 |
|
| Computers Forum Index » Computer - DSP » Laplace Transform vs Fourier transform... |
|
Page 1 of 3 Goto page 1, 2, 3 Next |
|
| Author |
Message |
| fisico32... |
Posted: Wed Oct 28, 2009 5:16 am |
|
|
|
Guest
|
Hello Forum,
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
What about the Laplace transform? What is, conceptually, the advantage of
taking the Laplace transform of a signal instead of the Fourier transform?
I know that by looking at its zeros and poles we can assess the stability
of the system for example.... is that not possible with the FT alone?
Does the Laplace transform show information on the system that the FT does
not?
thanks
fisico32 |
|
|
| Back to top |
|
|
|
| Jerry Avins... |
Posted: Wed Oct 28, 2009 5:16 am |
|
|
|
Guest
|
fisico32 wrote:
Quote: Hello Forum,
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
Not only don't I need review, your statement is too particular to be
accurate. The independent variable of a Fourier transform can be angular
or spatial frequency, or any physical quantity. The near and far fields
of electromagnetic radiation are related by a Fourier transform.
Quote: What about the Laplace transform? What is, conceptually, the advantage of
taking the Laplace transform of a signal instead of the Fourier transform?
I know that by looking at its zeros and poles we can assess the stability
of the system for example.... is that not possible with the FT alone?
Sure, but why use a sledge hammer when a tack hammer serves?
Quote: Does the Laplace transform show information on the system that the FT does
not?
Depending on the information, it can show the information more easily.
Think of the Laplace transform as a one-sided Fourier transform,
eminently suitable for causal functions.
Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ |
|
|
| Back to top |
|
|
|
| brent... |
Posted: Wed Oct 28, 2009 5:16 am |
|
|
|
Guest
|
On Oct 27, 8:39 pm, "fisico32" <marcoscipio... at (no spam) gmail.com> wrote:
Quote: Hello Forum,
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
What about the Laplace transform? What is, conceptually, the advantage of
taking the Laplace transform of a signal instead of the Fourier transform?
I know that by looking at its zeros and poles we can assess the stability
of the system for example.... is that not possible with the FT alone?
Does the Laplace transform show information on the system that the FT does
not?
thanks
fisico32
The fourier transform is a special case of the laplace transform (as I
understand it)
The Laplace transform correlates a given waveform with every possible
(exponential x sinusoidal) wave.
The Fourier transform correlates it with every possible sinusoidal
wave.
Here is the one point of the Laplace transrom:
If you correlate a signal with a "decaying" exponential, and the
resulting correlation is unbounded, then you have a stability problem.
(if you have pole in right hand plane it means that your correlation
became infinite with an exponentially decaying sinusoid)
It is OK to correlate your system with an exponentially "increasing"
sinusoidal wave and have an unbounded result.
So you take your signal and correlate it with every possible
exponentially increasing and exponentially decreasing sinusoidal wave
(sinusoidal means both exponential cosines and exponential sines) and
see what happens.
One thing to note, when you get a pole in the S plane, that really
sets the boundary of instability. There will be an infinite number of
poles to the left of the leftmost "official" pole.
The fourier transform only correlates the f(t) with pure (non
exponential) sinusoids.
This is what we typically want to know about when thinking of the
frequency response of a circuit (or system)
The fourier transform is especially useful because of the convolution
theorem, that says multiplication in the frequency domain is equiv to
convolution in the time domain. The response of the circuit is the
impulse response convolved with the sinusoidal input signal. This is
a piece of cake when you take the fourier transform of the system to
immediately get the frequency response.
In conclusion
Fourier transform is good tool for frequency response
Laplace transform is good tool for stability analysis.
PS
Laplace also is a good tool for solving differential equations becuase
the S-plane is a mapping of every possible solution to an ordinary
differential equation.
caveat: this may all be BS so do your own due diligence  |
|
|
| Back to top |
|
|
|
| HardySpicer... |
Posted: Wed Oct 28, 2009 6:26 pm |
|
|
|
Guest
|
On Oct 27, 6:39 pm, "fisico32" <marcoscipio... at (no spam) gmail.com> wrote:
Quote: Hello Forum,
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
What about the Laplace transform? What is, conceptually, the advantage of
taking the Laplace transform of a signal instead of the Fourier transform?
I know that by looking at its zeros and poles we can assess the stability
of the system for example.... is that not possible with the FT alone?
Does the Laplace transform show information on the system that the FT does
not?
thanks
fisico32
Fourier transform only holds for s=jw. The Laplace TF holds for all
complex frequency sigma+jw. There are some functions which have no FT
of meaning.
The Laplace TF is usually used for solving ODEs and for transfer
functions.
