hi mark
can you suggest to me a paper or text in which I can understand "taper, smoothing, band width' what is the difference betwen the windows typesin H/V toolbox
Can a wrong choice of 'Window type' in taper or in smoothing, produce a mistake?
Hou can I choose the right window type?
r
taper and smoothing window types
Re: taper and smoothing window types
Hi Riccardo,
Most of these topics are discussed in SESAME guidelines.
http://sesame.geopsy.org/Delivrables/De ... elines.pdf
The taper multiplies the original signals by a window function to ensure that both ends of the signals are null. Fourier transform applies only to periodic signals which is not the case for real signals. But there are finite in time (limited time window length), hence the periodicity can be gained by pasting the original window before and after, and infinitely. If the the value at the ending time is not the same as at the starting time, you may introduced biases in your transform. If you define a 50% long taper the transition from 0 to 1 will take half of your signal length, with a strong modification of your waveform. Usually 5% is enough and does not modifies too much the waveforms.
The most common smoothing method for H/V is Konno Ohmachi 1998
https://pubs.geoscienceworld.org/ssa/bs ... m=fulltext
Do you mean the 'band width' in the smoothing parameters?
Best regards,
Marc
Most of these topics are discussed in SESAME guidelines.
http://sesame.geopsy.org/Delivrables/De ... elines.pdf
The taper multiplies the original signals by a window function to ensure that both ends of the signals are null. Fourier transform applies only to periodic signals which is not the case for real signals. But there are finite in time (limited time window length), hence the periodicity can be gained by pasting the original window before and after, and infinitely. If the the value at the ending time is not the same as at the starting time, you may introduced biases in your transform. If you define a 50% long taper the transition from 0 to 1 will take half of your signal length, with a strong modification of your waveform. Usually 5% is enough and does not modifies too much the waveforms.
The most common smoothing method for H/V is Konno Ohmachi 1998
https://pubs.geoscienceworld.org/ssa/bs ... m=fulltext
Do you mean the 'band width' in the smoothing parameters?
Best regards,
Marc
Re: taper and smoothing window types
thank you Mark
R
R
Re: taper and smoothing window types
>>Do you mean the 'band width' in the smoothing parameters?<<
and in the active f-k toolbox fk Power?
and in the active f-k toolbox fk Power?
Re: taper and smoothing window types
Band width in the FK toolbox (passive and active) is the width of the filter for each frequency band. FK is calculated for several central frequencies as defined by the frequency sampling. The cross-spectrum matrix is evaluated over a band width around each central frequency with a Gaussian weighting function. If the band width is null, a single FFT value is considered. The band width is relative.
Band width for the smoothing parameters works in the same way. In this case the shape of window can be adjusted to any of the available window functions.
Band width for the smoothing parameters works in the same way. In this case the shape of window can be adjusted to any of the available window functions.