Hi Friends

I want to know is there any block that converts my Power Spectral Density Function in frequency domain to Time domain function.

or Please any one know how to convert my power Spectral Density Function in frequency domain to Time domain..as a combination of Math Blocks...Please suggest me some ideas in this regard..

On Aug 19, 5:46=A0pm, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> Hi Friends
>
> I want to know is there any block that converts my Power Spectral Density=
 Function in frequency domain to Time domain function.
>
> or Please any one know how to convert my power Spectral Density Function =
in frequency domain to Time domain..as a combination of Math Blocks...Pleas=
e suggest me some ideas in this regard..

You can't.
Calculation of the PSD destroys all information about phase.
You tell me what you want to do about phase, then I'll tell you how to
inverse transform.

And after you've told me what you're going to do about phase, you can
tell me what a Math Block is.

TideMan <mulgor@gmail.com> wrote in message <1b1791c1-226c-4673-b43a-c33733fc6513@b4g2000pra.googlegroups.com>...
> On Aug 19, 5:46 pm, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> > Hi Friends
> >
> > I want to know is there any block that converts my Power Spectral Density Function in frequency domain to Time domain function.
> >
> > or Please any one know how to convert my power Spectral Density Function in frequency domain to Time domain..as a combination of Math Blocks...Please suggest me some ideas in this regard..
> 
> You can't.
> Calculation of the PSD destroys all information about phase.
> You tell me what you want to do about phase, then I'll tell you how to
> inverse transform.
> 
> And after you've told me what you're going to do about phase, you can
> tell me what a Math Block is.


Hi friend..

I have a Power Spectral density function which is in Frequency domain and i need the Time domain of its form thats all my need...

The fn is  fn(f)= a*([ ln(b/c)^2) ]^-1)*([1+ (1.5*((f*a)/d))])^-5/3

where a,b,c,d are constants..how this function can be converted to time domain

where f is frequency
Hi sorry Buddy if u unable to understand..

TideMan <mulgor@gmail.com> wrote in message <1b1791c1-226c-4673-b43a-c33733fc6513@b4g2000pra.googlegroups.com>...
> On Aug 19, 5:46 pm, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> > Hi Friends
> >
> > I want to know is there any block that converts my Power Spectral Density Function in frequency domain to Time domain function.
> >
> > or Please any one know how to convert my power Spectral Density Function in frequency domain to Time domain..as a combination of Math Blocks...Please suggest me some ideas in this regard..
> 
> You can't.
> Calculation of the PSD destroys all information about phase.
> You tell me what you want to do about phase, then I'll tell you how to
> inverse transform.
> 
> And after you've told me what you're going to do about phase, you can
> tell me what a Math Block is.
I need to add four function to get a new time domain function in the four function 3 are in Time domain and one is A POWER SPECTRAL DENSITY function in frequency domain,which i need to convert to time domain so that i can add and get it a complete time domain function....
and math block is nothing is there any pre defined math block available or not...i just enquird..

Thank u buddy ...in advance

On Aug 20, 9:22=A0pm, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> TideMan <mul...@gmail.com> wrote in message <1b1791c1-226c-4673-b43a-c337=
33fc6...@b4g2000pra.googlegroups.com>...
> > On Aug 19, 5:46=A0pm, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> > > Hi Friends
>
> > > I want to know is there any block that converts my Power Spectral Den=
sity Function in frequency domain to Time domain function.
>
> > > or Please any one know how to convert my power Spectral Density Funct=
ion in frequency domain to Time domain..as a combination of Math Blocks...P=
lease suggest me some ideas in this regard..
>
> > You can't.
> > Calculation of the PSD destroys all information about phase.
> > You tell me what you want to do about phase, then I'll tell you how to
> > inverse transform.
>
> > And after you've told me what you're going to do about phase, you can
> > tell me what a Math Block is.
>
> Hi friend..
>
> I have a Power Spectral density function which is in Frequency domain and=
 i need the Time domain of its form thats all my need...
>
> The fn is =A0fn(f)=3D a*([ ln(b/c)^2) ]^-1)*([1+ (1.5*((f*a)/d))])^-5/3
>
> where a,b,c,d are constants..how this function can be converted to time d=
omain
>
> where f is frequency
> Hi sorry Buddy if u unable to understand..

mmmm S~f^(-5/3)
Kolmogorov's -5/3 law of turbulence energy decay, right?
So you're modelling turbulence?
And you're looking for an analytical solution in the time domain?
Good luck!!
Why not simply solve the Navier Stokes equations directly?

TideMan <mulgor@gmail.com> wrote in message <2f68a056-79f6-4f57-a3cf-827241795a0c@u4g2000prn.googlegroups.com>...
> On Aug 20, 9:22 pm, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> > TideMan <mul...@gmail.com> wrote in message <1b1791c1-226c-4673-b43a-c33733fc6...@b4g2000pra.googlegroups.com>...
> > > On Aug 19, 5:46 pm, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> > > > Hi Friends
> >
> > > > I want to know is there any block that converts my Power Spectral Density Function in frequency domain to Time domain function.
> >
> > > > or Please any one know how to convert my power Spectral Density Function in frequency domain to Time domain..as a combination of Math Blocks...Please suggest me some ideas in this regard..
> >
> > > You can't.
> > > Calculation of the PSD destroys all information about phase.
> > > You tell me what you want to do about phase, then I'll tell you how to
> > > inverse transform.
> >
> > > And after you've told me what you're going to do about phase, you can
> > > tell me what a Math Block is.
> >
> > Hi friend..
> >
> > I have a Power Spectral density function which is in Frequency domain and i need the Time domain of its form thats all my need...
> >
> > The fn is  fn(f)= a*([ ln(b/c)^2) ]^-1)*([1+ (1.5*((f*a)/d))])^-5/3
> >
> > where a,b,c,d are constants..how this function can be converted to time domain
> >
> > where f is frequency
> > Hi sorry Buddy if u unable to understand..
> 
> mmmm S~f^(-5/3)
> Kolmogorov's -5/3 law of turbulence energy decay, right?
> So you're modelling turbulence?
> And you're looking for an analytical solution in the time domain?
> Good luck!!
> Why not simply solve the Navier Stokes equations directly?


