Fig. 2From: Performance vs. hardware requirements in state-of-the-art automatic speech recognitionDifferent speech features obtained at different processing stagesBack to article page