VoxForge
The CMU_ARCTIC database was constructed at the Language Technologies Institute at Carnegie Mellon University. It consists of around 1150 utterances selected from out-of-copyright texts from Project Gutenberg.
The prompt file used in the CMU_ARCTIC database were originally designed as US English single speaker prompt file for Speech Synthesis research (i.e Text to Speech). Since it is phonetically balanced, we are using it to generate prompt files for the creation of Speech Recognition Acoustic Models.