Customisation and quality assessment of spoken language description

This paper describes a world-wide web based system which allows a speech data corpus developer to interact with a system for the comprehensive description of a spoken language data corpus. The interface at the browser allows the user to define the speaking environment that is to be described and then progressively to describe it and all the data files arising from it in a modular fashion. The quality of the description can be tested incrementally and the feedback generated used to further update the description. The user is effectively guided through the necessary modules in a way that reminds but does not demand adherence to a strict pattern. The complete data description can be displayed on screen and also downloaded in a simple text form to the user. Users may register as clients of the system in which case they can store descriptions on the system and upgrade these descriptions at any time