Dual path noise detection and isolation for an acoustic respiration monitoring system detects noise in an acoustic signal recording lung sounds using two discrete noise detection techniques. A first technique detects portions of the signal that exhibit long-term, moderate amplitude noise by analyzing cumulative energy in the signal. A second technique detects portions of the signal that exhibit short-term, high amplitude noise by analyzing peak energy in the signal. Noisy portions of the signal are isolated using the combined results of the dual path detection. A respiration parameter is estimated using the signal without resort to the noisy portions and information based at least in part on the respiration parameter is outputted.
