Segmental Models With An Exploration Of Acoustic And Lexical Grouping In Automatic Speech Recognition