Ya exactly..Buddy...ur truly eminent...
But i dont know about the navier stokes ...i am basically environmentalist...i am just need to study...
if u know about this kindly suggest me some books or link to any material  regarding how tackle this function and convert to time domain...
expecting ur reply .....
Thanks buddy...

On Aug 19, 1:46=A0am, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
> Hi Friends
>
> I want to know is there any block that converts my Power Spectral Density=
 Function in frequency domain to Time domain function.

No, there is not.

You can convert PSD to transform amplitude.

However, in order to get the correct time domain function
you also need the transform phase.

Prove it to yourself by transforming your favorite time
function and trying to invert without the phase.

Hope this helps.

Greg

On Aug 20, 9:45=A0am, Greg Heath <he...@alumni.brown.edu> wrote:
> On Aug 19, 1:46=A0am, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
>
> > Hi Friends
>
..> > I want to know is there any block that converts my Power Spectral
Density Function in frequency domain to Time domain function.
..>
..> No, there is not.
>
> You can convert PSD to transform amplitude.
>
> However, in order to get the correct time domain function
> you also need the transform phase.
>
> Prove it to yourself by transforming your favorite time
> function and trying to invert without the phase.
>
> Hope this helps.
>
> Greg

The OP hasn't asked for an invertible method from the time domain to
the frequency domain and back. Nor has the OP explicitly asked for a
unique time domain conversion from PSD. The truth is that there are
many time domain conversions available. For many modeling and testing
applications there is no requirement for a specific and unique time
waveform.

So, scale the PSD (if it actually is a PSD) to a power spectrum, pick
a phase function to use to convert power spectrum to complex
coefficients, then ifft.

There are still issues about data type and size and how to pick the
phase function that require more information about the application
than the OP has provided. For example, does the time series need to be
real only?

Dale B. Dalrymple

On Aug 20, 1:51=A0pm, dbd <d...@ieee.org> wrote:
> On Aug 20, 9:45=A0am, Greg Heath <he...@alumni.brown.edu> wrote:> On Aug =
19, 1:46=A0am, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
>
> > > Hi Friends
>
> .> > I want to know is there any block that converts my Power Spectral
> Density Function in frequency domain to Time domain function.
> .>
> .> No, there is not.
>
>
>
> > You can convert PSD to transform amplitude.
>
> > However, in order to get the correct time domain function
> > you also need the transform phase.
>
> > Prove it to yourself by transforming your favorite time
> > function and trying to invert without the phase.
>
> > Hope this helps.
>
> > Greg
>
> The OP hasn't asked for an invertible method from the time domain to
> the frequency domain and back. Nor has the OP explicitly asked for a
> unique time domain conversion from PSD. The truth is that there are
> many time domain conversions available. For many modeling and testing
> applications there is no requirement for a specific and unique time
> waveform.
>
> So, scale the PSD (if it actually is a PSD) to a power spectrum, pick
> a phase function to use to convert power spectrum to complex
> coefficients, then ifft.
>
> There are still issues about data type and size and how to pick the
> phase function that require more information about the application
> than the OP has provided. For example, does the time series need to be
> real only?

On Aug 20, 1:51 pm, dbd <d...@ieee.org> wrote:
> On Aug 20, 9:45 am, Greg Heath <he...@alumni.brown.edu> wrote:> On Aug 19=
, 1:46 am, "RAGAVENDAR " <raghuragaven...@gmail.com> wrote:
>
> > > Hi Friends
>
> .> > I want to know is there any block that converts my Power Spectral
> Density Function in frequency domain to Time domain function.
> .>
> .> No, there is not.
>
>
>
> > You can convert PSD to transform amplitude.
>
> > However, in order to get the correct time domain function
> > you also need the transform phase.
>
> > Prove it to yourself by transforming your favorite time
> > function and trying to invert without the phase.
>
> > Hope this helps.
>
> > Greg
>
> The OP hasn't asked for an invertible method from the time domain to
> the frequency domain and back. Nor has the OP explicitly asked for a
> unique time domain conversion from PSD. The truth is that there are
> many time domain conversions available. For many modeling and testing
> applications there is no requirement for a specific and unique time
> waveform.
>
> So, scale the PSD (if it actually is a PSD) to a power spectrum, pick
> a phase function to use to convert power spectrum to complex
> coefficients, then ifft.
>
> There are still issues about data type and size and how to pick the
> phase function that require more information about the application
> than the OP has provided. For example, does the time series need to be
> real only?

Excellent point.

I have modelled atmospheric scintillations of radar reflections
from satellites when the following were appropriate
a. uniform random phase (-pi,pi]
b. truncated random Gaussian phase (-pi,pi]

During that project I also investigated using the Hilbert
transform. However, the results did not match the data.

Hope this helps.

Greg