RESEARCH OUTPUTS
Speech-enabled web
Category: Technologies
| The web’s increasing importance and penetration into every-day activities makes the need for its accessibility more imperative than ever. Information repositories, communication means, and interactive services based on the internet are some of the most active development areas today. As true universal access and inclusive design are now becoming critical requirements, efforts to design and implement browsing helpers, encounter the dilemma of generality versus speciality. A general web enhancement tool (such as a screen reader or a text browser) is able to deal with virtually any web page based on uniform, systematic and widely applicable interaction patterns. On the other hand, a custom tool tailored to the specifics of a website will perform better in the sense that it will be able to exploit available a priori information on the site’s structure to convey information more coherently to the user. WebSpeech is a tool developed at ILSP that is intended for people with visual impairments. It tries to combine advantages from both customized and generic web enhancement tools. It consists of a generic engine and a set of case-specific filters. The engine is able to communicate with an external web browser, acquire and parse the content of web pages, synthesize bi-lingual text to speech using ILSP’s high quality speech synthesizer, and supports a set of common functionalities such as navigation through hotkeys, audible selection lists etc. The role of each filter is to capture the specifics of a website: the way to determine the title of the current page, the website navigation menus, the current location in the site’s overall map etc. However, the most significant task of the filter is to determine the structure of actual textual information in the current page, mainly its paragraph structure and any tabular formulations. WebSpeech poses no requirements on a page and introduces no overhead to the design and development of a website. Based on its engine/filter approach, it can deal with any website without any modifications in the tool’s engine. |
|