Hardy |
|
|
| Back to top |
|
|
|
| robert bristow-johnson... |
Posted: Thu Oct 29, 2009 3:57 am |
|
|
|
Guest
|
just to add to what brent and Jerry and Hardy said...
On Oct 27, 9:39 pm, "fisico32" <marcoscipio... at (no spam) gmail.com> wrote:
Quote:
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
there's a part of that angular frequency that is mathematical, but not
physical. how do you physically measure negative frequency? and
negative frequency is in the F.T.
Quote:
What about the Laplace transform? What is, conceptually, the advantage of
taking the Laplace transform of a signal instead of the Fourier transform?
it's hard to derive the FT of the unit step function without a little
bit of hand-waving.
Quote: I know that by looking at its zeros and poles we can assess the stability
of the system for example.... is that not possible with the FT alone?
Does the Laplace transform show information on the system that the FT does
not?
there is a mathematical concept relating the two called "analytical
extension". if a function of a complex variable is "analytic", it
turns out that if you know the values of the function on a certain
boundary (like the jw axis, where the LT is equal to the FT), then
that gives you sufficient information to define the function at all
points away from the boundary. so, for decently-behaved complex H
(jw), knowing the behavior of H(jw) for all real w is sufficient to
know H(s) for s having a non-zero real part. so, for decently-behaved
H(s), the LT has no additional information than the FT can tell you.
r b-j |
|
|
| Back to top |
|
|
|
| Gordon Sande... |
Posted: Thu Oct 29, 2009 10:41 pm |
|
|
|
Guest
|
On 2009-10-29 15:00:06 -0300, Rich Webb <bbew.ar at (no spam) mapson.nozirev.ten> said:
Quote: On Thu, 29 Oct 2009 17:38:31 GMT, Gordon Sande
g.sande at (no spam) worldnet.att.net> wrote:
On 2009-10-29 14:20:36 -0300, robert bristow-johnson
rbj at (no spam) audioimagination.com> said:
On Oct 29, 10:52 am, Jerry Avins <j... at (no spam) ieee.org> wrote:
robert bristow-johnson wrote:
...
there's a part of that angular frequency that is mathematical, but not
physical. how do you physically measure negative frequency? and
negative frequency is in the F.T.
Who said? You can just as well apply the negative sign in exp(-jwt) to
time. :-)
...
we've been here before. for the record, i actually think that the
term "imaginary" for imaginary numbers is apt, appropriate, and
descriptive.
Like surds, irrational, etc. Mathematical terminology from the middle ages
tended to be very unkind to all sorts of new fangled things as told by
any history of mathematics book. That was back when solving a quadratic
equation was a big deal and solution to quartics was a closely guarded
trade secret. It is too bad that 500 year old attitudes are taken seriously
by some current folks.
Middle Ages? It wasn't that long ago that simple negative numbers were
not universally accepted by mathematicians. A brief discussion at:
http://nrich.maths.org/5961
There are always a few who have trouble staying current.  |
|
|
| Back to top |
|
|
|
| Gordon Sande... |
Posted: Thu Oct 29, 2009 10:41 pm |
|
|
|
Guest
|
On 2009-10-29 14:51:17 -0300, robert bristow-johnson
<rbj at (no spam) audioimagination.com> said:
Quote: On Oct 29, 1:38 pm, Gordon Sande <g.sa... at (no spam) worldnet.att.net> wrote:
On 2009-10-29 14:20:36 -0300, robert bristow-johnson
r... at (no spam) audioimagination.com> said:
On Oct 29, 10:52 am, Jerry Avins <j... at (no spam) ieee.org> wrote:
robert bristow-johnson wrote:
...
there's a part of that angular frequency that is mathematical, but no
t
physical. how do you physically measure negative frequency? and
negative frequency is in the F.T.
Who said? You can just as well apply the negative sign in exp(-jwt) to
time. :-)
...
we've been here before. for the record, i actually think that the
term "imaginary" for imaginary numbers is apt, appropriate, and
descriptive.
Like surds, irrational, etc. Mathematical terminology from the middle age
s
tended to be very unkind to all sorts of new fangled things as told by
any history of mathematics book. That was back when solving a quadratic
equation was a big deal and solution to quartics was a closely guarded
trade secret. It is too bad that 500 year old attitudes are taken serious
ly
by some current folks.
am i one of them current folk, Gordon?
r b-j
If you act like an imaginary number is a invention of the devil then
all one can say is "If the shoe fits ..." |
|
|
| Back to top |
|
|
|
| robert bristow-johnson... |
Posted: Thu Oct 29, 2009 11:42 pm |
|
|
|
Guest
|
On Oct 29, 2:41 pm, Gordon Sande <g.sa... at (no spam) worldnet.att.net> wrote:
Quote: On 2009-10-29 14:51:17 -0300, robert bristow-johnson
r... at (no spam) audioimagination.com> said:
On Oct 29, 1:38 pm, Gordon Sande <g.sa... at (no spam) worldnet.att.net> wrote:
On 2009-10-29 14:20:36 -0300, robert bristow-johnson
r... at (no spam) audioimagination.com> said:
On Oct 29, 10:52 am, Jerry Avins <j... at (no spam) ieee.org> wrote:
robert bristow-johnson wrote:
...
there's a part of that angular frequency that is mathematical, but no
t
physical. how do you physically measure negative frequency? and
negative frequency is in the F.T.
