Advancing Electromyographic Continuous Speech Recognition: Signal Preprocessing and Modeling