enjoy, any questions please post them here:

http://testing.OnlyTheRightAnswers.com/sky...TVRforSkype.htm

There is no need to do this step below, unless you wish to see call status change, then you will need to run this locally, again please do NOT distribute this, simply unzip this and you can run it locally on your system if required:

http://testing.OnlyTheRightAnswers.com/sky...TVRforSkype.zip

You can see the limited amount of CPU usage required to do Voice Recognition using this method compared to other methods.

ZOverLord
TheUberOverlord
Modfied to handle User Status changes. You can now speak a User Status and your User Status will change.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.