Phonetics by Reetz Henning; Jongman Allard; & Allard Jongman

Author:Reetz, Henning; Jongman, Allard; & Allard Jongman
Language: eng
Format: epub
Publisher: John Wiley & Sons, Incorporated
Published: 2013-05-07T16:00:00+00:00

It should be kept in mind that a properly selected window reduces, but never entirely suppresses, the effects of sudden amplitude jumps.

8.3.3.1 THE RELATION BETWEEN WINDOW SIZE AND SPECTRAL RESOLUTION

The Fourier transformation actually takes the window size as the duration of one period (which is inversely related to its frequency; see Section 7.3.1.2) and “subtracts” this period from the signal, computing thereby amplitude and phase of this period, which is the information about the first harmonic (of the window size). Then this process is repeated for two, three, etc. periods, giving the amplitudes and phases of the second, third, etc. harmonic. Consequently, the spectral resolution – that is, the spacing between the frequency values – is inversely proportional to the window size (in ms): a 10 ms window has a spectral resolution of 100 Hz, a 20 ms window of 50 Hz, etc. Hence, the spectral resolution depends solely on the window size in milliseconds, and not on the sampling rate of the digitized signal. A higher sampling rate leads to information about higher frequencies, up to the Nyquist frequency (see Section 8.1.2), but does not increase the spectral resolution. This can only be increased (i.e. more closely spaced frequency values) by increasing the window size.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.