Speaking style is another key feature that defines the possible uses of the speech corpus. For instance a corpus containing spontaneous or non-prompted speech will not be useful for a dictation task.
Unfortunately many speech corpora contain only one speaking style and are therefore restricted in their re-use for different applications. This is a pity considering the fact that the recruitment and recording of speakers is the most expensive part of a corpus production. Therefore we strongly recommend specifying at least two different speaking styles for a corpus production. The following list gives an overview of the main speaking styles with rising complexity.
Please keep in mind that the chosen speaking style will interfere with other specifications like the recording setup, the speaker profiles etc.