-
Buckeye corpus of conversational speech
This dataset focuses on spoken English encompassing daily live situations. -
Switchboard
Human speech data comprises a rich set of domain factors such as accent, syntactic and semantic variety, or acoustic environment.