A large, free audio sample database (10M words pronounced), a test bed for voice activity detection algorithms and for single-syllable word recognition - View it on GitHub