The Voice Browser Working Group

The Voice Browser Working Group's mission is to support browsing the web by voice. The web is much more than just the web pages you can see, it is also the web pages you can hear and speak to. While end users are familiar with interacting with visual html web pages rendered in their browser of choice, many users might be surprised to realize that today they regularly interact with the voice web through VoiceXML (VXML) and other technologies developed and standardized by the Voice Browser Working Group. Just as many sites have an html presence on the web for visual browsing, most large companies have a vxml presence on the web for voice browsing, which is most often accessed by calling the companies phone number. Unlike most visual web browsers, voice web browsers are typically without chrome and run in the cloud, so they are often transparent to the end user. But otherwise, all the normal power of the web applies including taking advantage of web services, markup, linking, uris, cacheing, standards, accessibility, and cross-browser support.

There are a suite of independent standards that are also supported as a parts of VoiceXML. These standards can be, and are being, used alone in non-VXML contexts; however, they achieve a powerful synergy when used in support of VXML. The latest recommendation of these web standards are:

VoiceXML (VXML)

a language for for creating audio dialogs that feature synthesized speech, digitized audio, recognition of spoken and DTMF key input, recording of spoken input, telephony, and mixed initiative conversations;

Speech Grammar Recognition Specification (SRGS)

a document language that can be used by developers to specify the words and patterns of words to be listened for by a speech recognizer or other grammar processor;

Semantic Interpretation for Speech Recognition (SISR)

a document format that represents annotations to grammar rules for extracting the semantic results from recognition;

Pronunciation Lexicon Specification (PLS)

a representation of phonetic information for use in speech recognition and synthesis;

Speech Synthesis Markup Language (SSML)

a markup language for rendering a combination of prerecorded speech, synthetic speech, and music;

In addition to recommendations that are used as part of VoiceXML, there are a couple of powerful specifications that are used to interact and control voice sessions (as well as control many other types of sessions and processes). These are:

Call Control (CCXML)

a markup language to enable fine-grained control of speech (signal processing) resources and telephony resources to perform scenarios such as call screening, whisper call waiting, and call transfer;

State Chart XML (SCXML)

a markup language to simply and precisely represent the semantics of state machines;

The W3C Voice Browser Working Group (members only) is chartered to develop the next generation of the voice web. Currently the group's main two areas of focus in the near term are on driving SCXML to last call status and on the next version of VoiceXML (3.0).