Interpolation noise when we resample the irregularly sampled data regularly

Started by qaisar April 8, 2005
Hello to all,

Dear i am trying to resample the irregularly sampled data regularly, in
order to apply the standard spectrum analysis techniques on it. For this
purpose i am using linear interpolation. But there exist interpolation
noise, so kindly tells me the methods to reduce the interpolation noise.

I shall be thank full to you for your Help.

Regards.

Qaisar


		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com
"qaisar" <alsaeed86@yahoo.com> wrote in message 
news:b9GdnStK4_htJsvfRVn-1w@giganews.com...
> > Hello to all, > > Dear i am trying to resample the irregularly sampled data regularly, in > order to apply the standard spectrum analysis techniques on it. For this > purpose i am using linear interpolation. But there exist interpolation > noise, so kindly tells me the methods to reduce the interpolation noise. > > I shall be thank full to you for your Help. > > Regards. > > Qaisar
There was a long thread here on that subject maybe a year ago. You could Google for it.... Fred
http://groups-beta.google.com/group/comp.dsp/browse_thread/thread/8e8704781babe1e7/9621af7ea381f01c?q=resample+irregularly+sampled+data&rnum=3#9621af7ea381f01c

"Peter K." <p.kootsookos@iolfree.ie> wrote in message 
news:1113014323.367665.221520@g14g2000cwa.googlegroups.com...
>
http://groups-beta.google.com/group/comp.dsp/browse_thread/thread/8e8704781babe1e7/9621af7ea381f01c?q=resample+irregularly+sampled+data&rnum=3#9621af7ea381f01c
>
and: http://groups-beta.google.com/group/comp.dsp/browse_frm/thread/aa2be6819f584d85/4076738cb693e9f8?q=comp.dsp:+Fred+Marshall+interpolation+linear+audio&rnum=1#4076738cb693e9f8 Olli Niemitalo
In article <b9GdnStK4_htJsvfRVn-1w@giganews.com>,
qaisar <alsaeed86@yahoo.com> wrote:
>Dear i am trying to resample the irregularly sampled data regularly, in >order to apply the standard spectrum analysis techniques on it. For this >purpose i am using linear interpolation. But there exist interpolation >noise, so kindly tells me the methods to reduce the interpolation noise.
You could use a higher-order interpolation than linear. I seem to recall that someone in this newsgroup pointed out a paper which related higher order Lagrangian interpolation to band-limited windowed-sync interpolators (which are often implemented by table lookup in the so-called multiphase technique). I wonder if non-uniformly spaced higher-order interpolation might then relate to some sort of non-symmetric warped window-sync weighting functions? (with a warping function related to the distance(s) to n nearest neighboring sample points?) If so, maybe multiphase interpolation could still be used, but with some sort of warped phase table indexing algorithm? IMHO. YMMV. -- Ron Nicholson rhn AT nicholson DOT com http://www.nicholson.com/rhn/ #include <canonical.disclaimer> // only my own opinions, etc.
>In article <b9GdnStK4_htJsvfRVn-1w@giganews.com>, >qaisar <alsaeed86@yahoo.com> wrote: >>Dear i am trying to resample the irregularly sampled data regularly, in >>order to apply the standard spectrum analysis techniques on it. For
this
>>purpose i am using linear interpolation. But there exist interpolation >>noise, so kindly tells me the methods to reduce the interpolation
noise.
> >You could use a higher-order interpolation than linear. > >I seem to recall that someone in this newsgroup pointed out a >paper which related higher order Lagrangian interpolation to
band-limited
>windowed-sync interpolators (which are often implemented by table lookup >in the so-called multiphase technique). I wonder if non-uniformly
spaced
>higher-order interpolation might then relate to some sort of
non-symmetric
>warped window-sync weighting functions? (with a warping function >related to the distance(s) to n nearest neighboring sample points?) >If so, maybe multiphase interpolation could still be used, but with some >sort of warped phase table indexing algorithm? > > >IMHO. YMMV. >-- >Ron Nicholson rhn AT nicholson DOT com http://www.nicholson.com/rhn/
>#include <canonical.disclaimer> // only my own opinions, etc. >
Dear its a nice idea but the use of multiphase interpolation for irregularly sampled data makes then algorithm quiet complex and requires more and more computational efforts. Can someone tell me that how much interpolation error is acceptable, inorder to have a spectrum with appropriate accuracy. Is there some defined limit on interpolation error and spectrum quality. Thanks in advance for your ideas. This message was sent using the Comp.DSP web interface on www.DSPRelated.com