Key points are not available for this paper at this time.
An objective and accurate measurement and characterization of breath sounds was carried out by a fast-Fourier-transform frequency-domain analysis. Normal vesicular breath sounds, picked up over the chest wall of 10 healthy subjects showed a characteristic pattern: the power of the signal decreased exponentially as frequency increased. Since the log amplitude vs. log frequency relationships were linear, they could be characterized by the values of the slope and the maximal frequency. The average slope of the power spectrum curves was found to be (in dB/oct +/- SD) 13.0 +/- 1.4 over the base of the right lung, 12.6 +/- 2.4 over the base of the left lung, 9.8 +/- 1.4 over the interscapular region, and 14.4 +/- 4.3 over the right anterior chest. The maximal frequencies of inspiratory and expiratory breath sounds, picked up over the base of the right lung, were (in Hz +/- SD) 446 +/- 143 and 286 +/- 53 (P less than 0.01), over the base of the left lung 475 +/- 115 and 284 +/- 47 (P less than 0.01), over the interscapular region 434 +/- 130 and 338 +/- 77 (P less than 0.05), and over the right anterior chest 604 +/- 302 and 406 +/- 205 (P less than 0.05). Breath sounds picked up over the trachea were characterized by power spectra typical to a broad spectrum sound with a sharp decrease of power at a cut-off frequency that varied between 850 and 1,600 Hz among the 10 healthy subjects studied.
Gavriely et al. (Sun,) studied this question.