Filtering is a process that transforms an input signal or function into an output signal or function. Filters are designed systems that selectively transfer certain components of the input signal while attenuating others. There are various types of filters including low-pass, high-pass, and band-pass filters. Filtering can occur in both the time domain and frequency domain. In the frequency domain, filtering involves multiplying the input spectrum by the transfer function of the filter. This changes the amplitude and introduces a phase shift. In the time domain, filtering is performed by convolving the input signal with the impulse response of the filter. Filtering is used for applications like removing noise from signals or extracting specific frequency components.
Filtering is a process that transforms an input signal or function into an output signal or function. Filters are designed systems that selectively transfer certain components of the input signal while attenuating others. There are various types of filters including low-pass, high-pass, and band-pass filters. Filtering can occur in both the time domain and frequency domain. In the frequency domain, filtering involves multiplying the input spectrum by the transfer function of the filter. This changes the amplitude and introduces a phase shift. In the time domain, filtering is performed by convolving the input signal with the impulse response of the filter. Filtering is used for applications like removing noise from signals or extracting specific frequency components.
Filtering is a process that transforms an input signal or function into an output signal or function. Filters are designed systems that selectively transfer certain components of the input signal while attenuating others. There are various types of filters including low-pass, high-pass, and band-pass filters. Filtering can occur in both the time domain and frequency domain. In the frequency domain, filtering involves multiplying the input spectrum by the transfer function of the filter. This changes the amplitude and introduces a phase shift. In the time domain, filtering is performed by convolving the input signal with the impulse response of the filter. Filtering is used for applications like removing noise from signals or extracting specific frequency components.
Filtering is a process that transforms an input signal or function into an output signal or function. Filters are designed systems that selectively transfer certain components of the input signal while attenuating others. There are various types of filters including low-pass, high-pass, and band-pass filters. Filtering can occur in both the time domain and frequency domain. In the frequency domain, filtering involves multiplying the input spectrum by the transfer function of the filter. This changes the amplitude and introduces a phase shift. In the time domain, filtering is performed by convolving the input signal with the impulse response of the filter. Filtering is used for applications like removing noise from signals or extracting specific frequency components.
I. General meaning Filter to generate a new function (y(t)) from the old one (x(t)) Examples : a. Seismograph a b. Filter of electrical network
to transfer a land movement to be seismogram
to transform an input voltage to out put voltage
So : Filter is special design of system (initial lecture)
all of physics system have specific filter effect Example : to change spectral character of an input x(t) to be y(t) b d band-pass, l low-pass, hi h high-pass filt filters Filter is consist of : a. Analog Filter b. Digital filter Input TD FD
x(t)
Physical system that designed to simulate a filter
process that th t acquired. i d a numerically filtering process (by digital computer) h(t) Filter
X(f)
Out-put y(t) Y(f)
H(f)
Characters of filter (system)
I Input t
Out O t putt
x(t)
y(t)
Analog
xi((t))
yi((t))
Digital g
Linear
ai xi(t)
ai yi(t)
Stationer
x( t - )
y( t - )
| x(t) |certain
Stable Causal
x(t) = 0 x(t) ()0
t < t1 t0
explanation l ti
ai = constant
| y(t) |certain y(t) = 0
t < t1
Filter Transfer For linear filter :
H ( ) =
Y ( ) X ( )
In complex number H ( ) = H ( ) e
H ( ) = Amplitude-Transfer function i ( )
(Amplitude response)
( ) = phase-Transfer function (phase response)
In time domain (Filter transfer):
h (t ) =
1 2
H ( ) e i t d
Impulse Response
Filter operation :
x(t)
y (t ) = h(t ) * x(t ) Y ( ) = H ( )). X ( )
h(t)
y(t)
In time domain y domain In frequency
IMPULSE and STEP RESPONSE
1. If x(t) = (t) = Delta (Dirac) function, hence y(t) = h(t) = Impulse Response 2. If x(t) = (t) = Step function, hence t
y (t ) =
h ( ) d
Step response
Example : Transfer function of
low-pass filter
y (t ) =
h ( ) d
Separation of Transfer Function in Amplitude and Phase
I Input, t outt putt and d ttransfer f ffunction ti in i complex l :
X () = X () eix ( ) x(t)
Y ( ) = Y ( ) e
i y ( )
h(t)
H ( ) = H ( ) ei ( ) y(t)
Filtering process in Amplitude and phase:
Y () = H() . X ()
y () = () +x () + 2n
n = 0, 1, 2, 3,
It means : Filtering process H() is multiplication in spectrum amplitude
and addition in spectrum phase
Examples : 1.
