VoiceXML

The VoiceXML ( tensible Voice e' X' M arkup L anguage or language of vocal extensible beaconing ) " is a standardized language of programming of vocal application.

Vocal application

A vocal application is a Computer program of Téléphonie which makes it possible a person to discuss on the telephone with a interactive vocal Serveur. The service of consultation per telephone of the schedules of train in is an example: " I compose a number, I ask for a schedule, a machine announces the schedule souhaité" to me;.

A vocal application allows:

the communication of the user towards the system either by Codes DTMF (“musical keys”) using the keys of the telephone keyboard or by Voice recognition;
the communication of the system towards the user either by recorded advertisements, for example in studio, or by advertisements in Voice synthesis.

VoiceXML language

Definition

The VoiceXML is a dialect XML. As its name indicates it, it is founded on XML. It is a standard defined by the VoiceXML Forum , a consortium created by AT&T, Lucent, Motorola and IBM. It has had for summer recognized by W3C. It is resulting from the unification of technologies Phone Markup Language from Lucent and AT&T, VoxML of Motorola and SpeechML of IBM.

The VoiceXML is conceived to be included/understood by a particular system, called vocal Navigateur ( Voice browser in English), which will interpret the orders of the user and will answer by the means indicated.

Similarity between HTML and VoiceXML

There is analogy between the languages HTML and VoiceXML. The first is the language of creation of the Web pages and the second the vocal language of writing of the scenario. Thus the vocal Navigateur, homologous with the navigator Web used by the Net surfer is integrated in the vocal platform (SVI). A Web page posts text and images, whereas a page in VoiceXML diffuses audio files and/or messages in voice synthesis. Navigation on a Web site is carried out thanks to the keyboard and with the mouse. In a vocal service, it is done thanks to the keys of the telephone (DTMF) or by the word: voice recognition.

This similarity between the two languages makes possible the use of the same infrastructures as the Web servers, to reach the information systems.

Operation

A page VXML is generally distributed in HTTP by a Web server. This page can be dynamic as for a Web page, It can be generated “with stolen” by a waiter of application or more simply thanks to language PHP. The beacons constituting the page are interpreted by waiter VXML: it is about the phase known as of “parsing”. Waiter VXML sends to the customer on his telephone sound messages, resulting from files sound. In return the customer by interaction on the keys of his telephone or by his voice dispatches an answer to waiter VXML. This answer will make it possible the waiter to require a new page VXML of the Web server. It is just as for a Web page which calls another by a click of them on a URL.

VoiceXML versus code owner

The VoiceXML makes it possible to reduce the loads of development of the vocal applications.

To date, a very great number of actors of this industry agrees on the fact that the development of scenario vocal is up to ten times faster in VoiceXML than with a language of old generation. This factor is all the more large as the project is important. Within the framework of renewal of platform, several recent references in the banking world and of insurance made it possible integrators to deploy vocal services on the same webservices as those deployed for the Web. The vocal service which would have required 5 to 6 months of development specific and settling was carried out in 10 days in VoiceXML.

This speed of development is explained on the one hand by the use of standard and not of a language owner requiring of the specialists. In addition the “webisation” of the applications trades (access as light a customer mode Web) allows the use of the same infrastructures as those developed by the Web servers, to reach the information systems (safety, footbridge, DBMS, waiter of applications). All these services are re-used to reach the central data in real-time.

The integration of a VoiceXML program does not require any recompiling of the core because VoiceXML is a language interpreted in real-time. Because it is based on Web technologies, VoiceXML is adopted by the developers of the Web and the creation of vocal gates is carried out without particular competences in the vocal field. VoiceXML allows the use of an architecture 3 thirds which differentiate the elements from a vocal service: dated, description of the scenario (appliance server) and execution of scenario (SVI).

It is all the more simple to make function a SVI in mode ASP with language VXML because dated them and the distribution of pages VXML can always remain with the centres of the company, while the expensive and technical environment is externalisé.

External bonds

VoiceXML Weblog
Book of Jose Rouillard " VoiceXML the language of access to Internet by téléphone"
Forum VoiceXML
VoiceXML Specification on the site of the W3C and its French translation
W3C' S Voice Browser Working Group
W3C' S VoiceXML 2.0 Recommendation
VoiceXML Forum Tutorial
VoiceXML Review E-zine
OpenVXI, interpreter VoiceXML open-source
VXIasterisk, SVI - VoiceXML Interpreter for Asterisk PBX
publicVoiceXML, interpreter VoiceXML open-source

Random links: