The practice of using rising-falling chirps (i.e., those that sweep up
and down in frequency) in vibration testing is considered. This type
of chirp is shown to have an amplitude spectrum which exhibits severe
drop-outs in amplitude at many frequencies in the excitation bandwidth
. It will lead to Frequency Response Function (FRF) estimates that are
very sensitive to noise and should not be used. (C) 1996 Academic Pre
ss Limited