A number of woven fabrics an selected and their rustling sounds recorded. T
he spectra forms of these recorded sounds are then obtained by FFT. The lev
el pressure of total sound (LPT) is calculated to evaluate sound loudness o
f specimens. Level range (Delta L) and frequency differences (Delta f) are
used to quantify sound spectra shapes. Linear trends in frequency with auto
regressive errors are fitted to amplitude, and three coefficients (ARF, ARE
, ARC) Of the functions are obtained. KES-FB is used for mechanical propert
y measurements. Fabrics with the same fiber type show similar spectra shape
s. Silk and polyester satin fabrics generally make the quietest sounds with
low LPT, and they also have lower ARC values. Silk crepe de chine shows th
e highest level range value (Delta L). ARC values are significantly correla
ted to Lpr. The sound parameters of each fiber group are predicted with som
e mechanical properties such as bending, sheer, compression, and surface pr
operties, fabric thickness, and weight by stepwise regressions.