| Title | Optimization of Weighting Factors for Multiple Window Time-Frequency Analysis |
| Authors | Maria Sandsten, Johan Sandberg |
| Alternative Location | http://www.eurasip.org/Proc..., Restricted Access |
| Year | 2009 |
| Pages | 2283 - 2287 |
| Document type | Conference paper |
| Conference name | 17th European Signal Processing Conference (EUSIPCO 2009) |
| Conference Date | 2009-08-24 |
| Conference Location | Glasgow, Scotland |
| Status | Published |
| Quality controlled | Yes |
| Language | eng |
| Abstract English | This paper concerns the optimal weighting factors for multiple<br> window spectrogram estimation of different stationary<br> and non-stationary processes. The choice of windows are of<br> course important but the weighting factors in the average of<br> the different spectrograms are as important. The criterion for<br> optimization is the normalized mean square error where the<br> normalization factor is the spectrogramestimate. This means<br> that the unknown weighting factors will be present in the numerator<br> as well as in the denominator. A quasi-Newton algorithm<br> is used for the estimation. The optimization is compared<br> for a number of well known sets of multiple windows<br> and the results show that the number as well as the shape of<br> the windows are important factors for a small mean square<br> error. |
Questions: webmaster
Last update: 2013-04-11
Centre for Mathematical Sciences, Box 118, SE-22100, Lund. Telefon: +46 46-222 00 00 (vx)