To search, Click below search items. |
| All Published Papers Search Service |
|
|
|
Title |
Spectrum Estimation by Several Interpolation Methods |
|
Author |
Manabu Ishihara |
| Citation |
Vol. 6 No. 2 pp. 205~209 |
|
Abstract |
In this paper the experimental results on the spectrum estimation of the Japanese speech waveforms /a/, /i/, /u/, /e/, /o/ obtained by three different approaches of interpolation, that is Akima’s approximation, the spline function approximation and Lagrange’s formula, are reported. In our experiments, the speech waves are at first analyzed by FFT(Fast Fourier Transform) technique and then the FFT spectra are smoothed or fitted by varying the numbers of data points in the frequency region and applying three interpolation methods respectively. These methods are very available to extract the local peaks in the speech spectrum. The data values obtained by the investigation and comparison for the interpolated spectra indicate that Akima’s approximation method is fast in calculation of smoothing process and suitable to consider the spectrum envelopes. However spline approximation is useful expect for the case of all data points being used and Lagrange’s interpolation formula is not useful for the spectral estimation in our experiments. |
|
Keywords |
Spline’s function, Akima’s approximation, Lagrange’s formula, Digital Filter |
|
URL |
