The stimulus for the creation of a common speech database began with a research paper in which researchers collected speech recognition devices and used common speech data to compare their performance.