Using Speech Recognition to Receive a Selection

When enabled by the UseSR property in the configuration, Web telephony accepts DTMF digits together with speech input. Using speech recognition (SR), Web telephony tries to detect whether the spoken input sounds like one of the phrases sent to the SR engine. Web telephony uses the SRParameters property in the configuration to load an SR engine.

The phrases sent to the SR engine for each menu item (regardless of TTS mode) are:

• InnerText property for anchors created by the A element, or menu buttons created by the BUTTON and INPUT TYPE=button elements.
• VALUE attribute for submit buttons created by the BUTTON TYPE=submit and the INPUT TYPE=submit elements.
• ALT attribute for areas created by the AREA element.
• Option text for OPTION elements inside a list created by using the SELECT element.
• Label text for radio buttons created by the INPUT TYPE=radio element.

Each of these can have additional phrases if the GRAMMAR attribute is defined on them.

Phrases for stand-alone buttons and check boxes are taken only from the configuration.

If the WTEApplication.UseSR property is TRUE and the WTEApplication.OfferItems property is FALSE, WTE does not play the announcements inside a menu. However, WTE plays the announcement of the menu itself.