Another common KWS method is Acoustic KWS. Using this method, the engine does not attempt to transcribe the entire stream of speech. Like the LVCSR-based method, this method employs the Viterbi search.