Ecomars: Speech Database for External Communications in Maritime Settings

Abstract

Results obtained from different international research projects clearly show the feasibility of implementing simultaneous automatic translation systems to key areas when there is limited content. One of these controlled areas, in regards to vocabulary and syntactic structure, is that of routine oral conversations that take place in maritime settings between ships or between ships and shore services through the use of Standard Marine Communication Phrases.
Hence, the University of A Coruña has conducted a study titled “Language industries applied to oral communications in maritime settings” which had as its main goal to study the legal and technical possibilities and the commercial suitability of the development of an automatic language translator for oral communications in that area.
Among the many activities to be developed in this project, is the compilation of actual oral samples of such communications; samples that have given rise to the emergence of the speech database ECOMARS.
ECOMARS is the first and unique spontaneous speech database for external communications in maritime settings in Spanish and English that has been compiled and dealt with in accordance with the needs stated in the aforementioned project. The specifications corresponding to its design, realization and treatment will be broadly discussed in this article.