Exploiting synchrony spectra and deep neural networks for noise-robust automatic speech recognition | IEEE Conference Publication | IEEE Xplore