To determine if wheezing is a reproducible clinical sign, we presented recorded breath sounds from asthmatic patients to four groups of health professionals: pediatric residents, nurses, pediatricians, and physiotherapists. Their subjective assessments included scores of wheezing severity and estimates of wheezing duration. All participants repeated the test at least two weeks later. Results were compared to computer aided spectral analysis of the recorded breath sounds. Interobserver and intraobserver variability fell somewhat between chance and total agreement. In contrast, the computer analysis allowed an objective and reproducible characterization of wheezing in asthma.