End-to-end acoustic modeling using convolutional neural networks for HMM-based automatic speech recognition | Litlas