Prev: Possible new faster algorithms for FFT of 3^n and 6^n sizes.
Next: conventional wisdom how to upsample very large arrays, accurately?
From: JCH on 10 Jan 2010 05:12

"pnachtwey" <pnachtwey(a)gmail.com> schrieb im Newsbeitrag
news:4e8b1657-7e35-4630-9090-298b23d6b8ea(a)u7g2000yqm.googlegroups.com...
> On Jan 9, 1:55 am, "JCH" <ja...(a)nospam.arcornews.de> wrote:
>> "Laron" <jason.pi...(a)inbox.com> schrieb im
>> Newsbeitragnews:BOednfutGppmrtnWnZ2dnUVZ_uWdnZ2d(a)giganews.com...
>>
>> > Hi,
>> > When simulate the FIR filter response, run interp(Matrix,n) in
>> > matlab,
>> > the maximum of Matrix is 1,but the response is larger than 1?
>> > I wonder know why this could be happen and how to degrade this
>> > effect?
>>
>> You possibly have an ODE (system) of 2nd degree:
>>
>> See Page 1
>> *http://home.arcor.de/janch/janch/_control/20100109-overshooting/
>> 0,0001 u'' + 0,01 u' + u = w
>>
>> See Page 2: Degrading using higher damping
>> 0,0001 u'' + 0,018 u' + u = w
>>
>> --
>> Regards JCH
> This has nothing to do with the original question. Laron wants to
> know how to INTERPOLATE without OVERSHOOTING.


A 'possible' alternative explaination could be a bad DFT approximation
(Gibbs phenomenon). I don't like calling it interpolation in this context.

What you see here is approximation:

* http://home.arcor.de/janch/janch/_control/20100110-dft/


--
Regards JCH





From: jim on 10 Jan 2010 16:15


pnachtwey wrote:

> On Jan 6, 4:41 pm, jim <"sjedgingN0Sp"@m(a)mwt,net> wrote:
> > Laron wrote:
> >
> > > I just know there would be a lpf, not quite sure about the reconstruction
> > > process.
> > > I got an idea that the overshoot caused from the lpf, but the detail "how"
> > > is still not clear.
> >
> > A low pass filter with only positive values and a DC gain of 1 will
> > guarantee no overshoot.
> >
> > -jim
> But that is not interpolating.

And why would you say that?

Take his example of [1 0 1 1 0 1].

Stuff it with 0 so that it becomes

[1 0 0 0 1 0 1 0 0 0 1 0]


Low Pass Filter with [1/2 1 1/2]

and you get

[1 .5 0 .5 1 1 1 .5 0 .5 1]

In practice this will be done more efficiently by filtering the original with
[1/2 1/2] and interleaving the filtered result with the original sequence.


-jim

First  |  Prev  | 
Pages: 1 2 3 4 5
Prev: Possible new faster algorithms for FFT of 3^n and 6^n sizes.
Next: conventional wisdom how to upsample very large arrays, accurately?