On Mar 12, 3:03�pm, Vladimir Vassilevsky <antispam_bo...@hotmail.com> wrote:> Ron N. wrote: > >>>"Vladimir Vassilevsky" <antispam_bo...@hotmail.com> wrote in message > >>>news:_PzBj.15715$Ej5.2369@newssvr29.news.prodigy.net... > > >>>>Recently I run into the problem with a basic task: design of a polyphase > >>>>set of filters for the interpolation of a signal. The input is the sampled > >>>>signal; the output should be the interpolated values spaced at 1/10 of the > >>>>sample. > > >>>>So, I designed the LPF at 10 x sample rate by the Parks-McClellan > >>>>algorithm, and then decimated it into 10 subsets of the coefficients. > > >>>>But, although the original filter is well within the specs, the decimated > >>>>subsets (calculated to the original sample rate) are not! There is a > >>>>significant difference is in the passband flatness as well as in the > >>>>stopband. > > > I think the key difference might be that Vladimir was > > concerned about passband ripple after proper decimation or > > subsampling. > > My goal was a variable delay line with a resolution of 1/10 of a sample. > So I am using just one of the subsets of the coefficients of the x10 > interpolation filter at a time. It was quite unexpected to find out that > the frequency response of a subset filter in the passband could be very > different from that of the complete interpolation filter. > > I don't know what should be the exact design procedure, however it looks > like for the worst case the interpolator filter by N should have lower > stopband by the factor of N also. I guess you have to do several > iterations adjusting the stopband. > > Vladimir Vassilevsky > DSP and Mixed Signal Design Consultanthttp://www.abvolt.com- Hide quoted text - > > - Show quoted text -Hello Vlad, Yes you are on the right track. Years back I have to time delay equalize analog reconstructed channels on a E1 line. So I needed a bank of filters each with a different integral delay of 1/32 sample. So I designed a long low pass filter whose passband was a little less that 1/32 of the frequency range. And of course the stop bands had to be deeper by 32. The resulting filters when decimated by 32 worked wonderfully well. So you just have to look at where your aliasing is falling and make sure it does no harm. Clay
FIR interpolator
Started by ●March 11, 2008
Reply by ●March 13, 20082008-03-13
Reply by ●March 13, 20082008-03-13
On Wed, 12 Mar 2008 14:03:30 -0500, Vladimir Vassilevsky <antispam_bogus@hotmail.com> wrote:> > >Ron N. wrote: > >>>>"Vladimir Vassilevsky" <antispam_bo...@hotmail.com> wrote in message >>>>news:_PzBj.15715$Ej5.2369@newssvr29.news.prodigy.net... >>> >>>>>Recently I run into the problem with a basic task: design of a polyphase >>>>>set of filters for the interpolation of a signal. The input is the sampled >>>>>signal; the output should be the interpolated values spaced at 1/10 of the >>>>>sample. >>> >>>>>So, I designed the LPF at 10 x sample rate by the Parks-McClellan >>>>>algorithm, and then decimated it into 10 subsets of the coefficients. >>> >>>>>But, although the original filter is well within the specs, the decimated >>>>>subsets (calculated to the original sample rate) are not! There is a >>>>>significant difference is in the passband flatness as well as in the >>>>>stopband. >>> >> I think the key difference might be that Vladimir was >> concerned about passband ripple after proper decimation or >> subsampling. > > >My goal was a variable delay line with a resolution of 1/10 of a sample. >So I am using just one of the subsets of the coefficients of the x10 >interpolation filter at a time. It was quite unexpected to find out that >the frequency response of a subset filter in the passband could be very >different from that of the complete interpolation filter. > >I don't know what should be the exact design procedure, however it looks >like for the worst case the interpolator filter by N should have lower >stopband by the factor of N also. I guess you have to do several >iterations adjusting the stopband. > > > >Vladimir Vassilevsky >DSP and Mixed Signal Design Consultant >http://www.abvolt.comFWIW, my application was a delay-tracking filter to lock the symbol clock on a comm signal. If the receive reference and the transmit reference were not far off or locked, the polyphase would essentially be a variable-delay filter like you're describing, but in practice automatically tracked out the differences. I did, after the thing was in the field and working, have a sudden theoretical concern one day that perhaps it was really screwing up the phase response by doing it that way (and I had no real reason for concern, it ran about 0.3dB from theoretical in testing). I spent a bunch of time in MathCAD and with other tools worrying myself that these sorts of problems could be real, but in the end ran out of reasons for being concerned about it. That system had, IIRC, 256 phases. I was careful about the coefficient designs, but didn't go out of the way to do anything super-special. Perhaps I was lucky in accidentally having all the right steps in the design process. As I understand it that company is still selling those modems, and that was done about 13 years ago. I assume if it were broken it'd have shown up by now. ;) Eric Jacobsen Minister of Algorithms Abineau Communications http://www.ericjacobsen.org
