English speech synthesis based on multi-layered context oriented clustering; towards multi-lingual speech synthesis