We present a broad-band optimal ventilator waveform (OVW), the concept of which was to create a computer-driven ventilator waveform containing increased energy at specific frequencies (f). Values of f were chosen such that nonlinear harmonic distortion and intermodulation were minimized. The phases at each f were then optimized such that the resulting flow waveform delivered sufficient volume to maintain gas exchange while minimizing peak-to-peak airway opening pressure. Simulations with a linear anatomically consistent branching airway model and a nonlinear viscoelastic model showed that respiratory resistance (Rrs) and elastance (Ers) estimates at 0.1-2 Hz from the OVW are far superior to those from a standard step ventilator waveform (SVW) during healthy and obstructed conditions and that the OVW reduces the influences of harmonic interactions. Using a servo-controlled oscillator, we applied individual sine waves, an OVW containing energy at 0.15625-2.4 Hz, and an SVW to healthy humans and one symptomatic asthmatic subject before and after bronchodilation. The OVW was markedly superior to the SVW and always provided smooth estimates of Rrs and Ers. Before bronchodilation in the asthmatic subject Rrs was highly elevated and Ers was markedly increased with f; after bronchodilation the level of Rrs and the f dependence of Ers decreased. Although based on results from only one asthmatic subject, these data suggest a dominant influence of airway constriction and lung inhomogeneities during asthmatic bronchoconstriction that is alleviated by bronchodilators. These and other results indicate that the OVW approach has high potential for simultaneously probing f and amplitude dependence in the mechanical properties of clinical subjects during physiological breathing conditions and perhaps during dynamic bronchoconstriction.