Summary
This paper describes the planning and creation of the Mixer and Transcript Reading corpora, their properties and yields, and reports on the lessons learned during their development.