## 2aSP8. Short-time Fourier transform, Wigner--Ville analysis, and wavelet transform compared.

### Session: Tuesday Morning, December 3

### Time: 10:45

**Author: J. Romero**

**Location: Lab. de Acustica, Dept. de Fisica Aplicada, Facultad de Ciencias Fisicas, c/Doctor Moliner, Burjasot 46100 Spain**

**Author: Salvador Cerda**

**Location: Lab. de Acustica, Dept. de Fisica Aplicada, Facultad de Ciencias Fisicas, c/Doctor Moliner, Burjasot 46100 Spain**

**Abstract:**

Short-time Fourier transform (STFT), Wigner-Ville analysis (WVA), and
wavelet transform (WT) are possibly the more important methods of time-frequency
analysis. STFT makes use of the time-window function to locate Fourier analysis.
WT makes use of the time-frequency window to locate the analysis in the
time-frequency domain. WVA is different than STFT and WT in the sense that no
outer function is made that uses it in the analysis. Time-autocorrelation is
defined and Fourier transform is applied. Six signals were studied with
different methods. STFT was applied with three time windows: Gaussian, Hamming,
and rectangular. The time duration of the time window has been varied (23, 12, 6
ms), showing the uncertainty principle effect. WT was applied using the real
part of Morlet function. Improvements with respect to STFT were shown. Finally,
WVA was applied to the signals. The time-frequency resolution of WVA is very
good, but the presence of the 4(pi) factor at the time to calculate the Fourier
transform, reduces the frequency interval of analysis by a factor of two.

ASA 132nd meeting - Hawaii, December 1996