Pointers to clean, single-speaker speech in the original recordings, kept for a possible future voice model.
No voice model has been made. Arthur Penhale's rule for generated voice: never.