x (t ) = cos 0t
= harmonic oscillation with constant 0,
1 Amax and 0t phase corner
Filtering process of h(t) :
y(t ) = H (0 ) cos[0t + (0 )]
= Harmonic Oscillation
0 = constant ; Amax = multiply by |H(0)| ; phase = additional with
constant (0) It means : after filtering happen phase shift of
( 0 ) t0 = 0
y(t ) x(t t0 )
(0 ) =
>0 <0 0
t0 =
left
>0 SHIFT = <0 0
t0 called as phase delay
right
of x(t)
2. If x(t) = random function
Fourier theorem
So
1 x(t ) = 2
x(t) can be figured as superposition of
some (un-limited) harmonic oscillation
x( )(cost + i sin t )d
Filtered by H() : - Amplitude multiplied by |H()| - Phase shifted by () A lit d multiply Amplitude lti l and d phase h shifting hifti change h to t frequency f Hence, Filter change shape of x(t) to be y(t) F For analyzing l i on transfer t f off energy spectral t l only l (without ( ith t phase) h ):
Y ( ) = H ( ) . X ( ) 2
yy (t ) = hh (t ) xx (t )
Classification of filtering : 1 Ch 1. Character t off filt filtering i a. Filter without distortion
H ( ) = constant >0
b. Filter of Amplitude distortion (and phase delay t0)
H ( ) =
Function of
( ( ) = t0 )
c. Filter of phase distortion (and amplification by H0)
H ( ) = H 0
= constant
() = non linear in , if H0 =1
all pass-filter
2 Mechanism of filtering 2. a. Natural filtering : Earth transmits all of physical signal (seismic wave, EM, etc) b. Filter of Instruments : Electric and mechanic analog filter (example : Seismograph) c. Filter of Mathematic : Mathematical simulation of filtering process (Filter digital)
3. Objective of filter a. Filtering for eliminating noise
to get clean signal
b. Filtering for eliminating specific (certain) noise from some interference
signal 4. Filter parameters a. Frequency filter Filter is designed base on differences of peak spectral component (in different band frequency) b. Amplitude filter the discrimination based on Amplitude difference (Trigger Level Filter) 5. Method of Filtering Example: a. Multiplication in FD b Convolution b. C l ti iin TD c. Cross correlation in TD
FT and Convolution C Convolution l ti and d FT Matched filter
There is some other method that used for special function and special required such as : Inverse Filtering, Matched filter, Wiener filter (optimum filter), Shaping filter, Spiking filter and Prediction error filter
Filtering g technique q
not only y time function (t) () can be applied also in space function x, y and z (and its equivalent)
TD (t), sec
space Domain (x, s, y.etc), m, km,etc
FD (f), Hz
wave number D (k), 1/x (km), 1/(m)
And filtering can be also extended to multi-dimension :
It means that is not only single function ( time only or space only), l ) but b t it iis a sett off ffunction ti in i space and d time ti as subject for a multi-dimension filter array seismology We only deals with single function filter according to point of 4 and 5 in classification of filter
Example of Freq. filter
A record of x(t) is consist of 2 components x1(t) and x2(t) :
x(t ) = x1 (t ) + x2 (t )
X ( ) = X1 ( ) + X 2 ( )
Assumed that X1() and X2() is different in amplitude spectrum
X1() is acquire and X2() will be avoided
Needs filter of H() with character :
[X1() + X 2 ()].H () = X1()
So
X 1 ( ) H ( ) X 1 ( ) + X 2 ( )
Characteristic of Filter Characteristic transfer of filter is indicated by its spectrum (FD)
Example : Based on shape of A Amplitude lit d transfer t f function (FD)
