A preliminary investigation into the performance of a simple Time Encoded Speech (TES) isolated word recognition (IWR) direct voice input (DVI) system, using both normal and whispered (unvoiced) speech is described. Experimental conditions include evaluations with untrained military speakers in severe acoustic background noise (c. 70-90 dB SPL) with handheld omnidirectional microphones. System performance in both normal and whispered speech domains is presented.