| Re: Savitzky-Golay filter [message #68607] |
Thu, 12 November 2009 23:57 |
d.poreh
Messages: 406
Registered: October 2007
|
Senior Member |
|
|
On Nov 12, 10:03 am, wlandsman <wlands...@gmail.com> wrote:
> On Nov 12, 3:22 am, Dave_Poreh <d.po...@gmail.com> wrote:
>
>> What about wavelet? Does it smooth like Savitzky-Golay filter?
>> Dave
>
> Why don't you tell us what data you have and what you want to do with
> it?
>
> Yes, there are wavelet smoothing techniques. They differ from
> Savitzky-Golay smoothing because -- well, because they use wavelets.
> --Wayne
Good answer!!!!
Dave
|
|
|
|
| Re: Savitzky-Golay filter [message #68613 is a reply to message #68607] |
Thu, 12 November 2009 10:03  |
wlandsman
Messages: 743
Registered: June 2000
|
Senior Member |
|
|
On Nov 12, 3:22 am, Dave_Poreh <d.po...@gmail.com> wrote:
> What about wavelet? Does it smooth like Savitzky-Golay filter?
> Dave
Why don't you tell us what data you have and what you want to do with
it?
Yes, there are wavelet smoothing techniques. They differ from
Savitzky-Golay smoothing because -- well, because they use wavelets.
--Wayne
|
|
|
|
| Re: Savitzky-Golay filter [message #68617 is a reply to message #68613] |
Thu, 12 November 2009 00:22 |
d.poreh
Messages: 406
Registered: October 2007
|
Senior Member |
|
|
On Nov 11, 11:12 pm, Dave_Poreh <d.po...@gmail.com> wrote:
> On Nov 11, 10:18 am, wlandsman <wlands...@gmail.com> wrote:
>
>
>
>
>
>> On Nov 11, 5:12 am, Dav_Poreh <d.po...@gmail.com> wrote:
>
>>> Folks
>>> Hi;
>>> I am running Savitzky-Golay filter to take the derivations (first and
>>> second order). In comparison to derive function there is remarkable
>>> difference between Savitzky-Golay and routine derivation. I don’t know
>>> which one is correct. Does this back to Taylor approximation or
>>> something else?
>>> Any help kindly appreciated
>>> Cheers
>
>> First of all, there is no single "correct" answer. You don't have a
>> continuous function to compute a derivative, but rather a sampled,
>> finite set of points. In the Savitzky-Golay filter one uses a local
>> polynomial approximation at each point, and then takes a derivative
>> of the polynomial. (So the derivative depends on the order of the
>> polynomial approximation, among other things.) You don't say what
>> your other method of computing the derivative is, but deriv.pro uses
>> a 3 point interpolation.
>
>> I would expect the two methods to give a similar answer for a smooth
>> function, but wouldn't be surprised to see them differ for a poorly-
>> sampled, or non-smooth function.
>
>> --Wayne
>
What about wavelet? Does it smooth like Savitzky-Golay filter?
Dave
|
|
|
|
| Re: Savitzky-Golay filter [message #68618 is a reply to message #68617] |
Wed, 11 November 2009 23:12 |
d.poreh
Messages: 406
Registered: October 2007
|
Senior Member |
|
|
On Nov 11, 10:18 am, wlandsman <wlands...@gmail.com> wrote:
> On Nov 11, 5:12 am, Dav_Poreh <d.po...@gmail.com> wrote:
>
>> Folks
>> Hi;
>> I am running Savitzky-Golay filter to take the derivations (first and
>> second order). In comparison to derive function there is remarkable
>> difference between Savitzky-Golay and routine derivation. I don’t know
>> which one is correct. Does this back to Taylor approximation or
>> something else?
>> Any help kindly appreciated
>> Cheers
>
> First of all, there is no single "correct" answer. You don't have a
> continuous function to compute a derivative, but rather a sampled,
> finite set of points. In the Savitzky-Golay filter one uses a local
> polynomial approximation at each point, and then takes a derivative
> of the polynomial. (So the derivative depends on the order of the
> polynomial approximation, among other things.) You don't say what
> your other method of computing the derivative is, but deriv.pro uses
> a 3 point interpolation.
>
> I would expect the two methods to give a similar answer for a smooth
> function, but wouldn't be surprised to see them differ for a poorly-
> sampled, or non-smooth function.
>
> --Wayne
Thanks Wayne.
Cheers
|
|
|
|
| Re: Savitzky-Golay filter [message #68639 is a reply to message #68618] |
Wed, 11 November 2009 10:18 |
wlandsman
Messages: 743
Registered: June 2000
|
Senior Member |
|
|
On Nov 11, 5:12 am, Dav_Poreh <d.po...@gmail.com> wrote:
> Folks
> Hi;
> I am running Savitzky-Golay filter to take the derivations (first and
> second order). In comparison to derive function there is remarkable
> difference between Savitzky-Golay and routine derivation. I don’t know
> which one is correct. Does this back to Taylor approximation or
> something else?
> Any help kindly appreciated
> Cheers
First of all, there is no single "correct" answer. You don't have a
continuous function to compute a derivative, but rather a sampled,
finite set of points. In the Savitzky-Golay filter one uses a local
polynomial approximation at each point, and then takes a derivative
of the polynomial. (So the derivative depends on the order of the
polynomial approximation, among other things.) You don't say what
your other method of computing the derivative is, but deriv.pro uses
a 3 point interpolation.
I would expect the two methods to give a similar answer for a smooth
function, but wouldn't be surprised to see them differ for a poorly-
sampled, or non-smooth function.
--Wayne
|
|
|
|
| Re: Savitzky-Golay filter [message #68640 is a reply to message #68639] |
Wed, 11 November 2009 07:00 |
d.poreh
Messages: 406
Registered: October 2007
|
Senior Member |
|
|
On Nov 11, 2:12 am, Dav_Poreh <d.po...@gmail.com> wrote:
> Folks
> Hi;
> I am running Savitzky-Golay filter to take the derivations (first and
> second order). In comparison to derive function there is remarkable
> difference between Savitzky-Golay and routine derivation. I don’t know
> which one is correct. Does this back to Taylor approximation or
> something else?
> Any help kindly appreciated
> Cheers
Folks
Hi;
I am running Savitzky-Golay filter to take the derivations (first
and
second order). In comparison to derive function there is remarkable
difference between Savitzky-Golay and routine derivation. I don’t
know
which one is correct. Does this back to Taylor approximation or
something else?
Any help kindly appreciated
Cheers
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