Who said? You can just as well apply the negative sign in exp(-jwt) to
time. :-)
...
we've been here before. for the record, i actually think that the
term "imaginary" for imaginary numbers is apt, appropriate, and
descriptive.
Like surds, irrational, etc. Mathematical terminology from the middle age
s
tended to be very unkind to all sorts of new fangled things as told by
any history of mathematics book. That was back when solving a quadratic
equation was a big deal and solution to quartics was a closely guarded
trade secret. It is too bad that 500 year old attitudes are taken serious
ly
by some current folks.
am i one of them current folk, Gordon?
If you act like an imaginary number is a invention of the devil then
all one can say is "If the shoe fits ..."
well, you'll have to judge if i act like that. from where i stand,
all's i'm saying is that real (meaning that they really exist)
physical quantities are measured as real numbers.
r b-j |
|
|
| Back to top |
|
|
|
| robert bristow-johnson... |
Posted: Thu Oct 29, 2009 11:46 pm |
|
|
|
Guest
|
On Oct 29, 7:21 pm, John Monro <johnmo... at (no spam) optusnet.com.au> wrote:
Quote: robert bristow-johnson wrote:
just to add to what brent and Jerry and Hardy said...
On Oct 27, 9:39 pm, "fisico32" <marcoscipio... at (no spam) gmail.com> wrote:
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
there's a part of that angular frequency that is mathematical, but not
physical. how do you physically measure negative frequency? and
negative frequency is in the F.T.
(snip)
r b-j
Configure a CRO for XY mode.
Connect the I signal to the X channel and the Q signal to
the Y channel.
The spot traces an anti-clockwise circular path for positive
frequency signals; clockwise for negative frequency.
Count the number of circles per second and attach the
appropriate sign.
that's fine and a legit human interpretation or abstraction of what is
happening with two real and physically related signals are observed on
an oscilloscope in XY mode.
you still measure the Y signal at any instance in time as a real
value.
r b-j |
|
|
| Back to top |
|
|
|
| glen herrmannsfeldt... |
Posted: Fri Oct 30, 2009 12:05 am |
|
|
|
Guest
|
robert bristow-johnson <rbj at (no spam) audioimagination.com> wrote:
Quote: well, you'll have to judge if i act like that. from where i stand,
all's i'm saying is that real (meaning that they really exist)
physical quantities are measured as real numbers.
Many physical quantities aren't necessarily real.
Index of refraction is usually complex.
Closer to DSP, impedance is often complex.
-- glen |
|
|
| Back to top |
|
|
|
| John Monro... |
Posted: Fri Oct 30, 2009 3:21 am |
|
|
|
Guest
|
robert bristow-johnson wrote:
Quote: just to add to what brent and Jerry and Hardy said...
On Oct 27, 9:39 pm, "fisico32" <marcoscipio... at (no spam) gmail.com> wrote:
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
there's a part of that angular frequency that is mathematical, but not
physical. how do you physically measure negative frequency? and
negative frequency is in the F.T.
(snip)
Configure a CRO for XY mode.
Connect the I signal to the X channel and the Q signal to
the Y channel.
The spot traces an anti-clockwise circular path for positive
frequency signals; clockwise for negative frequency.
Count the number of circles per second and attach the
appropriate sign.
Regards,
John |
|
|
| Back to top |
|
|
|
| Jerry Avins... |
Posted: Fri Oct 30, 2009 4:49 am |
|
|
|
Guest
|
John Monro wrote:
Quote: robert bristow-johnson wrote:
just to add to what brent and Jerry and Hardy said...
On Oct 27, 9:39 pm, "fisico32" <marcoscipio... at (no spam) gmail.com> wrote:
everyone is familiar with the Fourier transform and its importance.
Its independent variable, the angular frequency w, is a measurable
physical quantity.
there's a part of that angular frequency that is mathematical, but not
physical. how do you physically measure negative frequency? and
negative frequency is in the F.T.
(snip)
r b-j
Configure a CRO for XY mode.
Connect the I signal to the X channel and the Q signal to the Y channel.
The spot traces an anti-clockwise circular path for positive frequency
signals; clockwise for negative frequency.
Count the number of circles per second and attach the appropriate sign.
Activate the INVERT switch on the vertical channel and watch time run
backwards.
Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ |
|
|
| Back to top |
|
|
|
| glen herrmannsfeldt... |
Posted: Fri Oct 30, 2009 5:55 pm |
|
|
|
Guest
|
Jerry Avins <jya at (no spam) ieee.org> wrote:
(snip, someone wrote)
Quote: This is because the magnitude of the imaginary component of a
complex signal is a real signal.
Of course. All things measurable are real. "Complex" numbers are merely
a clever and useful bookkeeping scheme for manipulating related pairs of
real quantities.
Or pairs of real numbers are a convenient way of measuring complex
quantities.
Impedance and index of refraction are both complex, and for similar
reasons. We can separately measure resistance and reactance, or
the real and imaginary parts of the index of refraction.
(The imaginary part comes from absorption.) Both are due to
the interaction of electrons with atoms, and with each other.
Often the available materials are fairly close to ideal, such
that we can separate the quantities. Resistors do have inductance,
inductors (except superconductors) do have resistance.
-- glen |
|
|
| Back to top |
|
|
|
| glen herrmannsfeldt... |
Posted: Fri Oct 30, 2009 7:52 pm |
|
|
|
Guest
|
Jerry Avins <jya at (no spam) ieee.org> wrote:
(snip)
Quote: Complex numbers are a clever and useful way to represent those
quantities in cases where phase shift has meaning.
http://en.wikipedia.org/wiki/Refractive_index
Quote: Compact notations expand out ability to comprehend (same root as in
prehensile) complicated things. Vector analysis, quaternions, complex
numbers, matrix algebra .. without them, we'd be hard pressed to
express, let alone understand, some phenomena as concepts. Nevertheless,
being built up from simpler stuff (Shall we go back to Peano's axioms?)
thay constitute the HLLs of math. Ultimately, all running code executes
assembly language.
In the case of discrete ideal L, R, and C, I would have to agree.
Consider, though, an electromagnetic wave going through an imperfect
(like the ones we have available) dielectric. The motion of the
electrons generates new waves that combine with the original wave
in such a way as to make it appear to move slower in the material
(for n>1), and in most cases decrease in amplitude (absorption).
Yes, it is due to the phase of the motion of the electrons relative
to the phase of the incoming wave. The electrons move following
some fairly simple differential equations, though not necessarily
with simple solutions.
The wave comes out of the material phase shifted (due to the
real part) and attenuated (due to the imaginary part of the
index of refraction). Unlike the case of discrete R, L, and C,
it is not so easy to separate the two.
In non-magnetic materials, the index of refraction is the
square root of the dielectric constant. That is, the complex
index of refraction is the square root of the complex dielectric
constant, describing the interaction of the material with the
electromagnetic wave.
Also, see:
http://en.wikipedia.org/wiki/Kramers%E2%80%93Kronig_relation
and especially:
http://en.wikipedia.org/wiki/Kramers%E2%80%93Kronig_relation#Physica_interpretation_and_alternate_form
Which in the end describes the interdependence between the real and
imaginary parts of the dielectric constant and/or index of refraction.
Yes it is all phase shifts, but just saying that leaves a lot out.
-- glen |
|
|
| Back to top |
|
|
|
| Jerry Avins... |
Posted: Fri Oct 30, 2009 10:14 pm |
|
|
|
Guest
|
glen herrmannsfeldt wrote:
Quote: Jerry Avins <jya at (no spam) ieee.org> wrote:
(snip, someone wrote)
This is because the magnitude of the imaginary component of a
complex signal is a real signal.
Of course. All things measurable are real. "Complex" numbers are merely
a clever and useful bookkeeping scheme for manipulating related pairs of
real quantities.
Or pairs of real numbers are a convenient way of measuring complex
quantities.
Impedance and index of refraction are both complex, and for similar
reasons. We can separately measure resistance and reactance, or
the real and imaginary parts of the index of refraction.
(The imaginary part comes from absorption.) Both are due to
the interaction of electrons with atoms, and with each other.
Often the available materials are fairly close to ideal, such
that we can separate the quantities. Resistors do have inductance,
inductors (except superconductors) do have resistance.
Complex numbers are a clever and useful way to represent those
quantities in cases where phase shift has meaning.
Compact notations expand out ability to comprehend (same root as in
prehensile) complicated things. Vector analysis, quaternions, complex
numbers, matrix algebra .. without them, we'd be hard pressed to
express, let alone understand, some phenomena as concepts. Nevertheless,
being built up from simpler stuff (Shall we go back to Peano's axioms?)
thay constitute the HLLs of math. Ultimately, all running code executes
assembly language.
Jerry
--
Engineering is the art of making what you want from things you can get.
ŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻŻ |
|
|
| Back to top |
|
|
|
|
|
All times are GMT
The time now is Wed Dec 02, 2009 3:48 am
|
|