Incorporating information from syllable-length time scales into automatic speech recognition | IEEE Conference Publication | IEEE Xplore