Internet application for collective realization of speech evaluation by listening tests

The paper describes the developed multiaccess internet application for collective realization of listening tests which are very important for evaluation of the synthetic speech quality used in text-to-speech systems or human-machine interfaces. This application together with testing speech and image corpora are stored on the server PC, where the listening tests are also executed. The application communicates with the user via the HTTP protocol. Obtained results of evaluation can be immediately processed and displayed in numerical and graphical representation. Control and administration of the application is also described in this paper. At present, in addition to the conventional types of listening tests, the special ones - type of emotional state or speaking style, influence of visual information on speech evaluation - are implemented. In the near future we plan to extend the application area for evaluation of audio-visual material.