VoxForge
My first upload...
Thanks for the submission,
Please note, we prefer segmented speech (i.e. one audio file for each prompt line) ... otherwise I have to segment it manually (because the training algorithms can't deal with large speech files).
thanks,
Ken