Abstract
The paper presents the upcoming corpus of records of professional radio speakers, whose creation has been supported by the grant of GA UK. This corpus of records with their transcriptions will serve as a source for the research of the pronunciation usage in Czech media.
The material has been gained from the audio archives of the Czech radio, from which records of many radio programmes in a good quality can be downloaded. The choice of speakers is focused on those speakers from whom a higher number of longer monological records are available, so that it would become probable that the observed phenomena will be found in a sufficient number of occurrences.