Multi-resolution phonetic/segmental features and models for HMM-based speech recognition | IEEE Conference Publication | IEEE Xplore