Voicing parameter and energy based speech/non-speech detection for speech recognition in adverse conditions | Synapse